01268173

8/7/2019 01268173

http://slidepdf.com/reader/full/01268173 1/6

IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 40, NO. 1, JANUARY/FEBRUARY 2004 11

Grounding and Ground Fault Protectionof Multiple Generator Installations on

Medium-Voltage Industrial and CommercialPower Systems—Part 1: The Problem Defined

Working Group ReportPrafulla Pillai , Senior Member, IEEE , Bruce G. Bailey, Jim Bowen, Gerald Dalke, Bruce G. Douglas, Jay Fischer,James R. Jones, Daniel J. Love, Charles J. Mozina, Neil Nichols, Clifford Normand, Lorraine Padden, Alan Pierce,

Louie J. Powell, David D. Shipp, Norman Terry Stringer, and Ralph H. Young

Abstract— This paper discusses typical grounding practices andground fault protection methods for medium-voltage generatorstators, highlighting their merits and drawbacks. Particular at-tention is given to applications of multiple generators connected toa single bus. The paper also provides an overview of the generatordamage mechanism during stator ground faults. Problem areasassociated with each type of grounding are identified and solutionsare discussed. The paper also provides a list of references onthe topic. The paper is intended as a guide to aid engineers inselecting adequate grounding and ground fault protection schemesfor medium-voltage industrial and commercial generators fornew installations, for evaluating existing systems, and for futureexpansion of facilities, to minimize generator damage from statorground faults. These topics are presented in four separate parts,Parts 1–4. Part 1 covers scope, introduction, user examples of stator ground failure, and theoretical basis for the problem. Part2 discusses various grounding methods used in industrial appli-cations. Part 3 describes protection methods for the various typesof grounding and Part 4 provides a conclusion and bibliographyof additional resource material.

Index Terms— Generator ground fault protection, generatorgrounding, ground fault protection, grounding, stator irondamage.

I. SCOPE OF PAPER

I N RECENT years, severe damage to bus-connected genera-tors from statorground faults hasbeen observed in numerous

industrial plants. Such generator failures may require extensivestator lamination repairs at the manufacturer’s premiseswith the

Paper ICPSD-IAS 48–01–P1, presented at the 2002 Industry ApplicationsSociety Annual Meeting, Pittsburgh, PA, October 13–18, and approved forpublication in the IEEE T RANSACTIONS ON INDUSTRY APPLICATIONS by thePower Systems Protection Committee of the IEEE Industry ApplicationsSociety. Manuscript submitted for review October 15, 2002 and released forpublication November 4, 2003.

The authors are members of the Grounding and Ground Fault Protectionof Medium-Voltage Generator Installations Working Group of the Medium-Voltage Protection Subcommittee, Power Systems Protection Committee, IEEEIndustry Applications Society.

P. Pillai, Chairperson of the Grounding and Ground Fault Protection of Medium-Voltage Generator Installations Working Group, is with KelloggBrown & Root, Houston, TX 77002-7990 USA (e-mail: [email protected]).

Digital Object Identifier 10.1109/TIA.2003.821638

associated down time. The primary objective of this paper is topresent methods of protecting medium-voltage industrial gener-

ators against extensive and expensive stator iron damage frominternal ground faults. This paper will review the issues associ-ated with various grounding practices and ground fault protec-tion methods to minimize iron damage.

The paper summarizes some basic considerations in selectinggrounding and ground fault protection of generators installed onmedium-voltage power systems with multiple ground sourcesand serving load directly at generator terminal voltage. The dis-cussions also apply to generators installed in parallel with utilitytransformers. However, the paper excludes installations withspecial grounding requirements such as independent power pro-ducer (IPP) connections and mining applications. Also, rotorground faults are outside the scope of this paper.

