2006 International Conference on Power System Technology

Power Grounding Safety: Copper Grounding

Systems vs. Steel Grounding SystemsYexu Li, Jinxi Ma, and Farid Paul DawalibiSafe Engineering Services & technologies ltd.

1544 Viel, Montreal, Quebec, Canada, H3M 1G4Tel.: (514) 336-2511 Fax: (514) 336-6144 Web: www.sestech.co

Jinsong ZhangJiangsu Electric Power Research Institute

243 Phenix West Road, Nanjing, P. R. China, 210036

Abstract--More and more conventional groundingsystems made of steel are being replaced with coppergrounding systems in Asian countries. This paper analyzesthe performance of grounding systems made of steel orcopper conductors, taking into account the impedances ofthe ground conductors. A series of computer modelssimulate grounding systems of different sizes in varioussoil structures. Numerical results such as groundimpedance, ground potential rise (GPR), touch and stepvoltages, and potential differences between copper groundconductors are compared with those pertaining to steelgrounding systems. The performance advantages ofgrounding grids made of copper are demonstratedthroughout the comparison. This paper also presentsbriefly a practical grounding performance before and afterthe replacement of steel with copper conductors.

Index Terms-Grounding, ground potential rise (GPR),ground potential difference (GPD), touch voltage, stepvoltage.

I. INTRODUCTION

The quality and performance of grounding systemsare the major concerns in today's power system design,due to the significant increase of short-circuit currentlevels associated with the need to provide the energyrequired by the phenomenal industrial growth ofdeveloping countries such as China. Durable, reliablepower system grounding is essential for maintaining thereliable operation of electric power systems, protectingequipment, and ensuring the safety of the public andutility personnel. A grounding system must be properlydesigned and its performance must be evaluated usingadequate software tools [1-6]. Furthermore, choosing asuitable conductor for the grounding grid is importantnot only to provide satisfactory performance, but also tomaintain the quality of grounding systems over the longterm.

In Western nations, copper has been universally usedin grounding grids for a long time. In China, due to thehigher cost of a copper grid, copper grounding systemswere not used in the past. Instead, grounding systemswere made of steel conductors that have higherpermeability and lower conductivity than copper. Thisraises some unique issues, particularly if the substationsize is large and the soil resistivity is low: it is likelythat there are significant potential differences betweendifferent parts of the grounding system, which could

interfere with the normal operation of electronicequipment inside a substation or even damage theequipment [8]. Corrosion of the steel conductors canresult in an impaired grounding system that could causeserious problems when a fault occurs. Indeed, since theend of the 1980's, many serious power-failure incidentshave been reported and it is found that inadequategrounding is one of the important causes in most of theincidents. Therefore, more and more aging groundinggrids made of steel are being considered for replacementwith copper conductors in China.

The focus of this paper is to analyze the performanceof grounding systems made of steel or copperconductors, taking into account the impedances of theground conductors. A series of computer models areused to simulate grounding systems of different sizes invarious soil structures. Numerical results such as groundpotential rise (GPR), touch and step voltages, andpotential differences between the conductors inside thecopper grounding grids are presented and comparedwith those for steel grounding systems. This paperdocuments the performance advantages of groundinggrids made of copper over those made of steelconductors. Furthermore, the grounding performance,i.e. ground potential rises and ground potentialdifferences (GPRs and GPDs), touch voltages and stepvoltages, are evaluated accurately during a phase-to-ground fault for practical examples before and after thereplacement of steel with copper conductors. Theanalysis and the discussions presented in this paper canbe used as a guide to estimate the groundingperformance for a grounding system before replacingthe steel conductors with copper ones.

II. COMPUTER MODELS & METHODOLOGE

The reference computer model used in this studyconsists of a square grounding grid. Different grids withvarious dimensions in different soil models areexamined. The burial depth of the grids is 0.5 m. Forsteel grids, resistivity of steel is 12 times that of copperand permeability of steel is about 250 times that of air(during faults). The steel conductors have a cross-section of 480 mm2 and an equivalent radius of 1.236cm. The copper conductors have a cross-section of 1602mm and an equivalent radius of 0.7137 cm. The systemoperating frequency is 60 Hz. The following is a list of

1-4244-0111-9/06/$20.00c02006 IEEE.

the parameter settings of the computer models used inthe study, with the reference model described above.

