1

200619th International Lightning Detection Conference24-25 April • Tucson, Arizona, USA 1st International Lightning Meteorology Conference26-27 April • Tucson, Arizona, USA

CONTINUING STUDIES INTO NEW LIGHTNING PHENOMENA

David Newton BSc, CEng, FIEE, FIMechE, AWE Aldermaston, UK Malcolm Jones PhD, F,Inst,P, CPhys, CEng, AWE Aldermaston, UK

Marvin E Morris PhD, Bolt, Inc, Albuquerque, NM, USA

Keywords: Lightning phenomena, lightning hazards, lightning protection

1. INTRODUCTION Lightning presents an ever present hazard to both individuals and high consequence operations.

Although the subject has been under study for many years, it is still not fully understood and continues to throw up new surprises, some of which lead to potentially disastrous circumstances. This paper deals with a topic which relates to aspects of lightning phenomenology, potential hazards, analysis methods and mitigation criteria. In recent years thunderstorm activity has exhibited phenomena, observed by a number of independent investigators that cannot be explained in terms of direct lightning attachment to structures. Very large currents, coupled with very large apparent strike rates, have been recorded within and external to building structures. This has occurred in the presence of thunderstorm activity but in the absence of evidence of direct lightning attachment. These phenomena continue to be investigated in terms of gathering further experimental data and data from structures exposed to thunderstorm activity. The work, to date, on understanding this new phenomenon continues to indicate that there may be additional thunderstorm related hazards, which are not fully understood.

2. BASIS FOR NEW PHENOMENA There have been reports in the literature about strong high frequency emissions during

thunderstorm activity, i.e. frequencies that are typically well above the range generally associated with lightning stroke characteristics. Further, the National Severe Storm Laboratories in the US, (Mazur and Rahnke, 2001), has reported significant phenomena which are related to thunderstorm activity but not to direct lightning attachment. They have identified damaging events to radar environmental covers when no direct strike was recorded. In the UK a number of munitions facilities (Storage and Processing) have had Peak Current Sensor cards attached to both dedicated lightning protection system down conductors and other grounding elements. Very large currents have been indicated on both external conducting elements as well as on some internal grounding components albeit in the absence of direct lightning attachment. The peak current sensor cards, that are attached perpendicular to the conducting element, have a ferromagnetic strip that records the peak intensity of the magnetic field when current flows through the conducting element. These Peak current sensor Cards have been exhaustively tested using a lightning simulator and show that when correctly installed accurately record the peaks current values to within a few percentage points of injected values.

2.1 TYPICAL RESULTS FROM UK MUNITIONS FACILITIES Table No 1 below provides Peak Current Sensor Card values for a typical Munitions Facility in the

North of the UK for a 3 month period March to June of 2003. Specific Peak Current Sensor Card locations are shown, those annotated (*) are for internal grounding element locations.

2

Defence Munitions Facility Period March to June 2003

Tp1 >120,000 amps 260 tp1 Ph1/1 57,000 amps Tp18 36,000 amps Tp12 41,000 amps Tp11 89,000 amps

* Ph1/20 Breakdown Area floor Tape >120,000 amps Tp8 28,000 amps Ph1/11 Lower Loop >120,000 amps

IWC 5 wacr 2 hi by door 41,000 amps * 486 HAP Cell Crane earth bond 98,000 amps

486 ext lps rod 4 97,000 amps 486 ext lps rod 7 104,000 amps IWC 5 wacr 4 horiz by door 41,000 amps

260 horiz low front door 41,000 amps * Elect Wksp earth bar ph1/13 >120,000 amps

Table 1 – (Test locations and Peak Current values)

Note: * denotes facility internal location point. The values indicated at Table No 1 are not outside of the expect range of lightning currents, however the values are for Peak Current on a single conducting element for which each facility will have many. EA Technology Ltd, who provide data on lightning activity in the UK confirm that there was thunderstorm activity in the area during the period albeit that no cloud-to-ground lightning was recorded for the specific area. Likewise there was no visual sighting of cloud-to-ground lightning from Site personnel, Site being manned 24 hours per day. It should also be noted that the Peak Current Sensor Cards saturate at 120kA, therefore at 4 locations the true Peak Values are unknown but greater than 120kA, two of which being inside facilities which carryout high consequence explosives operations.