The paper will:1) discuss factors requiring consideration in selecting

grounding and ground fault protection schemes formedium-voltage industrial generators;

2) identify problem areas associated with grounding andground fault protection of generators, especially formultiple units operating in parallel on medium-voltageindustrial power systems;

3) provide alternate solutions to the identified problems;4) identify items to be addressed in detail in future working

group papers.The paper is organized into four parts. Part 1 covers scope,

introduction, user examples of stator ground failure, and theo-retical basis for the problem. Part 2 discusses various groundingmethods used in industrial applications. Part 3 describes pro-tection methods for the various types of grounding, and Part 4provides a conclusion and bibliography.

II. INTRODUCTION

Many existing and new industrial facilities include multiplegenerators operating on plant distribution buses at the medium-voltage level (see Fig. 1). The trend of in-plant generation on acommon bus is increasing due to low cost and simplicity. Also,

0093-9994/04$20.00 © 2004 IEEE

8/7/2019 01268173


12 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 40, NO. 1, JANUARY/FEBRUARY 2004

Fig. 1. Medium-voltage industrial distribution system with multiple sources.

Fig. 2. Generator internal ground fault —current flow after opening generatorbreaker.

the average size of bus-connected industrial generators is largerin recentyears than in thepast. While the economicsof bus-con-nected in-plant generation is attractive, it imposes on the powersystem engineer concerns and added tasks of careful considera-tion regarding generator protection and equipment capabilities.

The fault type to which stator windings are most often sub-jected is a short circuit to ground. Many incidents of severedamage to bus-connected generators from stator ground faultshave been reported in recent years. It has been recognized byrecent studies that the generator damage is caused more by theground fault current contribution from the generator itself thanfrom the system. During a stator ground fault in a generator, thefault current persists even after opening the generator breaker,thereby causing more extensive iron damage (see Fig. 2). Thesignificant increase in such incidents has alerted users and in-surers. Also, multiple grounding of sources will result in higherground fault currents causing severe damage and coordinationproblems. Therefore, special attention must be given to gener-

ator grounding and ground fault protection.It should be noted that the method of ground fault protectionis directly related to the method of system grounding used.There are many decisions, considerations and alternatives thatshould be carefully examined while designing an adequateand reliable grounding system for increasing personnel safety,minimizing equipment damage and avoiding unwanted inter-ruptions in plant operation. Standards and other publicationswhich cover generator grounding and ground fault protectionare available, but they do not address specific problems asso-ciated with bus-connected multiple generator installations onmedium-voltage industrial systems. Therefore, concern andconfusion exist among engineers regarding the appropriate

Fig. 3. Simplified single-line diagram of example system.

application of grounding and ground fault protection for suchinstallations.

III. EXAMPLES OF STATOR GROUND FAILURES

A large paper mill had the experience of having two gen-erator failures approximately one year apart. Each of the twoair-cooled units, installed in 1971, was rated 15 625kVA, 13 800V. Each generator was wye-connected and grounded throughits own 400-A grounding resistor. See Fig. 3 for a simplifiedsingle-line diagram of the generator protection system. The pro-tective scheme included the standard electromechanical relayprotection as listed in Table I. The total system ground currentavailable was 2000 A from three generators and two utility tietransformers.

The first unit to fail tripped offline as the result of a windingfailure at a position approximately 25% electrically from theline terminals of the Phase 3 winding. The stator winding wasburned over an area of approximately 8 in in length. The coresteel was also burned approximately 8 in, thus requiring itsreplacement. The restacking of the core steel could only beaccomplished by removing the stator and shipping it to themanufacturer ’s plant.

The generator field was shipped to a service shop and theretaining rings were removed. There was splattered copper andsteel with burned insulation from the stator winding imbeddedin the field winding end turns. Copper contamination was alsolocated in the field ’s cooling passages.

The total cost to rebuild the stator core, rewind the stator,rewind the field, and upgrade the generator protection was ap-proximately $1 500 000. The incremental cost to remove, ship,replace the coresteel, and reinstall the stator contributed approx-imately $500 000 of this total cost.

