Document 4Qb44OrvJK51GMwJdLXw6ERQp

SUPPLEMENT 3 MINIMIZING CAPACITOR CASE RUPTURE - AN OVERVIEW J* John E. Harder Advisory Engineer PO O R Q U A LIT Y H R IG IN A L R.D. McClain, Manager Capacitor Equipment Engineering W estinghouse Electric Corporation Bloomington. Indiana George E. Mercier, Manager Capacitor Unit Engineering A bstract This paper discusses those factors of design, manufacture, and application which most affect the reliable perform ance and minimization of case rupture of capacitor units on electric transmission and distribution systems. It is the objective of the paper to help the casual capacitor user Identify Items which are most important in the selection and application of capacitors to assure reliable performance As an aid in this regard, we would like to share number of observations and experiences which we fe may be helpful to some users in improving the rellabili of their capacitor installations. These comments are primer directed at shunt capacitors for power factor correction i utffity systems at 2 4 0 0 volts or higher, in equipments I pole top installation, in metal enclosed equipment, and open stacks. Background During the past several years there have been numerous discussions relatino to capacitor reliability and case rupture performance. This high level of attention probably stems both from increasing emphasis on safety and environmental protection, and from several incidents of high capacitor failure rates. Most of these discussions have been related to capacitor case rupture curves, capacitor fusing, and testing of capcitor units to improve reliability. While the reHabllty of Westinghouse power capacitor units has probably been the best in the industry over the past several years, we nevertheless have had incidents of higher than desirable failure rates and have been involved In discussions of capacitor reliability and protection with many users. Several observations emerge from this experience: 1. There appear to be some misconceptions regarding teats to predict the reliability of caoadtor units. 2. Some of the field problems which have been ex perienced may have been aggravated by application oversights. 3. In a number of cases the bank protection has been defective or Inappropriately adjusted. 4. In some cases, a more thorough examination of the Quipment during installation and/or more thorough or regular maintenance examinations might have helped avoid problems. As a result of these observations, we have come to the conviction that users seriously interested in reducing case n a tu re probability need to review ail aspects of capacitor selection, application, instaiatlon. protection, and main tenance for their equipment installations. * Case R upture Capacitor casa rupture, if it occurs, may pose sev p roblem s: 1. Potential hazards to people and property from force expeKed narls and possible fire. 2. D ielectric fluids leaking onto painted surfaces may finishes. 3. Potential environmental hazard (partlculaily In the i of older apparatus containing polychlorinated tuot (PCBJ)and the necessity for special cleanup, disp and reporting. This vulnerability exists in addition to the cost of system vars and the costs of repair, maintenance replacem ent. The severity of a specific occurrence may deoen many factors. Rupture size may vary from a small fracture (with 'w e e p in g ' fluid loss) to disintegrate the capacitor and broadcast of parts and material. Factors affecting the severity include at least: 1. The cause and location of the initial dielectrir 2. Capacitor unit design and rating. 3. Fault protection employed for the unit and eqirir 4. Equipment arrangement (senes and parallel gt grounding, etc.). 5. System characteristics. For example, a fault at the too edge of a winding c by transient overvoltage might give a more violent r of energy than a puncture brought about by overt in the central portion of the winding. Or, ca container with large, flexible sides might accommod. slow development of gas pressure from a low * arc better than a capacitor with a smaller mor container. Available system fault current, or the of a nearby, parallel connected c s d a L ..^ w y m the power input and rupture violenci fe :? ! * NPC00026350 753687 Case rupture is not one uniform phenomenon with one ultimate consequence. It is. rather, the result of many varied sets of circumstances, with many possible outcomes. About the only two things that are uniform about capacitor case ruptures are: 1. The case if fractured at least to the point where fluid may (eakout. 