3. PEAK CURRENT SENSOR CARDS The Peak Current Sensor Card monitoring system used is manufactured by OBO Bettermann

GmbH & Co of Menden Germany, which utilises peak current sensor (PCS) cards. The PCS card is a sensor which, in the form of a magnetic card, detects and permanently stores the peak value of pulse currents. The cards are evaluated in a special magnetic card reader that gives a digital display of the recorded peak current value. The information given by OBO Bettermann states that the magnetic strip on the PCS has a special coding. Lightning current flowing in a conducting element builds up a magnetic field as a function of current magnitude, this magnetic field alters the coding of the magnetic strip. This change is evaluated and displayed by the card reader. The PCS cards have dimensions of 54mm x 85mm and shown at Figure 1.

3

Figure No 1 – Peak Current Sensor Card The special coding embedded into the magnetic trip of the PCS card as read by a Scopemeter attached to the reading head of the magnetic card reader is given at Figure 2 below.

Figure 2 – Scopemeter trace of PCS Card as supplied from OBO Bettermann. The image at figure 1 is a trace with a mV vertical scale that provides a reflection due to the travel of the Card Reader scanning head, i.e. left to right and return. The magnetic strip is initially magnetised in the plane of the card, directed along its length, and is saturated when all of the magnetic dipoles are aligned with the direction of the applied magnetising field, this occurs at a peak current of 120kA. Testing results have confirmed that the PCS cards are saturated at lightning currents of >120kA. Therefore the H field at the extremity of the PCS card (83mm) would be:

15 Am10X3.2H083.0.2kA120H

r.2kA120H

−=π

=

π=

r = 83mm

4

The trace at Figure 2 below is that for a current of 50kA which shows that the initial coding has been destroyed by the re-alignment of the magnetic dipoles for approximately 40% of the card length. This is consistent with full saturation (100%) at 120kA.

Figure 2 – 50kA nominal Peak Current.

It has been found that analysis of the Scopemeter trace provides a more accurate assessment of the peak current than reliance on the Card Reader digital readout. Analysis of the trace profile also provides assurance that the values are real and not caused by erroneous effects. After response, dipoles are aligned into a direction normal to the strip and the resultant magnetic field direction allows assessment of the current direction in the associated conductor.

4 LIGHTNING SIMULATOR TESTING

The lightning simulator test rig at Culham Lightning, Oxford, UK was used for testing the performance of the OBO Bettermann PCS Cards, it consists of a quasi-coaxial arrangement of four return conductors around a central conductor or an arrangement with supplementary conductors carrying the total injected current. The rig in its standard configuration with a so-called arc electrode installed is shown at figure 3 below. The testing required for the AWE trials shorted out the open gap beneath the central conductor with a metal conducting element capable of carrying up to 200kA. A variety of conductor arrangements were used in order to facilitate investigation into how the PCS card orientation with respect to conductors affected the readings obtained.

5

Figure 4 – Lightning Simulator Arrangements

Figure 5 – H-Field distribution within quasi-coaxial Lightning Simulator

H Field Magnitude (Amps/mm)

X Direction (mm)-400 -300 -200 -100 0 100 200 300

Y D

irect

ion

(mm

)

-400

-300

-200

-100

0

100

200

300

1 10 100

(28kA Up Central Conductor)

6

Figure 5 above models the H-Field magnitude within the quasi-coaxial simulator rig, it clearly shows that the rig return conductors can effect on the H-field magnitude particularly in the diagonal as opposed to the vertical. Therefore the vertical arrangement that maximises the distance of the return conductors was used throughout the simulator testing as this more accurately reflects the rig injection current values.

Lightning simulation employs the Component A/D waveform which is produced by a two-stage capacitor bank of up to 387.5 µF per stage that can be charged to a maximum of 40 kV per stage. Component A is produced by a two-stage capacitor bank configured for a discharge capacitance of 190µF. The bank resistance and inductance in the discharge circuit are set to give an essentially unidirectional pulse with the appropriate ratio of peak current to action integral. Component D is produced by a two-stage capacitor bank configured for a discharge capacitance of 50µF. Resistance and inductance in the discharge circuit are set to give a pulse exhibiting typically a 10% reversal with the appropriate ratio of peak current to action integral. Two waveform were used in the testing, standard waveform with 25µs rise time to peak and fast waveform with 11µs rise time to peak.

4.1 LIGHTNING TEST SIMULATIONS

A number of test configurations have been simulated with both the normal and fast waveforms. The lightning simulator was fitted with dedicated systems to measure the current injected by the arc attachment. The system for measuring high-amplitude components comprises a Rogowski coil transducer installed around the simulator’s low-potential terminal. The output signal is integrated, and then recorded by a transient digitiser. Figure 6 below shows a typical set-up arrangements together with the injected waveform data.

