Government If 1 of Canada Gouvernernent du Canada

Canadian General Office des normes Standards Board gdndrales du Canada Reaffirmed

May 1997

Advanced Manual For: Eddy. 7 - Current Test Method

~anonal aranaara of Canada

The CANADIAN GENERAL STANDARDS BOARD (CGSB). under whose auspices this National Standard of Canada has been developed is a government agency within Public Works and Government Services Canada. CGSB is engaged in the production of voluntary standards in a wide range of subject areas through the media of standards committees and !he consensus process. The standards committees are composed of representatives of relevant interests including producers, consumers and other users, retailers. governments, educational institutions, technical, professional and trade societies, and research and testing organizations. Any given standard is developed on the consensus of views expressed by such representatives.

CGSB has been accredited by the Standards Council of Canada as a national standards-development organization. The standards that it develops and offers as National Standards of Canada conform to the criteria and procedures established for this purpose by the Standards Council of Canada. In addition to standards it publishes as national standards, CGSB produces standards to meet particular needs, in response to requests from a variety of sources in both the public and private sectors. Both CGSB standards and CGSB national standards are developed in conformance with the policies described in the CGSR Policy Manual for the Development and Maintenance of Standards.

CGSB standards are subject to review and revision to ensure that they keep abreast of technological progress. Sugsestions for their improvement, which are always welcome, should be brought to the notice of the standards committees concerned. Changes to standards are issued either as separate amendment sheets or in new editions of standards.

An up-to-date listing of CGSB standards, including details on latest issues and amendments, and ordering instructions, is found in the CGSB Catalogue, which is published annually and is available without charge upon request. An electronic version, ECAT, is also available. More information is available about CGSB products and services at our Web site - http://www.pwgsc.gc.cdcgsb.

Although the intended primary application of this standard is stated in its Scope, it is important to note that it remains the responsibility of the users of the standard to judge its suitability for their particular purpose.

The STANDARDS COUNCIL OF CANADA is the co-ordinating body of the National Standards System, a federation of independent, autonomous organizations working towards the further development and improvement of voluntary standardization in the national interest.

The principal objectives of the Council are to foster and promote voluntary standardization as a means of advancing the national economy, benefiting the health, safety and welfare of the public. assisting and protecting the consumer, facilitating domestic and international trade, and furthering international co-operation in the field of standards.

A National Standard of Canada is a standard which has been approved by the Standards Council of Canada and one which reflects a reasonable agreement among the views of a number of capable individuals whose collective interests provide, to the greatest practicable extent, a balance of representation of producers, users, consumers and others with relevant interests, as may be appropriate to the subject in hand. I t normally is a standard that is capable of making a significant and timely contribution to the national interest.

Approval of a standard as a National Standard of Canada indicates that a standard conforms to the criteria and procedures established by the Standards Council of Canada. Approval does not refer to the technical content of the standard: this remains the continuing responsibility of the accredited standards-development organization.

Those who have a need to apply standards are encouraged to use National Standards of Canada whenever practicable. These standards are subject to periodic review: therefore, users are cautioned to obtain the latest edition from the organization preparing the standard.

The responsibility for approving National Standards of Canada rests with the:

Standards Council of Canada 270 Albert Street Suite 200 Ottawa, Ontario KIP 6N7

The testing and evaluation of a product against this standard may require the use of materials and/or equipment that could be hazardous. This document does not purport to address all the safety aspects associated with its use. Anyone using this standard has the responsibility to consult the appropriate authorities and to establish appropriate health and safety practices in conjunction with any applicable regulatory requirements prior to its use. CGSB neither assumes nor accepts any responsibility for any injury or damage that may occur during or as the result of tests, wherever performed.

Attention is drawn to the possibility that some of the elements of this Canadian standard may be the subject of patent rights. CGSB shall not be held responsible for identifying any or all such patent rights. Users of this standard are expressly advised that determination of the validity of any such patent rights are entirely their own responsibility.

Funher information on CGSB and its services and standards may be obtained from:

The Manager Strategic Standardization Group Canadian General Standards Board Ottawa, Canada KIA 1G6

How to order ccC~>a Publications:

by telephone - (819) 956-0425 or - 1-800-665-CGSB

(Canada only)

by fax - (819) 956-5644

by mail - CGSB Sales Centre Ottawa, Canada K I A I G 6

I in person - Place du Portage Phase 111, 6B 1 1 1 Laurier S m e t Hull, Quebec

by email - ncr.cgsb-ongc@pwgsc.gc.ca

on the W e b - http:Nwww.pwgsc.gc.ca/cgsb

NATIONAL STANDARD OF CANADA Reaffirmed May 1997

ADVANCED MANUAL FOR: EDDY CURRENT TEST METHOD

Notes:

This Manual is based on Eddy Current Manual Volume I: Test Method published by the Atomic Energy of Canada Limited. Chalk River Laboratories (Report AECL7523), written by V.S. Cecco, G. Van Drunen and F.L. Sharp.

Publication RC 1433, Innovations in Eddy Current Testing (Document Number RC 1433). complements this manual and is available through the Atomic Energy of Canada Limited, Engineering Technologies Division, Nondestructive Testing Development Branch, Chalk River Laboratories. Telephone (613) 584-3311 ext. 4623.

Prepared by the Approved by the Canadian General Standards Board Ccc- Standards Council of Canada @

Published February 1986 by the Canadian General Standards Board

Ottawa, Canada KIA 1G6

@Minister of Public Works and Government Services Canada - 1 9 9

No pan of this publication may be repmdud in any form without the prior permission of the publisher.

CANADIAN GENERAL STANDARDS BOARD

COMMITTEE ON NONDESTRUCTIVE TESTING, EDDY CURRENT METHOD

(Membership at date of reaffirmation)

Cecco, V.S. Chairperson

Bagley, W.

Dewalle, S.

Dewalle, R.

Dziuh, G.

Fiorito, G.

Kennedy, W.

Marshall, D.

Martin. D.G.

Mazurek, G.

Momson, G.

Newhury, J.

Pfeiffer, F.

Quirion, Capt A.

Reid, J.

Schnuhh. P.

Siehen, G.

Stasuk, D.G.

Szpakowski, E.

Szucs, J.R.

Taylor, D.

Tremblay. S.Y.

Wright. R.

S i i e n , E. Secretary

Atomic Energy of Canada Ltd.

Ingersoll-Dresser Pumps Canada Ltd.

Canadian N.D.E. Technology Ltd.

Andec Manufacturing Ltd.

Canadair Inc.

Collkge Ahuntsic

Canadian Welding Bureau

Dofasco Inc.

Ontario Hydro

Consultant

Air Ontario

Natural Resources Canada

Canadian Airlines International

National Defence

IndusVial Nondestructive Testing

First Air

NDT Management Association

Stasuk Testing & Inspection Ltd.

Transport Canada

Air Canada Base 025

Taylor NDE Services

Metaltec Inc.

Norsand Metals Inc.

Canadian General Standards Board

Acknowlrdg,nrnf is made for ihr fmnslution qf lhis Nnfionul Srandard of Canu& by the Tramlorion Bureau of Publir Works and Govrmmrn! Srwirrs Cunuda.

CANICGSB-48.14-M86

ADVANCED MANUAL FOR EDDY CURRENT TEST METHOD

TABLE OF CONTENTS

L, CHAPTER 1 - SCOPE AND INTRODUCTION

1.1 SCOPE 1.2 EDDY CURRENT TESTING 1.3 HISTORICAL PERSPECTIVE OF EDDY CURRENT TESTING 1.4 ORGANIZATION OF MANUAL

CHAPTER 2 - EDDY CURRENT FUNDAMENTALS

2.1 BASIC EQUIPMENT 2.2 GENERATION O F EDDY CURRENTS

2.2.1 Introduction 2.2.2 Magnetic Field Around a Coil 2.2.3 Equations Governing Generation of Eddy Currents

2.3 FUNDAMENTAL PROPERTIES OF EDDY CURRENT FLOW 2.4 SKIN EFFECT

2.4.1 Standard Depth of Penetration 2.4.2 Depth of Penetration in Finite Thickness Samples 2.4.3 Standard Phase Lag 2.4.4 Phase Lag in Finite Thickness Samples

L 2.5 SUMMARY 2.6 WORKED EXAMPLES

2.6.1 Standard Depth of Penetration and Phase Lag

CHAPTER 3 - ELECTRICAL CIRCUITS AND PROBE IMPEDANCE

3.1 INTRODUCTION 3.2 IMPEDANCE EQUATIONS AND DEFINITIONS 3.3 SINUSOIDS, PHASORS AND ELECTRICAL CIRCUITS 3.4 MODEL OF PROBE IN PRESENCE OF TEST MATERIAL 3.5 SIMPLIFIED IMPEDANCE DIAGRAMS

3.5.1 Derivation of Probe Impedance for ProbeISample Combination 3.5.2 Correlation Between Coil Impedance and Sample Properties

3.6 SUMMARY 3.7 WORKED EXAMPLES

3.7.1 Probe Impedance in Air 3.7.2 Probe Impedance Adjacent t o Sample 3.7.3 Voltage-Current Relationship

PAGE

CHAPTER 4 - INSTRUMENTATION

PAGE

4.1 INTRODUCTION 4.2 BRIDGE CIRCUITS

4.2.1 Simple Bridge Circuit 4.2.2 Typical Bridge Circuit in Eddy Current Instruments 4.2.3 Bridge Circuit in Crack Detectors

4.3 RESONANCE CIRCUIT AND EQUATIONS 4.4 EDDY CURRENT INSTRUMENTS

4.4.1 General Purpose Instrument (Impedance Methad) 4.4.2 Crack Detectors 4.4.3 Material Sorting and Conductivity Instruments

4.5 SEND-RECEIVE EDDY CURRENT SYSTEMS

4.5.1 Hall-Effect Detector 4.5.2 Send-Receive Coils and Lif t-Off Compensation

4.6 MULTIFREQUENCY EQUIPMENT 4.7 PULSED EDDY CURRENT EQUIPMENT 4.8 SPECIAL TECHNIQUES 4.9 RECORDING EQUIPMENT

4.9.1 Frequency Response

4.10 SUMMARY 4.1 1 WORKED EXAMPLES

4.1 1.1 Impedance a t Resonance

CHAPTER 5 - TESTING WITH SURFACE PROBES

5.1 INTRODUCTION 5.2 SURFACE PROBES

5.2.1 Probe Types 5.2.2 Directional Properties

5.2.2.1 Sensitivity a t Cent re of a Coil

5.2.3 Probe Inductance

PARAMETERS AFFECTING SENSITIVITY TO DEFECTS

5.3.1 Sensitivity with Lift-off and Defect Depth 5.3.2 Effect of Defect Length

COMPARAISON BETWEEN SURFACE AND THROUGH-WALL INSPECTION IMPEDANCE GRAPH DISPLAY

5.5.1 Effect of Resistivity 5.5.2 Effect of Permeability 5.5.3 Effect of Thickness 5.5.4 Effect of Frequency 5.5.5 Effect of Probe Diameter 5.5.6 Comparison of Experimental and Computer Impedance Diagrams

CHARACTERISTIC PARAMETER DEFINITION O F "PHASE" TERMINOLOGY SELECTION O F TEST FREQUENCY

5.8.1 Inspecting fo r Defects 5.8.2 Measuring Resistivity 5.8.3 Measuring Thickness 5.8.4 Measuring Thickness of a Non-Conducting Layer on a Conductor 5.8.5 Measuring Thickness of a Conducting Layer on a Conductor

PROBE-CABLE RESONANCE SUMMARY WORKED EXAMPLES

5.11.1 Effect ive Probe Diameter 5.11.2 Character is t ic Paramete r

L CHAPTER 6 - SURFACE PROBE SIGNAL ANALYSIS

6.1 INTRODUCTION 6.2 EDDY CURRENT SIGNAL CHARACTERISTICS

6.2.1 Defect Signal Amplitude 6.2.2 Defect Signal Phase

6.3 EFFECT O F MATERIAL VARIATIONS AND DEFECTS IN A FlNITE THICKNESS

6.4 COIL IMPEDANCE CHANGES WITH DEFECTS

6.4.1 Surface Defect Measurement 6.4.2 Subsurface Defect Measurement

6.5 COIL IMPEDANCE CHANGES WITH OTHER VARIABLES

6.5.1 Ferromagnetic Indications 6.5.2 Electrical Resistivity 6.5.3 Signals from Changes in Sample Surface Geometry

6.6 CALIBRATION DEFECTS 6.7 SUMMARY

PAGE

62 64

67 6 7 67 67 67 6 8

69 72 73

73 74 77 7 8 79

80 8 1 8 2

3 2 82

83 83

83 8 5

8 7 90

90 90

9 1

9 1 93 93

94 9 7

CHAPTER 7 - 'TESTING OF TUBES AND CYLINDRICAL COMPONENTS

PAGE

7.1 INTRODUCTION 7.2 PROBES FOR TUBES AND CYLINDRICAL COMPONENTS

7.2.1 Probe Types 7.2.2 Comparing Differential and Absolute Probes 7.2.3 Directional Properties 7.2.4 Probe Inductance 7.2.5 Probe-Cable Resonance

IMPEDANCE PLANE DIAGRAMS

7.3.1 Solid Cylinders

7.3.1.1 Sensitivity in Cen t r e of a Cylinder

7.3.2 Tubes 7.3.3 Character is t ic Frequency for Tubes 7.3.4 Computer Generated Impedance Diagrams

CHOICE OF TEST FREQUENCY

7.4.1 Test Frequency for Solid Cylinders 7.4.2 Test Frequency for Tubes

PROBES FOR DETECTING CIRCUMFERENTIAL CRACKS SUMMARY WORKED EXAMPLES

7.7.1 Calculate f / fg t o operate a t knee location, for a cylinder 7.7.2 (a) Calculate o p t ~ m u m test frequency for tube inspection

(b) Determine operating point for above frequency (c) Calculate frequency t o discriminate ferromagnetic indications

CHAPTER 8 - TUBE TESTING - SIGNAL ANALYSIS

8.1 INTRODUCTION 8.2 EDDY CURRENT SIGNALS

8.2.1 Defect Signal Characterisit ics 8.2.2 Effect of Test Frequency 8.2.3 Calibration Tubes and Simple Defects 8.2.4 Vectorial Addition and Defects a t Baffle Plates 8.2.5 Tube Inspection at Tubesheets 8.2.6 Testing Tubes with Internal Surface Probes

ANOMALOUS EDDY CURRENT SIGNALS

8.3.1 Ferromagnetic Inclusions and Deposits 8.3.2 Conducting Deposits

PAGE

8.4 MULTIFREQUENCY EDDY CURRENT TESTING

8.4.1 Background 8.4.2 Multif requency Testing of Dented Tubes

8.5 SUMMARY

CHAPTER 9 - METALLURGICAL PROPERTIES AND TESTING FERROMAGNETIC MATERIALS

9.1 INTRODUCTION 9.2 ELECTRICAL CONDUCTIVITY

9.2.1 Factors Affecting Resistivity 9.2.2 Material Sorting by Resistivity

9.3 MAGNETIC PROPERTIES

9.3.1 Magnetic Hysteresis 9.3.2 Magnetic Permeability 9.3.3 Factors Affecting Magnetic Permeability

9.4 TESTING MAGNETIC MATERIALS

9.4.1 Simplified Impedance Diagrams 9.4.2 Impedance Diagrams 9.4.3 Material Sorting by Magnetic Permeability

L 9.4.4 Testing for Defects in Magnetic Materials

9.5 SUMMARY

9.6 WORKED EXAMPLES

9.6.1 Calculate Conductivity 9.6.2 Calculate Magnetic Permeability 9.6.3 Calculate Standard Depth of Penetration

CHAPTER 10 - SUPPORTING INFORMATION

10.1 NOMENCLATURE 10.2 DEFINITIONS 10.3 ABBREVIATIONS FOR NONDESTRUCTIVE TESTING TERMS 10.4 REFERENCES 10.5 INDEX

CANADIAN GENERAL STANDARDS BOARD ADVANCED MANUAL FOR:

EDDY CURRENT TEST METHOD

CHAPTER 1 - SCOPE AND INTRODUCTION

1.1 SCOPE

This manual covers t h e principles of t h e eddy cur ren t method of nondestructive tes t ing including relevant e lect romagnet ic theory, instrumentation, test ing techniques and signal analysis.

I t is intended t o act as;

- a source of educational mater ia l t o r>ersons who a r e seekine certificatiorl according t o t h e requirements of 4 8 1 ~ ~ - 1 3 ~ - ~ e r t i f i c a t i G n of ond destructive Testing Personnel (Eddy Current Method), and

- as a guide and reference t e x t fo r educational organizations and training c e n t r e s t h a t a r e providing o r planning courses of instruction in Eddy Current Testing.

Note t h a t t h e degree of scientif ic deta i l in th is manual i s primarily d i rec ted towards Level I1 and Level 111 cert if ication applicants. I t is assumed t h a t t h e reader is familiar with basic e lect r ica l theory and t h e e lements of algebra and calculus. Many suitable textbooks and reference manuals a r e available to Level I applicants t o b e used prior to o r in conjunction with this publication. The handbook referenced in itern 5 of

L section 10.4 is part icularly suited to t h e requirements f o r Level I expertise.

1 .% EDDY CURRENT TESTING

Eddy cur ren t tes t ing (ET) is a nondestructive test technique based on inducing e lect r ica l cur ren t s in t h e material being inspected and observing t h e in teract ion between those cur ren t s and t h e material . Eddy cur ren t s a r e generated by e lect romagnet ic coils in t h e test probe, and monitored simultaneously by measuring probe e lect r ica l impedance. Since it's an electrornagnetic induction process, d i r e c t e lect r ica l con tac t with t h e sample is no t required; however, t h e sample mater ia l has t o b e conductive.

Eddy cur ren t test ing is a versati le technique. It's mainly used for thin materials; in th ick materials, penetration const ra ints l imit t h e inspected volume t o thin su r face layers. In addition t o flaw inspection, ET can be used t o indirectly measure mechanical and metallurgical character is t ics which cor re la te with e lec t r i ca l and magnet ic properties. Also, geomet r ic e f f e c t s such as thickness, curvature and probe- to-material spacing influence eddy cur ren t f low and can be measured.

T h e large number of potential ly significant variables in ET is both a s t rength and a weakness of t h e technique since effects of o therwise t r iv ia l parameters c a n mask important information o r b e misinterpreted. Virtually everything tha t a f f e c t s eddy

current flow or otherwise influence probe impedance has t o b e taken into account t o obtain reliable results. Thus, credible eddy current testing requires a high level of operator training and awareness. I

HISTORICAL PERSPECTIVE OF EDDY CURRENT TESTING

Electromagnetic testing - t h e interaction of magnetic fields with circulating electrical curents - had i t s origin in 1831 when M. Faraday discovered electromagnetic induction. He induced current flow in a secondary coil by switching a battery on and off. D.E. Hughes performed the first recorded eddy current t e s t in 1879. He was able t o distinguish between different metals by noting a change in excitation frequency resulting frorn e f fec ts of test material resistivity and magnetic permeability.

Initially, the extreme sensitivity t o rnany material properties and conditions made ET difficult and unreliable. Figure 1.1 illustrates this point. it took until 1926 before the first eddy current instrument was developed t o measure sample thickness. By the end of World War I1 further research and improved electronics made industrial inspection possible, and many practical instruments were developed. A major breakthrough came in the 1950% when Forster developed instruments with impedance plane signal displays. These made i t possible t o discriminate between different parameters, though the procedure was still empirical. During the 1960% progress in theoretical and practical uses of eddy current testing advanced the technology frorn an empirical a r t t o an accepted engineering discipline.

During tha t time, other nondestructive test techniques such a s ultrasonics and radiography became well established and eddy current testing played a secondary role, mainly in t he aircraf t industry. Recent requirements - particularly for heat e exchanger tube inspection in t h e nuclear industry - have constributed significantly t o further development of ET as a fast, accurate and reproducible nondestructive t e s t technique.

Until recently, eddy current testing was a technology where the basic principles were known only t o researchers, and a l'black box" approach t o inspection was of ten fol- lowed. The authors' objective in compiling this manual is t o draw upon research, laboratory and industrial inspection experience t o bridge tha t gap and thereby permit t he full potential of eddy current testing t o be realized.

ORGAMZATION OF MANUAL

The presentation moves from theory (including a review of basic electrical concepts) t o tes t methods and signal analysis. Simplified derivations of probe response t o tes t parameters a r e presented t o develop a basic understanding of eddy current tes t principles. Thus, eddy current signals a r e consistently illustrated on impedance plane diagrams (the display used in modern eddy current t e s t instruments) and t o aid explanation, t he parameter "eddy current phase lagu is introduced.

Since probes play a key role in eddy current testing, technical aspects of probe design a r e introduced with pertinent electrical impedance calculations. While knowledge of basic electrical circuits is required for a complete understanding of eddy current t e s t

principles, a good technical base for inspection can sti l l be obtained if sections of this manual requiring such a background a re skipped.

From an applications point of view, the material in this manual provides an inspector with the necessary background t o decide: 1) what probe(s) t o use, 2) what tes t frequencies a r e suitable, 3) what calibration defects or standards a r e required for signal calibration and/or

simulation, 4) what tes t s a r e required t o differentiate between significant signals and false

indications, 5) how t o est imate depth of real defects.

To supplement theory, practical examples a r e presented t o develop proficiency in performing inspections, and t o illustrate how basic principles a re applied t o diagnose real signals.

I t rnay be useful t o review sections 10.1, 10.2 and 10.3 before continuing and t o refer t o these sections a s necessary while reading this manual.

FIGURE 1.1 Misinterpreted Signals

CHAPTER 2 - EDDY CURRENT FUNDAMENTALS

L 2.1 BASIC EQUIPMENT

Basic eddy current test equipment consists of an alternating current source (oscillator), a probe containing a coil connected t o t he current source, and a voltmeter which measures t he voltage change across t h e coil, a s shown in Figure 2.1.

OSCILLATOR VOLTMETER

cuRREM PROBE

PROBE / - MOVEMENT

CRACK

I I t : I " i T PLATE Y

FIGURE 2.1 Eddy Current Test Equipment

The oscillator must be capable of generating a t ime varying (usually sinusoidal) current at frequencies ranging from about 1 kHz (1000 cycles per second) t o about 2 MHz (2,000,000 cycles per second). Oscillators which operate at higher or lower frequencies, o r with pulsed currents, a r e used for specialized applications.

The coil within the probe is an insulated copper wire wound onto a suitable form. The wire diameter, t h e number of turns and coil dimensions a r e all variables which must b e determined in order t o obtain the desired inspection results. Coil variables a re discussed in la ter chapters.

Depending upon t h e type of inspection, a n eddy current probe can consist of a single test coil, an excitation coil with a separate receive (sensing) coil, or an excitation coil with a Hall-effect sensing detector, as shown in Figure 2.2.

VOLTMETER

I OSCILLATOR

C O I L

VOLTMETER

TEST A R T I C L E

/ E X C I T A T I O N

C O I L

-0- OSCILLATOR

, SENSING co 1 L

VOLTMETER

OSCILLATOR

--

TEST A R T I C L E I E X C I T A T I O N

co 1 L

( A ) SELF- INIIUCTANCE (B) SEND-RECEIVE (C ) MAGNETIC REACTION

FIGURE 2.2 Eddy Current Inspection Systems

The voltmeter measures charges in voltage across t h e coil which result from changes in the electrical conditions and properties of t he conducting material tes ted and/or changes in relative position between the coil and the material tested. This voltage change consists of an amplitude variation and a phase variation relative t o the current passing through the coil. The reason for amplitude and phase changes in this voltage is discussed in Chapter 3.

GENERATION OF EDDY CURRENTS

Introduction

In this section the topic of the magnetic field surrounding a coil carrying current is introduced together with the mechanism by which eddy currents a r e induced and how they a r e measured.

Magnetic Field Around a Coil

Oersted discovered tha t whenever there is an electr ic current, a magnetic field exists. Consider e lectr ic current directed along a wire, a magnetic field is created in such a direction t h a t if your right-hand thumb points in t he direction of current, your curled fingers point in t h e direction of t he magnetic field. This is t h e "right-hand rule".

Associated with a magnetic field is magnetic flux density. I t has the same direction as the magnetic field and i ts magnitude depends upon position and current. I t is therefore a field vector quantity and is given t h e symbol 8. Its units in the SI system is t he tesla (T) o r webers per square met re ( u b / rZ ) . The B-field distribution around a long straight wire is shown in Figure 2.3(a). In Figure 2.3(b) the B-field in t h e axial direction of a single turn is shown as a function of radius. As more windings a re added, each carrying the same current, the flux density rapidly increases and i t s associated distribution is altered.

(a) Straight Wire (b) Single T u r n Coil C m & Flowing into page

FIGURE 2.3 Magnetic Flux Distribution

Flux density varies linearly with electr ic current in t he coil, i.e., if coil current doubles, flux density doubles everywhere. The total magnetic f lux,@p , contained within t h e loop is the product of B and t h e a rea of t he coil. The unit in t h e 51 system for magnetic flux is t h e weber (Wb).

2.2.3 Equations Governing Generation of Eddy Currents

In any electrical circuit, current flow is governed by Ohm's Law and is equal t o the driving (primary circuit) voltage divided by primary circuit impedance.

The eddy current coil is part of the primary circuit. The current passing through the coil normally varies sinusoidally with time and is given by:

where , I , is the peak current value in the circuit and w (omega) is the frequency in radiansls ( w equals2nfwhen f is frequency in hertz).

From Oersted's discovery, a magnetic flux ( exists around a coil carrying current (see Figure 2.4) proportional t o t h e number of turns in the coil ( N p and the current ( Ip) .

PROBE

( p r i m a r y circuit)

FIGURE 2.4 Coil Carrying Alternating Current Adjacent to a Test Sample

Faraday's Law states tha t a voltage (V,) is created or induced in a region of space when there is a changing magnetic field. When we apply this t o our coil,

d@ 2 is t he ra te of change in 4 with time. where d t

P Since coil current varies sinusoidally with time, total magnetic flux in t h e coil also varies sinusoidally,

+P = @, s i n ( u t )

where 4, is t h e magnetic flux corresponding t o I ,

The induced voltage a s described by equation 2.4 results in

which also varies periodically with time. If we bring t h e coil close t o a test sample, Ohm's Law s ta tes tha t if there is a driving voltage ( V , ) and the sample's impedance is finite, current will flow,

where I, is current flowing through the sample, V , is induced voltage and 2, is t h e sample's impedance or opposition t o t he flow of current.

These induced currents a r e known as eddy currents because of their circulatory paths. According t o Lenz's Law, they, in turn, generate their own magnetic field which opposes t h e primary field,

and

where 6~ is t he equilibrium magnetic flux surrounding the coil in t he presence of a t e s t sample.

The flow of eddy currents results in resistive (Ohmic) losses and a decrease in magnetic flux. This is reflected as a decrease in probe impedance. In equation form,

and

Equation 2.9 indicates a coil's impedance is a function of the magnetic field surrounding i t and in turn the magnetic field is governed by induced current in the specimen (equations 2.8 and 2.7). The relations between probe impedance and sample properties will be derived in Chapter 3.

To summarize, flux is s e t up by passing alternating current through the test coil. When this coil is brought close t o a conductive sample, eddy currents a r e induced. In addition, t he magnetic flux associated with the eddy currents oppose t h e coil's magnetic flux, thereby decreasing ne t flux. This results in a change in coil impedance and voltage drop. It is t h e opposition between the primary (coil) and secondary (eddy

current) fields tha t provides t he basis for extracting information during eddy current testing.

It should be noted tha t if a sample is ferromagnetic, equation 2.9 sti l l applies but t h e J

magnetic flux is strengthened despite opposing eddy current effects. The high magnetic permeability of ferromagnetic materials distinguishes them from non- ferromagnetic materials and strongly influences eddy current tes t parameters.

Ferromagnetic specimen inspection is discussed in Chapter 9 and unless specified t h e rest of t he manual is restricted t o non-ferromagnetic materials.

2.3 FUNDAMENTAL PROPERTIES OF EDDY CURRENT FLOW

Eddy currents a r e closed loops of induced current circulating in planes perpendicular t o the magnetic flux. They normally travel parallel t o t he coil's winding and parallel t o the surface. Eddy current flow is limited t o t he a r ea of t h e inducing magnetic field.

Test frequency determines depth of penetration into t h e specimen; a s frequency is increased, penetration decreases and the eddy current distribution becomes denser near t he specimen's surface. Test frequency also a f fec ts t he sensitivity t o changes in material properties and defects.

Figure 2.5(a) shows t h e algebraic relationships and Figure 2.5(b) t h e oscilloscope display of eddy current and magnetic field distribution with depth into the specimen. Both the eddy currents and magnetic flux ge t weaker with depth because of "skin effect". In addition t o this attenuation, t he eddy currents lag in phase with depth. Eddy currentst phase lag is t h e key parameter t h a t makes eddy current testing a useful NDT method. The parameters skin depth and phase lag a r e discussed in t he next section.

FIGURE 2.5 Eddy Current and Magnetic Flux Distribution With Depth Into a Conductor

SKIN EFFECT

Eddy currents induced by a changing magnetic field concentrate near t he surface adjacent t o the excitation coil. The depth of penetration decreases with tes t frequency and is a function of electrical conductivity and magnetic permeability of the specimen. This phenomenon is known as t h e skin e f f ec t and is analogous t o t he situation in terrestrial heat conduction where daily surface temperature fluctuations a r e not appreciable below t h e earth's surface. Skin e f f ec t arises as follows: the eddy currents flowing in t he t e s t object at any depth produce magnetic fields which oppose t h e primary field, thus reducing net magnetic flux and causing a decrease in current flow as depth increases. Alternatively, eddy currents near t h e surface can be viewed a s shielding the coil's magnetic field thereby weakening the magnetic field at greater depths and reducing induced currents.

The equation for flow of induced currents is

where J is current density, u is conductivity, IJ is magnetic permeability and ~2 is a differential operator of second order.

For a semi-inf inite (thick) conductor the solution t o t he above equation is

where J,/ J, is t h e rat io of eddy current density J , at depth x t o t h e surface density Jo , and e = 2.71 8 is the base of natural logarithms. B is given by x/ 6

where 6 = ( ~ f l J 0 ) -112 s

t he standard depth of penetration (see next section).

Equation 2.12(a) can be separated into two components:

which describes t he exponential decrease in eddy current density with depth, and

denoting the increasing t ime or phase lag of t h e sinusoidal signal with depth.

2.4.1 Standard Depth of Penetration

Figure 2.6 illustrates t he change in eddy current density in a semi-infinite conductor. - Eddy current density decreases exponentially with depth.

FIGURE 2.6 Eddy Current and Magnetic Flux Distribution With Depth in a Thick Plate

The depth at which eddy current density has decreased t o l/e o r 36.8% of t h e surface density is called t h e standard depth of penetration. The word 'standard' denotes plane wave electromagnetic field excitation within the tes t sample (conditions which a r e rarely achieved in practice). The standard depth of penetration is given by

, inches

where F is electrical resistivity in microhm-centimetres, f is test frequency in hertz, and ur is relative magnetic permeability (dimensionless)*.

* See Chapter 9 for a description of electrical and magnetic properties. IJ , ' V A , incremental permeability, a t zero biassing magnetization flux.

The skin depth equation is strictly t rue only for infinitely thick material and planar magnetic fields. Using t h e standard depth, 6 , calculated from t h e above equation makes i t a material/test parameter rather than a t rue measure of penetration.

2.4.2 Depth of Penetration in Finite Thickness Samples

Sensitivity t o defects depends on eddy current density at defect location. Although eddy currents penetrate deeper than one standard depth of penetration they decrease rapidly with depth. At two standard depths of penetration ( 2 6 ) , eddy current density has decreased t o ( 1 / c ) or 13.5% of t he surface density. At three depths ( 3 6 ) the eddy current density is down t o only 5% of t he surface density. However, one should keep in mind these values only apply t o thick samples (thickness, t > 5 6 ) and planar magnetic excitation fields. Planar field conditions require large diameter probes (diameter > 10 t ) in plate testing or long coils (length > 5 t ) in tube testing. Real test coils will rarely meet these requirements since they would possess low defect sensitivity. For thin plate or tube samples, current density drops off less than calculated from equation 2.12(b) as shown in Figure 2.7. For solid cylinders the overriding factor is a decrease t o zero at the centre resulting from geometry effects as shown in Fig. 2.7(c) and discussed in Section 7.3.1.

One should also note, tha t the magnetic flux is a t tenuated across t he sample, but not completely. Although t h e currents a r e restricted t o flow wihtin specimen boundaries, t he magnetic field extends into the air space beyond. This allows the inspection of multi-layer components separated by a n air space.

0 . 2 4 .8 .B 1 . 0 1. - t

( a ) PLATE (LARGE COIL . 0 > 101)

--- - EQUATION 2 I 2 ( b ) ACTUAL

0 .2 4 .6 .8 1 . 0 I - r 0

( c ) ROD (ENCIRCLING COIL . t > S r p )

1, = EDDY CURRENT OENSITY AT SURFACE

J , OR J , = EDDY CURRENT DENS l T I AT LOCAT l ON r OR r BELOW THE SURFACE

r t

PLATE GEOYETRY 1 1

TUBE U(O ROO GEOMETRY

( r , = O FOR ROO)

HGURE 2.7 Eddy Current Distribution With Depth in Various Samples

The sensitivity t o a subsurface defec t depends on the eddy current density at tha t depth, i t is therefore important t o know the effect ive depth of penetration. The effective depth of penetration is arbitrarily defined as the depth a t which eddy current density decreases t o 5% of t h e surface density. For large probes and thick d

samples, this depth is about th ree standard depths of penetration. Unfortunately, for most components and practical probe sizes, this depth will be less than 3 S t h e eddy currents being at tenuated more than predicted by the skin depth equation. The ef fec t of probe diameter on t h e decrease in eddy current density o r defect sensitivity with depth is discussed in Section 5.3.1.

2.4.3 Standard Phase Lag

The signal produced by a flaw depends on both amplitude and phase of the currents being obstructed. A small surface defect and large internal defect can have a similar e f fec t on t h e magnitude of test coil impedance. However, because of the increasing phase lag with depth, there will be a characteristic difference in t he test coil impedance vector. This e f fec t allows location and extent of a defect to b e determined.

Phase lag is derived from equation 2.12(c) for infinitely thick material. I t represents a phase angle lag of x / Gradians between the sinusoidal eddy currents at t h e surface and those below t h e surface. i t is denoted by the symbol 8 (beta) and is given by:

r a d i a n s

or B = x / b x 5 7 degrees (2.14b)

where x is distance below the surface in t h e same units a s 6 . 90 s i n ( w t )

5 7 " 1 1 4O 1 1

I - p (DEGREES) I

X p = -g x 5 7 , DEGREES

FIGURE 2.8 Eddy Current Phase Lag Variation With Depth in Thick Samples

When x is equal t o one standard depth of penetration, phase lag is 57" or one radian.

L This means tha t the eddy current flowing below the surface, a t one standard depth of penetration, lag the surface currents by 57'. At two standard depths of penetration they lag the surface currents by 114.O. This is illustrated in Figure 2.8.

