Io

Fatigue Properties o f NonferroustAlloys for Heat Exchangers,

Pumps, and Piping

Assignment 86 1o8MEL R&D Report 232/66

May 19656

By

M. R. Gross and R. C. Schwab

U. S. NAVY1

MARINE ENGINEER-ING LABORATORYI

II apolis, Md/O t P ~ ,~

.601

DedicatcM To- PROGRESS IN MARINE ENGINEERIN(e

Bet.st Available, Copy

Distribution of this document is unlimited.

~O6 4'?~b OO

Fatigue Properties of NonferrousAlloys for Heat Exchangers,

Pumps, and Piping

Assignment 86 108MEL R&D Report 232/66

May 1966

By

M. R. Gross and R. C. Schwab

M. R. GROSS

R. C. SCHWAB

Approved by:A

)istribution of thi~g"G~~umv,,,t is ur ime- d

46 t% T%

ABSTRACT

The fatigue behavior of 13 nonferrous

alloys used for corrosion-resistant heat

exchangers, pumps, and piping systems was

investigated over a broad life spectrum of

100 to 100-million cycles. Both cast and

wrought copper-base and nickel-base alloys

were studied. It is concluded that wrought

Monel* and forged Ni-Al bronze have the

highest fatigue strengths, whereas gun metal

and valve bronze have the lowest. The effect

of salt water on fatigue performance was not

found to be highly significant. The use of

Langer's equation to predict stress-cycle

relationships gave satisfactory results for

wrought alloys but appeared to be overly con-

servative for cast alloys.

II

*Registered trade name of the International Nickel Company,Incorporated. C

iii

MEL Report 232/66

TABLE OF CONTENTS

Page

!RIBUTION LIST ii'RACT iii:ODUCTION 1:RIALS INVESTIGATED 2[OD OF TEST 2oURE CRITERIA 4rLTS OF TESTS 5'ARISON OF FATIGUE STRENGTHS 6!LUSIONS:RENCES

8

OF FIGURESFigure 1 - Drawing, Rotating Cantilever Beam

Fatigue SpecimenFigure 2 - Drawing, Low-Cycle Fatigue SpecimenFigure 3 - Curve, Flexural Fatigue Curves, Gun Metal

(Cast)Figure 4 - Curve, Flexural Fatigue Curves, Valve

Bronze (Cast)Figure 5 - Curve, Flexural Fatigue Curves, Ni-Al

Bronze (Cast)Figure 6 - Curve, Flexural Fatigue Curves, Ni-Al

Bronze (Forged)Figure 7 - Curve, Flexural Fatigue Curves, Superston

40 (cast)Figure 8 - Curve, Flexural Fatigue Curves, 70-30

Cupronickel (Cast)Figure 9 - Curve, Flexural Fatigue Curves, 70-30

Cupronickel (wrought)Figure 10 - Curve, Flexural Fatigue Curves, 90-10

Cupronickel (Hard)Figure 11 - Curve, Flexural Fatigue Curves, Cufenloy

40 (Annealed)Figure 12 - Curve, Flexural Fatigue Curves, Cufenloy

40 (DSR)Figure 13 - Curve, Flexural Fatigue Curves, Cupro-

nickel 707 (Wrought)Figure 14 - Curve, Flexural Fatigue Curves, Monel

"E'T (Cast)Figure 15 - Curve, Flexural Fatigue Curves, Monel

(Wrought)

v

Introduction

Many copper-base and nickel-base alloys are used in the con-

struction of heat exchangers, pumps, and piping systems designed

to handle fresh or saline water. In the selection of materials

for such applications, consideration is given primarily to corro-

sion resistance, erosion resistance, and heat transfer charac-

teristics. In most applications the applied stress levels are

low. Accordingly, the structural strength properties of the

materials are relatively unimportant.

In recent years, more and more attention has been given to

the structural properties of these alloys because of (1) cost and

weight reduction programs, (2) conservation of strategic materials,

(3) development of new high-strength alloys, and (4) new applica-

tions which impose high stress levels. Typical of the latter are

sea-connected cooling systems for hydrospace vehicles.

One of the most likely modes of mechanical failure in systems

undergoing cyclic pressurization or thermal shock loading is

metal fatigue. The frequency of stress cycling in such systems

may vary from that of an occasional start-up and shutdown to

vibrational forces developed by the movement of the heat-exchanger

fluids. Little or no published information on the alloys used in

this type of service was found in reviewing the literature several

years ago. Accordingly, tests were conducted at the U. S. Navy

Marine Engineering Laboratory to establish the fatigue behavior of

1

ariety of corrosion-resistant nonferrous alloys. The results

these tests are presented in this paper.

.erials Investigated

The 13 alloys investigated are listed in Table 1, together

-h their chemical compocitions and tensile properties.