Figure 6

7

Multi-stranded conductor accurately centred between lightning simulator rig current return pillars. Figure 7 below is the Scopemeter trace and the trace profile analysis parameters that confirm the value of the injected lightning current.

Figure 7 – Scopemeter trace for PC Card attached to multi-stranded conductor

Testing of the PCS Cards included multiple pulse, positive and negative pulse, various card orientations together with different conductor arrangements, (multi-stranded and solid rectangular sections). The influence of the current in the rig return conductors, see figure 5 above, was assessed as part of the set-up arrangements.

4.2 TEST RESULTS

Trials using the Culham Lightning Simulator indicate that the PCS Cards accurately record the pulse peak current provided that the card is not influenced by adjacent conducting elements, typically errors of less than 8% were indicated. Erroneous readings can be obtained if current flowing in adjacent or supplementary conductors influences the PCS Card under test. PCS Card locations should avoid areas where there is multiple conducting elements or other equipment that can generate EM fields that are in close proximity to the card. However, where there are multiple conductors that would be sharing the lightning current the PCS Card will clearly read low. It also became apparent when testing the cards at various orientations to the conductor that different field magnitudes were required to change the dipole domains along specific directions. This suggests that the ferrimagnetic material crystals have both an easy and a hard magnetisation direction, this being consistent with the characteristics of ferrimagnetics such as NiFe2O4, a typical material that has been widely used in the card security industry.

5.0 COLLABORATIVE WORK

The National Severe Storm Laboratory in Norman Oklahoma is now working closely with AWE on this new phenomena, Dr Vladislav Mazur of the NSSL agreed to install PCS Cards on a 270 foot high tower in the Dallas area of Texas, USA, AWE would read and analyse the cards when removed following any significant thunderstorm activity in the area of the tower. Appendix 1 at the end of this paper gives details of the tower and the arrangements for installation of the PCS cards. Peak Current Sensor Cards were attached to a lightning protection system down conductor which was one of two that connect an air terminal at the top of the tower to the ground electrode system. This consisted of a Low level P C Card attached to the conductor adjacent to current transformer at tower base, and a High level P C Card attached to the conductor at approximately 1.5m above

Card 1A3.5 9.75

221.253

1.5

X i

833 i

Card 1A3.5 9.75

221.253

1.5

X i

833 i

221.253

1.5

X i

833 i

8

the lower level P C Card. The NSSL current probe saturates at approximately 10kA. Two sets of PCS Cards have been analysed to date, both sets following significant lightning activity in the area. In both cases information received from the US NLDN suggested that cloud-to-ground lightning activity was in the area but that in neither case was the tower directly struck. Figure 8 is the lower PCS Card from the first set, which is adjacent to a conventional current probe, which indicates a peak current of 13.3kA which is in good agreement with the current probe value of 10.5kA.

Figure 8 – Trace analysis for Lower PCS Card

Figure 9 below is the trace for the upper PCS Card from the first set which indicates a value of 83kA

Figure 9 - Trace analysis for Upper PCS Card

Analysis of the two trace profiles show that the peak current values are real (all the attributes of an expected profile), however, the information from the NLDN suggests that the cloud-to-ground strike was some 800m from the tower and that it was a negative stroke with a current peak value of only 22kA. Analysis of the dipole domains of the magnetic strip indicates that the current direction was from the top of the tower to ground, what might be expected from streamer/upward leader collapse.

The second set of PCS Cards analysed show similar results only that in this case the lower card

indicated a value of 43kA whilst the upper card suggested a peak current value of between 5kA and 10kA, the lower limit of resolution for the card. Information from the NLDN suggested that there were two near cloud-to-ground lightning strikes, one negative strike of peak current 86kA and a positive strike of 241kA, neither of which directly attached to the tower. Analysis of the

9.251.15

Card - Lower Dallas Tower 13.3kA

0 10 20 30 40 50 60 70 80 901000

4080

120160200240280320360400

�

�

306.974

1.5

Xi

833 i

9.251.15

Card - Lower Dallas Tower 13.3kA

0 10 20 30 40 50 60 70 80 901000

4080

120160200240280320360400

�

�

306.974

1.5

Xi

833 i

Card - Upper Dallas Tower 83kA

6.3 2.9

0 10 20 30 40 50 60 70 80 90 1000

153045607590

105120135150

�

�

108.796

0.983

X i

833 i

Card - Upper Dallas Tower 83kA

6.3 2.9

0 10 20 30 40 50 60 70 80 90 1000

153045607590

105120135150

�

�

108.796

0.983

X i

833 i

9

magnetic strip dipole domain showed that the current direction was from ground upwards, suggesting it may have been a negative streamer/upward leader collapse from response to the positive nearby cloud-to-ground strike. The analysis of the relationship between current and dipole domain direction requires knowledge of the PCS Card orientation with respect to ground, Figure 10 below describes the process.