Subsequent investigation revealed that the 400-A groundingresistor was installed in such a manner that the ground leadcould possibly short out 25% of the resistor grid. As a result,the actual ground fault current could have been as much as 20%greater than the design value. The second unit was tripped of-fline due to phase differential relay (device 87) operation. Thewinding failed in the middle of the Phase 1 coil, in a similarmanner to theprevious unit. Thecoil burning was approximately

8/7/2019 01268173


PILLAI et al. : GROUNDING AND GROUND FAULT PROTECTION OF MULTIPLE GENERATOR INSTALLATIONS —PART 1 13

TABLE I

Fig. 4. Generator winding failure.

Fig. 5. Core damage.

10 in long. Thefieldwinding also hadsplattered copperfrom thestator failure on the end turns. The stator core of this unit alsohad to be restacked due to the damage.

Investigation of the second unit revealed that the operatingtimes of the phase differential (device 87) and ground differen-tial (device 87GN) relays allowed the 87 device to pickup priorto the 87GN device. Both units utilized field breakers that wereopened when the fault was detected. It was felt that both of theseunits failed initially due to an internal turn-to-turn short in thecoil which quickly escalated to a phase-to-ground fault.

Figs. 4 and 5 show photographs of the generator parts thatfailed as described above. Fig. 4 shows the generator windingfailure as viewed from inside the unit. Fig. 5 shows the burning

Fig. 6. Typical one-line diagram.

of a stator lamination resulting from winding failure. The Fig. 5photograph was taken after removal of the laminations from thestator up to the area of the failure.

IV. THEORETICAL BASIS FOR THE PROBLEM

The one-line diagram, shown in Fig. 6, depicts a simplifiedindustrial system of a medium-voltage bus with one generatorand one utility step-down transformer. While any resistor ratingcould be chosen for this example, to make it as general as pos-sible, both resistors will be assumed to be rated 400 A, for amaximum system ground fault level of 800 A.

Faults inside generators and transformers will be limited bythe impedance of the generator or transformer winding and,therefore, will be lower in magnitude than faults on the bus.Therefore, the most severe fault condition for a generator (ormotor) is a fault directly at the terminals on the first turn of thestator winding. For the system illustrated in Fig. 6, this fault willhave a magnitude equal to the system maximum of 800 A, with400A flowing into thegenerator from external sources ( “systemsources ”) and 400 A generated within the generator itself. Thepotential damage associated with each of these currents can beconsidered separately and the total damage determined by su-perposition.

An intuitive expectation is that the damage caused by a faultinside a generator is proportional to the energy released in thearc at the fault point. A general expression for the energy re-leased in a fault is

Energy (1)

8/7/2019 01268173



Therefore, the damage associated with a fault is a function of two variables, the magnitude of current and the duration of thefault .

The value of in (1) is also a factor. A value of 2 would applyin the case of purely resistive heating. Various researchers havepredicted values for k for an arc in the range of 1 –2 [2], [3]. Thepurpose of this paper is to address the system design implica-tions of stator fault point damage, not to suggest an exact valueof . Therefore, it is sufficient for this analysis to arbitrarily pick a value for the purpose of illustration.

A. Stator Damage Due to Current Through the Transformer Neutral (System Current)

The technology of stator ground fault detection ranges fromthe conventional stator differential relay to modern detectionmodalities that can detect faults very close to the neutral endof the winding. For this hypothetical worst case scenario, there-fore, it is reasonable to assume that the fault will be detected

and tripping will be triggered with no intentional time delay.Allowing for one cycle of relay time with a five-cycle breaker,the 400-A current from the resistor on the utility step-downtransformer will persist for six cycles (0.1 s on 60-Hz systems).Therefore, a damage parameter associated with the externally-sourced ground fault can be determined by evaluating this in-tegral expression of (1) over the six-cycle period during whichthis current will flow.

The curves shown in Fig. 7 depict the damage indices viewedin two ways. Fig. 7(a) shows that the potential damage increasesas the current rating of the neutral grounding resistor on thetransformer becomes larger. Fig. 7(b) shows how the damageaccumulates with time for the singular case of a 400-A resistor

on the transformer neutral. This curve is plotted on semi-log-arithmic axes in order to depict more clearly the way that thedamage accumulated during the six-cycle period of time priorto opening the generator breaker. Note that all of the damage as-sociated with current from the resistor on the step-down trans-former takes place during this short period.