2. It is undesirable. Rather than analyzing the details of all of the possible case rupture phenomena, in this paper we focus on those factors which experience indicates wiH tend to minimize the numbers and violence of case ruptures. THE CAPACITOR UNIT The single most important factor in minimizing capacitor case rupture Is the reliability of the capacitor units them selves. The reliability Is determined by both the capacitor design and by the manufacturing quality control processes. Type T 2 3 4 5 6 Design AFPFA AFPF AFPFar AFPF AFPF AFPFA TABLE 1 Partial Oiseharg* Voltsgn t 25aC (Perant RUM Vi Flute Potycnonnai&d oohenyi PoiycNow*ad twmenyt PDIV 1 S8 200 PDEV 117 1 133 Isoprooyttnphenyl 250 167 Phtnaiaie e s te r* !rchfcjrob*rvzene 236 139 Butyfied monocworotohanytowae 208 167 Mineral Ori 170 65 A m conventional tori FPF 75ga. A n , 45ga. paoer. 75g a h*n> narrow foil wrte railed edges PCtv m Partial dacnarge inception vottage PDEV Partial oacnarge <j*t>ncbon voltage Resign and Testing Basically, capacitor designs are determined by matching the field performance requirements with sufficient design performance capabilities to insure adequate margins be tween the two throughout the anticipated period of use. The capacitor design must take into account a complex set of physical, chemical, electnce] and mechanical properties Involving the materials used in the capacitor as well as the manner in which they are combined to make u p the capacitor. Potential capacitor materials and designs are evaluated by utilizing test methods that have previously been found to be particularly useful and by continually devising new tests to make more accurate assessments of the materials or capacitor designs. When film-type capacitors were introduced more than 10 years ago. the problems associated with capacitor operation at high ambient temperatures were essentially eliminated due to the low temperature rise of this type of capacitors. The limiting design factor was shifted from a thermal to a partial-discharge-fnception-voltage problem. Therefore, more and more attention has been paid to overvoltage endurance testing and partial discharge testing of modem capacitors. Partial discharge testing is used in many types of capacitor evaluation tests. One of the traditional methods is to measure the partial discharge inception and extinction voltages of capacitors using either electronic or ultrasonic type detection equipment. The partial discharge performance is related to both the impregnating fluid used and to the design of the winding. (Table 1) Because of the Improvement in partial discharge inception and extinction due to the use of rolled foil construction, Westinghouse high voltage capacitors utilize this design improvement to provide an increased design margin. The partial discharge inception voltage (PDIV) and partial discharge extinction voltage (PDEV) ore also measured as a function of dielectric temperature (Figure 1) Figure 1 Partial Discharge Inception Voltage Typical curve: Variations occur with dielectric designs Another useful test that has been used is an overvottag cycling test. This is used with several variations. Type * is shown in Figure 2. Figure 2 Overvoltage Cycling Test: Type I In this test, the capacitor is tested for 24 hours (46 cyctes) at each V( level (raised in 0.1 Vt steps each 24 hours), until failure occurs. During this -test the partial discharge intensity is monitored and is a measure of the dielectric degradation that is occurring durmg the test(Tab!e 21 Design AFP APP AFP TABLE2 Fluid V( F a illira Lavala H aled V PCB PC8 Wemcol 1230 260 260 A - convfnbcnflJfori * narrowforiwitnroWdd9 F P 160 ga Mm. *S g a pap P C B potvchtoniwuatf UdflyrtnL- narrow u tk q m edge fori Warned isopropyl O w e 1n IIllliJUIJllJM IIlUM IlillM JilIJiiMWiMUIlMiiJ ff W m P NPC00026351 753666 , This lest is also pertormed to determine the effect of low temperature in a Type II lest (Figure 3). in this test the capacitor is cooled to a low temperature before the test and re-coofed after 15 cycles to maintain the dielectric within a narrow range of low temperature. Results are shown m Table 3 Still another overvoltage cycling test,Type iv. is made as shown in Figure 5. In this test, the samples are tested to failure (Table 5) In addition. Hie partial discharge intensify is monitored during the test (Figure 6) and Ihe length of time that the partial discharges persist after an overvoltage is measured (Table 6) Figure 3 Low Temperature Overvoltage Cycling Test: Type II Design AFPFA AFPFar TABLES Fluid PCB Wemcol Temp *C 25 40 V Failure Level S Rated V 215 250 A conventional toil FPF - 75 0a. Mm. <35(ja paper. 