2.4.4 Phase Lag in Finite Thickness Samples

For this samples, eddy current phase decreases slightly less rapidly with depth than stated above. See Figure 2.9(a), (b) and (c) for the plots of phase lag with depth for a plate, tube, and cylinder, respectively. The phase lag illustrated in these plots does not change significantly with coil diameter or length. For thick samples and practical probe sizes, equation 2.14 is sufficiently accurate.

( a ) FLATE

( c ) ROO

PLATE GEOMETRY I 1 - TUBE AN0 ROO ( I , = 0 ) GEOMETRY

= PHASE WITH OEPTH I, OR r ,RELATIVE Pa 'Pr 10 SURFACE CURRENT

- ACTUAL CURVES --- CALCULATED. E O U A T l O N 2 . 1 4 ( b )

FIGURE 2.9 Eddy Current Phase Lag in Various Samples

Phase lag can be visualized as a shift in t ime of t he sinusoidally varying current flowing below t h e surface. This was illustrated in Figure 2.5. Phase lag plays a key role in the analysis of eddy current test signals. I t will be used throughout t h e manual .- t o link theory and observations. I t should not be misinterpreted or confused with t h e phase angle between voltage and current in AC theory. Both the voltage and current (and magnetic field) have this phase shift or lag with depth.

SUMMARY

Eddy current testing is based on inducing electrical currents in t h e material being inspected and observing the interaction between these currents and the material.

This process occurs a s follows: When a periodically varying magnetic field intersects an electrical conductor, eddy currents a r e induced according t o Faraday's and Ohm's Laws. The induced current (known a s eddy currents because of their circulatory paths) generate their own magnetic field which opposes t he excitation field. The equilibrium field is reduced resulting in a change of coil impedance. By monitoring coil impedance, t he electrical, magnetic and geometrical properties of the sample can be measured. Eddy currents a r e closed loops of induced current circulating in planes perpendicular t o t he magnetic flux. They normally travel parallel t o t h e coil's winding and parallel t o t h e surface. Eddy current flow is limited t o t he a r ea of the inducing magnetic field.

Depth of penetration into a material depends on i t s electrical resistivity, magnetic permeability and on tes t frequency. The basic equation of ET is the standard depth of penetration given by

where p is electrical resistivity, microhm-centimetres; f is tes t frequency, hertz;

and IJ, is relative magnetic permeability , dimensionless.

It s ta tes tha t in thick materials eddy current density decreases t o 37% of t he surface density a t a depth of one standard depth of penetration. In most eddy current tests, especially with surface probes, t he actual eddy current density (at a depth equal t o tha t calculated by equation 2.13a) is much less than 37%.

Eddy currents also lag in phase with depth into the material. Phase lag depends on t h e same material properties tha t govern depth of penetration and is given by

B = x / 6 = x , r a d i a n s (2.14a) 50-

where x is distance below surface, mm.

Phase lag is t h e parameter tha t makes i t possible t o determine the depth of a defect. I t also allows discrimination between defect signals and false indications. It is the key parameter in eddy current testing.

2.6 WORKED EXAMPLES

2.6.1 Standard Depth of Penetration and Phase Lag

PRoBLEM:(a) Calculate t h e standard depth of penetration in a thick 304 sst sample, a t a test frequency of 100 kHz.

(b) Determine the eddy current phase lag at a depth of 1.5 mm in 304 sst at 100 kHz.

SOLUTION: 304 sst properties: P = 72 microhm - centimetres

and M r - 1

(a) from equation 2.13(a),

Therefore t h e standard depth of penetration is 1.3 mm.

(b) from equation 2.14(b),

8 = x / b x 5 7

= - x 5 7 - 6 4 degrees 1.3

Therefore the phase lag is 64 degrees.

CHAPTER 3 - ELECTRICAL CIRCUITS AND PROBE IMPEDANCE

3.1 INTRODUCTION

Eddy current testing consists of monitoring t h e flow and distribution of eddy currents in tes t material. This is achieved indirectly by monitoring probe impedance during a test. An understanding of impedance and associated electrical quantities is therefore imperative for a fundamental appreciation of eddy current behaviour.

The f i rs t two sections review the electrical quantities important in eddy current testing. This is followed by presentation of a model of a test coil coupled t o test material and i t s equivalent electrical circuit. The equivalent circuit approach permits derivation of simplified impedance diagrams t o show t h e e f f ec t of test and material parameters on probe impedance in graphical form. Once the simple impedance diagram concepts of this chapter a r e understood, t h e more complex diagrams of subsequent chapters should present l i t t le difficulty.

3.2 IMPEDANCE EQUATIONS AND DEFWITIONS

All information about a sample is obtained through changes in electrical characteristics of the coil/sample combination. Therefore a basic understanding of electrical quantities is important for eddy current inspection.

RESISTANCE: (symbol: R, units: ohm, R)

Opposition t o t he flow of electrical current is called resistance. It is constant for both direct and alternating current. The electrical - component is called a resistor.

V = IR Ohm's Law (3.1)

where, V is voltage drop across resistor (volt), and I is current through resistor (ampere)

INDUCTANCE: (symbol: L, units: henry, H)

The property of an electr ic circuit by virtue of which a varying current induces an electromotive force in tha t circuit (self) or in a neighbouring circuit (mutual) is called inductance. The electrical component is called an inductor.

L = total flux linkages current through coil

where, N is number of coil turns

O p is magnetic flux (weber)

I is current (ampere)

kl is a geometric factor 2

A is coil's planar surface a rea (arm )

a is coil's axial length (mm)

The self-inductance of a coil is proportional t o coil windings squared ( N~ ) and planar surface a rea (A), and inversely proportional t o coil length ( I t ) . INDUCTIVE REACTANCE: (symbol: XL, units: ohm, nj

Opposition t o changes in alternating current flow through a coil is called inductive reactance.

X L = W L (3.4a)

or *L

= 2 ~ f L (3.4b)

where, f is frequency of alternating current (hertz), and w is angular frequency (radiansfsecond)

CAPACITIVE REACTANCE (symbol: X c , units: ohm, R)

Opposition t o changes in alternating voltage across a capacitor is called capacitive reactance.

Eddy current coil capacitive reactance is normally negligible. However, capacitance can be important when considering impedance of probe cables (Section 5.9 and 7.2.5).

where, C is capacitance (farad)

IMPEDANCE: (symbol: 2, units: ohm, !J)

The total opposition t o alternating current flow is called IMPEDANCE. For a coil,

121 (3.6)

and

where 12 1 is magnitude of 2, and is phase of Z (described in next section).

SINUSOIDS, PHAX)RS AND ELECTRICAL CIRCUITS

In a direct current (DC) circuit, such as a bat tery and light bulb, current and voltage .4

a r e described completely by their respective magnitudes, Figure 3.l(a). Analysis of alternating current (AC) circuits is more complex. Since voltage and current amplitude vary with time, t he phase (or t ime delay) relationship between them must also be taken into account. A typical AC circuit, an inductor in series with a resistor, is presented in Figure 3.l(b). This is a simplified model of a probe assembly: t he inductor is t h e reactive part of t h e assembly (coil) while t he resistor models both coil wire and cable resistance. Figure 3 . lk) shows voltage across t he inductor (VL) leads the current (I) by 90°, while voltage across t h e resistor (VR) is in phase with current. Since t h e current is common t o both inductor and resistor, i t is possible t o use current as a point of reference. Hence, we deduce t h e voltage across t h e inductor leads t h e voltage across t he resistor by 90'.

If one measures t h e voltage drop, V T , across both t h e inductor and resistor, we find VT leads current ( o r V R ) by an angle less than 90°, a s shown in Figure 3.l(d).

To evaluate t h e to ta l voltage VT , we add vectorially t he separate voltages V R and vL,

where j is a mathematical operator (rotates a vector CCW by 90°)

or VT I R e i n ( w t + O ) + j IuL s i n ( w t + ~ / 2 ) (3.9b) 4

Representing voltage waveforms as in Figure 3.l(d) or equation 3.9(b) can be both t ime consuming and confusing. A simpler voltage representation is available by means of phasor diagrams. In phasor diagrams t h e voltage is represented by i t s peak value (amplitude) and phase shift ( 0 ) relative t o the current. The two terms in equation 3.9(b) both contain an amplitude and phase shift so they can be represented by phasors. The f i rs t term's amplitude is IR and i t s phase shift is 0. The amplitude of the second te rm is I w L and i t s phase shift is / 2 or 90'. Each phasor can be represented by an arrow starting at the origin. The phasorts amplitude is indicated by t h e length of t h e arrow OP and the phase shift by the direction of t he arrow, see Figure 3.l(e). Phasors a r e displayed graphically with the resistive component ( V R ) , having a phase shift 8 = 0 , along the horizontal axis. As e increases the phasor rotates counter-clockwise. The reactive cornponent ( V ) , having a phase shift 8 - 90°, will be represented along t h e vertical axis.

Current is common t o both voltage components and since V I E , t he voltage graph of Figure 3.l(e) can be converted t o an impedance graph display, a s in Figure 3.1 (f). If this approach is applied t o eddy current testing, i t is found tha t any changes in resistance or inductive reactance will cause a change in the position of the end of t he vector (point P) which represents t he total impedance vector.

DIRECT CURRENT

1' V = 1R "R

I CURRENT AND VOLTLBE C U I I

( 4 ) DIRECT CURRENT CIRCUIT

ALTERNATING CURRENT

CURRENT MUST BE DESCRIBED BY

AMPLITUDE AND PHASE

( b ) ALTERNATING CURRENT C I R C U I T

I NOUC 1 I VE REACTANC

WF,

VL LEADS I r v 90' P

FIGURE 3.1 Representation of D i m t Current and Altanating Current Circuit Parameters

To obtain t h e reactive and resistive components from this graph requires knowledge of trigonometry.

Reactive component: X L = W L = 121 sin e (3.10)

Resistive component: R = 121 c o s 8 (3.11)

Amplitude of impedance: I Z I = dlt2 + xL2

Phase angle: €3 = Arctan XL/R (3.7)

Note t he x axis component represents pure resistance (phase shift = 0°) while t he y axis component represents pure inductive reactance (phase shift = + 90°). In these calculations i t is assumed coil capaci tance is negligible.

3.4 MODEL OF PROBE IN PRESENCE OF TEST MATERIAL

The test probe contains a coil which when placed on or close t o a t e s t sample can be considered as t h e primary winding of a transformer. The field c rea ted by alternating current in t h e coil induces eddy currents in t h e test sample which acts as a single turn secondary winding, Ng = 1 Figure 3.2(b). Eddy currents align to produce a magnetic field which tends t o weaken t h e surrounding ne t magnetic flux 4 p * according to Lenz's Law. #E

,-------a - - I SECONDARY ' V RECEIVE COIL s

FIGURE 3.2 Model of a Coil with Test Object

There a r e two methods of sensing changes in t h e secondary current, Is .The "impedance method" of eddy current testing consists of monitoring the voltage drop across the primary coil ( v p ' I pZp ) The impedance is altered by t h e load of t h e secondary of t h e transformer. Therefore, changes in secondary resistance, Rs , or inductance L , can be measured a s changes in v

P

The "send-receive1' method of eddy current testing uses two coils. Eddy current flow in t h e sample is altered by defects and these variations a r e detected by monitoring the voltage across a secondary receive coil, see Figure 3.2(c).

3.5 SIMPLIFIED IMPEDANCE DIAGRAMS

3.5.1 Derivation of Probe Impedance for Probe/Sample Combination

We now consider how changes in t h e test sample a f fec t coil impedance on the impedance graph display.

From the previous section the probe and t e s t sample can be modelled a s a transformer with a multi-turn primary (coil) and single turn secondary (sample), Figure 3.3(a). This circuit can be simplified t o an equivalent circuit where the secondary circuit load is reflected as a resistive load in parallel with t he coil's inductive reactance, Figure 3.3(b). This circuit is an approximate model of a real coil adjacent t o a conductor. I t is assumed t h a t all of t h e magnetic flux from the primary coil links t he t e s t sample; the coupling is perfect (100%). I t is also assumed that there is no skin depth attenuation or phase lag across t h e sample thickness.

( c ) EQUIVALENT SERIES CIRCUIT

FIGURE 3.3 Equivalent Circuits

The equivalent circuit concept can be used t o obtain simplified impedance diagrams applicable t o eddy current testing. These diagrams serve as an introduction t o t h e more detailed diagrams which include variations caused by t h e skin effect. The coilfsample circuit model can be transformed into the simpler series circuit by t h e following mathematical manipulations. The load resistance R , can be transfered from the secondary back t o t he primary winding be multiplying i t by t h e turns ratio squared, ( N / N 1 , Figure 3.3(b).

The total impedance of this parallel circuit can be evaluated and transformed into an equivalent series circuit a s follows:

where z1 = N ~ R ,

and 22 ' ~ X O ,

where Xo- w L o , c o i l inductive reactance in air.

~ N ~ R x Therefore I

8 0

P n 2 n P + j x o p

which transforms t o

This can be viewed as a series combination, in t he primary circuit, of resistance RL and inductive reactance X p or

The series circuit in Figure 3.3(c) is therefore fully equivalent t o t he parallel one of Figure 3.3(b). R p can be considered a s coil wire and cable resistance while Z p - R L + j x p is t he total impedance of the probefsample combination.

When the probe is far from t h e sample (probe in air), R, is very large and by substituting R , into equation 3.12a results in

R L - 0 , X p - X o and Z p m X 0

The above results can be obtained by removing component N ~ R , from

L Figure 3.3(b), since R a m OD implies an open circuit.

One last transformation in t h e equation is required before impedance graphs can be obtained. Equation 3.12(a) can be simplified by setting

Co - XoG

2 where G - 1 /N is equivalent circuit conductance.

Substitution in 3.12(a) yields

Normalizirrg with respect t o X o , the coil's inductive reactance when far removed from the sample (coil in air) results in

'L By varying C,, in equation 3.13, from 0 t o infinity t h e impedance curve of Figure 3.4 is obtained. The impedance locus is t h a t of a semi-circle with center a t X / x ,= % and R L / X o - 0 ; i t s radius is 112. With t h e help of equation 3.13 and Figure 3.4, impedance changes can be related t o changes in t h e sample characteristics.

NORMAL l ZED R E S I STANCE

FIGURE 3.4 Impedance Graph Display

3.5.2 Correlation Between Coil Impedance and Sample Properties

The e f f e c t of test parameter variations on probe impedance can be derived from equation 3.13. Each paramater is substituted in turn into C o-X ,/N;R,; if a n increase in t h e parameter results in a n increase in C,, t h e operatlng point (position on impedance diagram) moves DOWN t h e impedance curve, if Co decrease, t h e operating point moves UP t h e impedance curve. These correlations a r e useful in obtaining a quali tat ive appreciation of t h e e f fec t of t h e various test parameters. I t is also useful t o know t h a t probe/sample e f fec t s can b e derived from t h e simple equivalent parallel circuit where t h e sample is t r ea ted as a resistor in parallel with a n inductor (coil). The complete e f fec t can then b e obtained by adding t h e e f f e c t of 'phase lag', which will be t r ea ted in l a te r chapters. Study of equation 3.13 reveals t h e following:

I. An increase in R E results in a decrease in Co.. Therefore a n increase in resistance t o eddy current flow moves t h e operatmg point, P, UP t h e impedance curve (along t h e semi-circle), see Figure 3.5(a).

where, P is electrical resistivity, 1 is eddy current flow distance and A is cross- sectional a r e a t o current flow.

Therefore, p - c o n s t a n t x R E

An increase in e lect r ica l resistivity will move t h e operating point UP t h e impedance curve. The opposite is t r u e fo r a n increase in electrical conductivity. See Figure 3.5(a).

3. For thin wall tubes o r plates of thickness t,

and for constant probe o r tube diameter, D, and coil width, w,

An increase in t u b e wall (or p la te thickness) will move t h e operating point DOWN t h e impedance curve, see Figure 3.5(b).

2 4. C o - u L o / N p R s = c o n s t a n t x U

fo r constant sample properties.

An increase in test frequency will move t h e operating point DOWN t h e impedance curve, see Figure 3.5k).

5. Lo - c o n s t a n t x D ; probe inductance increases proportional t o probe or tube diameter squared.

Also R , - pn D / t w - c o n s t a n t x D , for constant thickness, t, and coil width, w. Substituting Lo and R into C o - w L ~ / N ~ R , results in Co-cons t a n t x D. An increase in probe diameter or tube drameter will move t h e operating point DOWN t he impedance curve, see Figure 3.5(d).

6. In t h e equivalent circuit derivation perfect coupling was assumed for sake of simplification. However, i t can be shown tha t when mutual coupling between coil and sample is decreased, t he impedance point t races smaller semi-circles a s C increases from 0 t o infinity, see Figure 3.5(e).

DSURFACE @ PROBE .L,

0 0 . 5 ( d l

DECREASING F I L L FACTOR OR INCREASING L I F T - O F F f!L ( e ) /X (

FIGURE 3.5 Simplified Impedance Diagrams

SUMMARY

The impedance method of eddy current testing consists of monitoring the voltage d

drop across a test coil. The impedance has resistive and inductive components; t h e impedance magnitude is calculated from the equation

1 , ohms

and t h e impedance phase is calculated from

wL 0 = Arctan - R degrees (3.7)

The voltage across t he tes t coil is V = IZ where I is t he current through t h e coil and Z is t he impedance.

A sample's resistance t o t he flow of eddy currents is reflected a s a resistive load and is equivalent t o a resistance in parallel t o t h e coil inductive reactance. This load results in a resistive and inductive impedance change in t h e test coil. Coil impedance can be displayed on normalized impedance diagrams. These a r e two-dimensional plots with the inductive reactance displayed on the vertical axis and resistance on t h e horizontal axis as in Figure 3.6.

NORMAL l ZEO l NDUCTANCE REACTANCE

R L NORMAL1 ZED R E S I STANCE, - 4 7

FIGURE 3.6 Impedance Graph Display

With this display we can analyze the e f f ec t of sample and test parameters on coil impedance. The equivalent circuit derivation of coil impedance is useful for a qualitative understanding of the e f fec t of various t e s t parameters. I t is valid only for non-ferromagnetic material and for t h e condition of no skin depth attenuation o r phase lag across t he sample. (Ferromagnetic materials will be covered in Section 9.4).

Note tha t a l l t es t parameters result in a semicircle display as they increase o r decrease. A resistance increase t o t h e eddy current flow or increase of sample's electrical resistivity moves the operating point UP t h e impedance diagram, i.e., increase in coil inductance and a change in coil resistance.

An increase in a sample's electrical conductivity, thickness o r tube diameter, moves the operating point DOWN the impedance curve. An increase in tes t frequency or probe diameter also moves the operating point DOWN t h e impedance curve. Although not shown in the above figure, a decrease in fill-factor or increase in lift-off results in a decrease in semicircle radius and smaller change in coil impedance.

In some test requirements it is advantageous t o operate at specific locations on the impedance diagram. By choosing the appropriate t e s t parameters this is usually possible.

3.7 WORKED EXAMPLES

Probe Impedance in Air

PROBLEM: An eddy current tes t is carried out at a test frequency of 50 kHz. Coil resistance is 15 ohms while i t s inductance is 60 microhenries. a) What is the inductive reactance of the test coil? b) What is the impedance of the test coil? c) What is the angle, fl , between t h e total impedance vector and the

resistance vector?

SOLUTION:

a) XL - 2 nfL - (2 n ) ( S O x lo3) (60 x ) XL - 18.8 ohms

= A r c t a n 18 8 A r c t a n - = A r c t a n 15

8 = 51.4 degrees

Probe Impedance Adjacent t o Sample

PROBLEM: An eddy current test is carried out on brass using a surface probe at 4

50 kHz. Coil resistance in air is 15 ohms and i t s inductance in air is 60 microhenries. Probe impedance with the probe on t h e brass sample is measured as z p = 24.5 ohms and 0 = 35 degrees.

Calculate: a) X , inductive reactance

and b) ItL, resistive load

SOLUTION: a) X = Z s i n e P P

= 2 4 . 5 sin 3 5 ' - 1 4 . 1 ohms b) RL = Z c o s e -

P R~~

= 2 4 . 5 c o a 35' - 1 5 . 0 = 5 . 1 ohms

3.7.3 Voltage - Current Relationship

PROBLEM: For t h e above probe impedance problem calculate voltage drop across the probe if tes t current is 100 milliamperes.

SOLUTION: Probe impedance 1 2 1 = 24.5 ohms

Ohm's Law states tha t V = 11'2 I

theref ore, v = (0.10) (24.5) = 2.45 volts.

Voltage across t he probe is 2.45 volts.

CHAPTER 4 - INSTRUMENTArnN

4.1 INTRODUCTION

All t h e information about a test part is t r ansmi t t ed t o t h e test coil through t h e magnet ic field surrounding it. The impedance eddy current method monitors voltage drop across t h e primary coil, V p I p Z p ; as coil impedance changes so will t h e voltage across t h e coil if current remains reasonably constant. The send-receive eddy cur ren t method monitors voltage developed across a sensing coil (or Hall e f f e c t de tec to r ) placed close t o t h e excitat ion coil, see Figure 2.2.

In most inspections, probe impedance (or voltage) changes only slightly as t h e probe passes a defec t , typically less than 1%. This small change is difficult t o d e t e c t by measuring absolute impedance or voltage. Special ins t ruments have been developed incorporating various methods of detect ing and amplifying small impedance changes.

The main functions of a n eddy current ins t rument a r e i l lustrated in the block diagram of Figure 4.1. A sine wave oscillator generates sinusoidal current , at a specified frequency, t h a t passes through t h e test coils. Since t h e impedance of two coils is never exact ly equal, balancing is required t o e l iminate t h e voltage di f ference between them. Most eddy current ins t ruments achieve th is through an AC bridge or by subtracting a voltage equal t o t h e unbalance voltage. In general they can to le ra te a n impedance mismatch of 5%. Once balanced, t h e presence of a defec t in t h e vicinity of one coil c rea tes a small unbalanced signal which is then amplified.

TRANSFORMER

r---- Q 1 D s C a METER

FIGURE 4.1 Block Diagram of Eddy Curtent Instrument

Since the sinusoidal unbalance voltage signal is too difficult and inefficient t o analyse, i t is converted t o a direct current (DC) signal retaining the amplitude and phase characteristics of the A C signal. This is normally achieved by resolving the AC .& signal into quadrature components and then rectifying them while retaining the appropriate polarity. In general purpose instruments, these signals a r e normally displayed on X-Y monitors. Simpler instruments, such a s crack detectors, however, have a meter t o display only the change in voltage amplitude. To decrease electrical instrument noise, filtering is used at the signal output; however, this decreases t h e frequency response and thereby restricts t h e inspection speed.

The most troublesome parameter in eddy current testing is lift-off (probe-to- specimen spacing). A small change in lift-off c rea tes a large output signal. The various methods used t o decrease this e f f ec t a r e discussed in t h e individual sections on specific eddy current instruments.

BRIDGE CIRCUITS

Most eddy current instruments use an AC bridge t o sense the slight impedance changes between the coils or between a single coil and reference impedance. In this section t h e important characteristics of bridge balance a r e discussed.

4.2.1 Simple Bridge Circuit

A common bridge circuit is shown in general form in Figure 4.2, t he a rms being indicated a s impedance of unspecified sorts. The detector is represented by a voltmeter. Balance is secured by adjustments of one or more of the bridge arms. Balance is indicated by zero response of t he detector, which means tha t points A and C a re a t t he same potential (have t h e same instantaneous voltage). Current will flow u through the detector (voltmeter) if points A and C on the bridge arms a r e a t different voltage levels (there is a difference in voltage drop from B t o A and B t o C). Current may flow in ei ther direction, depending on whether A or C is a t higher potential.

FIGURE 4.2 Common Bridge Circuit

If t h e bridge is made up of four impedance arms, having inductive reac tance and resist ive components, t h e voltage f rom R t o A must equal t h e voltage f rom B t o C in both amplitude and phase fo r t h e bridge t o b e balanced.

At balance,

and 11Z3 - I ~ Z ~

from which t h e following relationship is obtained:

Equation 4.1 states t h a t t h e ra t io of impedances of one pair of adjacent a rms must equal t h e ra t io of impedances of t h e o ther pair of adjacent a r m s fo r bridge balance.

If t h e bridge was made up of four res is tance arms, bridge balance would occur if t h e magnitude of t h e resistors satisfies equation 4.1 (with 2 1 replaced with R l,etc.). However, if t h e impedance components a r e eddy current probes consisting of both inductive reac tance and resistance, t h e magnitude and phase of t h e impedance vectors must sa t is fy equation 4. I.

In practice, th is implies t h e ra t io of inductive reac tance of one pair of adjacent a r Ins must equal t h e ra t io of inductive r e a c t a n c e of t h e o ther pair of adjacent arms; t h e s a m e being t r u e fo r t h e resistive component of impedance.

Figure 4.2 and equation 4.1 can b e used to i l lus t ra te t h e character is t ics 'figure 8' signal of a differential probe. If

1 G'

Z3 point C is at a higher potential than point A. ' Zqs

This implies t h a t when 1 increases (i.e., coi l moving across a defec t ) with ~2,236 Z4constant, t h e bridge voltage unbalance increases, and t h e opposite happens when Z3 increases. I t is this bridge unbalance charac te r i s t i c t h a t results in a plus-minus o r 'figure 8' signal a s t h e differential probe moves across a localized defect . This signal occurs independent of whether t h e two coils a r e wound in opposition o r in addition.

Typical Bridge Circuit in Eddy Cur ren t Instruments

Figure 4.3 i l lustrates a typical AC bridge used in eddy current instruments. I t is similar t o t h e bridge in Figure 4.2 excep t fo r t w o additional arms. In this bridge t h e probe coils a r e placed in parallel with variable resistors. The balancing, o r rnatching of voltage vector phase and amplitude, is achieved by varying these resistors until a null is achieved. Potent iometer R 2 balances t h e react ive component of t h e coils t o make the phase angle of each coil c i rcui t equal. Potent iometer R balances the resultant voltage with an equal voltage amplitude t o null t h e instantaneous voltage between R 1 and R 2 *

FIGURE 4.3 Typical Bridge Circuit Used in Eddy Current Instruments

The inductive voltage drop across each coil is equalized by controlling the current passing through the coils. This is done by varying potentiometer R2. However, when t h e test coil inductance differs significantly from reference coil inductance, potentiometer 12 will have t o be rotated t o one extremity. This means less current passes through one coil making i t less sensitive than t h e other coil. When this occurs,

4

a distorted (unsymmetrical) signal results if a differential probe is used. In addition, t h e common cable lead carries t h e unbalanced current, resulting in cable noise, especially if the cable is not properly shielded and grounded.

In t he Figure 4.3 circuit, t h e output voltage for large ( > 10% ) off-null (of f-balance) conditions is a nonlinear function of the change in coil impedance. However, for defect detection, close to balance, this discrepancy is small.

Bridge Circuit in Crack Detectors

Portable eddy current instruments a r e of ten used t o inspect for surface defects. A typical crack detector circuit is shown in Figure 4.4. An oscillator supplies an alternating current t o an AC Bridge, containing a single eddy current probe coil a s one arm of the bridge. A capacitor is connected in parallel with t he coil so t h e L-C (inductance-capacitance) circuit is near resonance. When t h e coil is placed on a tes t sample, the bridge is unbalanced and the pointer on t h e meter swings off-scale. The bridge can be balanced by adjusting potentiometer R .

FIGURE 4.4 Simplified Circuit of Crack Detector

4.3 RESONANCE CIRCUIT AND EQUATIONS L

Probe-cable resonance must be considered when operating at high tes t frequencies and/or using long probe cables. In addition, crack detectors a r e purposely designed t o operate close t o resonance. This section contains basic information about resonant (tuned) circuits.

If a capacitor is connected in parallel with t he test coil (inductor), there is a unique frequency at which t h e inductance-capacitance (L-C) circuit resonates. At this frequency the circuit is said t o be tuned. Under this condition the output voltage, for a given rneasurement, is maximum. A capacitor in parallel with the eddy current probe converts t he circuit of Figure 3.3(c) t o t ha t of Figure 4.5.

FIGURE 4.5 Parallel LC Circuit

At resonance,

hence 2 - - when R - 0

If resistance, R, i s negligible compared to Xp and X c resonance occurs when

Since w = 2 n f , resonant frequency is

where L is coil inductance in henries and C is cab le capac i t ance in farads.

When resistance, R, is significant,

X \ /

where Q . - , q u a l i t y f a c t o r . R

The resonant frequency of a pract ica l parallel resonant c i rcui t ( R f 0) is t h e frequency at which t h e react ive power of t h e inductance and capac i t ance a r e equal, o r t h e to ta l impedance appears as pure resistance.

EDDY CURRENT INSTRUMENTS

General instrument functions were described using t h e block diagram of Figure 4.1. In th is sect ion specif ic instruments a r e covered. I t answers t h e questions: What i s t h e test frequency? How is lift-off compensated for? How i s balancing achieved? What t y p e of outputs do they have?

General Purpose Instrument (Impedance Method)

Figure 4.6 shows a typical eddy cur ren t ins t rument with various control functions. FREQUENCY control sets t h e desired t e s t frequency. Frequency is se lected by continuous control o r in d iscre te s t eps f rom about 1 kHz t o 2 MHz. The coils' impedances a r e normally balanced using a n AC bridge circuit . These bridges require two coils on adjacent bridge a r m s such as a r m s No. 2 and No. 4 in Figure 4.3. Coil impedance must b e compatible with in ternal bridge impedance.

CARBON STEEL

HONEL

SST TYPE 304

LEAD

BRASS ALUMINIUM

\ COPPER STORAGE MONITOR

OUTPUT 0 FREQUENCY SELECTO(I

PROBE CONNECTOR

FIGURE 4.6 Typical Eddy Current Instrument With Storage Mcmitur

Most bridges can tolerate a coil impedance between 10 and 200 ohms. The BALANCING controls, labelled X and R in some instruments, a r e potentiometers R~ and ~2 in Figure 4.3. They match coil impedance t o achieve a null when the probe is in a defect f ree location on the t e s t sample. Some instruments have automatic balancing.

The bridge output signal amplitude is controlled by t h e GAIN control. In some instruments i t is labelled a s SENSITIVITY. It controls the amplifier of t he bridge output signal, shown in Figure 4.1, and therefore does not a f fec t current going through the probe. However, some instruments control amplification by varying current through the coils. This is undesirable because i t could cause coil heating, and when testing ferromagnetic materials t h e magnetization level changes, resulting in signal distortion and non-repeatable signals.

Following amplification of the bridge unbalance signal, the signal is converted t o direct current signals. Since t h e AC signal has both amplitude and phase i t is converted into QUADRATURE X and Y components. The quadrature components of the bridge output a r e generated by samplin the sinusoidal signal at two positions 90" apart (one-quarter wave) on t h e waveform f or by using electronic multipliers). The DC voltage values (amplitudes) constitute t h e X and Y quadrature components. If phase is taken relative t o t h e resistive voltage component, then the X quadrature component i s R ~ (orvR) and the Y component, xL(orVL), in equation 3.12(b) or Figure 3.4. We now have an efficient way of analyzing bridge unbalance signals.

Eddy current instruments do not have a phase reference. To compensate for this, they have a PHASE SHIFT control (phase-discrimination control). Normal impedance diagram orientation with inductive reactance displayed vertically (+ Y) and resistive w

horizontally (+ X) can be obtained experimentally. This is achieved by adjusting t h e PHASE control until t he signal from a probe approaching a fer r i te sample (high IJ and very high P ) displays a vertical (+ Y) signal indicating an increase in probe inductive reactance, s e e Section 5.5.6 for examples. PHASE control can also be used t o minimize t h e e f f ec t of extraneous signals such as lift-off. The X-Y signal pattern is rotated until t h e lift-off signal is horizontal (X). Thus any vertical (Y) channel signal indicates defects, thickness variations, etc., with l i t t le e f fec t from probe wobble.

The output signal is normally fi l tered internally t o decrease instrument or system noise. This decreases frequency response of t h e instrument and reduces t h e maximum inspection speed; at faster inspection speeds signal distortion results. Instruments can have a frequency response of 30 t o 1000 Hz, although 100 t o 300 Hz is most common. At 300 Hz, t h e maximum attainable tube inspection speed, t o de tec t a n abrupt defect without signal distortion, is about 0.25 m/s.

Signals a r e commonly displayed on X-Y storage monitors with the X component on t h e horizontal axis and the Y component on t h e vertical axis. The writing speed or frequency response is greater than 1 kHz for a s torage CRT.

Analysis of recorded signals is normally done visually. The storage monitor display in Figure 4.6 shows t h e change in coil impedance as a surface probe was placed on various test samples illustrating the e f fec ts of resistivity, permeability and lift-off.

In t he "impedance" method of eddy current testing, t h e flow of eddy currents is monitored by observing t h e e f fec t of their associated electromagnetic fields on the 4 electrical impedance of the inspection coil(& This impedance includes coil wire and cable resistance.

Coil wire and cable resistance increase linearly with temperature according t o

where a is temperature coefficient of resistance

and AT is change in temperature.

If the probe and/or cable experience a change in temperature during inspection, t he output signal from the eddy current instrument changes; this is normally referred to a s temperature drift.

4.4.2 Crack Detectors

A typical crack detector circuit was shown in Figure 4.4. Crack detector probes contain only one coil, with a fixed value capacitor in parallel with the coil t o form a resonant circuit. At this condition t h e output voltage, for a given change in coil impedance, is maximum. The coil's inductive reactance, XL , must be close to the capacitive reactance, xC . In most crack detectors this is in t he range of 20 t o 100 ohms.

Crack detectors tha t operate at o r close t o resonance do not have selectable test frequencies. Crack detectors for non-ferromagnetic, high electrical resistivity materials such as Type 304 stainless s teel typically operate between 1 and 3 MHz; those for low resistivity materials (aluminum alloys, brasses) operate at lower frequency, normally in t he 10 t o 100 kHz range. Some crack detectors for high resistivity materials can also be used t o inspect ferromagnetic materials, such as carbon steel, for surface defects. Normally a different probe is required; however, coil impedance and test frequency change very little.

METER

OU TPU T

W I T H L I F T - O F F = 0 . 1 mm

PROBE W I T H L I F T - O F F = O mn

/ SAMPLE W I T H DEFECT

f O S C I L L A T O R FREQUENCY, - f r

FIGURE 4.7 Meter Output with Varying Oxillator Frequency

Crack detectors have a meter output and three basic controls: balance, lift-off, and sensitivity. BALANCING control is performed by adjusting t h e potentiometer on the adjacent bridge arm, until bridge output is zero (or close t o zero). GAIN control (sensitivity) adjustment occurs at t h e bridge output. The signal is then rectified and displayed on a METER. Because the signal is filtered, in addition t o t he mechnical inertia of the pointer, the frequency response of a meter is very low (less than 10 Hz). LIFT-OFF CONTROL adjusts t h e test frequency (by less than 25%) t o operate slightly off resonance. In crack detectors the t e s t frequency is chosen t o minimize the e f fec t of probe wobble (lift-off), not t o change the skin depth or phase lag. The set-up t o compensate for probe wobble can be described with the help of Figure 4.7. Frequency is adjusted by trial-and-error t o obtain t h e same output signal on the meter with the probe touching the sample and at some specified lift-off (normally 0.1 mm). At this frequency a deep surface defec t will give a different reading on the meter, a s shown in Figure 4.7.