Soil Resistivity: 1 Q-m, 5 Q-m, 10 Q-m, 50 Q-m, 100Q-m, 200 Q-m, 500 Q-m, 1000 Q-m.

Non Uniform: (See Table 1)

Dimension: 10x10 m, 20x20 m, 50x50 m, 100x100 m,200x200 m, 250x250 m, 500x500 m.

Grid Energization Current: 20 kA.

Fault Location: middle of the grid; corner of the grid.

In a conventional grounding analysis approach, agrounding system is generally assumed as anequipotential structure. This would be inaccurate for agrounding grid made of steel, especially if thesubstation size is large and the soil resistivity is low. Infact, the ground impedance of the grounding system hasa significant inductive component, which is not takeninto account by classical grounding analysis methods. Inorder to get accurate results, the analysis of thegrounding performance was carried out using theMultiGroundZ software package [8], which takes intoaccount voltage drops along conductors in a groundingsystem, therefore, eliminating the assumption that agrounding grid is an equipotential.

In order to compute the touch and step voltages, a setof profiles were defined on the earth surface, 1 mbeyond the grid edge with 1 m observation pointspacing.

III. EFFECTS OF SOIL AND GRID DIMENSION

A. Uniform SoilThe grounding performance is significantly

influenced by soil resistivity. Considering the gridimbedded in a uniform soil shown in the above section,the ground impedance, GPR, GPD, touch and stepvoltages are computed for a 200 m by 200 m grid.Figure 1 presents the computed results for both copperand steel grids.

Ground Impedance

EEEEEEEEEEE IEEEEEE fEEEEE .XEEE iEE.XEEEEiVVVVVVVVVVVVVVVVVV VVVVVVVVVV VVVVVVV000000 iViiii000iiiiiii

t: _~-$.Copper: Magnitude_Steel: Magnitude _

Copper: Angle-Steel: Angle

10 100Soil Resistivity (ohm-m)

Let us look at the results shown in Figure 1. For asoil resistivity of 5 Q-m, the ground impedance is0.0132 Q for the copper grid and 0.0234 Q for the steelgrid, a difference of 77.5%. While for a soil resistivityof 1000 Q-m, the ground impedances are 1.129 Q and1.130 Q for copper and steel grids, respectively. Thedifference of 0.12% is negligible. Furthermore, theground impedance is no longer a pure resistance. Theinductive component of the impedance can besignificant when the soil is low or the grid is large. Forexample, the phase angles of the impedance for copperand steel conductor grids are 47.40 and 36.90 with a 1 Q-m of soil resistivity, respectively.

The ground potential rise (GPR) is 261.5 V for acopper grid, 56% of the steel grid (466.04 V) when thesoil resistivity is 5 Q-m. When the soil resistivityincreases to 1000 Q-m. The ground potential rises of thecopper and steel conductors are almost the same, i.e.,45130 V vs. 45020 V.

Let us now consider the maximum touch and stepvoltages in the grid area. In a soil of 5 Q-m and 1000 Q-m soil resistivity, the computed maximum touchvoltages are 54.2 V and 11041 V for a grid consisting ofcopper conductors and 93.7 V and 10580 V for a gridconsisting of steel conductors, resulting in 72.9% and4.10% differences, respectively. Similarly, the stepvoltages have differences of 96.7% and 1.6% for thetwo soil models, respectively. Note that when thecomputation takes into account the potential drop alongthe conductor, the touch voltage is defined as thedifference between the potential on the earth surface andthe GPR of a conductor that is the nearest to the point atwhich the earth potential is computed while step voltageis determined by the difference between the twoobservation points on the earth surface.

From the above discussions, one can conclude thatground conductor characteristics do affect the groundingperformance of a grounding grid, especially when thesoil resistivity is low. Grids with copper conductorshave a better performance compared to the steel ones.