75 ga ftim ar m narrow loti with rotted edges Design AFPFA AFPFar AFPFA AFPFA AFPFarAFPFA AFPFA TABLE 5 FluW PCB PCB DEHP + 2 5 *. TCB Wemcol Wemcoi PIMP M C S123S Cycles Ip Failure 47 167 130 > 1340 > 1340 10 500 A* FPF DEHP * TCB CHNP MCS1238 conventions! foil 75 ga litm 45 ga. peoer. 75 gs Irim narrow rolled edge toil rtietnythexylpwnatate. tncWoroben2ene diisononytpnthaiate isopropytbipnenyl 4 sullone 305 ' In another variation, Type III, the V, value is maintained at a constant value and the test capacitors are cycled for a predetermined number of times (Figure 4) After the overvoltage cycling, the power factor is measured at 100*C as a means of detecting effects of dielectric degradation (Table 4). Oastgn AFPFA AFPFar AFPFar Fluid PCB PCB Wemcol TABLE 4 Initial Temp P.F. 100*C 25*C 050 25*C 040 25*C .047 A conventional lod FPF 7Sga fam, 45g a. paper. 75go film ar - narrow rolled edoetoi) Final P.F. I0 9 *C After 1000 Cycles OPR 044 .0 4 9 no -4H OI Figuro 5 Overvoltage Cycling Test: Type IV Figure 4 Overvoltage Cycling Test: Type III Figure 6 Oscillogram from Overvoltage Cycling .. le s t: Type IV . f NPC00026352 753689 tXnfgn A*PFA A F P F ir AFPFA AFPFaAFPFA Fluid PCB PCB W im cot Wemeot DMP 25% TCB TABLE6 P ittili O lsefiargi lirtin illy -Pleoeoulomto I Sec Owing OV 4000 SOQ 400 200 1500 2 C y d M After OV 200 80 150 <2 200 After OV <2 <2 <2 <2 50 A m conventton il tod FPF m 75 ga; Wm. 45 gi- pumt. 75 ga. fflm a r narrow ro ledad y* foa OEHP rDattiyfhaxyfaftlfialata TCB tneNocobenzw* Figure 8 Low Temperature Switching Tests A stater test to simulate effects of transients* oscillatory discharge stress test (Figure 9). Voltage capability is also measured by step-stress tests 3 (Figure 7). In this test the voltage is maintained at a fixed value for a specified period and then is increased to the next higher voltage step until failure occurs. -si1 'I Partial discharges or power factor may be monitored either during or after each step (Table 7). When this lest is used at low temperatures, the capacitors are re-cooled after each voltage step. y-- I f . ' W tow*, A n /* . n /, V V U ty ti V T -T R gure9 Oscillatory discharge Test POOR The capacitor Is charged with a dc power si QUALITY a voltage Vt and discharged at a pre-selected and reversal. Then an ac power source is swttc ORIGINAL at a voltage Vt for a spectfied time. The number ot cycles required to cause failure is measured (TaNt fMC WNUTtS TABLE 8 PM kV M H z . AMS m .c v Figure 7 Step * Stress Test D M iQ fl AFFFA A F P F ir Fluid . V K R M d PCB nVnr-iV--I-C--Q-1I 300 300 V . % R itid 100 100 lo FM 30 > lO.fi AFPFw wvmcQi 400 100 1.4! AFFFFw Wwncol 400 100 4 .1 ir TABLE 7 D u lg n AFPFA AFPFr A F P F ir Fluid PCB W im cd W im cel Twnp *C *40 4 0 55 V llig Step i t F itlu ri. % R itid 140 IS O 140 These are but a few of the overvoltage tests u * assess the voltage capability of capacitors, but the among the newer tests devised for this purpose. In ge these tests all predict that most norvFCB fluids P* A b conventionalfofl FPF b 75 Qt. ftm, 45 gt. piprr. 75 ga. 13m m m nmowreMdedgofoR much better than PCB fluids using the same fita dielectric system. They also indicate that, of the PCB fluids compared, the Wetncol (tsopropyibip impregnated systems perform with the best results exception. Low temperature switching tests ore also used to check the effect of transient overvoltages. Capacitors are switched for a specified number o f times at increasing voltage levels M anufacturing and Q uality Control im tl failure occurs. The switching transient is typically Design tests such as those discussed above are e a per unit voltage of 1.75 times the test voltage level: in establishing the capability of a system- to meet The continuous test voltages are maintained for approximately two minutes, then turned off for a few seconds and switched on again with the associated transient. This switching test is usually made at various temperatures to establish a relationship between switching voltage