However, the meter output is a complex function of signal phase and amplitude, and cannot be used t o reliably measure depth of real defects. Nor can they be used t o distinguish between real and false indications such a s ferromagnetic inclusions.

Material Sorting and Conductivity Instruments

Material sorting, or conductivity instruments, have a precalibrated meter output and - have a unique way of compensating for lift-off. Instruments for sorting of high resistivity materials (Type 304 stainless s teel) use a fixed, high test frequency normally between 200 and 500 kHz, and those for low resistivity materials (aluminum alloys), a low test frequency, between 20 and 100 kHz. They incorporate AC bridges and normally have two coils (one as reference). Coil impedance is in t h e range of 20 t o 100 ohms. They either have bridge balancing o r a zeroing control, t o keep the signal on scale. GAIN CONTROL or sensitivity adjustment occurs at t h e bridge output. The signal is then rectified and displayed on a METER.

LIFT-OFF compensation is normally pre-set. Figure 4.8 explains how t h e probe- wobble (lift-off) signal is eliminated. The bridge is purposely unbalanced (by pre-set internal adjustment)" such tha t t h e unbalance point, P, is at the cent re of curvature of t h e lift-off impedance locus, AB. The instrument meter reads a voltage proportional t o t he distance, PB' or PAt, from t h e chosen unbalance point t o the impedance curves. The amplitude of this voltage remains constant with probe wobble but changes significantly for wall thickness (and resistivity) variations. In f a c t any signal tha t t races an impedance locus different from lift-off will change meter output.

PRESET UNBALANCE

A I R

C

V E S l STANCE

FIGURE 4.8 Unbalanced Bridge Method Showing Selection of Operating Point

* This is achieved by subtracting a signal equal t o OP from t h e signal OA.

With this type of instrument only t h e magnitude of t he impedance change is measured. This instrument is effective for conductivity and wall thickness measurement (and deep defects) and is simple to operate. I t has only two basic controls: balance and sensitivity.

4.5 SEND-RECEIVE EDDY CURRENT SYSTEMS

The "send-receive" eddy current method eliminates t he temperature drift sensed by general purpose instruments. The flow of eddy currents is monitored by observing the e f fec t of their associated electromagnetic fields on the voltage induced in an independant coil(s), Figure 4.9. The excitation o r primary coil is driven with a sinusoidal current with constant peak-to-peak amplitude t o obtain a constant magnetomotive force,

E X C l T A T I O N C O I L 7 r R E C E I V E C O I L S

R L A R G E

T E S T A R T I C L E

FIGURE 4.9 Send-Receive Circuit

This makes t h e excitation ma net ic flux @ independent of primary coil resistance. The secondary or receive coil f s) is c o n n e d d t o a high input impedance amplifier, hence t h e induced voltagev, is not affected by receive coil resistance. 4

The wire resistance of both t h e excitation and receive coils can change, because of temperature, without affecting the output signals; temperature dr i f t has thus been eliminated. Temperature independence makes this method useful for measuring resistivity, wall thickness and spacing between components. It has no significant advantage over t he impedance method for defec t detection, except in t h e through- wall transmission system discussed in Section 5.4.

Hall-Effect Detector

Most send-receive circuits consist of one excitation (or driver) coil and one or more receive (or pick-up) coils.

However, the induced magnetic flux 0, can be measured with a Hall-ef f ec t detector rather than by monitoring the induced voltage V, across a pick-up coil, s ee Figure 2.2b and 2 .2~ .

FIGURE 5.10 Hall Detector Circuit

The induced voltage in a pick-up coil is proportional t o t he t ime ra te of change of the magnetic flux and theref ore is proportional t o t h e test frequency,

The Hall de tec to r instead responds to t h e instantaneous magnitude of t h e magnet ic flux, @,

This means t h e output voltage is independent of test frequency, making it useful f o r low frequency inspection (especially if t h e d e t e c t o r h a s t o b e small).

The Hall de tec to r works as follows: When d i rec t current is passed through a Hall e lement , voltage (electric potential) is produced, perpendicular t o current flow, see Figure 4.10. This voltage is proportional t o t h e component of magnetic flux perpendicular to t h e e lement and t h e e lement su r face area. This voltage is NOT f rom a change in e lement resistance. Hall e lements as smal l as 1 rnm square a r e com mercially available.

4.5.2 Send-Receive Coils and Lif t-Of f Compensat ion

General purpose "send-receive" instruments a r e similar to "impedancew instruments, as described in Section 4.4.1. The main di f ference is t h e method of balancing because of t h e di f ferent coil configuration. Most send-receive c i rcui ts consist of one excitat ion coil and two receive coils positioned symmetrically inside o r outside t h e exci ta t ion coil. They can e i ther b e differential where both coils sense t h e test specimen o r absolute where only one coil senses t h e test specimen, as shown in Figure 4.9. Although coil impedance is not important in send-receive instruments, t h e induced voltage is a function of number of windings and test frequency. Therefore thei r inductive reac tance tend t o b e similar t o coils used in impedance instruments.

The sensing coils a r e wound in opposition s o t h e exci ta t ion field induces no n e t voltage in t h e receive coils when they both sense t h e s a m e material. In t h e presence of a defec t , t h e voltage changes as e a c h coil moves over it. Figure 4.9 i l lus t ra tes a su r face ref lect ion type probe where both exci ta t ion and pick-up coils a r e on t h e s a m e side of t h e test sample. However, t h e exci ta t ion coil and pick-up coils c a n b e placed on opposite sides of t h e sample; t h e method is r e fe r red t o as through-wall transmission. The two methods a r e compared in Section 5.4.

The output signals in most send-receive ins t ruments a r e t h e quadrature components of t h e secondary voltage. However, in s o m e special purpose instruments, o n e output signal is proportional t o amplitude and t h e o ther t o phase of t h e secondary voltage (relat ive t o primary voltage). They compensate f o r LIFT-OFF as follows: if coil-to- sample spacing varies the re is a large change in amplitude of t h e secondary voltage but l i t t l e change in phase. The phase shi f t between t h e secondary and primary sinusoidal voltages is measured at a voltage level V, slightly larger than zero, Figure 4.11. A t th is voltage t h e sinusoidal voltages have t h e s a m e phase sh i f t fo r ze ro lift-off as for maximum (perhaps 0.1 mm) lift-off. The voltage discriminator in these phase-shif t measuring eddy cur ren t ins t ruments t r igger on t h e V, voltage point, and therefore, t h e output signal fo r lift-off between 0 and 0.1 mm is minimized. Measurement of resistivity, wall thickness o r deep d e f e c t s can be made without lift- off noise.

PROBE SIGNAL, L I FT - O f f = 0

PROBE SIGNAL, L I F T - O F F = 0 . i a m

PROBE SIGNAL, DEFECT I N TEST ARTICLE

t

FIGURE 4.11 Secondary Voltage Waveform f o r Various Test Conditions

4.6 MULTIFREQUENCY EQUIPMENT

The eddy current NDT method is sensit ive to many test parameters , making i t very versatile. However, o n e is usually only in teres ted in a single parameter such as defects. Insignificant pa ramete rs such as changes in e lect r ica l o r magnet ic properties, t h e presence of den t s o r support p la tes in t u b e inspection and lift-off in surface probe inspection can mask d e f e c t signals. The multifrequency eddy cur ren t method was developed t o e l iminate t h e e f f e c t of undesirable parameters.

The response t o various anomalies changes wi th test frequency. This allows a means of discriminating agains t unimportant changes. In multifrequency instruments, t w o or more frequencies a r e used simultaneously (through t h e s a m e coil(s)). Coil current consists of two o r more superimposed frequencies, i.e., t h e coil(s) is exci ted with more than one t e s t frequency simultaneously. A three-f requency multif requency instrument acts t h e s a m e way as t h r e e s e p a r a t e (single-frequency) eddy current instruments. Band-pass f i l t e r s s e p a r a t e t h e signals at e a c h frequency. The discrimination o r elimination process is accomplished by combining t h e output signals (DC signals) f rom individual frequencies in a manner similar to simultaneous solution of multiple equations. The elimination of extraneous signals is achieved by matching t h e signal at t w o test frequencies and subtracting. This process is continued f o r o the r unwanted signals using o t h e r test frequencies until t h e final output consists of only t h e d e f e c t signal. A discussion of inspection resul ts with multi-frequency is covered in Section 8.4.

Multifrequency instruments have the same controls and functions as general purpose "impedance" type instruments, described in Section 4.4.1, with the addition of mixing modules. These modules are used t o combine or substract the output signals from each combination of frequencies.

4.7 PULSED EDDY CURRENT EQUIPMENT

Faraday's Law states that eddy currents are induced in a conductor by a varying magnetic field. This magnetic field can be generated by passing sinusoidally varying current through a coil. However, the current can be of other waveforms such as a train of pulses. This method works only on the send-receive principle where the flow of eddy currents is monitored by observing the ef fec t of their associated electromagnetic fields on the induced voltage of the receive coil(s). The voltage pulse is analyzed by observing i ts amplitude with time, Figure 4.12.

To compensate for LIFT-OFF, the voltage is sampled a t a preset time, t 1 When the waveform is triggered (measured) a t t ime t 1, the voltage for zero lift-off and maximum lift-off is the same, whereas the voltage waveform in the presence of a defect is different. This method is quite similar t o the send-receive method described in Section 4.5.3. Therefore, by measuring the voltage a t the appropriate crossing point, lift-off effects can be drastically decreased.

1 DEFECT I N T E S T A R T I C L E

FIGURE 4.12 Voltage Across a Pulsed Eddy Current Pick-Up Cod as a Function of Time

The pulsed eddy current method offers another advantage. The pulsed driving current produces an inherently wideband frequency spectrum, permitting extraction of more selective information than can be determined from the test specimen by a single d

frequency method. IJnfortunately, there is at present no commercially available instrument tha t operates on this principle.

SPECIAL TECHNIQUES

Two old methods used t o measure large coil impedance variations (greater than 5%) a r e t h e ELLIPSE and SLIT methods. These methods analyse t h e AC signal directly on an oscilloscope (without converting i t t o DC). They were mainly used for material sorting. They a r e obsolete methods and a detailed description is not warranted in this manual; a full description is contained in Reference 5.

Another technique, MODULATION ANALYSIS, is also described in Reference 5. It works on t h e same principle as "frequency spectrum analysis" where a discrete frequency component of a waveform can be analysed without interference from lower o r higher frequency noise. The inspection must be performed at constant speed (in fact i t only works if t he re is relative motion between coil and sample). It is used in production-line testing at speeds up t o 2 m/s or higher. It is a very specialized and complicated method and a detailed description is not warranted in this manual.

RECORDING EQUIPMENT

During inspection, ed-dy current instruments and recording equiprnent a r e typically connected as in Figure 4.13. The eddy current signal is monitored on a storage CRT (cathode ray tube) and recorded on X-Y and two-channel recorders. Recording on a n FM tape recorder for subsequent playback is also common. - The important characteristic of these recording instruments is FREQUENCY RESPONSE, or speed response, which limits inspection speed. Section 4.4.1 indicated general eddy current instruments have a frequency response of 100 t o 300 Hz, limiting t h e inspection speed t o 0.25 m/s. To be compatible, recording instruments must have t h e same or higher frequency response.

X-Y STORAGE

MON l TOR

INSTRUMENT

x? Yy I I PROBE

x ; y 2-CHANNEL

CHA#T RECORDER

r

FIGURE 4.13 Block Diagram of Eddy Current Monitoring Equipment

X-Y Recorders

#b

Signal analysis for signal discrimination and defect depth estimation is normally done on X-Y signal patterns. The CRT storage monitors have a frequency response of at least 1 k H z and therefore do not restr ic t maximum inspection speed. However, to obtain a permanent visual record of the signal, i t must b e recorded on X-Y recorders. The fastest recorders have a speed of response of 8 H z for small signals. This drastically limits inspection speed if used on-line. I t is therefore only used in t he laboratory or t o record playback from t ape recorders (this is done by recording at the highest tape speed and playing back at t h e lowest, a factor of 8:l for most tape recorder). One solution t o on-line recording of X-Y signals is t o paragraph the CUT display; however, this is not practical for recording many signals.

1.

Another solution is t o use storage monitors with hard copy (paper output) capability. These exist commercially but require custom-made control units. They have a frequency response of 1 k H z or higher.

X Y FM TAPE

RECORDER

6 o X Y

X-Y RECORER

-

Strip Chart Recorders

o f 0

-

Recording X and Y signal components against t ime is useful in locating defec ts and determining their length.

Common two channel ink-pen s t r ip char t recorders have a speed response of approximately 100 Hz. At maximum inspection speed (0.25 m/s) t h e recorded signal will decrease in amplitude and be slightly distorted. 4

Ink-ejection s t r ip char t recorders have a speed response of 1 kHz. These recorders a r e not readily available in North America and use a lot of paper.

Ultraviolet light recorders have a speed response higher than 1 kHz, but require special paper. These recorders a r e rarely used in eddy current testing.

FM Tape Recorders

Tape recorders allow storage of eddy current signals (on magnetic tape) for subsequent retrieval. They have a frequency response proportional t o recording speed. The lowest recording speed is 24 mm/s (15/16 ips) giving a frequency response of 300 Hz, and t h e fastest, 380 mm/s (15 ips), will respond t o 4.8 kHz.

Frequency Response

Eddy current instruments and recording instrumentation have limited frequency response. This means they require finite t ime to respond to an input signal. Frequency response, sometimes called speed of response, is defined as t h e frequency a t which t h e output signal falls to 0.707 (-3 dB) of t h e maximum input signal.

A test coil with an effect ive sensing width w passing over a localized defec t a t a speed s will sense t he point defect for a duration of w/s seconds. This signal is approximately equal t o one wavelength with a frequency

f - s / w her tz (4.6)

where s is speed in mm/s and w is width in mm.

For example, a t a probe speed of 0.5 m/s and probe sensing width of 2 mm, f = 250 hertz. If t h e instrumentation has a frequency response of 250 hertz, t he output signal is reduced to 0.707 t h e input signal and t h e X-Y signal is distorted. If t h e instrumentation frequency response is 500 her tz , t h e output signal decreases only slightly. For this example, t h e eddy current instrument should have a frequency response equal t o or greater than 500 he r t z t o obtain undistorted signals. Or inversely, if t he instrument frequency response is only 250 hertz, t he maximum i n s p e c t i o ~ speed should be reduces t o 0.25 m/s.

SUMMARY

Basic eddy current equipment consists of an alternating current source (oscillator), voltmeter and probe. When the probe is brought close to a conductor or moved past a defect , the voltage across t he coil changes and this is read off t he voltmeter. The oscillator sets the tes t frequency and the probe governs coupling and sensitivity t o defects.

For effective purchase or use of an eddy current instrument, t h e following information is needed: (a) type of instrument: impedance, send-receive, crack detector, etc. (b) type of outputs: single (meter) or quadrature (X-Y) component outputs (c) t e s t frequency (d) type of lift-off compensation.

Most eddy current instruments use an AC bridge for balancing but use various methods for lift-off compensation. Send-receive instrument should b e used for accurate absolute measurements in t he presence of temperature fluctuations. Multifrequency instruments can be used t o simplify defect signals in the presence of extraneous signals.

Eddy current instruments and recording equipment have a finite frequency response limiting the inspection speed t o normally 0.25 m/s.

Most instruments tolerate probe impedance between 10 and 200 ohms.

Crack detectors operate close t o coil-cable resonance. The resonant tes t frequency is given by

f = 1 / 2 l r f i r (4.4a)

where L is coil inductance in henries and C is cable capacitance in farads. The lift- off signal is minimized by adjusting the frequency (slightly off resonance) until zero and a small probe lift-off gives zero output signal. High test frequencies a r e normally used t o inspect for shallow defects in high resistivity or ferromagnetic materials. Low test frequencies a r e used for detecting deep defects or inspecting g o d conductors. Crack detectors have a meter output, and cannot b e used t o reliably measure defect depth.

4.11 WORKED EXAMPLES

4.1 1.1 Impedance at Resonance

PROBLEM: In a parallel L-C circuit, inductance is 80 x henries, capacitance is 5 x low9 farads and resistance is negligible. Calculate (a) resonant frequency, (b) inductive reactance and (c) capacitive reactance.

SOLUTION:

(b) Inductive Reactance, XL = 2 r f L (3.4b)

X~ 3 = 2 n x 2 5 2 x 1 0 x 80 x = 1 2 6 . 5 ohm.

(=) Capacitive Reactance, Xc l /ZnfC (3.5)

I 1

3 = 1 2 6 . 5 ohms 2 n x 2 5 2 x 1 0 x 5 x

CHAPTER 5 - TESTING WITH SURFACE PROBES

5.1 INTRODUCTION

The goal of this chapter is t o present a practical approach t o eddy current inspections using surface probes. The emphasis is on tes t variables such as test frequency, probe size and type; these a r e normally t h e only variables an inspector has at his control. These selections a r e usually determined by skin depth considerations, defect size, and probe size.

Impedance graphs and the Characteristics Parameter a r e included because they a r e tools tha t an inspector should not be without. A thorough understanding of impedance graphs is essential t o manipulate t e s t conditions t o minimize and/or t o cope with undesirable tes t variables. Erroneous conclusions a r e often made by persons who do not have a working knowledge of impedance graphs.

The scope of t h e approach t o an eddy current inspection can be very broad; a successful outcome usually depends on t h e proper approach. When planning an inspection the f i rs t questions tha t must be answered before proceeding are; For what type of defects is the inspection being conducted? If the expected defects a r e crack, how big a r e they? Do they have directional properties? What is t he minimum acceptable defect size? Does the material have ferromagnetic properties? Other variables will, of course, influence t h e test but these questions must be answered in order t o select an appropriate probe s ize and t e s t frequency.

5.2 SURFACE PROBES L

The eddy current probe plays two important roles: i t induces eddy currents, and senses t h e distortion of their flow caused by defects. Sensitivity t o defects and other variables in t he t e s t article can be affected by probe design. This is achieved by controlling direction of eddy current flow, by controlling t h e coil's magnetic field, and by selecting an appropriate coil size. The e f fec ts of undesirable material variations and/or variations in probe t o tes t ar t ic le coupling (lift-off) can often be decreased by using multiple coils.

A surface probe, as the name implies, is used for inspecting surfaces, f la t o r contoured, for defects or material properties. Defects can be either surface or subsurface. (Surface defects a r e those tha t break through, or originate at t h e surface - typically cracks, voids, or inclusions: a subsurface defect does not break t h e surface and is therefore not visible).

Other names used for variations of surface probe designs a r e pancake probe, f la t probe, spring probe or coil, spinning probe, and pencil probe.

5.2.1 Probe Types

Simple Probes

Surface probe designs can vary frorn a simple, single coil attached t o lead wires, to complex arrangements, a s shown in Figure 5.1. Most eddy current instruments require two similar coils t o satisfy their AC bridge network a s discussed in Chapter 4. If only one coil senses t h e test rnaterial, i t is an absolute probe; if both coils sense t h e test

FIGURE 5.1 Surface Probes

material, i t is a differential probe. The simple probe in Figure 5.l(a) is therefore undesirable because a second coil o r electrical device with similar impedance will be necessary for bridge nulling. An exception would be in t h e use of Crack Detectors; these instruments operate with an internal balancing circuit (see Section 4.2.3).

A bet ter arrangement is shown in the pencil probe of Figure 5.l(b). This probe incorporates a second coil (reference) mounted far enough from the test ar t ic le tha t i t will not be influenced by it. The two coils have t h e same impedance when the probe is balanced in air, but will change relative t o each other when t h e t e s t coil is coupled t o a sample. However, t h e degree of coupling is usually small because of t h e inherent small size of pencil probes s o t h e coils sti l l match well enough f o r most instruments over a reasonable frequency range. The probe shown has fe r r i te cores; ferr i te is used for three reasons:

1. higher inductance frorn a given coil size, 2. small surface a rea in contact with the material, 3. t he coil can be further from the contact surface providing greater wear

protection.

A further improvement in reference coil arrangement is shown in Figure 5.l(c); i t is a t tached t o a disc whose properties a r e similar t o t h e test material. With this arrangement t he relative impedance of the two coils will not be affected by t e s t frequency.

The probe shown in Figure 5.l(d) is a spring loaded type designed t o minimize lift-off. The shoe provides a broad area for squarely positioning the probe on a f la t surface, while the spring maintains probe contact at constant force.

Figure 5.l(e) shows a probe used for inspecting large diarneter tubing. The probe can be rotated and/or moved axially. The design shown incorporates a replaceable wear cap.

Other Probe Designs

A multi-coil array as shown in Figure 5.2(a) is useful for inspecting tubes. This type of probe could detect defects that would not be detected by a conventional circumferential coil (discussed in Section 7.5).

( 8 ) D l FFERENTI AL SURFACE PROBE

YULT 1 SURFACE . CO I L PROBE

F ERROMICNET l C

COYPENSATING

F IELD

GAP PROBE L l FT OFF COYPENSATIN6 PROBE

FIGURE 5.2 Special Surface Probes

A gap probe, Figure 5.2(b), uses ferromagnetic material t o shape the magnetic field. The field is confined by the core causing eddy currents t o flow in circular loops perpendicular to the flux lines.

.4 differential configuration is shown in Figure 5.2k); t he two coils a r e placed side- by-side. Both coils have high sensitivity t o localized variations but tend t o cancel out t h e effect of lift-off, gradual material variations, o r ambient temperature changes.

A lift-off compensating probe is shown in Figure 5.2(d); this probe combines the signals from two coils t o effectively ro ta te t h e defect signal relative t o the lift-off signal. Therefore, even on Vough" surfaces, shallow defects can be detected.

EST ARTICLE

(DRIVER COIL) RECEIVER C O l L

DRIVER C O l L

( b )

TEST ART ICLE

P ICK-UP C O I L S (WOUND OPPOSING EACH OTHER)

TEST ARTICLE

ELECTRICAL CONNECT IONS

FIGURE 5.3 Send-Receive Probes

Send-Receive Probes

Figure 5.3(a) shows a through-transmission probe arrangement. Current flowing in the SEND coil produces a magnetic field, par t of which is transmitted through the tes t article. The field is detected by t h e RECEIVER coil, inducing a voltage. There will be no signal variation from the receiver coil when a defect-free test ar t ic le is moved anywhere between the two coils as long as the coil-to-coil spacing remains constant.

Figure 5.3(b) shows a reflection-type probe arrangement. The probe consists of a large send coil which generates a field, and two small receiver coils wound in opposite directions, as mirror images t o one another, as shown in Figure 5.3k). With

t h e probe in air, ne t output is zero. However, if one end is placed near a test article, t h e field differs a t t he two ends, and a ne t voltage appears across t he two coils.

L 5.2.2 Directional Properties

Eddy currents are closed loops of induces current circulating in a plane perpendicular t o t h e direction of magnetic flux. Their normal direction of t ravel is parallel t o t h e coil winding and parallel t o the surface. See Figure 5.4.Pancake type surface probes a r e therefore insensitive t o poor bonding of coatings and flaws parallel to t h e surface of a sample.

SURFACE CRACK I EDDY CURRENTS

\ I LLYINAR CRACK

*.yiy!Z 1 ' r C C f Z I TEST PLATE . ;

I EOOY WRREWT FLOWS PARALLEL TO COIL IIWDIWGS

POOR S E l S l T l V l l V TO LlYlNAl lOWS

SURFACE CRACK

@

ZERO S E N S l l l V l N LOW S E N S I l I V l W Y A X l Y W SEWS1 T I V l T Y AT CENTRE OF COIL PARALLEL TO WIYOI~CS ACROSS WINDINGS

FIGURE 5.4 D i r e c t i d Properties of a Surface Probe

When testing for flaws such as cracks, i t is essential tha t t he eddy current flow be a t a large angle (preferably perpendicular) t o t he crack t o obtain maximum response. If eddy current flow is parallel t o t h e defect there will be l i t t le or no disruption of currents and hence no coil impedance change.

When testing for flaws parallel t o t h e surface, such a s laminations, a horseshoe shaped probe (a gap probe with a very large gap) may have reasonable sensitivity.

5.2.2.1 Sensitivity at Centre of a Coil

Probe impedance changes with coil diameter, as will be discussed further in Section 5.5. A simplified derivation of this diameter e f f ec t is derived below, for t h e case of no skin depth attenuation o r phase lag and long coils. From Faraday4s Law,

3.5.2 Correlation Between Coil Impedance and Sample Properties

The e f f e c t of test parameter variations on probe impedance can be derived from equation 3.13. Each paramater is substituted in turn into C o-X ,/N;R,; if a n increase in t h e parameter results in a n increase in C,, t h e operatlng point (position on impedance diagram) moves DOWN t h e impedance curve, if Co decrease, t h e operating point moves UP t h e impedance curve. These correlations a r e useful in obtaining a quali tat ive appreciation of t h e e f fec t of t h e various test parameters. I t is also useful t o know t h a t probe/sample e f fec t s can b e derived from t h e simple equivalent parallel circuit where t h e sample is t r ea ted as a resistor in parallel with a n inductor (coil). The complete e f fec t can then b e obtained by adding t h e e f f e c t of 'phase lag', which will be t r ea ted in l a te r chapters. Study of equation 3.13 reveals t h e following:

I. An increase in R E results in a decrease in Co.. Therefore a n increase in resistance t o eddy current flow moves t h e operatmg point, P, UP t h e impedance curve (along t h e semi-circle), see Figure 3.5(a).

where, P is electrical resistivity, 1 is eddy current flow distance and A is cross- sectional a r e a t o current flow.

Therefore, p - c o n s t a n t x R E

An increase in e lect r ica l resistivity will move t h e operating point UP t h e impedance curve. The opposite is t r u e fo r a n increase in electrical conductivity. See Figure 3.5(a).

3. For thin wall tubes o r plates of thickness t,

and for constant probe o r tube diameter, D, and coil width, w,

An increase in t u b e wall (or p la te thickness) will move t h e operating point DOWN t h e impedance curve, see Figure 3.5(b).

2 4. C o - u L o / N p R s = c o n s t a n t x U

fo r constant sample properties.

An increase in test frequency will move t h e operating point DOWN t h e impedance curve, see Figure 3.5k).

Therefore, eddy current flow and i ts associated magnetic flux a r e proportional t o radial distance from the centre of a coil. Hence no current flows in t he cen t re (r = 0) and there is no sensitivity t o defects at the cen t re of a coil.

Probe Inductance I

The factor governing coupling and induced voltage in test material is t he magnetic flux surrounding the coil. The total magnetic flux ( $p ) is proportional t o probe inductance (L) and current (I), i.e., @ a L I . In most eddy current instruments excitation current is kept reasonably eonstant (in t h e milliampere range) but probe inductance could vary by a factor of one thousand. The most important aspect of inductance is that probe impedance, which is a function of inductance, must be compatible with the instrument and signal cable,

X~ z 1 * and B - Arctan - R

where X L = 2 7 f L when f is in hertz, L in henries and R is coil wire resistance in ohms.

- --

TABLE 5.1 SURFACE COIL IMPEDANCE

Do = 1.6 mm Do = 3.2 mm Do = 6.3 mm Do = 12.7 mm Do = 25.4 mm

L = 0.27 MH L = 0.54 p H L = 1.1 VH L = 2.1UH L = 4.3 pH R = 0.2 n R = o . 1 ~ R = 0.05 0 R = 0.02n R = 0.01 R

L N = 21 40 AWG 34 AWG 28 AWG 22 AWG 16 AWG (0.080 mm) (0.16 mm) (0.32 mm) (0.64 mm) (1.3 mm)

L = 1.5 L = 3.0 L = 6.1 L = 12 L = 24 R = l R = 0.5 R = 0.3 R = 0.1 R = 0.06

N = 50 43 AWG 37 AWG 31 AWG 25 AWG 19 AWG (0.056 mm) (0.1 1 mm) (0.23 mm) (0.45 mm) (0.91 mm)

L = 5.8 L = 12 L = 23 L = 47 L = 94 R = 4 R = 2 R = l R = 0.5 R = 0.3

N = 98 46 AWG 40 AWG 34 AWG 28 AWG 22 AWG (0.040 mm) (0.080 mm) (0.16 mm) (0.32 mm) (0.63 mm)

L = 11 L = 23 L = 45 L = 90 L = 180 R = 9 R = 3 R = 2 R = 0.9 R = 0.5

N = 136 48 AWG 41 AWG 36 AWG 29 AWG 23 AWG (0.031 mm) (0.071 mm) (0.13 mm) (0.29 mm) (0.57 mm)

- -

L = 24 L = 49 L = 97 L = 195 L = 390 R = 17 R = 8 R = 4 R = 2 R = 1

N = 200

L 49 AWG 43 AWG 37 AWG 31 AWG 25 AWG (0.028 mm) (0.056 mm) (0.11 mm) (0.23 mm) (0.45 mm)

The self-inductance of a long coil (solonoid) can be calculated from the equation

(5. la)

Lo is self-inductance in henries where vr is relative permeability of core (normally = 1.0)

A is coil's planar surface area, m i 1 1 imc t r c s L is coil length, millimetres.

This formula is a good approximation for coils of lengthidiameter rat io greater than 10.

For a short coil, end ef fec ts will reduce inductance because of lower flux at coil ends. The N~ term remains since N enters in N 4p(total number of flux linkages) and again since 4 itself is proportional t o N. The following approximate equation can be used t o calcurate inductance of short coils:

where 7 is mean coil radius Do + Di

4 * nm

and K = 0.112 (2!L+Do + Di), mm

Most eddy current instruments will operate over a fairly broad range of probe impedance (and probe inductance) without substantial reduction in signal-to-noise rat io and signal amplitude. An instrument input impedance of 100 ohms is typical, although any impedance between 20 and 200 ohtns is generally acceptable, unless tes t frequency is too close t o probe-cable resonance; see Section 5.9. Exact probe inductance calculations a r e therefore not essential. To facili tate impedance calculations, Table 5.1 has been prepared. This table lists coil inductance and resistance (with probe away from test material) for various outside diameters and number of coil turns, keeping both the inside diameter and coil length equal to 0.2 t imes t h e outside diameter. Wire diameter is chosen t o fill available coil cross- sectional space. Using this table and the knowledge tha t inductance,

where N is number of turns of wire and 6 is average coil diameter, one can usually make a reasonable est imate of wire size and number of turns required t o achieve a particular inductance.

( a ) L I F T -OFF OISTANCE (mm)

( b ) SUBSURFACE DEFECT DEPTH (mm)

FIGURE 5.5 Decrease in Sensitivity with

(a) Lift-off (b) Defect Depth

PARAMETERS AFFECTING SENSITIVITY TO DEFECTS

During eddy cur ren t inspection one must b e a w a r e of t h e l imitations of t h e technique J

and should t a k e maximum advantage of its potential. Although sensit ivity to d e e p sur face d e f e c t s is excellent , sensit ivity t o d e e p subsurface d e f e c t s is very poor. A subsurface de fec t only 5 mm f r o m t h e su r face is considered very d e e p f o r eddy cur ren t test purposes.

There a r e t w o fac to rs t h a t contr ibute t o th is limitation. The skin dep th e f f e c t causes eddy currents t o a t t e n u a t e with dep th depending on t h e mater ia l properties and test frequency. This e f f e c t is normally minor and c a n b e controlled (within l imits) by reducing test frequency. The predominant e f f e c t (rarely mentioned) is t h e decrease in magnet ic flux, and consequently eddy cur ren t density, with depth because of t h e small d iamete r of most pract ica l probes. O n e can increase penetra t ion by increasing probe diameter , but as a consequence sensitivity t o shor t d e f e c t s decreases. One could opt imize sensit ivity if d e f e c t length is known; however, t h e maximum depth of detectabi l i t y is s t i l l very small. IJnlike ultrasonic inspection where a d e f e c t is de tec ted many transducer d iameters away, eddy cur ren t tes t ing is limited t o de tec t ing d e f e c t s at a dep th of less than o n e probe diameter . I t is this e f f e c t of probe d iamete r t h a t l imi ts rnost volumetric eddy cur ren t inspection t o mater ia ls less than 5 mrn thick. In following subsections, l imitations a r e discussed and empirical examples presented.

Sensitivity with Lift-off and Defec t Depth

There is a decrease in sensitivity t o d e f e c t s as a coil is moved away f rom t h e surface. This is caused by t h e decrease in magne t ic f lux density wi th distance resulting f rom f ini te probe diameter. Figure 5.5(a) shows t h e e x t e n t of this decrease for t h r e e probes of d i f ferent diameters. Note, for example, t h e sensit ivity of t h e smallest probe (5 mm diameter) decreases a f a c t o r of four when moved about 1 mrn f rom t h e surf ace.

This loss of sensit ivity with dis tance also apply t o d e f e c t s in a solid, in addition t h e r e will be a decrease d u e to skin dep th attenuation.

Figure 5.5(b) i l lus t ra tes t h e decrease in si nal ampl i tude wi th subsurface d e f e c t depth without skin depth a t t enua t ion (solid lines f and with skin depth a t t enua t ion (dashed lines). With large skin depths (low test frequency) t h e decrease in subsurface d e f e c t sensit ivity with dep th is similar to t h e decrease in sensit ivity with distance for surface d e f e c t s shown in Figure 5.5(a). This im l ies magnet ic flux density decreases with dis tance f rom t h e coil in a i r as in a solid /' without skin depth attenuation).

A t a typical t e s t frequency, where o n e skin dep th equals de fec t depth ( 6 = 2 m m for t h e dashed lines in Figure 5.5(b)), a fu r the r decrease , by about a fac to r of th ree , in signal amplitude at x = 2 mm is a t t r ibu ted t o skin depth attenuation. This occurs since at one skin depth eddy cur ren t density is 37% of sur face eddy cur ren t density.

The decrease in d e f e c t sensit ivity with dep th in a f in i t e thickness sample, without skin dep th a t tenuat ion, is approximately t h e s a m e as in a n infinitely thick sample. However, with skin depth a t tenuat ion, d e f e c t sensit ivity decreases less rapidly than t h e dashed lines in Figure 5.5(b); t h e curve would fall somewhere in between t h e dashed and solid lines.

In general, the decrease in defect sensitivity with depth is determined by probe size rather than skin depth attenuation. Since most defects a r e not much longer than

L sample thickness, one cannot use probes with coil diameter much larger than sample thickness (because of loss in sensitivity with defect length, Figure 5.6). Eddy current testing with surface probe is therefore normally limited t o thicknesses less than 5 mm.

5.3.2 Effect of Defect Length

Eddy current flow is limited t o t h e a rea of t h e inducing magnetic field which is a function of coil geometry; defect sensitivity is proportional t o coil diameter in a surface probe, and t o gap width in a horseshoe probe. As a general rule, probe diameter should be equal t o or less than t h e expected defect length. The ef fec t of probe diameter and defect length is shown in Figure 5.6. For example, when defect length equals probe diameter, the signal amplitude ranges from one-third t o two- thirds of the amplitude for an infinitely long crack depending on probe diameter and test frequency.