B. Current From the Faulted Generator

Tripping the generator breaker does not interrupt the currentthat rises through the generator neutral. This current will flow aslong as the generator field remains excited as a forcing function.

Tripping the generator breaker should also trigger tripping thefield, but the excitation will decay gradually under thecontrol of the generator single-line-to-ground short-circuit time constant

. While this parameter does vary from one generator to thenext, it falls in the range of 0.8 –1.1 s. Thus, the damage indexassociated with current produced by the faulted generator itself can be calculated using an expression similar to (1) in which thecurrent is a decaying exponential. This integral must be evalu-ated over the entire period of time required for the current todecay to zero

Energy (2)

(a)

(b)

Fig.7. (a) Energydueto “systemcurrent ”—for variousmagnitudesof current.(b) Arc energy due to 400-A “system current ” over time.

Fig. 8(a) shows the damage associated with thegenerator cur-rent for various ratings of the generator resistor up to a max-imum of 400 A. Again, it is apparent that higher resistor ratingswill result in greater damage. But note that in this instance, themaximum value of this damage parameter is about 5200 W scompared with about 800 W s in Fig. 7(a).

The reason for this difference is apparent in examiningFig. 8(b). Note that the fault energy associated with currentthrough the generator neutral resistor accumulates for severalseconds of time, not just the 0.1 s depicted in Fig. 7(b). This isbecause the fault current continues to flow until the generatorfield demagnetizes; there is no circuit breaker to interrupt faultcurrent through the generator neutral itself.

Comparing Figs. 7 and 8 yields two very important observa-tions.

1) Inthis simplecasewithone generatorandonetransformer,each of which is low-resistance grounded, most of thedamage in the faulted generator is attributable to currentfrom the generator itself. That is, most of the generatordamage is self-inflicted. Therefore, changing generatorgrounding practices would have far more impact onreducingstator ground fault damage thanchangingsystem(transformer) grounding practices. Obviously, increasingthe number of “system sources ” will result in increaseddamage, and with enough external sources, the damage

8/7/2019 01268173


PILLAI et al. : GROUNDING AND GROUND FAULT PROTECTION OF MULTIPLE GENERATOR INSTALLATIONS —PART 1 15

(a)

(b)

Fig. 8. (a)Fault energy due to “generator current ”—for various magnitudes of current. (b) Arc energy due to 400-A “generator current ” over time.

due to system current could exceed the damage due tocurrent through the neutral of the faulted generator.

2) Most of the “self-inflicted ” damage takes place afterthe generator breaker trips. Thus, while the importanceof fast generator protection cannot be overemphasized,faster protection does not necessarily mean less damagebecause tripping the generator breaker does not interruptthe flow of current through the generator neutral.

Large generators are rarely bus-connected. Instead, the gen-erator is associated with a dedicated step-up transformer, andother than perhaps station auxiliaries, no load is served at

generator terminal voltage. This is known as a “unit-connected ”generator (see Part 2 of the paper for details). Because there isno distribution system selectivity requirement, these generatorsare almost always grounded through distribution transformersequipped with secondary loading resistors. In these applica-tions, the worst case ground fault current is typically limitedto 10 A. Fig. 9 shows how potential fault damage is greatlyreduced, assuming that the initial magnitude of ground faultcurrent is limited to 10 A.

It is interesting to observe that while the frequency of ironburning on bus-connected industrial generators is increasing,there are no anecdotal reports of iron burning on unit-connectedgenerators with stator ground faults. One researcher demon-

Fig. 9. Fault energy versus time with 10-A grounding.

strated that a generator can withstand fault currents up to 10 Ain magnitude indefinitely without iron burning [ 4].