and temperature (Rgura.8). performance requirements. However, when large Q< of a particular product design are actually marnri only strict adherance to the use of proper manul techniques and materials will insure that the a quality is produced. The verification that the proper te< and materials have been used comes from produ 3 -4 zsnm m sm NPC00026353 753690 spection and tests For a oroduct such as capacitor units, a heavy emphasis should be placed on production (routine) testing of every capacitor produced. Although this adds some initial cost to the product, it gives the maximum assurance of a reliable, trouble-free capacitor with lower long-term costs. This is a desirable goal from both the user's and the manufacturer's viewpoint. Capacitors are a static device which should require little or no maintenance and also should have a long, useful life with a fow field failure rate. CAPACITOR UNIT APPLICATION The rather targe number of field problems which h been aggravated by inappropriate unit selection has bi somewhat surprising to us. White we are not naive lene to believe that the basic application documents' (Af IEEE Standard 18-1969 or NEMA C P I) are throuj understood by all individuals involved in selecting capacit we are .omowhat surprised at the apparent 'lack awareness of factors which need to be considered i selection of capacitor units. There aie many manufacturing techniques and tests used to make high quality capacitors. Some of the, more critical of these are: In selection of units, basically four items need to considered: voltage, current, kVAC, and tempera! Voltage f The recommended overvoltage application limit capacitor units ta illustrated in Figure 10. This figure ap 1. Winding in a dean room with temperature, humidity to the voltage across the active element of the capa and particle control. (Note: for units not connected to the case, the 4 2. Dry overpotential tests on windings arid assemblies. applies to the insulation between the active element an 3. Dry capacitance and discharge resistor test. case. The *BIL* for units connected to the case r 4 . Dry leak test of container and bushings. only to the strength of the bushing(s).) ' 5. Batch-type drying and Squid filling process. 6. Thorough heat treatment to insure complete impreg nation. 7. Power factor measurement at maximum internal tem perature limit at rated voltage - all capacitors (routine test), 8 . Ultrasonic partial discharge measurement to assure com plete impregnation has been achieved all capa citors (routine test). 9. Overpotential test on a!) capacitors troutine test) at a dc voltage of 4.3 or 6.25 times rated voltage for 10 seconds (test level selected by the customer with regard to application requirements). 10. Final leak test by heating units to maximum dielectric temperature limit for a minimum of six hours. Some of the circumstances which have result exceeding the recommended limits include: ( 1. Restnking of capacitor breakers or switches produced voltages in excess of the recbmmi impulse limits. ( 2. Voltage rise across short-circuiMlmiting ior . reactors in series with the capacitor has rc in higher than recommended continuous v 3. Group fusing of three phase ungrounded ca banks has. at the time of unit short circuit, n in overvoltages for units in the unfaulted i exceeding the recommended limit. ! 4. If the substation bus voltage continuously runs than nominal, on banks with several series'gro increase in voltage due to the loss Iof units in one aeries group may result in a in excess of the recommended voltage These production tests on sQ capacitors win assure a customer that each capacitor received has been checked against potential problems in each of the following basic areas of concern: PO O R balance of the units in that group. A cons application would consider the maximum appec (including any addition to crest voltage from hai plus the maximum allowable unbalance. I 5. High harmonics, in addition to leading to UTY current and thermal overloading, can also a cienlly to peak-to-peak voltages across the i P otential Problem A ro* Th*rm* tatoeMcatrength Production A s tu rin c o T o i P F. at rated oOaga at maximum diatoctrie hotepor temperanua W ovarporttnttai last to bring about internal partial discharges a degradation of the dielectric. Such appjicati occur when standard shunt capacitors; arc to power filters on the basis o f fundamental