The sensing area of a probe is t he a r ea under t h e coil plus an extended area due t o the magnetic field spread. The effect ive diameter, Def f . of a probe is approximately equal t o the coil diameter, D, , plus four skin depths,

At high frequencies the 4 6 term will be small and the sensing diameter can be assumed t o be about equal t o coil diameter, but at low t e s t frequencies the rnagnetic field spread can be significant. In this case i t is common t o use ferr i te cups t o contain the field. This results in a concentrated field without affecting depth of penetration.

,- ,, 1 m PROBE O l l U E T t R

1.3 m PROBE O l l Y E T E R

I MHz = 0.36 Rm C r g 8 ~ ~ o ~ ~ z = 1.16 m

C W 0

I I I I I I I 1 I I 0 2 4 6 8 1 0 I 2 14 16 18 20 22

FIGURE 5.6 Effect of Defect Length

COMPARISON BETWEEN SURFACE AND THROUGH-WALL INSPECTION

The major limitation of conventional eddy current methods is lack of penetration. Figure 5.7(a) illustrates typical results obtained with the conventional eddy current method, where the probe is placed on one side only of a 4 mm thick, 100 m m diameter tube. Test frequency is 30 kHz and skin depth, 6 = 1 7 m m Note the drastic decrease in signal amplitude and the significant phase rotation of t h e defect signals with depth. A defect has t o be long and very deep before i t can be picked up from the opposite side of t h e tube wall. This decrease in sensitivity with depth is due t o both finite probe s i t e and skin depth attenuation.

Figure 5.703) illustrates typical results obtained with through-wall transmission equipment where excitation and receive coils were located directly opposite each other across t h e wall. The probes were conventional absolute pancake type surface probes. The output signal appears a s a 'figure 8' because t h e signal was filtered (differentiated).