An example was given earlier of an actual system that in-cluded three generators, each rated 15 625 kVA and groundedthrough 400-A resistors. In addition, the system included twowye-connected utility transformers with 400-A resistors, for atotal available ground fault current of 2000 A. The analyticaltechniques presented here can be applied to that instance to ret-rospectively predict the fault energy levels that may have oc-curred.

The faulted generator experienced a stator ground fault at apoint that would have resulted in a current in the affected statorwinding of about 75% of the theoretical current that would havebeen allowed by the nominal rating of its own neutral resistor.

However, because that resistor was partially shorted, the actualavailable current was about 20% higher than rated. Therefore,the actual fault from the generator was probably close to thetheoretical 400 A available for a terminal fault.

Because the fault occurred within the winding of the gener-ator, the sources on the system would have contributed less thantheir nominal currents. With a fault 25% of the way betweenthe terminals and the neutral, those sources would have con-tributed a maximum of 75% of their nominal currents. Takingthese factors into consideration, the curves in Fig. 10 show theaccumulation of fault energy versus time. Note that because of the large number of sources, the energy from “system sources ”is significant. However, the energy from the faulted generatorstill exceeds the energy from “system sources ” because of thetime required for the stator fault current to decay.

V. SUMMARY

This paper presented Part 1 of a four-part WorkingGroup Re-port on generator grounding and ground fault protection. Part1 has introduced the mechanism of generator damage duringstator ground faults. Actual examples are given where exten-sive damage occurred even after opening of the generator cir-cuit breaker. This damage is due to the extended time requiredfor the field to decay; thereby, maintaining the flow of currentto the fault.

8/7/2019 01268173



Fig. 10. Fault energy accumulation.

Part 2 of this Working Group Report discusses variousgrounding methods used in industrial applications, highlightingtheir advantages and limitations. Part 3 describes the protectionmethods for the various types of grounding. Part 4 of the reportprovides a conclusion and a bibliography of additional resourcematerial on the subject of generator grounding and ground faultprotection.

REFERENCES

[1] L. J. Powell, “The impact of system grounding practices on generatorfault damage, ” IEEE Trans. Ind. Applicat. , vol. 34, pp. 923 –927,

Sept./Oct. 1998.[2] L. E Fisher, “Resistance of low voltage arcs, ” IEEE Trans. Ind. Gen.Appl. , vol. IGA-6, pp. 607 –616, Nov./Dec. 1970.

[3] H. I. Stanback, “Predicting damage from 277 v single phase to groundarcing faults, ” IEEE Trans. Ind. Applicat , vol. IA-13, pp. 307 –314,July/Aug. 1977.

[4] D. Shipp and F. Angelini, “Characteristics of different power systemsneutral grounding techniques: Facts and fiction, ” in Conf. Rec. PCIC ,1990, pp. 107 –116.

Prafulla Pillai (M’88–SM’95) received the B.Sc.(Electrical Eng.) degree from Kerala State Univer-sity, Kerala, India, in 1967, and the M.Sc. (ElectricalEng.) degree from Queen ’s University, Kingston,ON, Canada, in 1977.

She was with Fertilizers and Chemicals Company,India, from 1967 to 1975, primarily designing elec-trical systems for chemical plants. She was involvedin engineering consulting services in Canada from1977 to 1980. She was with Ontario Hydro, Canada,from 1980 to 1993, initially designing utility substa-

tionsand later studying electrical power systems in theirResearch Division.Shejoined Brown & Root (now Kellogg Brown & Root), Houston, TX, in 1993, asSenior Electrical Engineer. She currently holds the position of Senior TechnicalAdvisor. Her primary responsibilities are in the areas of power system anal-ysis, protective relay coordination, and relay setting calculations for industrialinstallations.

Ms. Pillai is a Registered Professional Engineer in the State of Texas andin the Province of Ontario, Canada. She holds several positions in the Industrialand Commercial Power Systems Department of the IEEE Industry ApplicationsSociety and its committees and working groups.

01268173

Documents