100000

O100000

*_ 1000o

100

a-

- 101-

1000

GPR & GPD

10 100

Soil Resistivity (ohm-m)1000

Figure 1 Grounding performance in various uniform soils.

100

a) 10

c enX anQ (D

-s SDO

'acn 0.1

200a- 0.01

0.0011

Maximum Touch Voltages100000

10000

1000

4._

0_ 1001

10

10 100

Soil Resistivity (ohm-m)

10000

1000

S-a0)

._

~~~~~~~~~~~~~~~~~~~. I100

10o l1;-|0|0 -0 -4

I ~~~~I10 100

Soil Resistivity (ohm-m)

1000 1 1000

Figure 1 Grounding performance in various uniform soils (Continued).

B. Non- Uniform Soil

moloveresylay(Soi

M

Table 1 shows the soils modeled in the study. Soil Table 2 shows the grounding performance for a 200dels 1 & 2 have two horizontal layers with a low- m by 200 m grid for the soil models defined in Table 2.,r-high resistivity and a high-over-low resistivity, From the table, the following conclusions can be made:pectively. Soil models 3 & 4 have three horizontal - For both two-layer and three-layer soil models,ers. Soil 3 has a high resisitivity bottom layer and when the bottom layer resistivity is high, the grid4 has a low resistivity bottom layer. consisting of steel conductors has a similar

performance as the one consisting of copperTable 1 Non-Uniform Soil Models Used in the Study conductor. The differences between the computed

grid GPR are only 4% and 3% for soil models 1Soil Layer Resistivity Thickness and 3, respectively.[odel (ohm-m) (m) - For both two-layer and three-layer soil models,

1 Top 10 2 when the bottom layer resistivity is low, theBottom 500 Infinite influence of the steel conductor resistivity and

Top 500 2 permeability on the grounding performance is2 large. In other words, copper ground conductors

Bottom 10 Infinite are more effective. For example, the differenceTop 200 2 between the computed GPR for copper and steel

3 Middle 20 3 grids reaches 27.5% for soil model 2 and 42.10% for

Bottom500 Infinite ~~~~soil model 4.- Similar conclusions apply for touch and step

Top 10 2 voltages. For soil model 4, the difference between4 Middle 500 3 touch voltage can reach 80.6% and the difference

IBottom 20 1 in1inito between step voltages can reach 78.4%.

Table 2 Grounding Performance in Non-Uniform Soils

SoilModel

2

3

4

Impedance (ohms)

Copper Steel0.59X1.3° 0.62Z.2 °

0.09 10 0. 2Z19.7

0.70X1.3° (72z..00

0.07 110 OJ.I0z20.6V

GPR (V)

Copper s

11933. 11

1851.4 23

13997.0 1

1414.3 2(

C. Effects ofGrid DimensionIn this section, we investigate the performance of

different size of grids made of copper and steelconductors. Figure 2 shows the ground impedance,GPR, touch and step voltages of a copper grid and a

steel grid in a 10 Q-m uniform soil with different gridsizes. The energization point is at the corner of the gridin order to observe better the difference between thecopper and steel grids. It is clear that the difference

Touch Voltage (V) Step Voltage (V)

Copper Steel Copper Steel739.6 861.12 311.2 370.7

1445.6 1821.9 649.72 865.83

2424.4 2620.5 1111.8 1231.6

286.4 517.4 124.1 221.4

between the ground impedance of the copper grid andthat of the steel grid is very small for smaller grid(smaller than 50 m by 50 m). The difference becomeslarger and larger with larger grid sizes. It reaches 146%when the grid is 500 m by 500 m. The same trend canbe observed for GPR, touch and step voltages as well.

One interesting phenomenon causes the groundimpedance curve for steel conductors to increaseslightly for sizes larger than 200 m by 200 m. This is

Maximum Step Voltages

because at grid locations far away from the locationwhere the fault is discharged in the grid, the leakagecurrent is almost out of phase with that close to the faultlocation. Therefore, the ground impedance can be

Ground Impedance

Em

0

L 0. 1

'a)E

0.0110

10000I

=, 1000

100m

I 10x

1

5~

0V

z0~

100Grid Width (m)

1000

Touch Voltages

--+-Copper Conductor

-o--Steel Conductor

10

10000

1000a,

* 100

0.

a,

x

11.

100

Grid Width (m)1000

slightly larger. This phenomenon can also be observedfor copper grids but for larger dimensions, because thecopper has a lower resistivity and permeability.