r

I

O . D . S L I R F A C E 25% GROOVE

~~~c~~~~ > U R F A C k 5 U P F A C E

TUBE \' R O T A T I O N

75% FRO'. O U T S I D E

J U R F P C E S U R F A C E SROOVE

/ I N S I D E S U R F A C E 1 v o i r I\ , GROOVE 1 L

50% 75% I . D . SURFACE GROOVE

A M F L I T U O E O F D E F E C T S IGNAL, Y C O U P O N E N 1

0 D . GROOVE \ 25% 50% I . D . GQOOVE

3 . 3 " 'F DEEP v 0.8 qm D E E P

!3 m m L o r 6 H O L E S . 0.8 rc ~ I A . 13 mm LONG 13 LONG

X - Y DISPLAY OF DEFECT SICNALS

(a) Conventional Surface Probe Results

25% 50% 75% O , D , v I , D ,

GROOVE HOLES GROOVE

A M P L I T U D E OF DEFECT S I G N A L S , Y COMPONENT X-Y DISPLAY OF DEFECT S IGNALS (FILTERED)

(b) Through- Wall Transmission Results

FIGURE 5.7 Comparing Conventional and Through-Wall Transmission Techniques

The Y-amplitude presentation in Figure 5.7(b) shows d e f e c t signal ampl i tude does not change significantly with d e f e c t depth. I t is important t o n o t e t h e phase of t h e signals does not change with d e f e c t dep th when using t h e through-transmission method a s shown in t h e X-Y display.

5.5 IMPEDANCE GRAPH DISPLAY

Impedance graphs a r e a n indispensable a id in eddy current inspections. An understanding of these graphs provides a n opera to r a c lear p ic ture of al l variables and t h e ability for appropr ia te act ion t o minimize e f f e c t s of adverse conditions.

All information abou t t h e test a r t i c le is t ransmi t t ed t o t h e test coil via t h e magnet ic field. The variation of t h e magnet ic flux, 4 , with t i m e induces a voltage, V, across t h e test coil which, by Faraday's Law, depends on t h e magnitude and r a t e of change of $ and on t h e number of turns in t h e coil, N

= - L d I / d t since 4 = LI/N.

The variation in ampl i tude and phase of th i s vol tage vector indicates t h e condition of t h e test art icle. The voltage vector can b e resolved in to t h e two quadratures, in- phase V o , and out-of-phase V g o . Since V = IZ and I i s kep t approximately constant, t h e voltage graph c a n b e replaced with t h e impedance graph, as discussed in Section 3.3.

Impedance depends not only on t es t a r t i c le variables but also on probe parameters. 4

The probe parameters a r e coil d iameter , number of turns, length, and c o r e material . The instrument parameter t h a t a f f e c t s impedance is test frequency (since f a d $ / d t ). To overcome t h e necessity of plotting impedance graphs for e a c h t e s t coil, probe impedance is normalized. The graphs c a n then be used t o study t h e e f f e c t of t e s t a r t i c le variations without dependence of probe details.

The inductive reac tance component is normalized by dividing by t h e product of frequency and coil inductance ( w L o when t h e probe is away f rom test material (in air).

where w is angular frequency, radians/second L is inductance, henries Lo is inductance of coil in a i r , henries XL is reactance, ohms Xo i s r eac tance of coil in a i r , ohms

I A I R

TEST A R T I C L E l NDUCT I VE REACTANCE

ART

u RES l STANC

'I C L E

w L - WLo

I R O C I

( a) BEFORE NORMAL I Z4T I ON (b)

{ A I R I \

AFTER NORMAL I ZAT 1 ON

FIGURE 5.8 Coil Impedance Display

The resistive component is normalized by subtracting coil wire and cable resistance, RDC , a n d t h e n d i v i d i n g by wLo ,

where RL is coil resistive load due t o eddy currents in t e s t material.

The normalized components X / X o a nd R / X a r e dimensionless and independent of both coil inductance and coi 4 wire and c a b 4 e resistance. Changes in t he normalized parameters indicate variations in eddy current flow into the test article only. Figure 5.8 displays probe impedance before and a f t e r normalization. Changes in t he t e s t article a r e reflected by a change in impedance point P. Figures 5.9 t o 5.1 1 a r e normalized coil impedance graphs, produced by computer simulation, showing the change in the point P for the following sample variables: electrical resistivity permeability, and thickness. Figures 5.12 and 5.13 show ef fec ts of test frequency and coil diameter.

'1.12 (COPPER I

LIFT-OFF.

NORCULIZED RESISTANCE

FIGURE 5.9 Impedance Graph-Resistivity

1 .6 ---- CONSTANT - CONSTANT

, 1 . 2 U

p - 700rtl.cm

U

g 1.0 p - 170

w N

p = 53

P p ' 1 .7

PERCIEIBILITY. p ,

R E S I S T I V I T Y , P

FREQUENCY ' 5 0

LIFT-OFF ' 0 THICK PLATE

NORCULILED RESl STANCE

FIGURE 5.10 Effect Impedance Graph-Permeability Effect

0 . 9 \ FREQUENCY, k H z

NORWLIZED RESISTANCE

FIGURE 5.11 Impedance Graph-Thickness Effect

0 . 0 4 0 . 0 0 . 1 0 . 2 0 . 3 0 . 4

NORWLIZED RESISTANCE

FIGURE 5.12 Impedance Graph-Frequency Effect

5.5.1 Effect of Resistivity

Figure 5.9 shows the e f fec t of electrical resistivity for a range of conducting materials. The impedance point moves up t h e curve with increasing resistivity. Impedance points for s tep changes in coil t o test ar t ic le spacing between zero and infinity a r e also included. Note tha t a small increase in spacing (lift-off) produces a large impedance change. This results from decreased magnetic flux coupling t o t h e sample. There would be a relatively larger e f fec t on the impedance of a small coil than on t h e impedance of a large coil for t h e same change in spacing.

5.5.2 Effect of Per meability

Note in Figure 5.10 there is a large impedance increase for a small increase in permeability. Small permeability changes can obscure other tes t variables.

5.5.3 Effect of Thickness

Figure 5.1 1 t races the impedance point path a s sample thickness decreases from infinity t o zero. As tes t material becomes thinner, causing increased resistance t o eddy currents, t he impedance point moves up the curve. This was also the case in t he resistivity graph, Figure 5.9. This implies t ha t any condition causing an increase in resistance t o flow of eddy currents, cracks, thinning, alloying elements, temperature, etc., will basically move the impedance point up t h e curve towards the probe impedance in air, X L / X , = l .

The impedance curve in Figure 5.1 1, from t h e knee down, makes a reversal swirl a s t he probe moves across a conductor with increasing thickness. This is due t o skin depth and phase lag effects which overshadow a l l basic movements of the impedance point.

5.5.4 Effect of Frequency

Figure 5.12 shows the e f fec t of test frequency (an instrument parameter). As frequency is increased, eddy currents a r e sampling a thinner layer close t o the surface (skin depth effect , discussed in Chapter 2). When frequency is decreased eddy currents penetrate deeper into the material and the impedance point moves up the curve.

Towards the upper end of the curve, impedance is mainly composed of resistance which has a grea t dependency on temperature, both in t h e test article and in coil wire resistance (although the l a t t e r does not appear on this normalized curve). I t is therefore desirable, when possible, t o operate near t h e knee of the curve say, 20 t o 200 kHz in this example.

5.5.5 Effect of Probe Diameter

Figure 5.13 shows e f f ec t of coil diameter (a probe parameter). Note increasing coil diameter moves t h e impedance point down t h e curve, similar t o increasing frequency. When tes t conditions dictate use of a low frequency, t he operating point can of ten be brought down the curve t o t he desired knee region by increasing coil diameter (provided test conditions will permit a large probe).

rnm

LIFT-OFF 1 1 11 1- 7

Frequency = 50 kliz

FIGURE 5.13 Impedance Graph-Surface Coil Diameter Effect

5.5.6 Comparison of Experimental and Computer Impedance Diaprams

The impedance graphs shown in Figure 5.9 t o 5.12, produced by computer simulation, can be verified using a standard eddy current instrument. Figure 5.14 shows probe response t o various test variables: resistivity, permeability, lift-off, and test frequency. The sold lines a r e output voltage t races generated by varying probe-to- test ar t ic le spacing (lift-off) from infinity t o contact with various conducting samples, while keeping test frequency constant at 10 kHz, and again at 100 kHz. The dashed lines, connecting the points when the probe was in contact with the samples, were sketched in t o show t h e similarity between these graphs and the normalized impedance graphs in the preceding section. Note t h a t the points move down the curve with increasing conductivity and also with increased frequency. For example, t h e operating point for 304 sst moved from t h e top of t he impedance diagram at 10 kHz t o near t he knee a t 100 kHz.

l N O U C T l VE REACTANCE

R - RESISTANCE

w L 0

( a )

A I R

INDUCT l VE REACTANCE

BRASS

C u

SST

f = I00 kHz

R - RESISTANCE

wL0

( b )

FIGURE 5.14 Probe Response to Variws Test Parameters at Two Frequencies

5.6 CHARACTERISTIC PARAMETER

In Section 5.5 impedance graphs were normalized to make test article parameters independent of probe properties such as inductance. Another method, proposed by W.E. Deeds, C.V. Dodd and co-workers, combines frequency and probe diameter with test material parameters, to form one characteristic parameter (2). -

- where r is

w is ur is

and a is

mean coil radius angular frequency relative magnetic permeability ( ~ 1 . 0 for nonmagnetic materials)

d

e lec t tical conductivity.

Using this characteristic parameter, one impedance graph can be plotted to describe four tes t parameters with P, a s t h e only variable.

1 .o

0 . 9 L I FT -OFF CONSTA

7 = COIL MEAN RAOlU 0 . 8

0.7 PERMEABILITY

0 . 6 u = ELECTRICAL I ! I \ I / I CONDUCT l V l TY I

0 0 0 5 0 . 1 0 0 . 1 5 0 . 2 0 0 . 2 5 0 .30 0.35

NORMAL l ZED RES l STANCE

FIGURE 5.15 Impedance Diagram with Characteristic Parameter, PC

Consider Figure 5.15. The solid lines a r e generated by starting with P C equal t o zero and increasing the value t o infinity (while holding coil t o tes t ar t ic le spacing constant). The dashed lines a r e generated by starting with the coil infinitely far away from the tes t article and bringing the coil closer until i t contacts (while holding P C constant). Note the similarity between these curves and the impedance graphs shown in preceding sections (the horizontal scale is twice the vertical scale).

The usefulness of t h e character is t ic pa ramete r is t h a t it provides a modelling parameter. Conditions of similari ty a r e m e t when

Tes t 1 Tes t 2

S P E C I M E N 0-0

STORAGE O S C l LLOSCOPE

0 l SPLAY

NOMENCLATURE

V VOLTAGE \ R s I - CURRENT

w - ANGULAR FREQUENCY ( W = 2 w f )

L o PROBE INDUCTANCE I N A I R

R,, - PROBE W l R E L CABLE OC R E S I S T A N C E

R, - SRCIMEN AC RSISTANCE

SUBSCRI PTS : T TOTAL L - INDUCTANCE R . RESISTANCE P - PRIMARY S - SECONDARY

FICURE 5.16 Coil Impedance/Voltage Display

Tes t conditions with t h e s a m e P, value have t h e s a m e operat ing point on t h e normalized impedance graph. If, for example, test a r t i c le resist ivity measurements were required (for checking consistency of alloying e lements fo r instance), t h e bes t accuracy would b e achieved by operat ing near t h e knee of t h e curve where t h e r e is good discrimination agains t lift-off. (Equation 5.4 does n o t include skin dep th e f fec t s , which may b e a n overriding consideration).

To o p e r a t e a t t h e knee position in Figure 5.15 a probe diameter and frequency combination a r e se lected such t h a t P c. 5 10. T h e value of P in equation 5.4 is given in SI units; we c a n use t h e followmg version using more familiar units.

- where r is t h e mean radius, mrn

f is frequency, H z p is e lect r ica l resistivity , micohm-cent i m e t r e

( v, 1 for nonferromagnetic material)

I t should b e noted t h a t t h e charac te r i s t i c pa ramete r , PC, must b e used in conjunction with Figure 5.15 (obtained analytically); i t cannot b e used to obtain Figure 5.15.

DEFINITION OF "PHASE" TERMINOLOGY

This sect ion a t t e m p s to clar i fy t h e concept of phase. The voltageJirnpedance graphs, presented in Section 5.5, a r e used as a link between impedance diagrams and t h e display on a n eddy cur ren t instrument monitor.

In eddy cur ren t work t h e most confusing and o f ten incorrect ly used t e r m is PHASE. w

P a r t of t h e problem ar ises because of t h e exis tence of two eddy cur ren t methods, coil impedance and send-receive. In th is sect ion a n a t t e m p t is made t o clarify some of t h e multiple uses of the word.

Figure 5.16 shows t h e impedance of a probe touching test material. The t w o axes represent t h e quadrature components, v a n d V R, of voltage across a coil. In t h e

pa ramete rs oL/wLo and RL/ wL,. t absence of real numbers, t h e axes can a so b e c o n s ~ d e r e d as t h e normalized

The following l i s t summarizes uses of t h e t e r m PHASE. One o r Inore of t h e s e a r e o f t e n used without adequa te explanation because t h e t e r m will have a colloquial meaning.

WL 1. ol, - Arctan - L

, angle between t o t a l vol tage vector and resist ive voltage vector.

NOTE: An impedance bridge measures amplitude of t h e impedance vector Z and t h e angle Q where t h e res is tance includes %c This vector could therefore not b e shown on Figure 5.16. (I t is shown on t h e impedance diagram in Figure 5.8(a)).

2. AO1. Change in phase o f normalized resul tant voltage vector as probe is moved over a defect .

3. 0, .

4. AO,,

5. G3,

6 . B ,

7. 0 4 '

8. 0).

Phase between secondary voltage (induced voltage) and primary voltage (excitation voltage). Send-receive instruments measure secondary voltage.

Change in phase of secondary voltage as probe is moved over a defect. This is approximately the phase measured by some send-receive eddy current instruments without X-Y outputs.

Phase between t h e voltage signals obtained from LIFT-OFF and a crack o r void. I t is related t o PHASE LAC 0 explained bel0w.(0~ is about double t he phase lag.) o 3 is used t o estimate defect depth during ET.

PHASE LAG (not shown in Figure 5.16) of eddy currents below the surface relative t o those a t the surface. It was derived in the eddy current density equation Chapter 2, i.e. 6 = x/6 for semi-infinite plates, where x is the distance below the surface and 6 is in radians.

Many eddy current instruments have a PHASE knob by which the ent i re impedance voltage plane display can b e rotated. It is common practice t o ro ta te t he display t o make LIFT-OFF horizontal. (on an eddy current instrument display, absolute orientation of inductive and resistive axes rnay be unknown).

Phase between inductive voltage and current in a circuit; 8 - g o m 5

5.8 SELECTION OF TEST FREQUENCY

L 5.8.1 Inspecting for Defects

The f i rs t question tha t must be answered before proceeding with an inspection is: For what type of defects is t he inspection being done? If t h e defects a r e cracks: What is t he smallest defect t ha t must be detected? Are the cracks surface or subsurface? Are they likely t o be laminar cracks or normal t o t h e test surface? A single general inspection procedure t o verify the absence of any and a l l types of defects of ten has l i t t le merit. Inspections often require two o t more test frequencies and/or different probes t o accurately identify defects.

Test, f requency can be selected without knowledge of the character is t ic parameter, P, ,or the operating point on the impedance graph. I t should be chosen for good discrimination between defects and other variables. The most troublesome variable is LIFT-OFF variations, so separation of defec ts f rom lift-off is the forernost consideration.

Only t h e skin depth equation has t o be used,

A t es t frequency where 6 is about equal t o t h e expected defect depth provides good phase separation between lift-off and defec t signals. Figure 5.17 illustrates t h e display on an eddy current instrument monitor as a probe passes over surface and

L I F T -OFF

\ \ SUBSURFACE VOlD ( A )

SUBSURFACE SURFACE VOID (0 )

CRACK

/ SURFACE CRACK

SUBSURFACE VOlO ( A ) 2 5 9 - SUBSURFACE VOID ( 8 )

INCREASING L I F T -OFF

X - Y DEFECT SIGNALS

FIGURE 5.17 Typical Response Signals for Two Types of Defects

subsurface defects. Test frequency is such tha t 6 equals depth of deepest defect , and instrument controls a r e selected such tha t a signal from lift-off is horizontal. Note t he difference in signal amplitude and angle relative t o lift-off of subsurface voids A u and R. This results from skin depth attenuation and phase lag.

If, during inspection, a signal indicating a defect is observed, tes t frequency may be altered t o verify whether t he signal represents a real defect o r t he e f fec t of another variable. This discussion is expanded in t h e next chapter under Signal Analysis.

5.8.2 Measuring Resistivity

Resistivity can be measured a t small localized areas or by sampling a larger volurne of a test article t o determine bulk resistivity. The volume of material interrogated depends on probe size and t e s t frequency. For bulk measurements a large probe would be used and a low frequency t o maximize penetration. The skin depth equation is again used t o est imate depth of penetration at the t e s t frequency.

Electrical resistivity measurement is a comparative technique; reference samples of known resistivity must be used for calibration. Variables tha t a f fec t t h e accuracy of resistivity measurement a r e lift-off, temperature, and chan es in t h e flow of eddy currents in tes t articles not related t o electrical resistivity such as cracks, thickness and surface geometry).

'i For best discrimination between resistivity and other variables t h e operating point on an impedance graph should be considered. Figure 5.12 illustrated the e f fec t of tes t frequency on normalized probe impedance. At t h e top of t h e graph the angle, between lift-off variations and t h e resistivity curve, is small. Moving down the curve

t he angle, separating the two variables, increases towards the knee with no appreciable change beyond that. However, small lift-off variations, at t h e bottom of t h e curve, produce a large impedance change. The best operating point is somewhere between the two extremes, near the knee of the impedance curve.

FIGURE 5.18 Resistivity Measurement and t h e Impedance Graph

Figure 5.18 shows the method of manipulating test conditions t o best deal with lift- off. Figure 5.18(a) shows the resistivity impedance curve with a frequency and probe selected t o operate near the knee. Figure 5.18(b) is an enlarged section of the curve rotated so lift-off signals a re approximately horizontal. This is t he view on an eddy current instrument monitor.

Next consider temperature effects. First, test article resistivity will be a function of temperature so test sample and standards should be at uniform temperature. A greater potential error is in probe wire resistance, R E . The coil wire resistance is a part of the probe impedance circuit, so variations in temperature which affect coil resistance will appear as an impedance change. For greatest accuracy, the inductive reactance, X L , should be large compared t o coil wire resistance; X L / % ~ > 50 is desirable.

Obviously this condition is not easily satisfied at low t e s t frequencies where inductive reactance is low. One solution is t o use a large diameter probe cupped in ferrite. The large diameter and ferr i te cup will both increase X L / R E . Another solution is t o use a Send-Receive instrument. Such an instrument has a high input impedance, sensing only voltage changes in the receive coil. Coil wire

resistance is insignificantly small in comparison t o instrument impedance and therefore has no effect.

Consider next t he e f fec t of changes in eddy current path not related t o electrical resistivity. If t h e tes t is supposed t o be a measurement of electrical resistivity, thickness should not influence the signal. The skin depth equation must again be used. Test ar t ic le thickness should be equal t o o r greater than three skin depths, t 2 3 8 ,

t 2 3 x 0 , rum

or 22500 - r 2- , Hz t

2

where t is thickness, p is resistivity in microhm-centirnetres, and f is frequency

Other sources of signals a r e edge e f fec ts and surface geometry. When the tes t article's edge is within the probe's magnetic field, a n increase in resistance t o eddy current flow will be detected. Edge ef fec t can be reduced by probe design, such a s a ferr i te cupped probe, or by increasing t e s t frequency.

If t he surface of t h e test article is contoured, the magnetic flux coupling will differ f torn that of a f la t surface and a correction factor may be required.

Cracks or voids a r e usually less of a problem. The signal frorn a crack will be very localized whereas resistivity variations a r e usually more gradual. The best procedure t o determine if a localized signal is from a change in resistivity is t o rescan with a smaller probe at higher and lower frequency (at least three t imes and one third the tes t frequency). The angle between t h e signals frorn lift-off and resistivity should vary only slightly whereas the angle between lift-off and defect signals will increase with frequency.

An example of resistivity variations in a zirconium alloy, due t o a change in oxygen concentration, is shown in Figure 5.19.

TEST ARTICLE WIDTH

(35 cm)

X , VOLTS

( a ) X -Y OISPLAY OF C O l L IMPEOANCE FROM CHANGE I N ELECTRICAL R E S I S T I V I T Y

( b ) MOD1 F l ED C -SCAN D l SPLAY l NG Y -COY PONENT OF C O l L IMPEDANCE VECTOR FROM A CHANGE I N ELECTRICAL R E S I S T I V I T Y

FIGURE 5.19 Eddy Current Signals from a Change in Electrical Resistivity m the

Surface of a Zr-Nb Test Article. Test Frequency = 300 kHz.

5.8.3 Measuring Thickness

Test frequency should be chosen so 'lift-off' and 'change in thickness' signals are separated by a 90' phase angle, see Figure 5.20(a). This frequency can be calculated using the skin depth equation. A reasonable approximation for thin sections is obtained when

which converts to

f = 1 . 6 p / t 2 , kHz (5.7a)

where 6 is skin depth, mm t is test article thickness, mm P is electrical resistivity, p 8 a n f is frequency, kHz

Pr is relative permeability (u, = 1 for nonferromagnetic material).

In testing thick material, this equation can similarly be used t o choose a test frequency t o separate lift-off and subsurface defect signals by 90'. Equation 5.7(a) can be used by replacing t with x,

where x is depth of subsurface defect.

EDDY CURRENT INSTRUMENT MONITOR DISPLAY

I ' MONl TOR O I S I L 1 Y

FIGURE 5.20 Thickness Measurement and the Impedance Graph

Conventional thickness measurement is t o display the lift-off signal horizontal (along the X axis) and use the vertical signal (along the Y axis) t o measure thickness, see Figure 5.20(b). If t he signal on the instrument monitor is s e t t o move from right t o lef t a s the probe is moved away from the test article, a vertical movement up or down denotes decreasing and increasing thickness respectively.

5.8.4 Measuring Thickness of a Non-Conducting Layer on a Conductor

An insulating layer will not conduct eddy currents so measurement of i ts thickness is essentially a lift-off measurement (provided i t is non-ferromagnetic), i.e. the distance between the coil and test article. At high test frequency a small variation in

lift-off produces a large change in probe impedance a s shown in t h e impedance graph of Figure 5.9.

To minimize t h e signal from variations in t h e base material, t h e test should therefore be done at t h e highest practical frequency. The maximum frequency would be limited by probe-to-instrument impedance matching, cable resonance problems and cable noise.

The measurement is a comparative technique so standard reference thicknesses must be used for calibration.

5.8.5 Measuring Thickness of a Conducting Layer on a Conductor

Measurement of t h e thickness of a conducting layer on a conducting tes t article can be done provided there is a difference in electrical resistivity ( Ap) between t h e two. The measurement is essentially t h e same a s t h e thickness measurement described in Section 5.8.3. There is one important difference; variables in t h e base plate, in addition t o t h e variables in t h e layer, will affect t h e signal.

Figure 5.21(a) shows a computer simulation of a layer thickness measurement. The model shows t h e magnitude and direction of variables when attempting t o measure a layer (clad 11, nominally 0.75 mm thick, with resistivity P = 3 uf2.cm on a base (clad 2) with resistivity 5 ufl. cm. The plot is part of a normalized impedance graph. In addition t o material property variables, the parameter of space (gap) between the layers is shown as well a s t h e e f fec t of an increase in test coil temperature. At 10 kHz, t / 6 is 0.8 and, a s predicted, t h e angle separating signals from lift-off and layer (clad 1) thickness is about 90'. Unfortunately, so a re the signals from test coil temperature, gap, and resistivity of the base (clad 3). Some of these parameters can be discriminated against a t higher and/or lower tes t frequencies.

R I Y C E O F V I R l I B L E S SHOWN I N C O W P U T O R P L O T S

r~0Yt"c.l . I0 **I 101

FIGURE 5.21 Computer Simulation of a Multi-Layw Sample

PROBE-CABLE RESONANCE

Probe-cable resonance must be considered when operating a t high test frequencies and/or using long signal cables, e.g., frequencies greater than 100 kHz and cables longer than 30 m. Most general purpose eddy current instruments cannot operate at or close t o resonance.

Probe-cable resonance can be modelled as shown in figure 4.5. In simple terms, resonance occurs when inductive reactance of t h e coil equals capacitance reactance of the cable, i.e. when

where w is angular frequency, in radianstsecond. L is coil inductance in henries and C is to ta l cable capacitance in farads.

Transforming this equation and substituting w- 2rr f shows resonance occurs when frequency is

This approach is sufficiently accurate for most practical applications. A more rigorous approach t o resonance is presented in Section 4.3.

Resonance is apparent when a probe and cable combination, which balances a t a low frequency, will not balance as frequency is increased. At t h e approach of resonance, the balance lines on t h e eddy current storage monitor will not converge t o a null. The two blancing (X and R) controls will produce nearly parallel lines rather than the normal perpendicular traces, on the s torage monitor. A number of steps can be taken t o avoid resonance:

1. Operate a t a test frequency below resonance, such that f is less than r 2. Select a probe with lower inductance. (Since f , is proportional t o 1 /

inductance must be decreased by a factor of four t o double resonant frequency). 3. Reduce cable length or use a cable with lower capacitance per unit length (such

a s multi-coax cables). This will raise t h e resonance frequency since capacitance is proportional t o cable length and f , is proportional t o I/ /F:

4. Operate a t a t e s t frequency above resonance, such that f is greater than 1. 2 f r 0 However, above resonance the sensitivity of all eddy current instruments decreases rapidly with increasing frequency because capacitive reactance ( X ,=I / w C) decreases, and current short circuits across the cable, rather than passing through t h e coil.

5.10 SUMMARY

Test probes induce eddy currents and also sense t h e distortion of their flow caused by defects. Surface probes contain a coil mounted with i t s axis perpendicular t o the test specimen. Because i t induces eddy currents t o flow in a circular path i t can be used t o sense all defects independent of orientation, as long as they have a component perpendicular t o t he surface. It cannot be used t o de tec t laminar defects.

For good sensitivity t o short defects, a small probe should be used; probe diameter should be approximately equal or less than t h e expected length. Sensitivity t o short subsurface defects decreases drastically with depth, even a 'thint 5 mm sample is considered very thick for eddy current testing.

The analysis of eddy current signals is t h e most important and unfortunately the most difficult task in a successful inspection. A thorough understanding of impedance graphs is essential t o manipulate test conditions t o minimize undesirable tes t variables. The characteristic parameter for surface probes is used t o locate the operating point on the impedance diagram. I t is given by

where is mean coil radius, mm; f is t e s t frequency, Hz; and p is electrical resistivity, microhm-centimeters.

The criterion for defect detection with impedance plane instruments is phase discrimination between lift-off noise and defec t signals. Test frequency is chosen such tha t 'lift-of ft and 'change in wall thickness' signals a r e separated by a 90" phase angle. This can be derived from the following equation:

2 f - 1 . 6 p l t , k H z (5-7)

where t is sample thickness, mm.

If inspection is performed a t high t e s t frequencies and/or with longcables, i t is desirable t o operate below probe-cable resonance frequency. This is norrnally achieved by using a probe of sufficiently low inductance.

To optimize test results, t he inspector has control over probe size and test frequency. In choosing probe diameter t h e following must b e considered:

operating point on impedance diagram probe inductance and resistance sensing a rea sensitivity t o defec t length sensitivity t o defec t depth sensitivity t o lift-off sensitivity changes across coil diameter (zero at centre) sensitivity changes with fe r r i te core o r cup.

Choice of test frequency depends on:

(a) depth of penetration (b) phase lag

(c) operating point on impedance diagram (d) inductive reactance (e) probe-cable resonance

5.11 WORKED EXAMPLES

5.11.1 Effective Probe Diameter

PR0BLEM:Determine sensing diameter of a 5 mm probe when (a) testing 316 stainless s teel ( P = 7 2 microhrn-centimetres) a t

2 MHz, and

(b) testing brass ( P = 6 .2 microhm-centimetres) a t 10 kHz.

SOLUTION:

Characteristic Parameter

PROBLEM: If an available probe had coil dimensions of 10 mm outer diameter and 4 mm inner diameter, determine the best frequency for resistivity measurements of a zirconium alloy ( P - 5 0 ~nicrohrn- centimetres).

SOLUTION: The best frequency for resistivity measurenents is when the operating point is at the knee location on the impedance diagram. This occurs when t h e characteristic parameter P ,110. Using equation 5.5,

P C - 7.9 x 10 -4 10 .0+ 4.0 4

j 2 £150 = lO

therefore, f = 50 kHz.

(This calculation places no emphasis on skin depth effect, which may be a n overriding consideration).

CHAPTER 6 - SURFACE PROBE SIGNAL ANALYSIS

6.1 INTRODUCTION

Manufacturing and preventive maintenance inspection of "flatw components with surface probes is one of t he oldest and most important applications of eddy current testing. Manufacturing inspection of small s teel components for defects and hardness is almost exclusively performed by eddy current methods. For safety reasons and preventive maintenance (savings on replacement costs and downtime) inspection of aircraft components for cracks and hea t t rea tment effects has been performed since commercial a i rcraf t first went into service. Eddy current testing is one of t h e most effect ive NDT methods for t he above applications because i t doesn't need couplants, i t is fast , and 100% volumetric inspection is of ten possible.

This chapter describes how t o maximize signal-to-noise by proper choice of test frequency and minimizing ttlift-offtt noise. Emphasis is given t o signal analysis and how t o recognize and discriminate between defect signals and false indications. An a t tempt is made throughout this chapter t o illustrate discussions with real or simulated eddy current signals.

6.2 EDDY CURRENT SIGNAL CHARACTERISTICS

Defect Signal Amplitude

A defect, which disrupts eddy current flow, changes test coil impedance as the coil is scanned past a defect. This condition is showm pictorially in Figure 6.1 which portrays eddy currents induced by a surface probe in a defective plate. Eddy currents flow in closed loops as illustrated in Figure 6.lta). When a defect interferes with the normal path, current is forced t o flow around o r under i t o r is interrupted completely. The increased distance of the distorted path increases t he resistance to current just a s a long length of wire has more resistance than a short length,

Eddy currents always take the path of least resistance; if a defect is very deep but short, current will flow around the ends; conversely, if a defect is very long (compared t o t h e coil diameter) but shallow, t he current will flow underneath. In summary, defect length and depth (and width t o some degree) increase resistance t o eddy current flow and this, in turn, changes coil impedance. (The ef fec t of defect size on flow resistance in tube testing is derived in Section 8.2.1).

In terrns of t h e equivalent coil circuit of a resistor in parallel with an inductor and i ts associated semi-circular impedance diagram (Section 3.51, a defect moves the operating point up t h e impedance diagram. Increasing resistance in a test ar t ic le changes both probe inductance and resistance.

In t he preceding discussion t h e defect was considered t o disrupt the surface currents closest t o t h e coil. Consider t h e difference between surface and subsurface defects. When a surface probe is placed over a deep crack of infinite length, t h e surface currents must pass underneath t h e defect if they are t o form a closed loop, see Figure 6.2(a). This is not t h e case with subsurface defects as shown in Figure 6.2(b). Although the void in this picture is not as far from t h e surface as t h e bottom of the crack, t he void may not be detected. Eddy currents concentrate near t he surface of a conductor, and therefore, tests a r e more sensitive t o surface defects than internal defects.

I EMV CURRENTS

TEST PLATE

EDDY CURRENTS FLOW I N CLOSED PATHS. A DEFECT

INTERFERES WITH THE NOKUAL PATH

TEST PLATE

I \ I EDOV CURRENT OlSTDRTlOU

AT CRACK

( b ) EDDY CURRENTS TAKE THE PATH OF LEAST RESISTANCE

UNDER OR AROUND A DEFECT

FIGURE 6.1 Eddy Currents in a Defective Plate

The skin depth equation helps in t he understanding of this phenomenon. In Chapter 2 i t was shown tha t current density decreased with distance from t h e surface in t h e following proportions: - 63% of the current flows in a layer equivalent in thickness t o the skin depth, 6 , - 87% flows in a layer equivalent t o two skin depths, 2 6 , - 95% flows in a layer equivalent t o three skin depths, 3 6 .

S U R F A C E C O I L

T E S T P L A T E

/ '"-/'7 ,,/

role

w U rr I Y = = YI w

C

I = 0 a - L

C w n

r r

C Y

YI 0 - ( a ) EDDY C U R R E N T FLOW U N D E R A CRACK ( b ) E D D Y C U R R E N T FLOW AROUND A

S U B S U R F A C E role

FIGURE 6.2 Eddy Current Flow in t h e P r e r a m of (a) Surface and (b) Subsurface Defect

Since only 5% of the current flows at depths greater than t h e 3 6 there is no practical way t o de tec t a subsurface defect at this distance from the surface. But in the case of a long surface defect 3 6 or greater in equivalent depth, most of the current is flowing under t he defect. Surface cracks will be detected and depth can be estimated even if eddy current penetration is a small fraction of the defect depth. Once eddy currents a r e generated in a metal surface, they will follow the contour of a crack because a potential is set-up about the crack.

Defect Signal Phase

From the above description one cannot predict a defec t signal in detail, only i t s relative amplitude and direction on the impedance diagram. A more complete explanation requires inclusion of phase lag. Consider t he cross section of a surface probe a s shown in Figure 6.3(a). This pictorial view shows t h e distribution of magnetic field magnitude and phase around a coil as derived by Dodd(2). The solid lines a r e contours of constant magnetic field strength; t h e dashed lines represent constant phase. Since t h e magnetic field and induced eddy currents have approximately the same phase, t h e dashed lines will also represent the phase ( B ) of t he eddy currents. Amplitude drops off exponentially with distance and eddy current flow increasingly lags in phase (relative t o eddy currents adjacent t o t h e coil) both with depth and with axial distance from the coil. Skin depth e f f ec t occurs in both radial and axial directions.

Figure 6.3(a) permits an approximate derivation of eddy current signals for the shallow surface, subsurface and deep surface defects illustrated. One needs to establish a reference phase direction as star t ing point; the LIFT-OFF direction is convenient and can be defined a s the signal resulting from increasing t h e space between the coil and t e s t article, starting from t h e point when t h e space is minimum.

The signal or e f fec t of defects can be imagined as t h e absence of eddy currents which were flowing in t he area before t h e defect existed at this location. As the defects approach t h e coil from positions 0 t o 5 in Figure 6.3(a), the signal on the eddy current storage monitor moves from point 0 t o 5, tracing the curves illustrated in Figure 6.3(b). This procedure is reasonably straight forward for shallow surface and subsurface defects since they a r e localized and only intersect one phase and amplitude contour at any given position. For t h e deep defect one has t o divide the defect into sections and determine weighted average values for amplitude and phase at each position.

The shallow surface defect in Figure 6.3(b) has a large component in t he lift-off direction; primarily i ts approach signal makes i t distinguishable from lift-off. As defect depth increases, signals ro ta te clockwise due t o increasing phase angle. The angle indicated in Figure 6.3(b) is - not t he value calculated from the phase lag equation,

where k? is phase lag (radians), x is distance of defect below the surface (mm) and 6 is skin depth (mm).

The angle between lift-off and defect signals is about 2 8 . Although probably not strictly true, one can imagine defect phase angle as t h e sum of a lag from t h e coil t o the defect and the same lag back t o t he coil.

D E E P D E F E C T

SHALLOW D E F E C T

S U B S U R F I C E DEFECT

DEFECT P O S I T l O N

D E E P C E F E C T

\ SUBSURFACE O C F E C I

SHALLOW D E F E C T

L I F T - O F F

FIGURE 6.3 Derivation of Eddy Current Signal Appearance for Three Types of Defects

The foregoing discussion assumes t h a t t h e d e f e c t is a to ta l barrier t o t h e flow of current. Although this assumption is valid fo r rnost cracks o r discontinuities, some c racks a r e part ial conductors. Fat igue cracks, fo rmed when t h e test a r t i c le is under a tensile s t ress , can become tightly closed when s t r e s s is released. The result is t h a t some fract ion of eddy currents could b e conducted across t h e c rack in terface and t h e magnitude of t h e coil impedance change d u e to t h e d e f e c t will b e less. The phase lag argurnent is s t i l l valid; a deep c rack will s t i l l b e distinguishable f rom a shallow c rack by t h e shape of t h e eddy current signal, but t h e sensit ivity t o such a c rack will b e reduced because of smaller amplitude.

6.3 EFFECT OF MATERIAL VARIATIONS AND DEFECTS IN A FINITE THICKNESS

For each test, one must decide on t h e test frequency t o use and on t h e phase setting. The conventional way of se t t ing phase on a n eddy cur ren t instrument is t o display t h e "lift-off" signal horizontally (on t h e X-axis) with t h e impedance point moving f rom right-to-left as t h e probe is raised. All mater ia l variables will then display a n eddy current signal at a n angle clockwise to t h e lift-off signal.

LIFT-OFF 1 . 5 mm I p = 72 p a cm 2.0 mm

pr = 1 .O

L I FT -OFF L I FT 4 F F

FREQUENCY = 1 0 kHz FREQUENCY = 50 kHz FREQUENCY = 2 0 0 kHz

FIGURE 6.4 Probe Response to Various Test P a r a m e t e r s at Three Frequencies

Discrimination between defec t s and o ther variables is accomplished through pa t t e rn recognition and varying test frequency. Figure 6.4 displays t h e change in coil impedance loci for various parameters at di f ferent test frequencies. The e lec t r i ca l resistivity ( A D ) signal angle, re la t ive t o lift-off, increases only slightly a s frequency

FIGURE 6.5 Computer Simulation of Probe Response to Various Test Parameters

i s increased, whereas a change in pla te thickness ( A t ) signal angle continually increases with frequency. The angle, between t h e signal from lift-off and pla te thickness change, equals about twice t h e phase lag across t h e plate thickness. The signal from a change in magnet ic permeabil i ty ( Au of t h e pla te is approximately 90" to t h e lift-off signal at low frequency and decreases only slightly with increasing frequency.

Figure 6.5(a) i l lustrates a computer simulation of coil response t o various test parameters. The simulation is based on t h e s a m e probe and test sample used in t h e previous figure. Comparison of these t w o figures reveals computer simulation gives very realist ic results.

Note at 50 kHz t h e increase in magnetic permeabil i ty signal (Au ) is t o t h e r ight of t h e e lect r ica l resistivity signal for t h e 7 m m probe. For t h e 25 mm probe at 50 kHz i t is t o t h e l e f t of t h e d p signal. As t h e operat ing point moves down t h e impedance curve with increasing probe diameter , a resist ivity signal r o t a t e s CW relative t o a permeability signal. Note also t h a t t h e permeabil i ty signal i s not perfectly parallel t o t h e inductive reac tance axis. This is due t o t h e skin depth and phase lag changing with permeability, ro ta t ing t h e signal CW.

During general inspection for a l l parameters in a thin pla te test frequency is normally chosen such t h a t 'lift-off' and 'change in p la te thickness' signals a r e separated by 90° on t h e impedance plane. This frequency is empirically derived by se t t ing ra t io between pla te thickness and skin depth equal t o approximately 0.8,

Substituting in equation 2.1 3 yields

f = 1 . 6 p l r 2 , kHz

where p is e lect r ica l resistivity (microhm-centimetres), and t is p la te thickness (mm).

This frequency has been proven in pract ice on various conductivity samples and various probe diameters. The 90' phase angle increases only slightly with increasing probe diameter, see Figure 6.5(b). All d e f e c t signals (from surface o r subsurface defects) will fal l inside th is 90" band. Shallow defects , c racks o r pits, on t h e opposite side of t h e pla te will produce a signal whose angle approaches t h a t of wall thickness, i.e 90". Shallow defec t s on t h e su r face neares t t h e probe will produce a signal whose angle is c lose t o t h a t of lift-off.

The two methods of discriminating between d e f e c t s and o ther variables, pa t t e rn recognition and varying test frequency, complement each other. Consider signal pa t t e rn behaviour due to nominal wall thickness and resistivity variations. These variables normally change gradually along a sample. Whereas cracks, pits, and subsurface voids o r inclusions exhibit a s t e p change. Discrimination between t h e s e variables is enhanced by analyzing the i r behaviour at di f ferent test frequencies, a s shown in Figures 6.4 and 6.5. An ex t remely important point t o remember is t h a t a l l de fec t s will fal l between t h e 'lift-off' signal angle and t h e 'decrease-in-wall- thickness' signal angle regardless of frequency. (For pract ica l applications th is s t a t e m e n t is valid; however, t h e signal f rom a shallow defec t with length g r e a t e r than a probe diameter may dip slightly below t h e lift-off signal).

CAL l BRAT l ON CRACKS

SAMPLE: p = 50 p a cm

p, = 1 '00

CRACK CRACK 2 mm DEEP NOTCH

2 m DEEP NOTCH * LIFT -OFF LI FT -OFF I

0.5 mm DEEP NOTCH 0 . 5 mm DEEP NOTCH

FREQUENCY = 50 kHz FREQUENCY = 300 kHz

FIGURE 6.6 X-Y Display of Coil Impedance Vector from Calibration Grooves

and a Real Crack. Estimated Depth = 1.3 mm.

6.4 COIL IMPEDANCE CHANGES WITH DEFECTS

6.4.1 Surface Defect Measurement

Figure 6.6 illustrates t he method used t o predict depth of surface defects. Pat tern recognition is used where coil impedance response from t h e defect is compared with calibration defects. To est imate defect depth by pat tern recognition, the real and calibration de fec t signals must be comparable in amplitude. This can be achieved by changing the gain of the display (normally by decreasing the calibration defect signals). Defect depth is estimated by interpolation.

Amplitude of defect signals is not a reliable parameter for estimating defect depth. Amplitude is affected by length and t h e degree of contact across t h e two interfaces (e.g., crack closure). Whereas t he coil impedance locus (the X-Y display of coil impedance) depends mainly on t h e integrated response with depth of t he eddy current phase lag.

6.4.2 Subsurface Defect Measurement

Signals from subsurface defects, Figure 6.10(b), have an average phase angle relative t o lift-off of approximately 2 B where 0 is t he phase lag of t he eddy currents at

dep th x. This signal i s similar t o a change in wall thickness signal and i t s phase was denoted by 0 3 in Figure 5.16.