.luded are the strength coefficient, K, and strain-hardening

)onent, n, contained in the true-stress/true-strain relaticn-

Lp:

a = K en

re are some deficiencies in the tensile properties of the cast

:erials with respect to the governing specifications. This is

be expected inasmuch as the specification requirements are

aally based on separately cast test coupons, whereas the values

ven in Table 1 were obtained on specimens removed from cast

ates.

thod of Test

Two types of flexural fatigue specimens were used in the

vestigation. The high-cycle fatigue tests were performed with

tating cantilever-beam specimens having the dimensions shown in

gure 1. These were constant deadweight load tests with a cycle

equency of 1450 cpm. The smooth test lengths were circum-

rentially and longitudinally polished to a metallographic

nish.

2

T -1 P4 14 -~ -Y CY - - u

w 0.0 0 I % C 0 O

.0 0 ~ now ~ 1 - .. 4 .. C u Cu CCu

& .1A 0 0 ' 0' 0A . 4 1 4 j a ' L CYa01

Cu .4 0NI-

-

0 0 0 0 0 0\ 00

A, e\ 0 04 D-4 1N 4.4 .-4 ,.4 N gr

0 0N

M) . Z-. 0 - ') 0%t- 0 X

0 '.4 '.4 '4 Cu Cuoo W LV.C R 0 N :

41 0 ' '0a

I.,44.. 1, 0-4 0. 0 0 'Do "$4.~C Cu 0 .

4) 0 Ca -C '. Cu

04 CQ

o4 14 - 05-4 CC, 14 n

r-4 f. IC-0 0 0

CC 0 0 0

A 5 c u. D, N, c0,

P1 00 0' 0 0

CQ U OCC a\. 0D 0~1~0 0 0~ 0 01

0 0\ N ' r -4 -4 0 r N

o4 = 0 0c U-% C' v 0 0% 0 C' Cu LC C'co -4' q,\ Cl C.x

U)r

ok 0 0 aC)\ * r N' '

c- 10 -q14 0 -4 -4 C\J- c'cr 4 ,I-0

r4 -0 I B ?' I '

C P .4 - 4mP~ m- 4-0F 0 -- - - - -- 14

0 C 0 u0 x0 '0 2 :u

-A44 41 Ai 41 1 4 14 1 1 4 1 1 0'00 U* ta toU U 1

u u & u U ) $4 Im41 4cCC_ __ oc 1 - 1 4 41 41

-4 C

410 10 0 '

E 51 c4-I L 0000 004C C 041 -4 j"0 E> cE C' 4 c0 1 14 X C 14 _14 4.4 X 4 Q 41 4D 41 0ru w1 I'.,- w4 1.3 0 : -0 0 ~0 0 0I0 :3 0 al0 0o UE-CJ______ > m > zZ C tC U - a' U L(. 4 al z m_ *)-

t5 U u ,Itu -t Qc j

The low-cycle fatigue tests were performed with equipment

ibed previously.' Flat flexure-type specimens having the

sions shown in Figure 2 were used. The short end of the

men was held stationary, while the long end was flexed

en mechanical stops by a hydraulic piston. One or more

n gages (0.25-inch gage length) were attached to the mini-

est section to record the longitudinal strain. The applied

ng force was measurpd with a load cell. The total strain

ACT, was obtained from strain gage readings, and the

al bending-stress range was calculated from elastic stress

las using the measured load range. Specimens were cycled

^pm.

All of the fatigue tests were of the completely reversed

(Fatigue ratio = -1). Whereas most of the specimens were

d in air, a few were tested with Severn'River water con-

usly wetting the test surface. Severn River water is a

ish estuary water containing 1/6 to 1/3 the salt content

tural seawater, depending on the season and the tide.

ous fatigue tests in both Severn River water and natural

ter havF. shown no significant differences in the effects

e two media.

re Criteria

Failure in the high-cycle, rotating cantilever-beam tests

sted of complete fracture. Failure in the low-cycle fatigue

4

tests was defined as one or more surface cracks 3/16 to I/4 inch

in length.

Results of Tests

The results of the tests are plotted in log-log form in

Figures 3 through 15. Two methods have been used in analyzirvj

the data. The top graph in each figure is the SR vs N relation-

ship for the data, where SR is the nominal reversed bending

stross and N is Lhe number of cycles to failure. SR was calcu-

lated from the elastic stress formula

SR =21~

where AM = bending moment range, in-lb.

c = distance from neutral axis to outermost fiber at

minimum cross section, in.

I = moment of inertia of minimum cross section, in.4

The bottom graph in each figure is the SPE vs N relationship

for the data, where SPE is the reversed pseudoelastic or apparent

elastic stress calculated as follows:

SPE = 6ET.E ..... (2)

2

where ACT = total strain range as determined from strain gages

on the test section, in/in.

E = modulus of elasticity (Table 1), psi.

5 -

urvilinear relationship between Spr and N was obtained by

fitting the following relationship to the data.

_PE = c + SE .....

Nm

C and m = best-fit constants

SE = endurance limit or fatigue strength at 10s

cycles, psi.

_st-fit equation for the SPE vs N data is given in each

a.

Equation (3) is a generalization of the following equation

sed by Langer 2 for predicting the SPE vs N fatigue curve

tensile test data.

_ E in ( 00 +S ..... (4