10000 I .~~~~~~_

1000

+GPR: Copper

-- GPR: Steel

10

GPR

100

Grid Width (m)

1000

Step Voltages

7N

-+Copper Conductor

-V--Steel Conductor

10 100

Grid Width (m)1000

Figure 2 Grounding performance of grids with different sizes

IV. PRACTICAL EXAMPLES

In this section, we compare the grounding performanceof two existing steel grounding grids that were recentlyreplaced with copper ones. Both substations are located inChina. Note that the power frequency is 50 Hz instead of60 Hz.

A. Case 1

The dimension of the 220 kV substation grounding gridis 130 m by 110 m with 8 m by 8 m meshes. The orginalgrid is made of 60 mm by 8 mm steel conductors and isburied at a depth of 0.6 m. It is later replaced by a gridmade of 150 mm2 copper conductors. Several soilresistivity measurements were made inside and around thesubstation in order to sample the shallow depth soilresistivities as well as the deep ones. Table 3 provides thefinal interpreted soil model that was used in the groundingcalculation.

Table 3 Equivalent 3-Layer Soil Model Used in the Study

To adequately design and evaluate the performance ofa substation grounding system, the ground impedance

should be computed and then validated by groundimpedance measurement. It is also important to determinehow much of the fault current returns to remote sourcesvia the overhead ground wires and neutral wires of thetransmission or distribution lines connected to thesubstation. Table 4 shows the computed currents that wereused in the grounding performance computation. Note thatthe difference between the computed ground impedanceand the measured one was about 9% for this case.

Table 4 Computed Ground Impedance and Fault CurrentDistribution

Ground Total Transformer GroundImpedance Cult Circulating Wires(ohms) (kA) Current (kA) (kA)__kA (kA )

Steel0.16/4.3 0 28.37 0.935*2 10.413Conductors

Copper 0. 15/2.2 0 28.37 0.93 5*2 10.200Conductors

The calculation of GPR, GPD, touch and step voltageswere carried out using the MultiGroundZ softwarepackage [8] which takes into account voltage drops alongconductors in a grounding system, therefore eliminatingthe assumption that a grounding grid is equipotential.Furthermore, during a fault on the secondary side of atransformer located in a substation, some current can flowthrough the grounding system from the fault location tothe transformer feeding the fault, resulting in largepotential differences between different locations of the

Resistivity ThicknessLayer (ohm-m) (m)

IUU 'i'

_

I~ ::

L-

I

6--.7r777777..I.

v

grid. The circulating current must be considered undersuch conditions. Table 5 shows the computed results.From the table, we can see that the maximum GPR, GPD,touch and step voltages for the copper grid are 93%, 35%,94% and 80% of the steel grid, respectively. Therefore,the grounding performance has been improvedsignificantly by replacing steel conductors with copperconductors.

Table 5 Grounding Performance

Maximum MaximumMaxi. Maximum Touch StepGPR (V) GPD (V) Voltage Voltage

(V)Steel 2650 403 234 147Grid

Copper 2475 141 220 118

B. Case 2In this case, the substation grid is about 132 m by 110

m with 10 m by 10 m meshes. The original grid consistsof steel conductors that have a 60 mm by 8 mm cross-sectional area. This grid was replaced with a new 20 m by20 m mesh grid that is made of copper conductors with a150 mm2 cross section area. The soil model (see Table 6)was obtained based on soil measurements that wereconducted inside and around the substation area.

Table 6 Equivalent 3-Layer Soil Model Used in the Study

Layer Resistivity

(ohm-m) (m)

The ground impedance was measured and calculated.The difference between the computed ground impedanceand the measured one was about 16%. Table 7 providesthe computed results for the ground impedance and thefault current distribution.

Table 7 Computed Ground Impedance and Fault CurrentDistribution

GroundTotal Transformer Ground

Impedance Fault Circulating Wires(ohms) Currents Current Current(ohms) (kA) (kA) (kA)

Steel rid 0.085Z110 27.15 4.26

Copper 0.074Z100 27.15 3.99G rid