L 6.5 COIL IMPEDANCE CHANGES WITH OTHER VARIABLES

Ferromagnetic Indications

In eddy current tes t ing t h e test coil is sensit ive to many test parameters. One variable t h a t o f ten causes problems is magnet ic permeability. A t common test frequencies o n e c a n easily mis take a signal due to increased permeability (ferromagnetic indication) for a serious defect . The following discussion briefly outlines t h e problem and shows how one can di f ferent ia te between defec t s and ferromagnet ic indications.

I t is generally recognized t h a t magnet ic sa tura t ion is required f o r eddy current test ing of ferromagnet ic alloys. Conversely, sa tura t ion is not usually employed when test ing %on-magnetic" alloys such as austeni t ic stainless s t e e l s and nickel base alloys. Unfortunately, these alloys and a n y alloys containing iron, nickel o r cobal t can display variations in magnetic permeability. This is caused by t h e s t rong dependence of magnet ic properties on metallurgical variables such as composition, grain size, thermal processing, cold work, contamination and segregation.

The following a r e examples of ferromagnet ic indications in nominally nonmagnetic alloys which have been encountered:

- Ferromagnetism associated with manufacturing d e f e c t s in Inconel 600 extrusions (possibly f rom chromium depletion at t h e surface). - Ferromagnetism associated with EDM calibration grooves in Type 304 stainless steel . - Permeabil i ty variations occuring in austeni t ic stainless s t ee l castings probably due t o segregation (or possibly contamination). - Ferromagnetic inclusions in zirconium alloys resulting f rom pick-up during forming.

- Magneti te (Fe 3 0 4 1 deposits on h e a t exchanger tubes due to s tee l corrosion somewhere e l se in t h e cooling system.

T h e f i r s t t w o types of de fec t s would have made d e f e c t depth predictions seriously inaccurate, and t h e las t t h r e e types of ferromagnet ic indications could have been mistaken fo r d e f e c t s such as c racks o r pitting.

Some of t h e anomalous ferromagnet ic indications listed above could b e suppressed by saturating t h e test a r e a with a permanent magnet possessing a flux density of a few kilogauss. If sa tura t ion is not possible (or incomplete) the re is another way t o determine if a n indication is due to a d e f e c t o r a magnetic ef fect . The method involves re tes t ing at a much lower frequency. I t i s illustrated in Figure 6.7 for t h e case of a surface probe passing over d e f e c t s and a ferromagnetic inclusion.

At typical t e s t frequencies (100-500 kHz) there is l i t t le phase separation between the signal from defects and magnetic inclusions. As test frequency is reduced, t h e operating point moves up t o t h e impedance curve and defect signals ro ta te a s shown.

FERMlYADlET lC FERRWIAGIIET I C

NOTCH

2 m DEEP

LIFTOFF 0.5 m DEEP

I00 hHZ - 2 nm OEEP

FERRO

\c, L 0 . '

0.5 m OEEP

FIGURE 6.7 Coil Impedance/Voltage Display at Three Frequencies

The important point t o note is t h a t relative t o lift-off, defect signals ro ta te CCW whereas the magnetic inclusion signal rotates CW and approaches 90' at low frequency (approximately 10 kHz or lower for t he above probe and sample). On t h e impedance diagram of Figure 6.7 t h e direction of the ferromagnetic signal would not vary appreciably with frequency; increased permeability primarily increases coil inductance.

When a magnetic inclusion is not on the surface - if i t is subsurface o r on the opposite side of a thin test plate - there is t he added complication tha t t h e angle of the signal will be rotated relative t o t he angle of a ferromagnetic indication on t h e surface adjacent t o t h e coil. This arises from phase lag across the plate thickness. The previous approach of retesting a t reduced frequency will also serve t o distinguish between defects and magnetic inclusions. If t he phase of t he signal from the indication increases t o 90° relative t o 'lift-off', i t is a ferromagnetic anomaly; if i t decreases t o nearly 0°, i t is a defect.

To summarize: (a) Many nominally %on-magnetic" alloys can exhibit ferromagnetic proper t ies and

almost any alloy can pick-up magnetic inclusions or contamination during manufacture or service.

(b) At normal eddy current test frequencies magnetic indications will often appear similar t o defects

(c) Magnetic indications can be distinguished from defects by retesting at a reduced test frequency.

6.5.2 Electrical Resistivity

Electrical resistivity is a material parameter which, unlike a defect, usually varies over significant area. However, if i t is localized, and t h e eddy current signal is small, i t could be be mistaken for a small defect. The best means of distinguishing the two is t o rescan with a smaller probe a the same test frequency, at three times the test frequency, and a t one third the t e s t frequency. Unlike a defect signal, the angle between resistivity and lift-off changes l i t t le with frequency. See impedance graph in Figure 5.9.

As with the detection of any signal source, resistivity is affected by skin depth. At high frequency, when skin depth is small, there will b e greater sensitivity t o surface resistivity variations. At lower test frequency, eddy currents penetrate deeper into the material so the measurement will represent a larger volume.

6.5.3 Signals from Changes in Sample Surface Geometry

Abrupt changes in surface curvature result in eddy current signals as probes t raverse L them. I t causes changes in coupling creating a large lift-off signal and the curvature

also changes eddy current flow distribution creating a n effective resistance change, yielding a signal at an angle t o t he lift-off direction. The combined e f f ec t may be a complicated signal, a s shown in Figure 6.8. The appearance of this type of signal will not change significantly when rescanned at higher and lower test frequency.

Such signals can be difficult t o analyze because they depend on how well t he probe follows complicated surface curvatures. Basically the direction of the impedance change obeys t h e following rules when using surface probes:

- decreasing radius of curvature on an external surface, e.g., ridge, produces a change in t he direction of increasing resistivity,

- decreasing radius of curvature of a n internal surface, e.g., groove, produces a change in t h e direction of decreasing resistivity.

Figure 6.8(a) illustrates t h e signal as a probe traverses a shallow groove (decrease in surface radius) on the internal surface of a 100 mm tube. Figure 6.8(b) shows the signal a s a probe traverses a f la t (increase in surface radius). The t e s t was done with a 9 mm diameter probe at a test frequency of 300 kHz.

1 V O L T I ( a ) W l DE SHALLOW GROOVE ( b ) L O C A L F L A T S P O T

FIGURE 6.8 X-Y Display of Surface Coil Impedance for Internal Surface Variations

in a 100 m m Diameter Tube

CALIBRATION DEFECTS

H

1 V O L T

Analysis of eddy current signals is, for t h e most part, a comparative technique. Calibration standards are necessary for comparing signal amplitude and phase (shape) of unknown defects t o known calibration defects. Calibration signals a r e also used for d

standardizing instrument settings, i.e., sensitivity and phase rotation.

Existing national specifications and standards only supply broad guidelines in choice of tes t parameters. They cannot b e used t o establish reliable ET procedures for most inspections. Figure 6.9 shows a calibration plate proposed by the authors for general application. The e f f ec t of t h e following can b e established using this plate:

1. Varying Electrical Resistivity 2. Varying Thickness 3. Surface Geometry (Curvature) 4. Defect Length for Constant Depth 5. Defect Depth for Constant Length 6 . increasing-subsurface Defect ~ k e for Constant Defect Depth 7. Increasing Distance of Subsurface Defects from the Surface with Constant

Defect Size 8. Varying Thickness of a Non-conducting Layer (lift-off) 9. Varying Thickness of a Conducting Layer

1 0 . Ferromagnetic Inclusions

I I I I 1 I I

on ~CoNoucrlNc

0.2 m 1.0 mm 0.1 I""

0.1 m 0.5 m .05 "m

0.05 m 0.1 m .Dl I""

t a t FRONl SIDE

FIGURE 6.9 Calibration Standard

More than one calibration plate would be required t o cover a complete range of materials. A group of three would normally suffice, comprising base materials: aluminum alloy, p-4 a. cm; bronze, p = 2 5 W . cm ; and Type 316 stainless steel, P -74 uR. cm.

Figure 6.10(a) illustrates eddy current signals obtained with an absolute surface probe from some of the calibration block defects. Figure 6.10(b) illustrates signals fro:n the same defects using a differential surface probe, similar t o t h a t in Figure 5.2(c).

0 . 5 mm DEEP 4 mm DEEP 4 mm DEEP

L IFT-OFF L I FT -OFF

SURFACE DEFECTS

0 . 7 mm DEEP I 0 . 7 mm DEEP

L IFT-OFF

SUBSURFACE DEFECTS

FIGURE 6.10 Eddy Current Signals With (a) Absolute and (b) Differentia1 Surface Probes

6.7 SUMMARY

Defect signal amplitude is a function of defec t length, depth and closure (if a crack). Signal phase is primarily a function of defect depth. For volumetric inspection of thin material the following t e s t frequency should b e used:

f = 1.6 p / t 2 , kHz (5.7)

where p is electrical resistivity, microhm-centimetre, and t is wall thickness, mm.

At this frequency there is good discrimination between defects and lift-off signals but not between defects and ferromagnetic signals. Magnetic indications can be distin uished from defects by retesting at reduced frequency. Defect signals rotate CCW 'i approaching 0') whereas ferromagnetic signals ro ta te CW (approaching 90°) relative t o lift-off signals.

There a r e few national standards governing eddy current inspections with surface probes. For effective inspection, a calibration block should simulate t h e test piece and contain appropriate surface and substrate defec ts along with ferromagnetic inclusions. Basic knowledge of phase lag and impedance diagrams is also required for reliable analysis of eddy current indications.

CHAPTER 7 - TESTING OF TUBES AND CYLINDRICAL COMPONENTS

7.1 INTRODUCTION

Tubes or rods up t o about 50 mm diameter can be inspected for defects with encircling coils. Defect sensitivity in larger diameter components decreases because the inspected volume increases while defect llvolume" remains the same for a given defect. For larger diameters, surface probes should b e used t o obtain higher defect sensitivity, see Chapter 5.

The components can be in t h e form of wire, bars or tubes and round, square, rectangular or hexagonal in shape, as long as appropriate coil shapes a r e used. Inspection is f a s t and efficient since a n encircling coil samples t h e complete circumference of t h e component, allowing 100% inspection in one pass.

Defect detectability depends on disruption of eddy current flow. Therefore, the best probe is t h e one which induces highest possible eddy current density in t h e region of material t o be inspected, and perpendicular t o t h e defect.

When planning an inspection, t he following questions must first be answered: - For what type of defects is the inspection t o be performed? - If cracks a r e expected, do they have directional properties? - Does the material or components in close proximity have ferromagnetic

proper ties?

Once these questions have been answered one can decide on suitable probe design, tes t frequency and calibration standards. With the proper procedures one can - discriminate between defect signals and false indications as well a s determine depth once a defect is located. These procedures a r e based on a knowledge of impedance diagrams and phase lag.

7.2 PROBES FOR N B E S AND CYLINDRICAL COMPONENTS

Probe Types

Four common probe types for testing round materials a r e illustrated in Figure 7.1: (b) and (d) a r e differential probes, (a) and (c) show absolute probes. Each type contains two separate coils t o satisfy AC bridge circuit requirements, which is t he typical mode of operation of most eddy current instruments, see Chapter 4. These bridges require matching coils on two separate legs of t he bridge t o balance, thus permitting amplification of t h e small impedance differences between the two coils. If t he two coils a r e placed side-by-side, both equally sensing the t e s t material, t he probe is "differentialt1. If one coil senses t h e test article, t h e other acting only as a reference, t h e probe is absolute.

Figure 7.l(a) and (c) show effect ive designs for absolute probes; t he piggy-back reference coil is separated from t h e test ar t ic le by t h e test coil and therefore couples only slightly to t h e test ar t ic le (fill factor < <1) .

r TUBE 1-- CENTERING D I S C S

REFERENCE C O l L

REFERENCE C O l L

( A ) NClRCLlNG PROBE. ABS LUTE f P I G w - B A c K REFERENCE?

(D) INTERNAL PROBE, D l FFERENTIAL

(0 ) ENCIRCLING PROBE, D IFFERENTIAL

FIGURE 7.1 Tube Probe Types

Coil Size

The best compromise between resolution and signal amplitude is obtained when coil length and thickness equal defect depth. See Figure 7.2 for a labelled diagram of a probe cross section.

As a general guideline for tube inspection, coif length and depth should approximately equal wall thickness. However, t o improve coupling, a rectangular cross section with thickness reduced t o one-half the length can be used. For greater sensitivity t o srnall near surface defects, coil length and thickness can both be reduced further. Unfortunately this will result in a decrease in sensitivity t o external (far surface) defects.

Coil spacing, in differential probes, should approximately equal defect depth o r wall thickness for general inspections.

FIGURE 7.2 Probe Coil Nomenclature

-

/ / / A / / / / . / / t

C O I L THICKNESS -TX T U B E - C O I L CLEARANCE

For increased sensitivity t o near surface defects, spacing can be reduced at the expense of a reduction in sensitivity with distance from the coil.

Probe-to-tube clearance or gap should be as small a s possible. In most internal tube inspections, a gap equal t o half the wall thickness is common. A larger gap (smaller f ill-factor o r coupling) results in a small decrease in near surface defect resolution and a large decrease in signal amplitude for a l l types of defects.

/ / / y / / / / / / / / / / / / / / / / / / L I

I: C O I L SPACING

Comparing Differential and Absolute Probes

-CI

Absolute probes with a fixed reference coil a r e essential t o basic understanding. They enable study of all physical properties of a test a r t icle by plotting characteristic impedance loci.

L C O I L WIDTH J / / / / / / / / / / /

When an absolute coil signal is plotted as a function of distance (as t h e probe travels along a tube axis) dimensional variations and discontinuities can be separated. See the example of Figure 7.3(b). The signal is a function of effect ive cross-sectional a r ea of eddy current flow, i.e., wall thickness in t h e case of tubes, and can be analyzed like a surface roughness t r ace with the ex t ra advantage tha t subsurface flaws can be sensed.

In tube testing with an internal coil, absolute probe signals from defects and supports a r e simple and undistorted; signals from multiple defects and defects under support plates a r e of ten vectorially additive.

1

A

.-

t

- D (AVERAGE COl L DIAMETER)

Differential probes have two act ive coils usually wound in opposition (although they could be wound in addition with similar results). When t h e two coils a r e over a flaw- f r e e area of test sample, there is no differential signal developed between the coils since they a r e both inspecting identical material. However, when first one and then the other of the two coils passes over a flaw, a differential signal is produced. They have t h e advantage of being insensitive t o slowly varying properties such a s gradual dimensional variations and temperature: the signals from two adjacent sections of a test ar t ic le continuously cancel. Probe wobble signals a r e also reduced with this probe type. However, there a r e disadvantages; t he signals may be difficult t o interpret, even t o t he extent of being misleading. Defect signals under support plates can be extremely complicated. The signal from a defect is displayed twice: once as t h e f i rs t coil approaches the defect and again for t h e second coil. The two signals from a mirror image and the signal direction from the f i rs t coil must be noted. If a flaw is longer than t h e spacing between t h e two coils only t h e leading and trailing edges will be detected due t o signal cancellation when both coils sense the flaw equally.

I I I I SUPPORT PLATE P O S I T I O N I .

L SHOWING CORRODED AREA

a

O I F F E R E N T I b L C O I L S ABSOLUTE COIL

L

TRACE Wl TH L0SOLUTE PROBE

O I S T I N C E +

TRACE l l T H O l F F E R E N T l h L

I PROBE

W ~ L L LOSS V COMPONENT

I C I

FIGURE 7.3 Eddy Current Y-Channel Recordings from a Brass Heat Exchanger Tube

OD = 26.9 mm, t=l.lmm, fgO = 21 kHz

An even more serious situation occurs with differential probes when t h e ends of a flaw vary gradually; t h e defect may not be observed at all. An example of this is shown in Figure 7.3; this brass hea t exchanger tube suffered general corrosion as well a s localized corrosion on ei ther side of a support plate. The gradual upward trend of the Y-DISTANCE recording in Figure 7.3(b) shows t h e pronounced grooves a t A and B a r e superimposed on a n a rea of general wall thinning in t h e vicinity of t h e support plate. Note t h e response of a differential probe t o t h e same defect in Figure 7.3k). The differential probe senses t h e localized grooves but t h e Y-DISTANCE recording shows no indication of the gradual wall thinning which was apparent in Figure 7.3(b). Table 7.1 compares advantages and disadvantages of t h e two probe types.

TABLE 7.1 COMPARISON OF ABSOLUTE AND DIFFERENTIAL PROBES

ADVANTAGES: DISADVANTAGES:

ABSOLUTE PROBES

respond t o both sudden and gradual - prone t o dr i f t from temperature changes in properties and dimensions instability combined signals a r e usually easy t o - more sensitive t o probe wobble than separate (simple interpretation) a differential probe show total length of defects

DIFFERENTIAL PROBES

not sensitive t o gradual changes in - not sensitive t o gradual changes h a y properties or dimensions miss long gradual defects entirely) immune to drift from temperature - will only de tec t ends of long defec ts changes - may yield signals difficult t o less sensitive t o probe wobble than an interpret absolute probe

Directional Properties

When inspecting for defects, i t is essential t ha t flow of eddy currents be as perpendicular a s possible t o defects t o obtain maximum response. If eddy currents flow parallel t o a defect there will be l i t t le distortion of t he eddy currents and hence l i t t le change in probe impedance.

The eddy current flow characteristics of circumferential internal or external probes a r e listed and illustrated in Figure 7.4.

EOOV CURRENTS EOOV WRRENTS

EDOV CURRENTS FLOW IN CLOSEO rrrns - EOOV CURRENT FLOWS PARALLEL T O EOOV CURRENT FLOW OIYINISHES TO L I M I T E O TO C O N W C T I N C M I T E R I A L C O I L WIMOINGS - NOT S E N S I T I V E ZERO AT THE C E N l R E Of A S O L I D ROO .

1 0 N R E L V C I R C U Y F E R E N T I I L CRACKS NO S E N S I T I V I T Y AT CENTRE

EDDY CURRENTS

EOOV A .-.

CURRENTS

EDOV CURRENT FLOWS PARALLEL E:IOV CURRENTS CONCENTRITE WEIR THE TO TUBE SURFACE NOT S E N S I T I V E SURFICE CLOSE TO THE C O I L - DEPTH TO L A M I N A R SEPARATIONS. OF P E N E T R A l l O N I S COUTROLLEO BY TEST FREOUENCV

FIGURE 7.4 Directional Properties of Eddy Currents in Cylindrical Test Articles

In addition t o considerations of eddy current flow direction t h e following a re important:

- Magnetic flux is not bounded by t h e tube wall but will induce eddy currents in adjacent conducting material, e.g. tube support plates in hea t exchangers.

- Eddy current coils a r e sensitive t o ferromagnetic material introduced into a coil's magnetic field. The ferromagnetic material need not be an electrical conductor nor need i t form a closed path for eddy currents.

- Eddy currents coils a r e sensitive t o all material variations tha t affect conductivity or permeability.

7.2.4 Probe Inductance

The equations quoted in Section 5.2.3 t o calculate inductance for surface probes a r e also used t o calculate inductance of probes for testing tubes and cylinders. The important aspect of inductance is t ha t probe impedance, which is a function of inductance, must be compatible with eddy current instrument and signal cables,

where XL - 2 'IF f L when f is in hertz and L in henries and R is coil wire resistance in ohms.

TA

BL

E 7

.2

EN

CIR

CL

ING

OR

IN

TE

RN

AL

CO

IL I

MPE

DA

NC

E

Do =

8.9

mm

Do =

12.

7 m

m

Do =

15.

9 m

m

Do =

19.

1 m

m

Do =

22.

2 m

m

Wir

e S

ize

L =

6.1

Va

L= 11

UH

L =

15 un

L

=2

5 U

H

31 A

WG

L

= 2

0 ,,H

N

= 2

5 (0

.23

mm

) R

= 0

.3

R =

0.4

"

R =

0.5

n

R =

0.6

n

R =

0.7

n

L =

23

L =

42

L =

59

L =

77

L =

96

34 A

WG

N

= 4

9 (0

.16

mm

) R

=l

R

= 1

.5

R=

2

R=

2

R=

3

L =

64

L =

110

L

= 1

60

L =

210

L

= 2

60

37 A

WG

N

=8

1

(0.1

1 m

m)

R =

3

R=

5

R =

6

R=

8

R=

9

L =

200

L

= 36

0 L

= 5

10

L =

660

L =

830

39

AW

G

N =

144

(0

.089

mm

) R

=9

R

= 14

R= 18

R

= 2

2 R

= 26

L =

490

L

= 8

80

L =

1.2

4 m

H

L =

1.6

2 m

H

L =

2.02

mH

41 A

WG

N

= 2

25

(0.0

71 m

m)

R =

24

R =

35

R =

45

R =

55

R =

64

Most eddy current instruments will opera te over a fairly broad range of probe impedance without a substantial reduction in signal-to-noise ra t io or signal amplitude. An instrument input impedance of 100 ohms is typical, although a probe impedance between 20 and 200 ohms is normally acceptable, unless t h e test frequency is t o o close t o probe-cable resonance frequency, see Section 7.2.5. Exact probe inductance calculations a r e therefore no t essential.

To fac i l i t a t e impedance calculations Table 7.2 has been prepared. This t ab le lists coil inductance and resistance (with probe in air) fo r various diameters and wi re sizes while keeping coil cross section constant at 1.2 mm x 1.2 mm. (These dimensions a r e fair ly typical of t u b e wall thickness in h e a t exchangers). With t h e aid of th is table, and knowledge t h a t inductance is proportional t o t h e square of number of turns and

2-2 t h e square of mean coil d iameter ( L a N D ) , one can usually make a reasonable es t imate of wire s ize and number of turns for a particular probe.

Probe-Cable Resonance

Probe-cable resonance must b e considered when operating at high test frequencies and/or using long signal cables, e.g. frequencies over 100 kHz o r cables longer than 30 m. Most general purpose eddy cur ren t instruments cannot opera te at or close t o resonance.

Probe-cable resonance can be modelled as shown in Figure 4.5. In simple terms, resonance occurs when inductive reac tance of t h e coil eauals capacit ive reac tance of t h e cable, i.e. when

where w is angular frequency, radians/second L is coil inductance, henries C is to ta l cable capaci tance, f a rads

Transposing this equation and substi tuting w = 2nf shows frequency is

resonance occurs wh

This approach is sufficiently a c c u r a t e fo r most practical applications. A more rigorous approach t o resonance is presented in Section 4.3.

Resonance is apparent when a probe and cab le combination, which balances a t a low frequency, will not balance as frequency is increased. At t h e approach of resonance, t h e balance l ines on t h e eddy cur ren t s to rage monitor will not converge t o a null. The two balancing (X and R) controls will produce nearly parallel lines, ra ther than t h e normal perpendicular t races , on t h e s to rage monitor. A number of s teps can be taken t o avoid resonance:

1. Opera te at a test frequency below resonance, such tha t f t,, is less than 0.8 f . 2. Select a probe with lower inductance. (Since f , is proport~onal to 1/ fi ,

inductance mus t be decreased a fac to r of four t o double t h e resonant frequency).

3. Reduce cable length o r use a cable with lower capacitance per unit length (such as multi-coax cables). This will raise t h e resonance frequency since capacitance is proportional t o cable length and f , is proportional to 1 / 8 , 4

4.. Operate at a test frequency above resonance, such tha t f t, is greater than 1 . 2 f , . However, above resonance the sensitivity of all eddy current instruments decreases r a idly with increasing frequency because capacitive reactance (X, = l/wC P decreases, and current short circuits across the cable rather than passing through t h e coil.

7.3 IMPEDANCE PLANE DIAGRAMS

Eddy current probe for testing cylindrical components differ mechanically f rorn those for plate testing, but coil impedance can b e t reated similarly for both test coil configurations. The impedance display t reatment introduced in Chapter 5 applies for internal and external circumferential coils with t h e following changes:

i) Lift-off becomes "fill-factorn. Fill factor is a measure of coupling between the coil and test object. In general, i t is t h e fraction of magnetic field tha t crosses t he tes t object; for a long coil, this is t he fraction of the t e s t coil a rea filled with tes t material. Fill-factor, (eta), is t h e ratio

for an encircling coil,

and

(7. la )

for a bobbin type internal coil,

where -0 D is cylinder diameter D is average coil diameter

and D~ is tube internal diameter

Fill-factor is always a quantity less than or equal t o one ( q < 1 . 0 ) For a coil inside a tube the impedance change due t o decreasing is t h e s a m e as an increase in D (with constant wall thickness). For a coil around a tube o r cylinder, decreasing 0 is t h e same a s decreasing Do.

ii) Probe diameter in plate testing is replaced by tube o r cylinder diameter in ETof cylindrical components. They have a similar e f fec t on t h e operating point on the impedance diagram.

Figure 7.5 summarizes t h e e f f ec t of test and material variables on a simple semicircular impedance diagram. Note the similarity of changes in resistivity, tes t frequency, diameter and fill-factor with the surface probe results of Figures 5.9 t o 5.13.

FREDUENCY ( 1 ) and

NORMALIZED RESISTANCE

- -.-.-

THIN . WALL TUBE

FIGURE 7.5 Simplified Impedance Diagram of a Long Coil Around a Nan-magnetic Thin-wall

Tube Showing Effect of Test and Material Variables

Impedance diagrams presented in t h e l i terature a r e often only strictly valid for long coils (much longer than material thickness), coil lengths for inspection a r e normally only a fraction of material diameter. Decreasing coil length has an effect similar t o decreasing fill-factor, i t causes t he impedance diagram t o be smaller than expected (but similar in shape) from coil and tes t material geometry. Following sections will present impedance diagrams for tubes and solid cylinders. For simplicity a fill-factor of unity will be used.

7.3.1 Solid Cylinders

The impedance diagram for a solid cylinder (diameter, Do ) inside a long coil is shown in Figure 7.6. As in Figure 7.5 an increase in test frequency or diameter moves the operating point (the point on the impedance diagram tha t specifies the normalized inductive reactance and resistance of t h e test coil) down the curve while an increase in resistivity moves i t up the curve. This diagram applies t o both wires and round bars.

CREASING R E S I S T I V I T Y

DECREASING F ILL -FACTOR

FIGURE 7.6 Impedance Diagram for a Solid Cylinder

The shape of impedance diagrams for cylinders differ markedly from a semicircle, particularly at higher tes t frequencies. The shape difference is due t o skin effect and phase lag, factors which were not included in arriving at the semicircular shape in Chapter 3. At high t e s t frequencies t h e curve approaches the X and Y axes at 45'.

In testing cylinders with an encircling coil i t should be recognized t h a t sensitivity t o defects a t t h e centre of bar or wire is zero, regardless of tes t frequency. The reason for this is illustrated schematically in Figure 7.7 which shows plots of eddy current density across a cylinder. Defects have t o disrupt eddy current flow in order t o a f fec t probe impedance. I t is apparent from Figure 7.7 tha t eddy current density is always zero a t t he centre of a cylinder resulting in no sensitivity t o defects.

0 LOW FREOUENCI 8 > 2 4

FIGURE 7.7 Schematic of Eddy Current Distribution in a Cylinder

Surrounded by an Encircling Coil

7.3.1.1 Sensitivity in Centre of a Cylinder

I t was s tated in the previous section tha t eddy current density in the centre of a cylinder is zero and hence there is no sensitivity t o defects. The relationship of current flow with depth into a cylinder is derived (very approximately) below, for the case of no skin depth attenuation and long coils. From Faraday's Law,

The magnetic flux density, 8, is approximately constant inside a long coil, hence

@ = BA

= (B) (nr2) where r is radial distance from centre of cylinder;

therefore,

Resistance t o flow of current is proportional t o path length and resistivity and inversely proportional t o cross-sectional area, Ac,

2 n r p U s = = 2 n r p *c u n i t l e n g t h x u n i t d e p t h

Since by Ohm's Law

and Z = = Is a t low test irequency and no skin depth effect ,

theref ore,

Therefore, eddy current flow is proportional t o radial distance from centre of a cylinder. Hence no current flows at the centre (at r=O) and there is no sensitivity t o defects. I

7.3.2 Tubes

The impedance diagram for an extremely thin-wall tube with either an internal or external circumferential coil is a semicircle. This shape is only obtained when wall thickness, t, is much less than skin depth (t <<6 ) , i.e. skin effect and phase lag are negligible. This situation will rarely be encountered in practice, especially at intermediate and high test frequencies, but the concept is useful since i t defines one of the coil impedance limits.

With an external coil the other limit is defined by the impedance curve for a solid cylinder (maximum possible wall thickness). The impedance diagram for any tube tested with an external coil, hence, has t o lie between the two broken curves in Figure 7.8, for example the solid line applies t o a tube with internal diameter 80% of

ENCIRCLING COIL

nu u z - - u - ," w - > - CVLINOER ( O i = 0 ) + u = 0 = TUBE ( D , /OD = 0 8 ) - D .d - . = w 0 =

NORIALIZED R E S I S T Y C E

FIGURE 7.8 Impedance Diagram for a Tube with Encircling Coil Showing Effect

of Decreasing Wall Thickness

the outside diameter i.e., D i / D , - 0 . 8 . Tubes with D i / D , greater than 0.8 would lie to the right of the solid line. The dotted lines in Figure 7.8 trace the shift in operating point as wall thickness decreases (Do constant, D i increasing). Note the spiral shape of the wall thickness locus. The thick wall end of the curve deviates from a semicircle locus.

This is attributed to phase lag across the tube wall and forms the basis for eddy current signal analysis which will be treated in detail in Chapter 8.

Figure 7.8 also illustrates the dependence of the terms "thick-wall" and "thin-wall" on tes t frequency. Near the top of the diagram (low frequency) a tube with D l / D o ' 0 - 8 qualifies as thin wall, there is no phase lag across the tube wall, t <<6. Near the bottom (high frequency) the same tube becomes thick-wall because thickness becomes much greater than skin depth, for eddy current purposes the tube now appears as a solid cylinder.

When a tube is tested with an internal circumferential coil t he impedance diagram for a thin-wall tube remains semicircular but that for a thick-wall tube differs markedly from a solid cylinder; compare Figures 7.8 and 7.9. The impedance locus for

N O W 1 L I Z E P RESISTANCE

FIGURE 7.9 Impedance Diagram for a Tube With Internal Coil Showing Effect

of Decreasing Wall Thickness

any given tube will again fall between the dashed curves a t intermediate frequencies and approach the thin-wall curve a t low frequency and the thick-wall curve a t high frequency as shown for tubes with D i / D o = 0.8 and 0.9. As in the previous figure, a change in wall thickness produces a coil impedance change along the dotted lines tracing a spiral shaped curve. Again, this departure from a semicircle is attributed t o phase lag across the tube wall.

7.3.3 Characteristic Frequency for Tubes

Section 5.6 described how the Characteristic Parameter PC- f 2 w ~ ~ , introduced by Deeds and Dodd, enables presentation of the effects of changes in r, w , p and on a single impedance diagram. This allowed test coil impedance t o be specified in terms of a single quantity rather than four independent variables. One could use this

parameter in testing cylinders and tubes. However, most eddy current l i terature refers t o a similar variable, t he characteristic o r limit frequency, f usually attributed to Forster. I t differs from pc because probe radius, F, is replaced with tube or cylinder dimensions.

By definition, f is t he frequency for which the Bessel function solution, t o Maxwell% magnefic field equations for a f ini te test object, equals one. (Bessel functions a r e similar to, but more complex than trigonometric sine and cosine functions). For a solid cylinder or thick-wall tube tested with an encircling coil,

I 5 . 0 7 ~

fg 2 , kHz "rDo

with P in microhm-centimetres and Do in millimetres.

For a thick-wall tube with an internal coil,

f = g

s.oze , kHz 2 "rDi

For a thin-wall tube with internal or external circumferential coils,

The ratio f / f defines t he operating point on impedance diagrams. For non-

magnetic materials ( 1 , frequency rat io for cylinders and thick-wall tubes tested with external coils is given by

where f is t es t frequency in kilohertz.

For a thick-wall tube tested with an internal coil,

For thin-wall tubes tested with internal or external coils,

f/fg - f ~ ~ t l S . 0 7 p

0 .? 0 4 0 . 6

NORMAL l ZED RESl STANCE

FIGURE 7.10

5 0 1 1 0 CYLINDER ( E X T E 4 w L C O I L ) f / f , = f o o l / s or p

TWIN.@ALL TUBE ( I N T E R I A L I EXTERNAL C O I L S )

' / t o = f O , ' / S o1p

Impedance Diagrams f o r Tubes and Rods with Long Coils and Unity Fill-factor Showing Variations of f/f Along Impedance Loci

g

Figure 7.10 shows impedance diagrams fo r thin-wall tubes, solid cylinders and thick- wall tubes with values of f / f (from 0 t o infinity) on t h e curves. The impedance plots a r e both di f ferent in shape and have drastically di f ferent f / f ratios. For example, at t h e "knee" in t h e curves a thin-wall t u b e h a s f / f -1, for a cylinder f / f g=6 and a thick-wall t u b e has f / t g- 4 . These di f ferences or iginate in t h e defining equations which conta in 0 , * , D: and D t . T o find t h e operat in point on an impedance diagram using frequency ra t io one has t o know t h e geometry f t u b e o r cylinder). For tubes which do not sa t is fy t h e conditions fo r e i the r thin o r thick wall, calculations of f I f is not possible excep t near t h e t o p and bot tom of impedance diagrams where curves for in termediate wall tubes converge with t h e thin-and thick-wall curves, respectively.

In addition t o defining operating point, frequency r a t i o can also b e used for extrapolation o r scale modelling using t h e similari ty condition. This condition states if two objects have t h e s a m e f I f then eddy cur ren t distribution is identical in each. Hence if test frequency f 1 m e e t s test requirements for a r t i c le No. 1, one can calcula te f for a r t i c le No. 2 f rom t h e following:

For cylinders, 2 3

for thin-wall tubes,

f l D i l t l P 2 = f ~ ~ i 2 ~ 2 P l

and for thick-wall tubes (internal inspection),

7.3.4 Computer Generated Impedance Diagrams

As indicated in the previous section, exact analytical solutions (Bessel function solutions) for impedance loci of test coils around or inside tubes a re only possible for limiting cases. These solutions have the additional drawback tha t they are only strictly t rue for long coils. An alternative was made available by C.V. Dodd and his co-workers (2) at Oak Ridge National Laboratories. They developed computer programs t o calculate coil impedance. These are valid for all coil lengths, internal and external coils and all tube wall thicknesses. Such computer programs permit paper experiments t o define operating point a s well at the effect of variations in coil size and shape, resistivity, wall thickness and test frequency.

Figure 7.11 is an example of computer generated impedance display for a short internal coil in an Inconel 600 tube at various test frequencies. Fill-factor and the ef fec ts of small changes in resistivity ( A p ) , wall thickness ( ~t ) and magnetic permeability ( A p ) were examined a t each frequency. Note the similarity with the impedance plots of Figure 6.5 obtained for a surface probe. The angular (phase) separation between fill-factor, A p , A t and bu provides the basis for eddy current signal analysis which will be t reated in Chapter 8.

FIGURE 7.11 Computer Simulation of Probe Response to Various Test Parameters

7.4 CHOICE OF TEST FREQUENCY

Tes t frequency is o f ten t h e only variable over which t h e inspector has appreciable control. Material properties and geometry a r e normally f ixed and probe choice is o f t e n d ic ta ted by test material geometry and probe availability. Choice of a suitable test frequency depends on t h e type of inspection. Test ing for d iamete r variat ions normally requires maximum response t o fill-factor which occurs at high frequencies. Testing for de fec t s requires penetration t o possible de fec t locations; su r face d e f e c t s can be de tec ted at higher frequencies than subsurface defects. Maximum penetra t ion requires a low frequency which st i l l permits c lear discrimination between signals f rom harmless variations in material properties and serious defects. T h e above fac to rs show choice of test frequency is usually a compromise.

Test frequency fo r Solid Cylinders

As discussed in Section 7.3.1, t h e sensitivity at t h e c e n t r e of a cylinder, wi th a n encircling coil, i s zero at all test frequencies. Therefore, t h e r e is no advan tage in using a very low test frequency t o increase penetration.

'vlaximum test sensit ivity is obtained when t h e impedance diagram operat ing point is near t h e knee of t h e curve. This condition occurs when f I f ' 6 At this point balanced sensit ivity to defects , resistivity and dimensions is obtained. A t this test frequency, D ~ I ~ a 3.5. Increasing t h e frequency r a t i o f I f g t o 15 o r 20 improves discrimination between surface de fec t s and fill-factor variat ions (probe wobble), a t t h e expense of reduced sensit ivity t o subsurface defects. Maximum sensit ivity t o diameter variations is obtained at higher test frequencies, f / f = 100 or more.

A frequency ra t io lower than 6 will result in a decrease in phase l ag and the re fore less phase discrimination between defec t s and fill fac tor . T o distinguish between ferromagnet ic variations (or inclusions) and defects , t h e operating point should be on t o p quadrant of t h e impedance diagram. A frequency ra t io of approximately t w o ( € 1 f = 2) would achieve this.

7.4.2 Test Frequency fo r Tubes

When inspecting tubes for defects, t h e cr i ter ion to sat is fy is (a) phase discrirninatiorl between d e f e c t signals and o ther indications and (b) good phase separation between internal and external d e f e c t signals. A test frequency, proven in pract ice on many types and sizes of tubes, is t h e frequency f 90 which yields 90" phase separat ion between f ill-factor variations (and internal d e f e c t signals) and external d e f e c t signals. The frequency f go is empirically derived f rom t h e ra t io between thickness and skin depth, slightly larger than one,

and conver ts t o

fgo - 3 p / t 2 kilohertz (7 .4)

where p i s resist ivity in microhm-centimetres and t is tube wall thickness in milli- metres. This equation is valid for both in ternal and external coil inspection and is roughly independent of t u b e diameter. A t f 90 , t h e r e is good sensit ivity t o internal

and external de fec t s and l i t t l e sensitivity t o magnet i te deposit and ferromagnetic support plates.

The character is i tc frequency ra t io f I f cannot b e used t o satisfy t h e cr i ter ion of phase discrimination, because t h e f equation is no t a function of phase lag. I t would a lso b e wrong t o use i t for de fec t detect ion because i t is a function of tube diameter. The l a t t e r would require d i f fe ren t test frequencies for d i f ferent d iameter tubes t o keep f 1 f constant.

If one desires t o distinguish ferromagnet ic signals f rom o ther indications, t h e operating point should be on t h e t o p quadrant of t h e impedance diagram for thin-wall tubing, Figure 7.10. This point is located by calculating t h e test frequency t o make t h e charac te r i s t i c frequency ra t io equal to o r less than 0 . 5 ( f / f 5 0 . 5 ) . Inspection Standards and Specifications

A number of industrial codes cover eddy cur ren t t u b e inspection. The various ASTM specifications a r e E-215 (aluminum alloys), E-243 (copper and copper alloys), E-426 (stainless s teels) and E-57 1 (nickel alloys). None of t h e ASTM standards specify t e s t frequencies, they somet imes present normal ranges such as 1 t o 125 kHz fo r aluminum alloys. Such numbers a r e of l i t t l e use in deciding on a suitable test frequency for a part icular test. The ASME Boiler and Pressure Vessel Code, Section V, Article 8 (1980) specifies t e s t frequency in t e r m s of t h e angle between through- wall and external de fec t indications f rom a calibration tube. The procedure specified will normally yield a frequency higher than f g o , perhaps as high as 2 f g b

Most calibration tubes consist of drilled holes of various diameters and/or various depths f rom t h e external surface. Some calibration tubes have EDM (electric discharge machining) notches in t h e c i rcumferent ia l and axial directions and on both internal and external surfaces.

PROBES FOR DETECTING CIRCUMFERENTIAL CRACKS

A conventional internal circumferential (bobbin) probe induces a flow of eddy cur ren t s parallel t o t h e coil windings and the re fore circumferential in direction (Figure 7.4). T o sense a defect , coil impedance must change; th is will occur only if t h e eddy current flow path is disturbed. Circumferent ia l de fec t s parallel t o th is current , which present no a rea perpendicular t o th i s path, will therefore not b e sensed.

FIGURE 7.12 (a) Probe No. 1 - Multi-pancake Coil Probe (b) Probe No. 2 - Zig-zag Coil Probe

To de tec t circumferential defects the coil must induce currents at an angle t o t he cracks. Two possible types of probes a r e (a) surface probes and (b) zig-zag probes. Figure 7.12 shows examples of such probes. The surface probe induces currents in a - circular pat tern whereas t h e zig-zag probe induces currents t o follow the 30" coil angle. The probes shown in Figure 7.12 a r e differential. In t h e surface probe configuration a multi-coil array is used; t h e four surface coils in each row a r e connected in series and the two rows a r e connected differentially. A single absolute surface coil can also be used, provided the probe maintains contact with the tube surface by spring force or other means (otherwise lift-off noise would be intolerable). See Figure 7.13 for t h e cross section of a typical spring-loaded internal probe for tube testing.

CABLE

CONNECTOR

REFERENCE S P R I N G C O I L

FIGURE 7.13 Spring Loaded Internal Surface Probe for Tube Inspections

A single surface probe is unquestionably t h e easiest t o use; signal analysis is discussed in Chapter 6. The main disadvantage is t he partial circumferential coverage; multiple passes or helical scanning a r e necessary for 100% coverage. Another disadvantage of the surface probe configuration (single or multiple) is the loss of sensitivity with distance from t h e coil. If surface coils a r e small, a s will be the case for most tube inspections, t he reduction in sensitivity with distance from the surface will be greater than with circumferential coils, see Section 5.3.1. The sensitivity t o small localized defects originating frorn t h e outside surface could be a s much a s 10 times lower than the sensitivity t o internal defects. A zig-zag coil has less attenuation t o outside defects, i t falls into t h e circumferential class in this respect. Neither zig-zag nor surface coil probes will give uniform sensitivity around their circumference. There will be peaks of maximum and minimum sensitivity depending on t h e angle between eddy current path and defect orientation. This can best be visualized by considering a short circumferential crack passing over the coils: there will be areas, such as at the peaks of t h e zig-zag, where eddy current flow is almost parallel t o t he crack, resulting in poor sensitivity.