6 RESULTS OF STUDIES The work carried out in the UK and the collaborative work with NSSL indicate that the high

currents recorded on the peak current sensor cards are real, extensive testing with the lightning simulator shows that the valid trace on the ferromagnetic strip can only be created by current flowing in the conductor that the PCS Card is attached to. Influence from other nearby conducting elements can be identified and screened out, PCS Card reader will normal reject any abnormal current traces but outputting a default value. Analysis of a number of PCS Cards attached to a single facility does show that these high currents are specific to a location and that the current does not appear to be evenly distributed throughout the conducting elements, however, when two PCS Cards are placed adjacent to each other the results concur when inductances are taken into account. In all cases evidence from National Lightning Detection Networks indicates that these effects are associated with indirect lightning activity. One facility, instrumented with PCS Cards, that was subject to direct lightning attachment only indicted typical lightning peak current values when distribution and inductances of conducting elements where assessed.

Figure 10 – Magnetic dipole domains and H-Field

Current Flow

Down Conductor

Upper P C Sensor Card attached to down conductor and oriented as shown below

North Pole North Pole

Lower P C Sensor Card attached to down conductor and oriented as shown below

Rear face of PC Card

Ferrimagnetic Strip

10

7 CONCLUSIONS Lightning phenomenology is still not fully understood and continues to raise new concerns. This

paper deals with the topic of apparent extraordinarily high currents on facility conducting elements due to indirect consequences of lightning activity. The OBO Bettermann peak current sensor cards have been extensively tested on a lightning simulator and found to be remarkably accurate, albeit that improper installation can lead to erroneous peak current values, such erroneous readings being generally rejected by the PCS Card reader or screened out in the trace analysis process. The nature of the ferromagnetic material used for the PCS Cards is such that it will exhibit spontaneous magnetism and consequently will record very fast high frequency events. The apparent absence of current distribution throughout facility grounding arrangements suggests that causal events may indeed be extremely fast, and not detectable by conventional measuring instrumentation. The testing of the PCS Cards to date together with the trace analysis technique would suggest that the recorded high currents are real. A number of mechanisms in combination may be causal such as atmospheric charge re-distribution, streamer/upward leader collapse, inter/intra-cloud discharges, etc. Research into this phenomenon is to continue and, until proven to the contrary, the effects should be considered as potentially hazardous to high consequence operations. Therefore appropriate safeguards should be put in place such as shielding and stand-off distance from conducting elements of facilities.

8 BIOS David Newton is UK’s Atomic Weapons Establishment (AWE) Chief Engineer and Senior

Electrical Authority, He is also a member of the Distinguished Specialist Group at AWE Phone: UK 0118 982 6950 Fax: UK 0118 982 4821 E-Mail: David.Newton@awe.co.uk Bldg C21.1, Aldermaston, Reading RG7 4PR England Malcolm Jones is a Senior Scientific Advisor at AWE. He is also the Head of the Distinguished

Scientist Group at AWE. Phone: UK 0118 982 4747 Fax: UK 0118 982 7589 E-Mail: Malcolm.Jones@awe.co.uk Bldg D2.1, Aldermaston, Reading, RG7 4PR England

Crown Copyright (2006) “This document is of United Kingdom origin and contains proprietary information which is the property of the Secretary of State for Defence. It is furnished in confidence and may not be copied, used or disclosed in whole or in part without prior written consent of the Intellectual Property Rights Group (IPRG) Ministry of Defence, Abbey Wood,

Bristol, BS34 8JH, England”.

11

APPENDIX 1

270ft High Tower and PCS Card Arrangements

Tower Leg

Upper P C Card

Lower P C Card

NSSL Current Probe

Down Conductor

Tower Leg

Upper P C Card

Lower P C Card

NSSL Current Probe

Down Conductor