The computed GPR, GPD, touch and step voltages forboth the steel grid and the copper grid are shown in Table8. As can be seen, the maximum GPR, GPD, touch andstep voltages for the copper grid are 86%, 64%, 80% and46% of the steel grid, although the copper grid has lessconductors (a mesh size of 20 m by 20 m for the copper

grid vs. a mesh size of 10 m by 10 m for the steel grid).

Table 8 Grounding Performance

Maximum MaximumMaximum Maximum Touch StepGPR (V) GPD (V) Voltage Voltage

(V)

Gridl 2042 542 214 84Copper 1755 349 172 39

V. CONCLUSIONS

The analysis of the performance of grounding systemsmade of steel or copper conductors has been carried out.The performance advantages of copper grounding gridsover steel grids have been demonstrated both with a seriesof computer models and with practical examples. Themain results are as follows:1. For uniform soils with high resistivity, grounding

systems consisting of steel conductors haveperformance that is similar to those consisting ofcopper conductors. For uniform soils with lowresistivity, grounding systems consisting of steelconductors have a much smaller effective size thanthose consisting of copper conductors. Therefore, acopper grid will provide a better performance than asteel grid.

2. For a two-layer soil, when the bottom layer resistivityis high, the grounding systems consisting of steelconductors behave much like copper grids. When thebottom layer resistivity is low, the effective size of agrounding system consisting of steel conductors willbe significantly reduced. In this case, a groundinggrid made of copper conductors will have a muchbetter performance than a grid consisting of steelconductors.

3. When the size of a grounding grid is small, theperformance of a copper grounding grid is notsignificantly improved compared to that of a steelgrid.

4. Due to the complexity of real soil structures, theperformance of individual grounding systems must beevaluated correctly using adequate software tools toavoid costly over-designs or dangerous under-designs.

VI. ACKNOWLEDGEMENTS

The authors would like to thank Safe EngineeringServices & technologies ltd. for the financial support andresources provided during this research work. Graphs andComputations depicted in this paper were generated by theCDEGS software package.

VII. REFERENCES

[1] G. Yu, J. Ma, and F. P. Dawalibi, "Effect of Soil Structures onGrounding Systems Consisting of Steel Conductors," Proceedingsof the International Conference on Electrical Engineering(ICEE'2001), Xian, China, July 22-26, 2001.

[2] J. Ma, and F. P. Dawalibi, "Latest Anlytical and Computationaltechniques in Grounding," Electrical Equipment, Vol. 2, Specialissue, Oct. 2001, pp.38-44. Beijing, China, October, 2000.

[3] R. D. Southey and F. P. Dawalibi, "Improving the Reliability ofPower Systems with More Accurate Grounding SystemResistance Estimates," Proceedings of the IEEE-PES/CSEEInternational Conference on Power System Technology,

PowerCon 2002, Kunming, China, October 13-17, 2002, Vol. 1,pp. 98-105.

[4] J. Ma, F. P. Dawalibi, W. Ruan R. D. Southey, R. Waddell, and J.K. Choi, "Measurement and Interpretation of Ground Impedancesof Substation Grounding Systems Connected to Ground Wires andMetallic Pipes," Proceedings of the 60th Annual Meeting of theAmerican Power Conference, Vol. 60-I, Chicago, April 14-16,1998, pp. 490-493.

[5] J. Ma and F. P. Dawalibi, "Extended Analysis of GroundImpedance Measurement Using the Fall-of-Potential Method,"IEEE Transactions on PWRD, Vol. 17, No. 4, Oct. 2002, pp. 881-885.

[6] J. Ma and F. P. Dawalibi, "Modem Computational Methods forthe Design and Analysis of Power System Grounding,"Proceedings of the 1998 International Conference on PowerSystem Technology, Beijing, August 18-21, 1998, Vol. 1, pp. 122-126.

[7] F. P. Dawalibi, J. Ma, Y. Li and Y. Yang, "DongShanQiaoSubstation Grounding Analysis," Report 129 (414), Project No.129 (414), Prepared for Jiangsu EPRI, China (DongShanQiaoGrounding Project), October, 2003.

[8] CDEGS Software Package, Safe Engineering Services &technologies ltd., Montreal, Quebec, Canada, November 2002.