Figure 7.14 shows examples of signal response t o real circumferential fatigue cracks with the probes discussed above.

( I ) MULTl -PANCAKE COl L PROBE

( b ) 216-ZAG COIL PROBE +

+y

( c ) BOBBIN m l L ROBE +

* - Y

FIGURE 7.14 Eddy Current Scans of Circumferential Cracks in Inconel Tubing

(Signal Amplitude Normalized to a 1.6 mm Diameter Through Hole). f = 400 kHz.

7.6 SUMMARY

Test coils induce eddy currents and also sense the distortion of their flow caused by defects. Encircling or bobbin probes have test coil(s1 mounted with their axes parallel t o t h e tube or rod axis. Since the coils a r e wound circumferentially the induced eddy currents also flow circumferentially. They cannot b e used t o detect circumferential cracks, laminar defects, nor defects in the center of a rod.

As a general guideline for tube inspection, probe coil length, depth, and spacing (if differential) should approximately equal wall thickness.

An absolute bobbin probe (single test coil) should b e used fo r general in-service h e a t exchanger inspection. However, f o r shor t localized defects , d i f ferent ia l probes ( two test coils side-by-side) are normally preferred. - Analysis of eddy cur ren t signals is t h e most important and unfortunately t h e mos t difficult task in a successful inspection. A thorough understanding of impedance diagrams and e f f e c t of phase lag is needed t o manipulate test conditions t o minimize undesirable test variables. The Charac te r i s t i c Frequency for t u b e inspection is used to loca te t h e operat ing point on t h e impedance diagram. I t is given by

kHz

where p is e lect r ica l resistivity and D is t u b e internal d iameter (for bobbin probe) and ex te rna l d iamete r (far encircling probe); t is tube wall thickness.

One needs t o know t h e operat ing point on t h e impedance diagram t o de te rmine ef Eects of f ill-factor, e l ec t r i ca l resistivity, and magnet ic permeability. The optimum sensitivity t o f ill-factor is near t h e bot tom of t h e impedance diagram, in t h e rniddle for e lec t r i ca l resistivity and at t h e t o p for magnet ic permeability.

'When inspecting tubes for defects , c r i t e r i a t o sa t is fy a r e (a) phase discrimination be tween d e f e c t signals and o ther indications and (b) good phase separation between internal and external d e f e c t signals. For general purpose tes t ing t h e frequency given by

kHz

1

is used where t is wall thickness in mm. This frequency yields 90" phase separat ion between internal and external d e f e c t signals and l i t t l e sensit ivity t o magnet ic deposits and ferromagnet ic support plates.

Special probes a r e needed t o inspect for c i rcumferent ia l c racks o r d e f e c t s c lose t o tubesheets. Single, spring loaded, su r face probes a r e effective.

7.7 WORKED EXAMPLES

7.7.1 PROBLEM: Calcula te frequency t o o p e r a t e at t h e knee location of t h e impedance diagram for a cylinder 5 mm in d iamete r and e lec t r i ca l resist ivity p = 1 0 rnicrohm-centimetres.

SOLUTION: f D& O - 6

f l f l 5 . 0 7 p

the re fore f 6 x 5 . 0 7 x 10

5 2 = 12 kHz

7.7.2 (a) PROBLEM:

L

SOLUTION:

7.7.2 (b) PROBLEM:

SOLUTION:

7.7.2 (c) PROBLEM:

SOLUTION:

Calculate the tes t frequency t o inspect Inconel 600 tubing with D i - 1 0 . 2 mm, t = 1.1 mm a n d p - 98rnicrohm- centimetres.

Best tes t results a r e obtained when there is sufficient phase separation between internal and external defect signals. A phase separation of 90" allows good discrimination between the two and reasonable defect depth estimates. To achieve 90" phase separation, the test frequency is determined by

(derived from t / 6 = 1.1)

= 3 x 9 8 = 245 kHz

( 1 . 1 1 ~

Therefore 245 kHz is the required frequency.

Determine the approximate operating point on the impedance diagram, for problem (a).

Since ' 1- 1 this tube cannot be considered thick or thin walled. Therefore, neither equation 7.2(b) nor 7.2(c) is strictly valid. However, for t / 6 > 1, equation 7.2(c) for thick-wall tubing will yield an approximate solution.

2 f / f g = fDi/5.07 P (7.3~) 2

= 245 x (10.2) 1 5 . 0 7 . ~ 98

This would place the operating point on the lower quadrant (much lower than the knee location) of t he thick-wall curve of Figure 7.10.

Calculate a test frequency for t h e above tube suitable for discriminating between ferromagnetic inclusions and defects, when testing with an internal probe.

The operating point should be on the top quadrant of t he impedance diagram for thin-wall tubing, Figure 7.10. This point is located by calculating t h e test frequency to make the ratio of Forster's characteristic frequency equal t o or less than 0.5.

= 0.5

therefore

f (0.5) (5.07p)/Dit

= 0.5 x 5.07 x 98/10.2 x 1.1 = 22 kHz

Therefore, at 22 kHz (9% of f ) , i t should be possible to 93 discriminate between defects an ferromagnetic indications.

CHAPTER 8 - TUBE TESTING - SIGNAL ANALYSIS

8.1 INTRODUCTION

Manufacturing and in-service inspection of tubes is one of the most important applications of eddy current testing. For in-service inspection of small-bore tubing in particular, eddy current is by f a r the most frequently used method. Access is usually limited t o tube ends which makes other NDT techniques difficult or impossible t o apply.

This chapter emphasizes in-service testing of tubes using internal probes. This approach is taken because testing of solid cylinders and tubes with external coils (manufacturing inspection) is generally less complicated. If t he reader understands in- service inspection he should encounter no problem applying similar principles t o other tes t situations.

Reasons for the appearance of impedance plane eddy current signals a re presented first. Repetition from previous chapters is intentional, i t was desired t o keep this chapter as independent as possible without excessive cross-referencing. Discussion of simple defect indications is followed by superimposed signals which a re frequently encountered during in-service inspection such as defects a t baffle plates and tubesheets. A section dealing with surface probe internal tube inspection is included, difficult test situations have been resolved with this technique. Signals which could be mistaken for real defects (anomalous indications) a r e the subject of another section. The chapter concludes with a discussion of multifrequency testing, including its advantages and limitations.

An a t tempt is made throughout this chapter t o illustrate discussion with real or simulated eddy current defect signals.

8.2 EDDY CURRENT SIGNALS

8.2.1 Defect Signal Characteristics

A defect, which disrupts eddy current flow, changes test coil impedance as the coil is scanned past the defect. A non-rigorous derivation of this effect can be obtained using Figure 8.1 which portrays eddy currents induced in a tube with either an internal or external coil. Consider a unit length of tube a s being the secondary winding of a transformer (similar t o treatment in Chapter 3). The resistance of a conductor of length a , cross-sectional area A and resistivity P is

R = P , p / A , ohms

Without a defect, resistance around this tube is

R~ = 2 n T p l t @.la)

Introduction of a long defect, of depth h, which constricts eddy current flow over the distance (in radians), increases total resistance t o

- 124-

or R = R, (defect f r e e resistance) + A R (due t o defect).

FIGURE 8.1 Schematic Illustration of Eddy Cur r a t Distribution

A r o d a Defect in a Tube

A shor t de fec t will also increase res is tance but by a smaller AR s ince current can flow both under and around it. Note t h a t i t i s width of a f f e c t e d zone, A8 , ra the r than ac tua l de fec t width which determines e f f e c t of t h e de fec t on resistance. In summary, t h e above argument i l lustrates t h a t d e f e c t length, dep th and width ( to some ex ten t ) all increase res is tance to current flow and hence defec t signal amplitude.

In t e r m s of t h e equivalent coil c i rcui t of a resistor in parallel with a n inductor and i t s associated semicircular impedance diagram (Chapter 3), a d e f e c t moves t h e operat ing point up t h e impedance diagram. Increasing res is tance in a specimen changes both probe inductance and resistance.

T h e above discussion does not predict a defec t signal in detail , only i t s approximate amplitude and direction on t h e impedance diagram. A more comple te explanation requires inclusion of phase lag. Consider a n absolute coil around a cylindrical sample as in Figure 8.2(a). (The t r e a t m e n t for a di f ferent ia l coil would be similar but more complicated because t h e twin coil configuration genera tes t w o mirror image signals and cross-coupling between t h e t w o coils causes fu r the r complications). Figure 8.2(a) shows t h e distribution of magnet ic field ampl i tude and phase around a coil as derived by Dodd (2). The solid lines a r e contours of constant magne t ic f ield strength; t h e dashed lines a r e constant phase. Since magnet ic f ield and induced eddy currents have about t h e same phase, t h e dashed lines also represent t h e phase of t h e eddy

currents. Similar diagrams could be derived fo r coils inside or around tubes. Amplitude drops off exponentially with dis tance and eddy current flow increasingly lags in phase (relat ive t o eddy currents adjacent t o t h e coil) both with depth and with axial distance f rom t h e coil. Skin depth e f f e c t occurs in both radial and axial directions.

Figure 8.2(a) permits derivation of eddy cur ren t signals fo r t h e surface, subsurface and deep defects illustrated. One needs to establish a reference phase direction as s tar t ing point, t h e fill-factor direction is convenient and can be defined as t h e signal resulting from a very shallow sur face d e f e c t which only decreases coupling without changing phase l ag distribution. Hence choosing t h e phase contour which just touches t h e surface under t h e coil as t h e 0" contour f ixes fill-factor direction as in Figure 8.2(b). T h e signal o r e f f e c t of de fec t s can be imagined as t h e absence of eddy currents which were flowing in t h e a r e a before t h e de fec t existed at this location. On moving t h e coil (or defects pas t t h e coil) f rom positions 0 t o 5 in Figure 8.2(a), one observes t h e change in amplitude and phase sketched in Figure 8.2(b). This procedure is reasonably s t ra ight forward f o r t h e su r face and subsurface defects since they a r e localized and only in tersect one phase and amplitude contour at any given position. For t h e deep defec t , o n e has to divide t h e de fec t in to sections and determine weighted average values fo r amplitude and phase at e a c h position.

The surface de fec t in Figure 8.2(b) has a l a rge fill-factor component, primarily i t s approach signal makes i t distinguishable f rom fill-factor. As defec t depth increases, signals r o t a t e clockwise due t o increasing phase angle.

The angle between fill-factor and defec t signals in Figure 8.2(b) is about 2 0 , where 0 - x / 6. Although probably not s t r ic t ly true, one can imagine defec t signal

phase angle as t h e sum of a lag of 8 f rom t h e coil t o t h e de fec t and t h e s a m e lag back t o t h e coil.

Ef fec t of Tes t Frequency

W e can now combine Figure 8.2 results with impedance diagrams from Chapter 7 t o i l lustrate t h e e f f e c t of test f requency on d e f e c t signal appearance. Figure 8.3(a) shows par t of Figure 7.9, t h e impedance diagram for a t u b e with D I / Do = 0.8 tes ted with a shor t internal coil. The do t ted lines t r a c e t h e impedance change with decreasing Do. An external d e f e c t (OD defec t ) in a t u b e is essentially a decrease in D, with D held constant, the re fore t h e d o t t e d lines t r a c e t h e change in impedance a s a coil is scanned past an OD defect . Note the similari ty between t h e subsurface defect in Figure 8.2(b) and t h e O D d e f e c t at 2 f go in Figure 8.3(a). The display is normally ro ta ted counter-clockwise to make a signal from fill-factor approximately horizontal. This i s achieved by rota t ing t h e phase control knob on t h e eddy current instrument.

\ S U B S U R F A C E D E F E C T 1 1 2 1

F I L L - t l C T O l . 1

FIGURE 8.2 Derivation of Eddy Current Signal Appearance f w Three Types of Defects

With this phase se t t ing and at f gg a n OD d e f e c t shows wall loss (tY) in a tube without a chan e in fill-factor as in Figure 8.3(b). An ID d e f e c t consists of wall loss (+Y component f as well as a large fill-factor (-X component) because of decreased coil/tube coupling. The through-wall de fec t (hole) signal contains e lements of both ID and O D defec t s and hence yields a signal which fa l ls between t h e two. Note t h a t - all de fec t signals must fall between decreasing fill-factor and O D defec t signals.

IIWMPL IZED RESI SIIINCE

FIGURE 8.3(a) Relation Between Impedance Diagram and Defect Signal Appearance

OD DEFECT THROUGH-WALL \

DEFECT

I D DEFECT

DECREASING F I L L FACTOR - Y

FIGURE 8.3 (b) Defect Signal Appearance at f gg

Figures 8.3(a) and 8.4 show what happens t o de fec t signals with changing test frequency. Reduced frequency results in rota t ion of d e f e c t signals towards t h e fill- f a c t o r direction. At very low frequencies (less than f go 1 4 ) signals f rom d i f fe ren t types of de fec t s become difficult t o distinguish due t o small phase angle separation.

Increasing t es t frequency increases phase separation between ID and OD d e f e c t signals as predicted by phase lag. A t f go the ID and OD defect signals a r e separated by about 90" with low sensitivity t o tube supports and external deposits. At higher t e s t frequencies, 2 fgo and above, higher sensitivity t o probe wobble and dents i s obtained and t h e increased angular separation of de fec t signals makes i t diff icult t o discriminate between OD defec t s and probe wobble o r fill-factor variations, see Figure 8.4k).

PROBE PROBE WOBBLE

PROBE WOBBLE

FIGURE 8.4 Appearance of Calibration Defect Signals at Di f fe ren t Test Frequencies

8.2.3 Calibration Tubes and Simple Defec t s

Both manufacturing and in-service inspection require calibration tubes with ar t i f ic ia l de fec t s for initial instrument set-up and subsequent signal analysis and interpretation. These tubes should be identical in mater ia l and s ize t o tubes t o be tested. Minimum calibration requirements include ID, OD and through-wall de fec t s (see also t h e ASTM and ASME codes c i t ed in Section 7.4.2). For in-service inspection, expected signal sources such as baffle plates, magne t i t e deposits and dents a r e useful and o f ten essential for reliable signal analysis. Figure 8.5 shows typical signals, at f g 0 , from a calibration tube suitable for in-service h e a t exchanger inspection. Both absolute and differential probe signals a r e shown. T h e 90" phase separation between ID and OD defec t s also exists fo r differential probes. Note t h e similari ty with t h e signals derived in t h e previous section.

STEEL SUPPORT PLATE

, . . ...

OUTSIDE l NS l DE THROUGH MAGNETITE GROOVE GROOVE HOLE

I I

OUTS l DE

I E C R E A S I N G F I L L FACTOR -

MAGNETITE

SUPPORT PLATE

HO!E O U T S I D E

D I F F E R E N T I A L

FIGURE 8.5 Eddy Current Signals from a Typical Calibration Tube.

Test Frequency fgo = 250 kHz.

Qualitat ive reasons for t h e appearance of ID, OD and through-wall defects w e r e presented in Section 8.2.2. The other signals in Figure 8.5 c a n be explained in a similar fashion. Magnetite is fer romagnet ic non-conductor, i t s signal i s due t o i t s high permeability. As indicated in Figure 7.1 1 increasing permeability of tube mater ia l yields a signal which falls between OD and through-wall defects. The magne t i t e signal in Figure 8.5(b) i s essentially such a signal ro ta ted about 90" clockwise because of phase lag across t h e tube wall. A den t places t u b e mater ia l in close proximity t o t h e coil resulting in improved coupling (increased fill-factor) and hence yields a signal opposite t o decreasing fill-factor. Probe wobble yields a signal very close t o t h e fill- f a c t o r direction because radial displacernent of t h e coil reduces t h e coupling to t h e tube. The reason for baff le p la te signal appearance is due t o a combination of factors. For carbon s teel baffles, t h e e f f e c t s of high magnetic permeability and in termediate resistivity partially cancel result ing in small signal amplitude. Phase lag across t h e tube wall ro ta tes th is signal clockwise.

S t 10 GROOVE 1 0 % 0 0 GROOVE 1 d nm 0.15 m

2.5 nm W I D E 2 , s IMI l l D E HOLE DENT

CARBON STEEL SUPPORT

7 I I

C t

Y CHANNEL ' I -

FIGURE 8.6 Appearance of Quadrature Components on a Chart

R e a d i n g for a Calibration Tube

In eddy current tube tes t ing one normally records t h e quadrature components (vertical, Y; horizontal, X) of coil impedance on a two-channel s t r ip char t recorder as shown in Figure 8.6. With phase adjusted as shown, any real de fec t will exhibit a Y component. T h e X-channel information is required f o r deta i led signal analysis to decide type and depth of de fec t s which can only be performed reliably through phase analysis. Accurate phase analysis can be done on-line by monitoring t h e signals on an eddy current instrument s torage monitor. Alternatively a n X-Y recorder o r similar device permits hard-copy s to rage of quadrature signals.

A f l aw indication on an X-Y monitor is normally a curved locus; i t does not have a simple and unique phase angle. If a n absolute probe is used t h e significant angle t o measure is t h e tangent angle at t h e de fec t signal t ip, see Figure 8.7(b). If a di f ferent ia l probe is used, t h e phase angle is t h e slope of t h e s t ra ight l ine joining t h e end points of t h e "figure-8" signal, see Figure 8.7(c). Figure 8.7(a) i l lustrates t h e change in phase angle with de fec t depth. This curve should be used only as a guide since de fec t signal phase angle can change with d e f e c t and probe geometry.

0 . 0 . DEFECT l .D. 0 .D. DEFECTS I DEFECTS THROUGH I

( b ) ABSOLUTE THROUGH

S l GNAL PATTERN PHASE ANGLE (0) , DEGREES

FIGURE 8.7 Eddy Current Phax AngIelDefect Depth Calibration Curve at fqo

When an eddy current signal source is located i t is o f t en useful t o r e t e s t at other frequencies t o confirm a defec t exists and/or to improve depth es t imate . Defect depth is es t imated f rom signal pa t t e rn recognition and verified by comparison with calibration defect signals at various test frequencies. Normally, frequencies of one- half and twice f g o a r e sufficient . However, t o check fo r magnet ic deposits or inclusions a frequency of one-tenth f o r less may be required (see Sections 7.4.2 and 8.3.1). Figure 8.4 shows e f f e c t of changes in frequency on calibration signals. Increasing t e s t frequency increases phase separation between ID and OD defec t s as predicted by phase lag. I t also increases sensit ivity t o probe wobble and dents but lowers sensitivity t o tube supports and external deposits. O n e might question t h e validity of comparing machined holes and grooves in calibration tubes with real de fec t s t o es t imate type and depth. The following examples justify this approach.

5 0 0 ~ 3 . 2 rnm CONCEPJTQ I C H O L E - GQOOVE

H O L E

C A L I B R A T I ON

D E F E C T S

FIGURE 8.9 Stress Corrosion Cracking in Type 316 Stainless Steel Tubing

(Do = 19.1 mm, t = 1.8 mm, fgO = 68 kHz)

8.2.4 Vectorial Addition and Defects a t Baffle Plates

During in-service inspection of tubes in heat exchangers, tube supports (baffle plates) a r e frequently defect prone regions. Inspection for defects at baffles is possible because eddy current signals a r e often vectorially additive. This permits analysis of superimposed signals; the signals can be (mentally, or graphically) subtracted from the total indication with resultant separated signals appearing similar t o calibration defects. Vectorial addition provides the basis for multifrequency eddy current testing (Section 8.4).

Figure 8.10 illustrates how signals from a steel baffle plate and an external groove a re added t o obtain a superimposed indication. The difference between the end points of the baffle plate and baffle and groove signals equals the indication obtained from the groove by itself.

I OD GROOVE

FIGURE 8.10 Vectorial Addition of Eddy Current Signals

Figure 8.1 1(a) shows a sect ion of stainless s t ee l tube removed from a power plant h e a t exchanger with par t of t h e carbon s teel support p la te still in place. The support w

shows considerable corrosion; originally t h e r e was about 0.25 mm clearance between t h e tube and t h e hole in t h e plate. Corrosion products have completely filled t h e gap leading t o crevice corrosion evident in Figure 8.1 1(b) which is a similar tube with t h e pla te removed. Calibration signals a r e presented in Figure 8. I l k ) . The eddy current signal from t h e baffle p la te region of Figure 8.1 l(a) is shown in Figure 8.1 l(d). This seemingly simple signal i s actually qui te complex. T h e upward component is due t o external pitting similar t o t h a t in Figure 8.1 1(b). The presence of a support p la te should result in -X, -Y signal components; in f a c t a + X deflection is observed. This is t h e result of denting of t h e tube. Denting is c i rcumferent ia l constriction of tubes d u e to compressive s t resses exer ted by baff le p la te corrosion products such as magneti te. The presence of magne t i t e can also contr ibute t o signal distortion particularly at low test frequencies. Tube denting is of concern because, in addition t o complicating eddy current signal analysis, i t c a n lead t o fur ther tube damage such as s t ress corrosion cracking o r thermal fa t igue because tubes a r e no longer f r e e t o expand and con t rac t during the rmal cycling.

FIGURE 8.1 1 Corrosion and Denting Under a Steel Baffle Plate

(Do = 15.9 mm, t = 1.25 mm, fm = 80 kHz)

Another example of defects near a carbon s teel tube support is shown in Figure 8.12. These were obtained from a brass, thermal power plant condenser tube which suffered erosion/corrosion on either side of supports. This is the same tube as in Figure 7.3. Defect signals from the baffle plate vicinity a r e so large the support signal is obscured. The main point of this example is t h e advantage of using phase angle, rather than amplitude, t o judge defect severity. Defect B with both differential and absolute probes has a phase angle approaching tha t of a through-wall hole, i.e., i t probably extends at least 75% through t h e wall. Defect A on the other hand is vertical and hence is probably no deeper than 50% even though i t exhibits greater amplitude than B.

- -- L X P L N S I O N S IGNAL

FIGURE 8.14 Schematic of Tube Geometry at Rolled Joint in Tubesheet

and Associated Eddy Current Signals

The end of t h e rolled joint at t h e inboard edge of a tubesheet is a defec t prone a r e a because of high residual and service s t resses and also because deposits t end t o accumulate at this location which can lead t o corrosion. Eddy current indications with bobbin-type probes f rom defec t s in this region can be difficult t o in terpret because of excessive signal distortion from t u b e expansion. Sensitivity may be improved by employing a spring loaded sur face probe as discussed in next section.

5.2.6 Testing Tubes with Internal Surface Probes

During in-service inspection of tubes, si tuations a r i se where conventional circumferential probes (both differential and absolute) prove inadequate. The case of circumferential cracks was t r ea ted in Section 7.5. Surface probe designs have also been found t o yield improved test results in t h e case of de fec t s at non-magnetic baffle plates and at hea t exchanger tubesheets.

Surface probes have several advantages over bobbin-type probes. They can be made much smaller than tube diameter and hence sample a smaller volume of tube periphery, th is provides inherently g rea te r sensit ivity to small defects. Spring loading of a surface probe against t h e tube wall e l iminates much of t h e fill-factor (lift-off) distortion caused by t u b e expansion in tubesheets. T h e main drawback t o su r face probe tube tes t ing is t h a t a number of scans have t o be made for complete

circumferential coverage. Conventional probes sample t h e en t i re tube in a single scan.

TUBE SHEET

70% OD DEFECT

1 . 6 mm HOLE\

t ' 70% OD DEFECT

A

1 . 6 mm

TUBESHEET

CONVENTIONAL PROBE SURFACE

PROBE

FIGURE 8.15 Comparison of Eddy Current Test Results in Heat Exchanger Tubesheet

Region with Conventional and Surface Probes

(Do = 12.5 mm, t = 1.2 mm, fgo = 200 kHz)

Figure 8.15 i l lustrates surface probe tes t ing at t h e tubesheet region of a power plant s t eam generator. I t compares signals, f rom what is believed t o be OD corrosion damage at t h e end of t h e rolled joint, obtained with conventional and surface probes. The reason for t h e character is t ic A'B'C' sur face probe signal i s as follows. As t h e probe is withdrawn from t h e t u b e (direction of arrow) it encounters t h e s t a r t of t h e expanded area. Failure of t h e probe to follow this contour exactly results in an increasing lift-off signal, A'R', superimposed on t h e impedance change, A'C', due t o t h e presence of t h e tubesheet. Both d e f e c t signals were obtained f rom t h e same tube, no te t h e considerable improvement in sensit ivity obtained with t h e surface probe. This tube was in f a c t leaking.

50% OD ECCENTRIC GROOVE

GAP

CALI RR4TIOIJ -

50% GROOVC

BAFFLE (MAXIMUM GAP) + 50% GROOVE

BAFFLE (NO GAP)

FIGURE 8.16 Internal Surface Probe Testing for Fretting Wear mder a

Non-Magnetic Baffle Plate. (Compare with Figure 8.13 Results)

A second example of improved sensit ivity with a n in ternal surface probe involves f re t t ing wear under non-magnetic baffle plates. Figure 8.16 shows results. Compare with Figure 8.13(a) which shows test results for t h e s a m e d e f e c t obtained with an internal circumferential probe. With no gap, t h e 50% groove was barely detectable with a conventional probe, while Figure 8.16 shows this d e f e c t is easily de tec ted with a surface probe.

ANOMALOUS EDDY CURRENT SIGNALS

Some eddy current signals can be mistaken fo r d e f e c t indications; these a r e called fa lse or anomalous signals. They a r i se because of t h e high sensitivity of eddy currents to many variables and demons t ra te t h e need f o r thorough analysis before concluding t h a t every eddy current signal represents a defec t . The following examples i l lus t ra te more common ones which have been encountered in practice.

8.3.1 Ferromagnet ic Inclusions and Deposits

Materials with re la t ive magnetic permeabil i ty g r e a t e r than 1.0 a f f e c t eddy current response drastically. Skin depth and probe inductance a r e both a f fec ted by permeability; permeability values of 50 to several hundred a r e typical.

Before ci t ing specific examples consider t h e general approach t o identifying signals f rom magnet ic materials. Such signals can b e distinguished f rom real de fec t s by reducing test frequency t o move t h e operat ing point near t h e t o p of t h e impedance diagram. Figure 8.17 i l lustrates t h e procedure where 1, 2 and 3 represent ferromagnetic mater ia l on t h e inside, in t h e t u b e wall and on t h e outside respectively. I t may be difficult to achieve a sufficiently high operating point with some instruments and probes when tes t ing low resistivity, large diameter tubes. However, if a low enough frequency is achieved, real d e f e c t indications will fall nearly parallel t o fill-factor whereas high permeabil i ty indications are nearly perpendicular t o fill-factor. At 240 kHz ( f ) in Figure 8.17, 1 and 2 could easily have been mistaken for ID defects. There is no confusion at 10 kHz since i t i s known t h a t all de fec t indications must fall between fill-factor and a n OD defec t signal. The following two examples demonstra te t h e procedure t o discriminate fa lse de fec t (ferromagnetic) indications.

FERROPAGNETI C

ANOMALIES

I I I.lmn GROOYC GROOVE I

NORMAL 1 ZED RESISTANCE. % C,

J @ 1 . 0 .

OECREASlNG F I L L FACTOR

OW DECREASING F I L L FACTOR @

FIGURE 8.17 Coil Impedance Display at Two Test Frequencies

Ferromagnet ic inclusions a r e occasionally encountered during eddy cur ren t tes t ing of non-magnetic materials . These ar ise f rom chips o r filings f rom s tee l tooling and handling equipment which a r e embedded during manufacture. The sur face of nominally non-magnetic stainless s t ee l s and nickel-base alloys can also become magnet ic as a result of cold working o r through alloy depletion f rom oxidation o r corrosion.

0.0. DEFECT I .D. DEFECT m -/'

250 kHz FERROMAGNETIC f

INCLUSION

\ FERROMAGNETIC INCLUSION

50 kHz

INCLUSION

- 10 kHz

FIGURE 8.18 Defect and Magnetic Inclusion Signals Obtained from a New Inconel 600 Tube (Do = 13 mm, t = 1.1 mm) with an Absolute External Coil.

f go = 250 kHz

Though one might consider a magnet ic inclusion a defec t , t h e r e a r e several reasons why it is important t o identify t h e origin of a n indication. Even very small , perhaps insignificant, magnet ic inclusions can yield s izeable eddy current signals because of t h e e x t r e m e sensitivity t o magnet ic permeability. A second reason t o determine defec t origin is so measures can be t aken t o minimize fu r the r damage; magnet ic inclusions a r e nearly always manufacturing defects . Figure 8.18 shows t h e signal from a magnet ic inclusion in new Inconel 600 tubing at various test frequencies. These results were obtained with a n external encircling probe; this explains t h e reversal in appearance of ID and OD defec t s f rom previous examples. The magnet ic inclusion yields a signal whose angular separation f rom t h e fill-factor direction increases as test frequency is reduced. The response of real de fec t s is just opposite.

I 0 .0 . DEFECT

1 . 0 . DEFECT T l NTERNAL

/

MAGNET l TE

250 kHz

MAGNET l TE

50 kHz

10 k H z

MAGNETITE '

FIGURE 8.19 Defect and Magnetite Signals from an Inconel 600 Tube

(Do = 13 mm, t = 1.1 mm) Obtained with an Absolute Internal Probe. fgO = 250 kHz)

Figure 8.19 shows eddy current response t o magnet i te ( Fe O4 ) deposits inside a n Inconel 600 t u b e at various test frequencies. As in t h e previous example, t h e existence of ferromagnetic mater ia l is verif ied by lowering test frequency; magnet i te signals r o t a t e clockwise whereas de fec t signals r o t a t e counter-clockwise. One could easily mistake t h e magnet i te signals for rea l de fec t s at 250 kHz and 50 kHz. Reducing test frequency can also be used t o verify t h e presence of magnet i te on t h e outside of a tube. This approach has been used to measure t h e height of sludge deposits (containing magnet i te) above tubesheets during in-service inspection of vert ical heat exchangers.

Figure 8.20 shows t h e eddy cur ren t signals from a Monel 400 s t e a m generator tube with external wall thinning near a tube support. The tube was inspected with a n absolute sa tura t ion probe and t h e signals recorded with wall thinning giving a vertically upward signal. A t 50 kHz t h e ver t ica l component of t h e complex signal is f rom wall thinning and t h e horizontal signal i s primarily f rom magnet ic deposit. A t 200 kHz ( 2 f t h e ver t ica l component is again from wall thinning but t h e horizontal signal is primarily f rom a n increase in tube magnetic permeabil i ty because of incomplete magnet ic sa tura t ion under t h e carbon s tee l tube support. A t 400 kHz eddy currents just barely penetra te through t h e wall. In this case t h e signal is primarily from tube magnetic permeabil i ty variations.

O.D. GROOVE

DENT

+ A~ MAGNET l TE

BAFFLE

BAFFLE BAFFLE PLATE

PLATE

f 2 = 200 kHz f ;= SO kHz

CAL l BRAT l ON TUBE

SIGNALS

MAGNET l TE

f , = s o kHz f * = 200 kHz

f 3 = 400 kHz

BAFFLE PLATE

MAGNET l TE

f j = 400 kHz

ACTUAL DEFECT S l GNAL

FIGURE 8.20 Eddy Current Signals from Monel 000 Tube at Baffle Plate Location.

(f 90 = 100 kHz)

Conducting Deposits

T h e most probable conducting deposit which may b e encountered during in-service tube tes t ing is copper. Copper taken in to solution in o n e par t of a cooling circuit , f rom brass tubes for example, c a n re-deposit at another location at t h e expense of a less noble metal such as iron. An example is shown in Figure 8.21 which is a copper- alloy t u b e f rom an a i r conditioner h e a t exchanger. Copper deposits occur near tube supports, maximum thickness was 0.05 mm. Even such a thin deposit yields a large eddy current signal s ince copper is a good conductor. Figure 8.21 shows response f rom both absolute and differential internal probes. T h e absolute probe gave eddy current

frequencies between f and 2 f . This appears t o b e a weakness in t h e code which may lead t o revision if copper deposits prove more common as boilers age. Inspection of Figure 8.22 reveals t h a t c learer discrimination between copper and defec t s is achieved at f / 2 ra the r than at f g 0 . Optimum test frequency for copper coa ted tubes appears t o be t h e frequency which just leaves signals below t h e horizontal fill-factor direction.

4 - 7Q% OD ECCENTRIC LROOYT B - 10% ID CONCENTRIC GROOVE C - 0.13 rnm THICK COPPER AROIJUD 'Ij9:: D - 0.05 rnrn THICK COPPER AIIOUND TlJEC

1.0

FIGURE 8.22 Eddy Current Signals Obtained with an Internal Circumferential Probe

from Simulated Copper Deposits on Tubes

8.4 MULTIFREQUENCY EDDY CURRENT TESTING

Background

Successful in-service Eddy Cur ren t inspection relies on eddy current probes t h a t can sense de fec t s and an analysis of eddy current signals. Both aspects a r e equally important. While scanning e a c h tube, eddy cur ren t signals a r e obtained f rom baffle plates, magnet i te deposits, dents, tubesheets, tube expansion, etc. and maybe defects. One must, therefore , discriminate between defec t s and insignificant signals and even more important, e s t i m a t e d e f e c t sever i ty when i t occurs together with other signal sources. I t would b e much eas ier if t h e d a t a could be processed t o contain only de fec t signals; Multifrequency ET can do this.

In multifrequency testing, t w o or more sinusoidal signals of different frequencies a r e fed simultaneously t o a single eddy cur ren t probe. Gain and phase of t h e output signal from each frequency can be separate ly controlled.

THROUGH BAFFLE WALL HOLE PLATE

MAGNET1 TE 1.3 mm

15.5 mm I

CALIBRATION TUBE

f , =20 kHz f , =I00 kHz f , =500 kHz

FIGURE 8.23 Internal Probe Response to Various Test Parameters.

fgo= 130kHz.

f , = 2 0 kHz 1 , = 100 kHz

( a )

FIGURE 8.24 Eddy Current Signal at Baffle Plate Position in Tube of Figure 8.1 1

fg0 = 130 kHz.

These signals c a n then be combined to el iminate unwanted signals and leave only t h e de fec t signal. This method is only e f fec t ive if a defec t signal d i f fers characterist ically f r o m unwanted signals and if signals a r e vectorially additive. T h e J

f irst condition makes detect ion of internal defects , in t h e presence of in ternal variations, impossible. T h e second requirement makes t h e method ineffect ive for detect ion of f re t t ing wear under non-ferromagnetic baffle plates (Section 8.2.4). As a consequence of combining signals f rom t h r e e di f ferent frequencies, d e f e c t signal amplitude decreases and instrument noise increases.

Eddy current penetration and phase lag a r e a function of frequency; increasing test frequency reduces penetra t ion and increases phase lag. Since an eddy current signal is a function of cur ren t density and phase lag, i t is possible to change t h e response t o various signal sources by changing test frequency.

If one simulates a h e a t exchanger tube with de fec t s , deposits, dents and support plates, o n e obtains t h e following results: (a) at high frequencies, only internal d e f e c t s and dents a r e detectable , Figure

8.23(c). (b) at in termediate frequencies, a l l f ea tu res a r e de tec tab le and t h e r e is phase

discrimination between internal and external de fec t signals (because of phase lag across t h e wall) and o ther signals, Figure 8.23(b).

(c) at low frequencies, baff le p la tes and magne t i t e deposits yield predominant signals with l i t t l e phase separation between in ternal and external de fec t signals, Figure 8.23(a).

With this background in mind, one can decide which combination of frequencies should be used t o e l iminate extraneous (unwanted) signals. The following two examples i l lus t ra te these effects .

For t h e dented tube example described in Section 8.2.3 (Figure 8.1 11, t h e extraneous signals making up t h e composite signal at f = 100 kHz can be determined by re- inspecting t h e tube at higher and lower t e s t frequencies. If t h e signals from t h e actual d e f e c t in Figure 8.24 a r e compared with t h e corresponding calibration signals in Figure 8.23, o n e can see at 500 kHz t h e signal is primarily f rom a den t while t h a t at 20 kHz contains a large baff le p la te signal component.

Multifrequency Testing of Dented Tubes

With a single frequency eddy cur ren t inspection, t u b e supports and dents tend t o mask signals f rom t u b e d a m a e under t u b e supports. This makes detect ion and es t imat ion of severi ty difficult an d time-consuming. In t h e remaining sect ion w e show how multifrequency simplifies t h e inspection of t h e dented t u b e described previously.

Figure 8.25 i l lustrates t h e t u b e stripping sequence; o n e o r more signals a r e removed by each mixing of t w o frequencies. By proper manipulation of t h e signals f rom t h e t w o lower frequencies, baffle p la te and magne t i t e deposit signals can be eliminated. However, t h e resul tant eddy cur ren t signal is st i l l d is tor ted by t h e 'denting' signal. Again, by combining th is resul tant signal with t h e signal f rom a higher test frequency, t h e dent signal c a n also be eliminated. The tube now looks bare. If a de fec t existed under t h e baffle plate, it would be very easy t o d e t e c t , t h e resul tant signal contains only information f rom t h e OD corrosion. This process of unwanted signal elimination is l ike solving t h r e e simultaneous equations with th ree unknowns

- 149 -

and solving for the parameter XI = defect .

f I

20 kHz

I

2 f a 100 kHz 500 kHz

FIGURE 8.25 Tube Stripping Sequence by Multif requency

As shown in Figure 8.23, the signal at each baffle plate is a composite signal comprising a baffle plate, magnetite deposit (or baffle plate corrosion products), dent and defect signal. Figure 8.26 illustrates elimination of baffle plate and magnetite signals. The probe is moved back-and-forth under t he baffle plate and the signal is monitored on the storage scope in the chopping mode, where both frequency signals a r e displayed simultaneously.

BPFFLE

PLPTE

B P F F l E PLPTE

1 % RESIOUPL B l F F L E PLATE SIGNPL

FIGURE 8.26 Suppression of Baffle Plate and Magnetite Signals

The f 2 signal is f i rs t ro ta ted t o match t h e f 1 signal orientation. Then f 1 amplitude is changed t o match, as nearly as possible, t h e f lsignal size. In this case, this method by itself doesn't work. However, by decreasing t h e vert ical component of t h e f 1 baff le p la te signal, one obtains a good match. On substracting t h e signal, through a n e lect ronic mixer ( C1 1 , t h e signals f rom t h e baffle p la te and t h e magnet i te deposit both nearly disappear. A small residual signal remains due t o di f ferent approach signals at t h e t w o test frequencies, indicated in Figure 8.26 by t h e two open circles. Although t h e baff le p la te signals a r e identical, t h e two points do not coincide; t h e baffle p la te is sensed ear l ier at t h e lower test frequency. This residual signal is insignificant f o r this application though i t can become qui te serious when tes t ing f o r small c racks under non-ferromagnetic baffle plates.

DENT

RESIDUAL DENT S I G N A L

FIGURE 8.27 Suppression of Dent Signal

Figure 8.27 i l lustrates how one can eliminate t h e 'denting' signal from t h e resul tant ( c = f 2 -f 1) signal. This is achieved by f i r s t matching t h e phase and amplitude of he cl and f 3 'dent' signals and then using a second mixing module ( C ) for sub t rac t~on .

Figure 8.28 t r a c e s t h e above sequence fo r two defect ive tubes, and shows t h e eddy current signal becoming simpler t o analyze with each step. On comparing defec t ive t u b e signals with those f rom a calibration tube, o n e observes t h e f 2 defec t signal is distorted by t h e baffle plate, dent and/or magne t i t e deposit. The C1 signal is only distorted f rom t h e dent signal, and C 2 is a clear signal indicating outside diameter (OD) pits approximately 50% deep. Even a n inexperienced inspector could analyse these results.

FIGURE 8.28 Multif requency Eddy Current Signals from Defective Tube

When using multi-frequency t o e l iminate "internal d iamete r (ID) noise", such as signals from cyclic internal d iameter variat ions ("pilger noise or die chatter"), dents and probe wobble, t h e signal amplitude f rom internal de fec t s is drastically reduced. However, signal amplitude f rom external de fec t s i s not a l t e red significantly. Multifrequency is more e f f e c t i v e for external de fec t detect ion than for detect ion of internal de fec t s in tubes.

8.5 SUMMARY

Defec t signal ampl i tude is a function of i t s axial and circumferential e x t e n t as well a s depth. Defect signal phase is primarily a function of depth. For general purpose volumetric inspection of h e a t exchanger tubes, a suitable test frequency is

f g O - 3 p / t 2 , kHz (7.4)

where p is e lect r ica l resistivity and t is wall thickness.

Inspection at f g o allows defec t depth t o be es t imated on t h e basis of signal phase. T o discriminate between defec t s and ferromagnet ic deposits a lower test frequency should be used; normally 10 o r 20% of f go.

Signal response from most significant service induced defects is usually comparable in amplitude t o tha t from a 1.6 mm diameter through hole. Stress corrosion cracking, general corrosion and frett ing wear give large signals whereas pitting corrosion and fatigue cracks give small signals.

Testing for frett ing wear under non-ferromagnetic support plates is difficult and unreliable with bobbin type probes, because defect and support plate signals a r e not vectorially additive. A surface type probe should be used.

Multifrequency equipment can be used t o eliminate unwanted components from complex signals such as support plates and probe wobble. This greatly simplifies signal analysis.

CHAPTER 9 - METALLURGICAL PROPERTIES AND TESTING FERROMAGNETIC MATERIALS

INTRODUCTION

O n e c a n find numerous re fe rences in NDT publications dealing with eddy current measurement of mater ia l properties, such as chemical composition, hardness, s t rength , corrosion damage, degree of cold work and e x t e n t of both carburization and decarburization. In fac t , none of these properties and mater ia l conditions a r e measured directly. Eddy cur ren t tes t ing is sensit ive t o mater ia l properties through their e f f e c t of resistivity and magnet ic permeabil i ty. As such, eddy currents only provide indirect measurement of mater ia l properties and c a r e must be taken t o ensure t h a t some unforseen mate r ia l variation does no t lead t o fa lse conclusions. Two precautions will help avoid fa l se test results:

(a) a sound basic understanding of ET as outlined in previous chapters (b) use of suitable standards for any part icular test; t h e condition of such standards

should be verified by independent methods, e.g., hardness tests, tensile tes ts .

A comple te t r e a t m e n t of mater ia ls property evaluation by eddy current t e s t ing is beyond t h e scope of this manual. The basics a r e covered and a few examples presented.

ELECTRICAL CONDUCTIVITY

Fac tors Affecting Resist ivity

All mater ia ls possess intrinsic res is tance t o e lect ron flow (current) which is t e rmed resistivity ( P , microhm-centimetres). The res is tance of a conductor i s given by

R = O ~ / A ohms

where 2 is length (cm) and A is cross-sectional a r e a ( cm2 ) . Resistivity values for various materials a r e l isted in Table 9.1.

Conductivity ( 0, siemenslmetre)" i s t h e ease with which electrons can move through a material . I t is t h e reciprocal of resistivity. In eddy current test ing, conductivity is frequently given as a percentage of t h e International Annealed Copper Standard (% IACS). In th is system conductivity of pure, annealed copper at 20°C is s e t t o 100% and conductivity of o ther mater ia ls is given as a percentage of copper. Conductivity of a mater ia l can b e calcula ted from i t s resistivity,

X I A C S 1 7 2 1

MATERIAL

Increasing t empera tu re normally increases resistivity (decreases conductivity) a s shown in Figure 9.1. Over a limited t empera tu re range t h e variation is usually linear according t o t h e relat ion

P = p 0 ( 1 + a A T )

TABLE 9.1

ELECTRICAL RESISTIVITY OF COMMON CONDUCTORS AT 20°C

Silver Copper Cold Aluminum 7075-T6 (A1 Alloy) Zinc Magnesium Admiralty Brass Iron Phosphor Bronze Lead 70 Cu-30 Ni Monel Zirconium Ti taniurn 304 SST Zircaloy-2 Inconel 600 Hastelloy X W aspaloy Ti-6A 1-4V

RESISTIVITY ( ~ n .cm)

CONDUCTIVITY (siemensjm)

CONDUCTIVITY (% IACS)

where P is resistivity at t empera tu re T ( ' C ) , Po is resistivity at a reference t empera tu re To, a ( @ C ) is thermal coeff ic ient of resistivity and AT is t h e temperature di f ference ( T-T . For common meta l s and alloys values of a range f rom less than 0.001 t o over 0.01, 0.004 is fairly typical.

Alloying normally increases resistivity. Figure 9.2 shows even small alloy additions t o aluminum can increase resistivity appreciably. T h e conductivity of binary Cu-Ni

COPPER

200 400 600 800 1000 1200 1400 TEMPERATURE ('K )

FIGURE 9.1 - - - -

Effect of Temperture on the Resistivity of Copper, Platinum and Titanium

3L I MANGANESE

1.0 2.0 3.0 4.0 5.0 6.0 ALLOY CONTENT (mass %)

FIGURE 9.2 Effect of Alloying Elements on the Electrical Resistivity of Aluminum.

alloys is shown in Figure 9.3. The dependence of conductivity on composition provides one basis for eddy current sort ing of mixed alloys. Oxygen impurity in zirconium and titanium alloys changes resistivity considerably. Figure 5.19 showed a non-uniform oxygen distribution in a zirconium-niobium de tec ted by eddy current testing.

COPPER/NICKEL ALLOYS

0 2 0 4 0 6 0 80 100 MASS % COPPER

100 80 6 0 40 20 0 MASS % NICKEL

FIGURE 9.3 Variation in Electrical Conductivity of Nickel-Copper Alloys with Composition

Cold work increases resistivity through introduction of la t t ice de fec t s in metals. A t normal temperatures, cold work has a relatively small e f f e c t on conductivity ( (10 % ) and can usually be ignored. The degree of cold work in some austeni t ic stainless s teels can be determined by ET, this is possible because cold work makes them ferromagnetic, not because of a resistivity change.

9.2.2 Material Sorting by Resist ivity

This is normally a n eddy current su r face probe method. Two instrument types are commonly used. Impedance display instruments o f fe r a comparative method as t rea ted in Section 5.8.2; t h e lift-off curves f o r unknown materials a r e compared with those of known standards and t h e resistivity of t h e unknown is e s t imated by interpolation. Meter readout instruments a r e also available with built-in "lift-off" compensation which are cal ibra ted di rect ly in % IACS. Both types of instruments require c a r e on t h e par t of t h e operator t o insure meaningful results. Ef fec t s which can contribute t o erroneous results follow (for more details see Section 5.8.2):

(a) t o o low a test frequency can make mater ia l thickness appear similar to resistivity changes.

(b) sample curva tu re affects coil coupling and hence i t s response (edge and other geometry e f f e c t s have a similar response).

(c) too high a test frequency could sense alloy changes a t t h e su r face of oxidized or corroded materials .

(d) conducting and nonconducting coatings a f f e c t test coil impedance. (el ambien t t empera tu re variations result in changes in sample resistivity and t e s t

coil resistance.

T h e above potential e r ro r sources can largely be overcome through use of suitable s tandards which duplicate materials t o be tested.

--

I 10 100 1000 TIME AT TEMPERATURE ( h )

FIGURE 9.4 Variation of Mechanical Properties and Conductivity in 7075-T6

Aluminum Exposed at 205OC

An example of eddy current tes t ing t o determine h e a t t r e a t e m e n t state of a n aluminum alloy is shown in Figure 9.4. These results a r e f rom Pellegrini (10) who indicates t h a t t h e technique c a n be used t o judge t h e f i tness of o v e r h e a t e ~ m a t e r i a l f o r fu r the r service. A similar approach has been used to assess hea t t r e a t condition of t i tanium alloys.

MAGNETIC PROPERTIES

For eddy current purposes one can classify mater ia ls as ferromagnet ic (magnetic) or non-ferromagnetic (nonmagnetic). Diamagnetic and paramagnetic materials can be considered nonmagnetic. Ferromagnetism has i t s origin in a quantum mechanics e f f e c t , t h e "exchange interaction". I t occurs in t h e e lements iron, cobalt , nickel and s o m e of t h e r a r e e a r t h metals. These elements have partially filled d and f e lect ron shells. Alloying with e lements which have a higher e lect ron to atom ra t io fills t h e s e d and f shells and makes t h e resulting alloys less magnetic, e.g., copper added t o nickel (Monel) and chromium added t o iron (stainless steel).

The main f e a t u r e separating magnet ic f rom nonmagnetic materials is magnetic permeability, P which is a measure of a material 's intr insic ability t o conduct magnet ic flux. I t is defined as t h e induced magnet ic flux density, 0, divided by external rnagnetic field intensity (magnetizing force), H,

For a i r and nonmagnetic materials P is a constant,

Po - 4n x lo- ' weberrlampere-metre 2 when B is in teslas* (T) o r webers lmetre and H i s in ampere lmet re ( ~ / m ) .

Simplification results if one uses re la t ive permeability, which is defined as

P, = p l u o (dimensionless)

Rela t ive permeability has t h e same value in all magnet ic systems of units. For magnet ic mater ia ls P, can be very large, whereas for nonmagnetic material Pr = 1 . 0 .

9.3.1 Magnetic Hysteresis

When a mater ia l i s magnetized in a coil, t h e magnetic f ield intensity, H, i s proportional to coil current. If a l ternat ing current is applied t o a magnetizing coil a a-H loop results as shown in Figure 9.5. As H increases f rom zero fo r the f i r s t t ime, B increases along t h e DC curve, path No. 1. When H decreases , B also decreases but along path No. 2. The difference between paths 1 and 2 is t e rmed hysteresis. When H has fal len to ze ro and residual flux density remains in t h e material , B, , called re tent iv i ty or residual flux density. On decreasing H fu r the r (reverse or negative current) flux density decreases t o zero at Hc which is t h e coercive magnet ic intensity o r coercive force. Decreasing H sti l l fu r the r drives t h e curve to point S 1. Additional AC cycles will r e t race t h e loop. A t point S 2 t h e material is sa tu ra ted , f rom S 2 t o S 3 t h e B-H curve is l inear with slope lJO . Flux density at saturation depends on material ; carbon s tee l sa tu ra tes a t about B = 2 tes la whereas Monel 400 sa tu ra tes at about 0.3 tesla.

*Conversion: 1 t e s 1 a - 1 o4 gauss; I A/m = 0.012566 oersted.

FIGURE 9.5 Hysteresis (or B-H) Loop

9.3.2 Magnetic Permeabil i ty

For eddy cur ren t inspection of ferromagnet ic mater ia ls several kinds of permeabil i ty play an important role. Normal permeability, U r , is a measure of a material 's abil i ty to conduct magnet ic flux; it is an important f a c t o r when determining t h e ease with which a magnet ic mater ia l can be sa tura ted.

Another permeability of concern in ET is incremental o r recoil permeability, P A .It is defined as

v A - ABlAH where A B is t h e change in flux density which accompanies a change in magnetizing force , A H , crea ted fo r example by a n eddy cur ren t coil's al ternating current . An incremental A H c a n be superimposed at any point on a DC magnetization curve as i l lustrated in Figure 9.6.

M A G N E T I Z I N G F O R C E I A l m l

FIGURE 9.6 DC Maptization Curve and Recoil Permeability for Iron

A t H=O w e have t h e relat ive initial permeability, Pi . In a magnet ic material without a biasing DC magnetic field, t h e normal permeability is equal t o the incremental p r m e a b i l i t y ,

In eddy current testing, test coil inductance and depth of penetration a r e influenced by incremental permeability not normal permeability. However, throughout this repor t i t is assumed t h a t t h e e d d y current test is performed without DC bias and with a low magnetizing f o r c e (low al ternat ing coil current). In this case, V, = V A , and fo r simplification purposes , is used in t h e skin depth and inductance equations and impedance diagrams; ur is used throughout t h e manual t o deno te incremental permeability ( uA) unless otherwise s ta ted. In addition, al l permeabil i t ies described herea f te r in t h e t e x t of t h e manual a r e relat ive permeabilities and a r e therefore dimensionless.

When an increasing DC magnetizing field is applied, a nonlinear E H relationship results as shown in Figure 9.7. The incremental permeability continuously decreases until sa tura t ion is achieved. A t saturation U A - 1.0. T h e normal permeability, instead, f i rs t increases t o a maximum value and then decreases gradually, see Figure 9.7; at sa tura t ion i t can still be very large.

FIGURE 9.7 Magnetization Curve, Incremental Permeabil i ty and Normal Permeabil i ty

f o r a 3Re60 Tube Sample

9.3.3 Fac to rs Affect ing Magnetic Permeabil i ty

Ferromagnet ic mater ia ls do not have unique magnetization curves but depend strongly on fac to rs such as

- thermal processing history, - mechanical processing history, - chemical composition, - internal stresses, - t empera tu re (if close t o Cur ie temperature).

The following examples i l lus t ra te t h e e f f e c t of some of these variables.

Figure 9.8 shows B-H curves, at room temperature , for t h r e e supposedly identical Monel 400 t u b e samples. The differences a r e a t t r ibu ted t o variations in nickellcopper con ten t within t h e normal alloy specification range.

Figure 9.9 shows variation of magnet ic permeability with cold work in Type 300 ser ies stainless s teels (2). In these "nonmagnetic" austeni t ic s tee ls a ferromagnet ic mar tens i t e phase form;during cold working increasing t h e magnetic permeability. I n contras t , most normally ferromagnet ic mater ia ls exhibit a decrease in permeability

as a result of cold work. The 300 series stainless steels can also become ferromagnetic as a result of welding, a magnetic delta ferrite phase forms during solidification.

I I I - - - - - - - - - - - -,,-,---- ----=-I ____-- - - - - . 400 - 800- 1200 --------- _____ - - - - - - - --'--- rust r . 2 3 ,

FIGURE 9.8 Magnetization Curves for Various M o d 400 Samples

l0OL I I I I - - - - A U S T E N I T I C S T A I N L E S S STEEL - - -

I COLD WORK

FIGURE 9.9 Variation of Relative Permeability with Cold Reduction

for Various Austenitic Stainless Steels (2)

6 MPa NO S T R E S S

' 24 MPa

-

A Y N E A L L E O 1 R O N

I

2 5 5 0 7 5 100

M A G N E T I Z I N G F O R C E ( A , m

FIGURE 9.10 Effect of Elastic Strain on the Magnetization of Iron (9)

Figure 9.10 shows changes in B-H curves for iron with internal stress. Note t h a t these s t ress levels a r e purely elastic, well below t h e yield strength. The changes in B-H (and permeability) a r e due t o magnetostr iction.

The above examples i l lustrate t h e inherent variability of B-H and hence permeability of ferromagnetic materials. Incremental permeability a f f e c t s a n eddy current coil's inductance as well as depth of eddy current penetra t ion in to a material. The large variations in permeability shown above make conventional eddy current tes t ing for de fec t s in magnetic materials very difficult if not impossible.

The best solution t o eddy current tes t ing of a magnetic mater ia l for de fec t s is t o bring i t t o a condition where U A = 1.0 . A few slightly magnetic materials can b e heated above thei r Cur ie temperature to make t h e m nonmagnetic. Monel 400 heated t o between 50° and 70°C has been tes ted in th is manner. Most materials have t o o high a Curie t empera tu re t o b e t e s ted by th is approach. The only other way to decrease P A t o unity is by magnet ic saturation. This top ic is t r e a t e d in a subsequent section.

9.4 TESTING MAGNETIC MATERIALS

9.4.1 Simplified Impedance Diagrams

A quali tat ive understanding of t h e e f f e c t of permeabil i ty on coil irnpedance can also b e obtained by t h e equivalent c i rcui t and i t s associated setnicircular impedance diagram t rea tment of Section 3.5. Coil inductance is a function of magnetic flux through it; f lux increases in t h e presence of a magnet ic material. For a cylinder surrounded by an encircling coil, coil inductance is proportional to both the cylinder's permeability and i t s cross-sectional area,

where L is primary coil (probe) inductance, lJr U A is t h e cylinder's incremental permeabil i ty and D, i t s diameter. An increase in permeability o r d iameter will increase coil inductance. By a similar t r e a t m e n t to t h a t presented in Chapter 3, one can genera te t h e impedance diagrams of Figure 9.11. Figure 9.1 1 (a) is obtained by plotting t h e encircling coil impedance normalized to t h e inductive reac tance in air. I t i l lustrates t h e e f f e c t of permeability and cylinder diameter. As permeability or cylinder d iamete r increases (with constant coil d iamete r ) coil impedance increases drastically. (This explains t h e good response to ferromagnetic inclusions and deposits discussed in Sections 6.5.1 and 8.3.1). There is no phase separation and hence no discrimination between variations in permeabil i ty and cylinder diameter. However, the re is about 90° phase separation and hence excellent discrimination between variations in permeabil i ty and resistivity.

( a 1 C Y L I N D E R ( b ) C Y L I N D E R

FIGURE 9.1 1

( c ) PLATE

Simplified Impedance Diagrams for Ferromagnetic Cylinders and Plates

Figure 9.1 1(b) is obtained by plotting t h e encircling coil impedance normalized t o i t s inductive reac tance with t h e ferromagnet ic cylinder inside t h e coil. This f igure indicates t h e e f f e c t of permeability and cylinder d iameter on operation point location. An increase in both permeability and cylinder d iameter moves t h e operating point DOWN t h e impedance curve (for constant f i l l factor) .

Surface probe inductance also depends on test sample permeability ( L is proportional t o lJr ) . An increase in permeability moves t h e operating point UP t h e impedance locus as shown in Figure 9.1 1k) . However, unlike curves for a cylinder where t h e semicircle increases drastically in s ize , t h e curve for a sur face probe increases only a small amount as previously shown in Figure 5.10. This results from much less eff ic ient coupling with su r face probes as compared t o encircling coils. A sur face probe with a f e r r i t e c o r e (or cup) coil permits b e t t e r magnet ic coupling (decreased magnet ic reluctance) and hence yields a larger impedance diagram than a similar a i r core coil. An additional observation can be made f rom Figure 9.11 (c); magnet ic permeability has t h e s a m e e f f e c t as elect r ica l resistivity and hence these t w o parameters cannot b e separated when using a sur face probe.

70 1 I I I

3 2 9 S T A I N L E S S S T E E L

10 20 30 9 0

N O R M A L I Z E D R E S I S T A N C E

FIGURE 9.12 Experimental Normalized Impedance Diagrams for Three Types 329

Stainless Steel Samples Tested with a Long Encircling Coil

9.4.2 Impedance Diagrams

Figure 9.12 shows experimental impedance curves fo r t h r e e di f ferent Type 329 stainless s teel samples t e s ted with long encircling coils. These curves differ markedly from a semicircle at t h e lower section of t h e impedance diagram, where t h e curve approaches t h e Y-axis at 4 5 O ra ther than 90'. These curves a r e nearly identical in shape t o t h a t presented in Figure 7.6 fo r a nonmagnetic cylinder. But, while the nonmagnetic curve in tersects t h e reac tance axis (Y-axis) at 1.0, t h e Figure 9.12 curves in tersects this axis at thei r respective V r values. Magnetic saturation of these samples would reduce them t o a common curve intersecting t h e axis at 1.0. This f igure is another example of typical permeabil i ty variations which may be encountered in supposedly "identical1' samples.

I N C R E A S I N G P R O C E C I A M E T E R

INCREASING F R E O U E N C Y

I

P E R M E A E I L I T Y

I N C R E A S I N G R E S I S T I V I T Y

N O R M A L I Z E D R E S I S T A N C E

FIGURE 9.13 Impedance Diagram for Ferromagnetic Material Showing

Effect of Material and Test Parameters

Figure 9.13 shows a n ac tua l surface probe impedance diagram for magnet ic material . T h e shape differs appreciably f rom a semicircle. Most test variables have a similar e f f e c t on t h e impedance diagram as for surface probes on nonmagnetic mater ia l (Section 5.5). To measure magnet ic permeability in t h e presence of lift-off noise, probe diameter and test frequency should be chosen to opera te in region A.

Eddy current inspection of magnet ic mater ia ls for de fec t s is diff icult or impossible because of random permeability variation as discussed in Section 9.3.3. In addition t h e r e a r e skin depth limitations. Without saturation, t h e initial permeabil i ty can range f rom 50 t o over 500. Since depth of penetration is inversely proportional t o t h e square root of permeability and test frequency,

t o obtain equal penetration requires a reduction in frequency by t h e s a m e fac to r of 50 t o over 500. Unfortunately, lowering frequency moves t h e operating point t o Region B in Figure 9.13 where the re is poor signal separation between lift-off, permeability and resistivity as well as reduced sensitivity t o defects.

Before leaving Figure 9.13 consider t h e character is t ic parameter , r2WJr0 (Section 5.6). Figure 9.13 shows t h e parameter is no t generally valid for ferromagnetic materials. I t indicates an increase inlJrshould move t h e operating point down t h e impedance curve l ike increasing frequency or probe diameter. In pract ice exact ly t h e opposite occurs. The character is t ic parameter should only be used fo r finding operating point of surface probes on nonmagnetic materials.

9.4.3 Material Sorting by Magnetic Permeability

Detailed t rea tment of this topic is beyond t h e scope of this manual. This section is essentially a warning,

Many properties of magnetic materials a f f e c t permeability as discussed in Section 9.3.3. Eddy current test ing has been used to s o r t mixed alloys as well a s measurement of hardness, decarburization, carburization, degree of cold work, strength, ductibility, etc. A standard, ASTM E566-76, of fe r s broad guidelines on this eddy current application.

Meaningful results with such test ing requires at leas t t h e following: - understanding of t h e variables affect ing a material 's electrical and magnet ic

properties - a sound knowledge of eddy current tes t ing - adequate standard samples verified by dest ruct ive examination or o the r

independent methods.

Testing fo r Defects in Magnetic Materials

Previous sections explained why satura t ion is required t o suppress e f fec t s of usually harmless permeability variations which could be mistaken fo r , o r obscure, de fec t signals. We only consider test ing of cylindrical materials; similar techniques can, at leas t in theory, be applied t o su r face probe testing.

Manufacturing inspection of rods, wires and tubes is accomplished fairly simply by external , water cooled magnetizing coils through which t h e mater ia l is passed. ASTM standard E309 covers such testing. In-service inspection again presents t h e most difficult si tuation due t o access and space limitations.

Figure 9.14 compares Y-channel eddy current signals from a Monel 400 tube a t f g o without and with magnetic saturation. Saturation results in good d e f e c t detection. Permeability variation due to cold work and internal stresses at a slight bend in t h e tube a r e completely suppressed by saturation. This tube was satura ted by superimposing t h e AC eddy current signal on DC magnetization power. Saturation of Monel 400 is also achieved by incorporating permanent magnets in t h e probe (8). -

0. D. SUPPORT FLAT PITS

DEFECT 1

I CALIBRAT ION TUBE

EDDY CURRENT TEST WITHOUT - SATURATION

SLIGHT BEND I N TUBE

EDDY CURRENT TEST W lTH , , , MAGNETIC SATURATION (10 X ABOVE GAIN)

FIGURE 9.14 Eddy Current Signals from a High Magnetic Permeability Monel 400 Tube.

Test Frequency = 50 kHz

Saturation with DC magnetization is limited by coil heating. H e a t dissipation is proportional to current squared and coil wire resistance ( P owe r = I R ) . To increase magnetization (H is proportional t o I) pulse sa tura t ion is used. The saturation current (DC) is switched on-and-off at regular intervals thereby reducing t h e heating e f fec t . T h e test current (AC) is superimposed on t h e sa tura t ion current and t h e eddy current signal is sampled only at maximum saturation. O n e commercial instrument, operating on this principle, is currently available. Testing speed is a function of pulse ra te , in general i t is much slower than conventional test ing.

If magnetic saturation at defects is not complete, a n eddy current test becomes a test for permeability, not eddy current tes t ing as described in previous chapters. This can be understood f romFigure 9.15 which i l lustrates t h e change in eddy current signals from calibration defects in a magnetic stainless s t ee l tube as degree of sa tura t ion is increased. The eddy current signals were obtained with an absolute bobbin type probe. Since defect signal amplitude decreases as satura t ion is approached, instrument gain was doubled fo r t h e 20 and 40 ampere saturation results. Magnetization was achieved with an external , wa te r cooled coil; 10 amperes produced about 2 . 8 x 10 A 1 m or 350 oersteds. Figure 9.15 shows one has t o be sa tura ted well past t h e knee in t h e magnetization curve (over 20 amperes) before eddy current d e f e c t signals appear normal, like those f rom nonmagnetic materials.

The reason for t h e charater is t ic eddy current signals f rom partially sa tura ted tubing is more clearly apparent in t h e eddy current impedance display of Figure 9.16 which includes impedance response as magnetization level increases. This figure shows, at part ial saturation (less than 10 amperes), de fec t signals consist nearly entirely of increasing and decreasing permeability. The initial increasing permeability signal component is a t t r ibuted t o less saturation on e i ther side of machined calibration defec t s while t h e decreasing permeability component is due t o more intense saturation in t h e reduced tube-wall region at defects.

Similar results a r e obtained with internal saturation using DC magnetization or permanent magnets. A single rare-earth permanent magnet was found t o be equivalent t o about 5 amperes of an external magnetizing cur ren t for this tube s ize material.

THROUGU noLE EXTERNAL MAGNETIZING COIL

/

-A-- A PROBE WOBBLE cC

C 0 0 GROOVE -/-- --- \AIR ---- D 1.0 GROOVE

I I I 1 I I I

5 10 I5 20 25 30 35 4 0 MAGNETIZING CURRENT ( A 1

FIGURE 9.15 Eddy Current Signals from E-&ite 26-1 Tube With Increasing Saturation,

(fgo = 100 kHz at Complete Saturation)

Eddy current tes t ing at part ial sa tura t ion may seem a t t r a c t i v e since d e f e c t sensit ivity is very high, i t may in f a c t develop in to a useful YDT technique. However, the re a r e drawbacks; M A is g rea te r than one and is variable. This means eddy cur ren t penetration is not defined and conventional phase analysis is impossible. Testing tubes for de fec t s at magnet ic supports could be a very questionable procedure s ince large permeability signals would be encountered which could be mistaken for or obscure defects. Even t h e bes t available sa tura t ion methods st i l l encounter problems in detect ing defec t s at s tee l baff le plates in some Monel 400 tubes which a r e only slightly magnetic.

Eddy current tes t ing at part ial sa tura t ion only measures permeability in a thin surface layer adjacent t o t h e test coil. This classifies t h e technique with NDT methods such as magnetic part icle inspection and leakage flux test ing. Leakage flux tes t ing responds t o t h e distortion of magnet ic flux at defec t s in a magnetized mater ia l using pickup coils o r Hall e f f e c t sensors. Par t ia l saturation ET with su r face

probes has an advantage over encircling (or internal) probes in the ability t o separate permeability from lift-off variations (Figure 9.13).

PROBE

A - PROBE WOBBLE B - THROUGH HOLE C - 0. D. GROOVE D - I. 0. GROOVE

FIGURE 9.16 Eddy Current Signals from E-Brite 2 6 1 Tube with Increasing Saturation,

fgo = 100 kHz

An example of t h e dangers of ET ferromagnet ic mater ia ls at part ial sa tura t ion is i l lustrated in Figure 9.17. I t shows eddy current signals f rom calibration defec t s in a 3Re60 h e a t exchanger tube t es ted with a differential probe. (3Re60 requires a flux - density of about 0.6T for complete saturation). Calibration d e f e c t s yield signals which change in phase with increasing depth leading t o t h e conclusion one may have a viable test technique. However, e las t i c deflection of t h e t u b e at a support p la te gives change of permeability signals nearly identical t o serious (50% and 75%) defects. This is due t o magnetostriction: changes in magnet ic properties due t o e las t i c s t ress such as shown in Figure 9.10.

- I I

T \ \ \

"i H O L E 3 R e 6 0

\ T U B E ' 3 c r s s B A F F L E

P A F F L E P L A T E

S I G N A L S

E L I S T I C

D E F L E C T I O N 5 m n Y n m 3 . m 2 m m 0

FIGURE 9.17 Eddy Current Signals from 3Re60 Tube With Partial Saturation for Various

Levels of Elastic Stress. Test Frequency fgo = 230 kHz.

The problem of Figure 9.17 was overcome with a multimagnet probe similar t o t h a t developed for Monel 400 tubing (8). This el iminated t h e false de fec t signals at tube supports and made these hea t exchangers inspectable by conventional ET techniques. I t was for tunate these particular heat exchangers had nonmagnetic, Type 304 stainless s teel , support plates. This permits tube sa tura t ion in t h e vicinity of supports. If t h e supports had been magnet ic they would have provided a low reluctance a l ternat ive path t o t h e saturation field leaving t h e tube only partially sa tura ted. Nonmagnetic support materials improve inspectability of ferromagnet ic tubes even though f re t t ing wear may be difficult t o d e t e c t with a conventional bobbin-type probe as discussed in Section 8.2.4.

SUMMARY

Eddy current test ing can be used t o measure e lect r ica l resistivity and magnetic permeability. This parameter , in some cases, can be corre la ted t o a material 's chemical composition, hardness, h e a t t r ea tment , etc. and therefore provide an indirect measurement of material properties. Material sort ing by electrical resistivity can be done with general purpose eddy current instruments o r with special instruments with mete r output calibrated in % IACS. C a r e must be taken t o obtain reliable results. Material sort ing by magnet ic permeabil i ty is not simple. I t requires a sound knowledge of magnetic properties and eddy current testing. Most of t h e commercial equipment make use of hysteresis distort ion and t h e method is empirical. I t is more reliable t o use general purpose eddy cur ren t equipment t o roughly measure magnet ic permeability and then cor re la te t o mater ia l property.

Testing ferromagnet ic materials for surface de fec t s i s possible but often unreliable. If mater ia l can be magnetically sa tura ted, i t appears a s non-ferromagnetic mater ia l t o t h e eddy currents. Testing at part ial saturation results in good sensitivity t o de fec t s and t o ferromagnetic anomaiies but can result in f a l s e indications. I t is possible t o magnetically sa tu ra te some ferromagnet ic tube alloys in unsupported tube sections, but nearly impossible under ferromagnet ic baff le plates.

Magnetic permeability a f fec t s t h e following: - depth of penetration - probe inductance - operating point on impedance diagram - character is t ic de fec t signal is no longer dependent on phase lag - drastically decreases signal-to-noise ratio.

9.6 WORKED EXAMPLES

9.6.1 PROBLEM: Convert resistivity of 5.5 microhm-centimetres t o conductivity in 36 IACS.

SOLUTION: %IACS = 1 7 2 1 ~

PROBLEM: Pure annealed iron under a magnetizing force, H, of 40 A/m results in a magnetic flux density, 8, of 0.028T. Determine magnetic permeability and relative permeability.

SOLUTION:

ll = B I H - 0 . 0 2 8 1 4 0 = 7.0 x henrylm

- NOTE: Tesla and ampere/metre a r e the preferred metr ic units for magnetic flux density and magnetic field strength respectively but gauss and oersted (nonmetric units) a r e still of ten used. To complete problem 9.6.2 in the gauss, oersted system requires t h e following calculations:

4 B = 0.028 T x 10 gauss = 280 gauss

1 T

H = 40 A/m x 0.012566 oersted = 0.503 oersted I A/m

" r = p / 11, = 557/1.0 = 557 (dimensionless)

PROBLEM: Calculate standard eddy current depth of penetration in carbon steel at a test frequency of 10 kHz (a) without saturation and (b) with 4

complete saturation. P = 1 5 microhm-centimetres, u = 300

SOLUTION: P A - U r - P i

(a) From Equation 2.13(a)

- p~ - 1.0 a t saturation

QUALITY SYMBOL

- 177 -

CHAPTER 10 - SUPPORTING INFORMATION

NOMENCLATURE

QUANTITY NAME

Cross-Sectional a r e a Radius Length Thickness Width Diameter Magnetic flux density Capaci tance Tes t frequency Optimum tube tes t ing frequency Character is t ics o r Limit frequency Resonant frequency Magnetic field intensity (Magnetizing force) Current Current density Self Inductance Number of turns (Windings) Character is t ic Paramete r

Resistance Resistive load Electric potential Depth below t h e surf ace Inductive Reactance Capacit ive Reac tance Impedance Standard Depth of Penetra t ion Permeabil i ty Resistivity Conductivity Magnetic flux Fill Fac to r Phase Lag Angular frequency Angle between Z & R

m e t r e 2

m e t r e m e t r e m e t r e m e t r e m e t r e 2 weberlmetre o r tesla farad h e r t z h e r t z

h e r t z he r tz amperes lmet re o r lenze ampere amperes tmet re 2

henry dimensionless dimensionless

ohm ohm volt met re ohm ohm ohm m e t r e henrytmetre microhm-centimetre s iemenstmetre weber dimensionless radian radianstsecond degree

UNIT SYMBOL

In 2

m r n r n rn

I n 2 Wb/m , T F Hz Hz

Hz Hz

A/m

A ~ / m *

H .. - s-2 n v m 52 n s-2 rn

Htm p Recm

Slm Wb -

rad radls

0

DEFINITIONS

This section lists the most common terms covered in the manual. For each term, t he J

symbol, the SI units and the section where the topic is covered is given, followed by t h e definition.

Absolute probe

- See Sections 5.2 and 7.2. - A probe having a single sensing coil.

Alternating current

- I ~c , amperes; see Chapters 2 and 3. - A current flow changing in amplitude and direction with time.

Anomaly

- See Sections 6.5 and 8.3. - An unexpected, unclassified eddy current signal. - A false defect indication.

Bridge

- SeeSection4.2.1. - Electrical circuit incorporating four impedance arms.

Calibration standard

- A t es t standard used t o est imate defect size and set-up instrument.

Capacitive reactance

- X , , ohms; see Section 3.2. - The opposition t o changes in alternating voltage.

Characteristic parameter - - r2uop , dimensionless, see Section 5.6.

- It allows tes t coil operating point t o be specified in terms of a single quantity rather than four independent variables.

Characteristic or limit frequency

- f g , hertz, s ee Section 7.3.3.

Characteristic frequency ratio

- f / f - dimensionless, s ee Section 7.3.3. - I t allows the test coil operating point t o be specified in terms of a single quantity rather than four independent variables.

Circumferential coil

- see encircling and internal probes.

Conductivity

- a (sigma), siemensfmetre; s e e Sections 2.4 and 9.2. - Measure of the ability of a material t o conduct current (alternating or direct

current).

Conductor

- Material capable of carrying electrical current.

Coupling

- The coil's magnetic field couples t o the tes t sample. - The change in probe impedance is directly proportional t o probe-sample

coupling.

Current

- I, amperes, s ee Section 3.3. - Flow of electrons.

Depth of penetration (standard)

- 6 (delta), millimetres; s ee Section 2.4. - The depth a t which the eddy current density has decreased t o l/e or 36.8% of

the surface density. - Also referred t o as skin depth.

Defect

- A flaw or discontinuity tha t reduces a material's integrity or load carrying capacity -may involve a loss of material.

Differential probe

- see Sections 5.2 and 7.2. - A probe having two sensing coils located side-by-side.

Direct current

- I DC ,amperes; see Section 3.3. - A current flow tha t is constant in amplitude and direction with time.

Discontinuity

- A defect.

Eddy currents

- see Chapter 2 and Sections 5.2.2 and 7.2.3. - A closed loop a l ternat ing current flow induced in a conductor by a varying

magnet ic field.

Eddy cvrent method

- An elect romagnet ic NDT method based on t h e process of inducing e lect r ica l currents in to a conductive mater ia l and observing t h e in teract ion between t h e currents and t h e material . In F r a n c e i t is known as t h e 'Foucault currents' method.

Edge effect

- see Section 5.8.2. - Signal obtained when a sur face probe approaches t h e sample's edge.

Effective depth of penetration

- see Section 2.4. - Depth at which eddy current density drops off t o 5% of t h e su r face density.

End effect

- see Section 5.8.2. - Signal obtained when an internal o r encircling probe approaches t h e end of a tube or rod (similar to edge effect ) .

Encircling probe (Coil)

- see Section 7.2. - Also referred t o as a feed-through coil. - A probe which completely surrounds test tnaterial; c a n be absolute or

differential.

- see encircling probe.

Ferrite

- Ferromagnet ic oxide material . - Used for cores in high frequency transformers.

Ferromagnetic

- see Section 9.3. - A mater ia l with a re la t ive magnet ic permeabil i ty g rea te r than 1.0.

Fill-f actor

- 0 (eta), dimensionless; see Section 7.3. - I t is a measure of coupling between t h e coil and t e s t object. - Fract ion of t h e test coil a r e a filled by t h e test specimen.

Flaw

- A d e f e c t .

Foucault currents method

- In France t h e Eddy Current Method is known as t h e 'Foucault currents' method.

- f , her tz , see Section 2.4. - Number of cycles of a l ternat ing current per second.

Frequency (angular)

- w (omega), radianslsecond; see Section 3.2. - Angular velocity, where w - 2 n f .

Hysteresis

- See Section 9.3.1. - Magnetizationcurve.

I ACS

- International Annealed Copper Standard, see Sect ion 9.2. - a

IACS % IACS - Conductivity as a percentage of t h a t of pure copper.

Inductance

- L, henries, see Section 3.2. - Rat io of t h e to ta l magnetic flux-linkage in a coil t o t h e current flowing through

t h e coil.

- 2, ohms, see Section 3.2. - The total opposition in an electrical c i rcui t t o f low of a l tenat ing current. - Represents t h e combination of those e lect r ica l properties t h a t a f f e c t t h e flow

of current through t h e circuit.

Impedance method

- Eddy current method which monitors t h e change in probe impedance; both phase and amplitude.

Inductive reactance

- X =, ohms, see Section 3.2. - The opposition to a change in a l ternat ing current flow.

Inductor

- A coil.

Internal probe (Coil)

- see Chapters 7 and 8. - A probe f o r test ing tubes (or holes) f rom t h e inside. The coil(s) is

circumferential ly wound on a bobbin.

Lift-off

- L.O., mm, s e e Sections 5.5 and 5.8.4. - Distance between t h e coil of a sur face probe and sample. - I t i s a measure of coupling between probe and sample.

Magnetic flux

- @ , webers, see Section 9.3.

Magnetizing force

- H, amperesfmetre , see Section 9.3.2. - Magnetic field intensity.

Magnetic flux density

- B, tesla, see Section 9.3.

Modulation analysis

- An instrumentation method which separa tes desirable f rom undesirable frequency signals f rom t h e modulating envelope of t h e carr ier frequency signal. - Tes t sample must move at constant speed.

Noise

- Any undesired signal t h a t obscures t h e signal of in teres t . - I t might be elect r ica l noise or a signal f rom specimen dimensional or property

variations.

Null balance

- see Section 4.2.1.

Ohm's law

- Electromotive f o r c e across a circuit is equal t o t h e current flowing through t h e circuit multiplied by t h e to ta l impedance of t h e circuit .

Operating point

- see Sections 3.5, 5.6 and 7.3.3. - Point on t h e impedance diagram t h a t specifies t h e normalized inductive

reac tance and resistance of a coil.

Oscillator

- The electronic unit in an eddy current instrument tha t generates alternating probe excitation current.

Parameter

- A material property or instrument variable.

Performance standard

- Also referred t o a s Reference Standard. - A t es t standard used t o qualify and calibrate a tes t system for a particular test.

Permeability (Magnetic)

- (mu), henrylmetre; see Sections 2.4 and 9.3. or P , dimensionless, relative magnetic permeability.

- Ratio between flux density, B, and magnetizing force, H. Permeability describes the intrinsic willingness of a material t o conduct magnetic flux lines.

Phase lag

- B (beta), radians or degrees; see Section 2.4. - A lag in phase (or time) between the sinusoidal currents flowing at the surface

and those below the surface.

- see Section 3.3. - A vector describing sinusoidal signals; i t has both amplitude and phase.

Primary field

- The magnetic field surrounding the coil due t o the current flowing through it.

Probe

- Eddy current transducer.

Reference coil

- Coil which enables bridge balancing in absolute probes. Its impedance is close to tes t coil impedance but does not couple t o test material.

Resonance

- See Sections 4.3, 5.9 and 7.2.5. - A circuit having an inductor and capacitor connected in series or parallel.

When inductive reactance equals capacitive reactance the circuit is tuned or in resonance.

Resistance

- R, ohms, see Section 3.2. - The opposition t o t he flow of electrical current. - Applies to DC and AC.

Resistivity

- p , microhm-centimetre; see Sections 2.4 and 9.2. - Reciprocal of conductivity (p -1 )"a ) . Saturation (Magnetic)

- A condition where incremental magnetic permeability of a ferromagnetic material becomes 1 .O.

Secondary field

- The magnetic field produced by induced eddy currents.

- See Sections 3.4, 4.5 and 5.4. - The variations in t h e test object which a f f ec t current flow within the test

object can be detected by observing their e f f ec t upon t h e voltage developed across a secondary receive coil.

Signal

- A change in eddy current instrument output voltage; i t has amplitude and phase.

Signal-t~noise ratio

- Ratio between defect signal amplitude and tha t from non-relevant indications. Minimum acceptable ra t io is 3: 1.

Skin depth

- See depth of penetration.

Skin effect

- See Section 2.4. - A phenomenon where induced eddy currents a r e restricted t o t he surface of a

test sample. Increasing test frequency reduces penetration.

Suface probe

- See Chapters 5 and 6. - A probe for testing surfaces, which has a f ini te coverage. The coil is usually pancake in shape.

Test coil

- Coil coupled t o tes t material. I t senses geometrid, e lectr ic and magnetic changes in test material.

Voltage

- V, volts, s ee Section 3.3. - Electric potential or driving force for current. - Output signal from an eddy current instrument.

Voltmeter

- The instrument used t o measure voltage.

- see Section 3.3. - A quantity having amplitude (magnitude) and direction. Normally represented as a line whose length represents t he quantity's magnitude and whose angular position, t he phase (relative t o some reference).

ABBREVIATIONS FOR NONDESTRUCTIVE TESTING TERMS

The following abbreviations of nondestructive testing terms a r e used where required in nondestructive testing standards published by CGSB:

Terminology

Nondestructive testing

Eddy current testing

Hardness testing

Leak testing

Magnetic particle testing

Penetrant testing

Radiographic testing

Ultrasonic testing

Visual testing

Abbreviation

NDT

ET

H

LT

MT

PT

RT

UT

VT

References:

Recommended Practice SNT-TC- IA: Personnel Qualification and Certification in Nondestructive Testing

American Society for Nondestructive Testing - 1980 Edition

Symbols for Welding and Nondestructive Testing American Welding Society

REFERENCES 10.4

L I . H.S. Jackson, "Introduction t o Electric Circuits", 2nd edition, Prentice-Hall Inc., Englewood Cliffs, New Jersey (1965).

2. C.V. Dodd, "The Use of Computer-Modelling for Eddy Current Testing", Research Techniques in NDT, Vol. 111, edited by R.S. Sharpe, Academic Press Ltd., London, p. 429-479 (1977).

3. H.L. Libby, "Introduction t o Electromagnetic Nondestructive Test Methods", Wiley-Interscience, New Y ork (1 97 1 ).

4. "Nondestructive Testing Handbook", Vol. 11, edited by R.C. McMaster, Ronald Press, New York, p. 36.1-42.74 (1963).

5. "Eddy Current Testing, Classroom Training Handbook", General Dynamics/Convair Division, San Diego, California (1979). CT-6-5 Second Edition.

6. W.J. McConnagle, "Nondestructive Testing", 2nd edition, Gordon and Breach, New York, p. 346-390 (1961).

7. F.R. Bareham, "Choice of Frequency for Eddy Current Tube Testing", British J. Applied Physics, - 1 1, p. 21 8-222 (1960).

8. V.S. Cecco, "Design and Specifications of a High Saturation Absolute Eddy Current Probe with Internal Reference", Materials Evaluation, 37, p. 51-58 - (1979).

9. J . Stanley, "Electrical and Mangnetic Properties of Metalsu, American Society for Metals, Metals Park, Ohio (1963).

10. H.V. Pellegrini, "Assessing Heat Damage in Aluminum Alloys with an Eddy Current Testing Technique", Metals Progress, E, p. 60-63 (1980).

11. ASME Boiler and Pressure Vessel Code, Section V, Article 8, Appendix 1, "Eddy Current Examination Method for Installed Non-Ferromagnetic Steam Generator Heat Exchanger Tubing" (1 978).

12. "Nondestructive Inspection and Quality Control", Metals .Handbook, Vol. 1 I, 8th edition, American Society for Metals, Metals Park, Ohio, p. 75-92 (1976).

13. R, Hochschild, llElectromagnetic Methods of Testing Metals", Progress in Nondestructive Testing, Vol. 1, MacMillan Co., New York, p. 59- 109 (1959).

10.5 INDEX

Absolute Probe - 51, 98-102 Alternating Current - 8, 16, 20-22 Anomaly - 91, 140 Bridge - 32-36 Bridge Balance - 32-36 Calibration Standard - 94-96, 1 17 Capacitive Reactance - 19 Characteristic o r Limit Frequency - 112-1 17, 120 Characteristic Frequency Rat io - 112-1 17, 121 Characteristic Parameter - 51, 69-71, 81, 82, 112 Circumferential Coil - 98, 102, 1 17 Conductivity - 11, 154-157 Coupling - 23, 27, 51, 100, 106 Current - 5-10, 20-22 Defect - 51, 60,61, 73, 83-90, 94, 123-138, 169 Depth of Penetration (Standard) - 12-17, 74 Differential Probe - 53, 54, 98-102 Direct Current - 20, 21 Discontinuity - 179 Eddy Currents - 5-17, 55, 102, 103, 123 Eddy Current Method (Testing) - 1, 18, 51, 83, 9 1, 123, 154 Edge Effect - 76 Effective Depth of Penetration - 14 Encircling Probe (Coil) - 98, 106, 109, 113, 142 End Effect - 180 Excitation Coil - 6, 1 1, 4 1, 62 Faraday's Law - 8, 16, 45, 55, 64, 109 Faraday, M. - 2 Farad - 19 Feed-Through Coil - 180 Ferr i te - 37, 180 Ferromagnetic - 10, 9 1, 1 59 Fill-Factor - 27, 106-108, 141 Flaw - 181 Forster - 2, 113 Foucault Currents Method - 181 Frequency-5 ,8 , 12, 16,67, 113, 116, 120, 121 Frequency (Angular) - 8, 18 1 Frequency Response - 48 Hall Detector - 6, 31, 42, 172 Henry - 18 Hysteresis (8-H curve) - 159-1 63 IACS - 154-157 Impedance - 8,9, 19,23, 30 Impedance Diagrams - 23-29 Impedance Method - 23, 3 1 Inductance - 18, 57, 58, 103, 105 Inductive Reactance - 19,25,64, 167 Inductor - 18 Internal Probe (Coil) - 99 Lenz's Law - 9, 22

Lift-off - 39-43, 78 Limit Frequency - 1 12-1 17 Magnetic Field - 6, 7 Magnetic Flux - 7-10 Magnetic Flux Density - 7, 159 Magnetic Permeability - 1 I, 12, 66, 67, 89, 91, 92, 141, 159-165 Magnetic Saturation - 159-161, 169-1 75 Magnetizing Force - 159, 161 Modulation Analysis - 46 Noise - 32, 34, 37, 46, 81, 152, 182 Non-ferroma netic - 10, 91, 142 Null Ralance f Bridge Balance) - 32, 33 Oersted - 6, 8 Ohm's law - 8, 16, 56, 1 10 Operating Point - 25-29, 7 1, 92, 1 13-1 15, 124, 141 Oscillator - 5, 31, 39 Parameter - 60, 183 Performance Standard - 183 Permeability (Magnetic) - 11, 13, 66, 67, 89, 91, 92, 141, 159-165 Phase - 72. 73 Phase Lag - 2, 14-17, 73, 85 Phasor - 20 Primary Circuit - 8, 24 Primary Field - 183 Probe - 51-57, 98-106 Receive Coil - 6, 23, 62, 75 Reference Coil - 32, 52, 53, 99 Resistance - 18, 26-29, 123-125 Resistivity - 12, 16, 66, 67, 74, 93, 154-158 Resonance - 35, 36, 80, 105 Saturation (Magnetic) - 159-161, 169-175 Secondary Field - 10, 184 Secondary Voltage - 73 Send-Receive - 6, 23, 3 1, 41-44, 75 Sensing Coil - 6, 23 Signal - 184 Signal-to-Noise Ratio - 58, 184 Similarity Condition (Law) - 71, 1 14 Sinusoidal - 5, 1 1 Skin Depth - 13, 14, 16, 117 Skin Effect - 1 1 Speed of Response - 48 Standard Depth of Penetration - 12-17 Surface Probe - 51-55 Test Coil - 52, 53 Vector - 20 Voltage - 8, 9, 20, 31 Voltrneter - 6