VIII. BIOGRAPHIESMs. Yexu Li received the B.Sc. degree inGeophysics from Beijing University and the M.Sc.degree in Seismology from the Chinese Academy ofSciences in 1986 and 1989, respectively. Shereceived the M.Sc. degree in Applied Geophysicsfrom Ecole Polytechnique of the University ofMontreal in 1996 and the Graduate Diploma inComputer Sciences from Concordia University in1998.

From 1995 to 1998, she worked as a Geophysicist with the SIALGeosicences Inc. in Montreal, and was involved in geophysical EMsurvey design, data acquisition and processing as well as interpretation.

She joined Safe Engineering Services & technologies ltd. in Montrealin March 1998 as a scientific researcher and software developer. She ispresently working on AC interference studies and software development.

Ms. Li has coauthored more than ten papers on geophysics andelectromagnetic interference analysis.

Dr. Jinxi Ma (M'91, SM'00) was born in Shandong,P. R. China in December 1956. He received the B.Sc.degree from Shandong University, P. R. China, andthe M.Sc. degree from Beijing University ofAeronautics and Astronautics, both in electricalengineering, in 1982 and 1984, respectively. Hereceived the Ph.D. degree in electrical and computerengineering from the University of Manitoba,Winnipeg, Canada in 1991. From 1984 to 1986, hewas a faculty member with the Department of

Electrical Engineering, Beijing University of Aeronautics andAstronautics. He worked on projects involving design and analysis ofreflector antennas and calculations of radar cross sections of aircraft.

Since September 1990, he has been with the R & D Dept. of SafeEngineering Services & technologies in Montreal, where he is presentlyserving as manager of the Analytical R & D Department. His researchinterests are in transient electromagnetic scattering, EMI and EMC, andanalysis of grounding systems in various soil structures.

Dr. Ma has authored and coauthored more than one hundred paperson transient electromagnetic scattering, analysis and design of reflectorantennas, power system grounding, lightning and electromagneticinterference analysis. He is a senior member of the IEEE PowerEngineering Society, a member of the IEEE Standards Association, and acorresponding member of the IEEE Substations Committee and is activeon Working Groups D7 and D9.

Dr. Farid P. Dawalibi (M'72, SM'82) was born inLebanon in November 1947. He received a Bachelorof Engineering degree from St. Joseph's University,affiliated with the Uni ersity of Lyon, and the M.Sc.and Ph.D. degrees from Ecole Polytechnique of theUniversity of Montreal.

From 1971 to 1976, he worked as a consultingengineer with the Shawinigan EngineeringCompany, in Montreal. He worked on numerousprojects involving power system analysis and

design, railway electrification studies and specialized computer softwarecode development. In 1976, he joined Montel-Sprecher & Schuh, amanufacturer of high voltage equipment in Montreal, as Manager ofTechnical Services and was involved in power system design, equipmentselection and testing for systems ranging from a few to several hundredkV. In 1979, he founded Safe Engineering Services & technologies, acompany specializing in soil effects on power networks. Since then hehas been responsible for the engineering activities of the companyincluding the development of computer software related to power systemapplications.

He is the author of more than two hundred papers on power systemgrounding, lightning, inductive interference and electromagnetic fieldanalysis. He has written several research reports for CEA and EPRI.

Dr. Dawalibi is a corresponding member of various IEEE CommitteeWorking Groups, and a senior member of the IEEE Power EngineeringSociety and the Canadian Society for Electrical Engineering. He is aregistered Engineer in the Province of Quebec.

Mr. Jinsong Zhang was born in Liaoning, China inNovember, 1968. He received his Bachelor Degreefrom Ha'ebin Technology University in Ha'ebin,China. He joined Jiangsu EPRI in 1991. Hisresearch interests are in high voltage technologyand over-voltage protection.