ii PwttiiUcharoes Rutotookag* Ultrasonic Mtoaf dfeeharoa measuamant Mgti (moeratur leak,0 * * requirements and without consideration of short time, and continuous harmonic currents i In determining the coordination between jarre capacitors, the protective level of the arrester fc When these production techniques ana tests are employed 1.2 microseconds and longer can be compare and are backed up by oroper design tests, the capacitor (without margin) with the 4 times per year rece u* * f wflt have assurance that he w fi aaftraiiy receive impulse voltage application limit of the capacitor capacitors that w il perform reliably,on his system. capacitor tends to act as a filter, eliminating tti NPC00026354 753691 Y*l ri'livj consider fronl-of-wave sparkover for the arrester. The *m njgin' is inherent in the selection of the 4 times per yai limit along with the fact that arrester values are maximum Also, the presence of the capacitors at the arrester location eliminates the need for considering the s to tt time reflections which need to be considered in trattntoim er and switchgear protection. In addition, the presence of the capacitor on the system tends to ac* as short drcuit to high frequency transients on the system, end win often reduce otherwise excessive voltages to bttiow arrester sparkover levels. Applied voltages in excess of the recommends^ limas make the unit vulnerable to damage. T$ that the reliability of the installation is in excess of the recommended applcafl& be avoided, to minimize the risks possible increase in failure probability. u 4 OPERATIONS 3.4 U 1.0 40 OPERATIONS 2J 2.1 00 OPERATION t \ 000 OPERATION IS \^ L h ilGfNAL k; t * - * V $ .9 POWER FREQUENCY " * 1 DURATION------SECONDS B - ------------- -- T 3-- PIQURE 10 CAPACITOR OVERVOLTAQC APPLICATION GIXOE NPC00026355 ._ i 763692 ilion 1 i Current The recommended overcurrent application limit tor capa citor units is shown in Figure 11. The subject of overcurrent application has been under consideration in both NEMA end IEEE standardizing committees, the recommendations may cnange in the near future. Also, the overcurrent capability of tuses, circuit breakers, bus work, C fs , etc. used In capacitor equipments may not always be the same as the capacitor units. For these reasons, applications falling in the shaded area of Figure 11 should be checked with the proposed manufacturer of the equipment. The situations for which overcurrents most often need to be considered are: 1. Large rectifier installations 2 Large arc furnaces or welders 3. Large SCR controlled equipment 4 Tuned raters 5. Resonant system conditions 6. Small capacitor banks frequently switched in with large banks Experience indicates very tittle problem to capacil from current alone. The most common *current* p have been baked-out fuses and burned-out currer formers due to harmonic overcurrents. Caoacitc are more commonly limited by total kVAC than by alone. POOR UALITY OR1G [NAC $ 1000 e m< u0. 5 100 u tz s* 1 10 > 3 sBE V ""4 0pfJVr. AOOpcrJYr. M O p ofJY r. 4000 OporJYr, Thl* curra In ttrrrShorn oram riteg* applk4iton ~i *IMPULSE CURRENT Tj g-jOOOPEft/V^ cowtiwDaur '* 1 __5 10 100 1000 10 100 lO t MILLISECONDS M C O N D f^ HOUSE 11 CAPACITO OVERCURRENT APPLICATION QUlDC KVAC The total continuous kVAC in a capacitor unit should be limited to 1.35 times nominal. In determining the total WAC, factors to be considered include' system voltage as a Per cent on nominal, capacitor unit capacitance as a per c* nt o f nominal (standard fimits *0. +15% ), In d harmonic Whrent content. The system conditions which may lead to exces are essentially the same as those listed undi (above), with the exception of switching. These Ir w it usually result in the selection of higher the voltage capacitor units to meet both the voltage requirements of the installation. A failure to the units appropriately has resulted In units be* in excess of their recommended kVAC limit witl excessive temperatures. N P C 00026356 753693 - ................ ........... & i m s i a !0 & ii& im . <h Am bient Temperature Standard Individual capacitor units ere intended to be tnermaily stable at 4 6 *C and are intended for a normal annual average ambient temperature of no higher than 3 5 'C . More realistic limits for equipments are 35-40*C for the hottest day expected at the location and an annua) average of 20*25*C . Large capacitor installations inside of buildings ki hot climates (and especially those close to large furnaces) need careful consideration of ambient temperature. The heat dissipated by the units themselves can add to the ambient temperature In an enclosed area. For such high temperature applications, an ambient temperature alarm and trip relay is appropriate. Capacitor unit voltage, current, kVAC end ambient tern* perature are important considerations in avoiding vulnerability to capacitor units and equipment failure. Avoiding unit failure Is the first and most important step In minimizing the probability of capacitor ease rupture. PROTECTION Sophistication of protection is related to the size of the capacitor installation and the effect on the system tn the event of unit or equipment failure. Individual unit or group fusing serves as the first line of protection for larger installations. Unbalance relaying quickly removes the bank from the system in the event of excessive numbers of unit fuse operations or other failure within the bank. Major fault protection Is provided by system overcurrent relaying or overcurrent relaying in the capacitor switching circuit breaker. A number of observations have been made concerning each of these areas of protection, which wttt be discussed. Fusing The optimum fuse Is the smallest, fastest fuse which wHI not result in nuisance operations and which will meet the voltage, fault current, and energy requirements of the location. Fuse-case rupture coordination has been the subject of much discussion and confusion. Fuse total clearing curves are, relatively speaking, the result of a rather well standard ized test procedure, and result from a rather uniform melting phenomenon from fuse fink to fuse fink at a given current level. Over the years rather consistent procedures have been developed tor fuse-fuse and fuse-relay coor dination. Even in this rather mature area of coordination, there are occasions! reports of mis-coordination. Capacitor case rupture Is not one rather uniform sequence of events as in fuse fink melting and clearing. There are many modes of failure resulting in different probabilities of case rupture. Even in one kVACrating, the physical volume of the can. Internal connections, operating temperature, as well as many other variables probably have some influence on the case rupture characteristics. Note that the normal current in a 2 4 kv unit is more than eight times the current of a 20 kV unit of the same kVACrating. The normal current through a 2.4 kV capacitor is the same magnitude as the current in a severely faulted 20 kV capacitor. This distinction is not normally made at the low current end of the case rupture curves. The great majority of case rupture tests reported have been started with essentially completely failed units, tn the field, there is at (east a possibility that the unit fault will develop gradually, pack by pack. Perhaps ttw presence or absence of parallel capacitors will have ar effect on toe internal gas generation during this oartia&y faulted condition. Capacltivety limited fault currents in higti voltage shunt capacitor banks may have different arcing and gas generation characteristics than inductively limiter fault currents on a distribution feeder or in toe laboratory There may be a great many differences between th conditions surrrounding a faulted capaeitor in service thar the conditions surrounding a pre-faulted capacitor under going case rupture tests in the laboratory. We understand that some users occasionally re-fuse caps citors that are not noticeably bulged, without testing then (a practice which we would discourage). This may rosul in energizing fautted units a second time, increasing thi probability of case rupture. Depending on the netur of toe fault, toe fuse may clear before toe unit show any visible signs of butging. Several users have commented that their observed cas rupture probability is higher than toe expected 101 based on coordinating a fuse with the capacitor cas rupture curves (which are plotted at the 10% probability < case rupture level}- Considering all of the possible diffe ences between field conditions and laboratory test cone tions, this is not entirely surprising. We regard the cas rupture curves as as giving an indication of capacitor cas rupture performance under a stated set of condition The process of case rupture testing has led to maf improvements in case rupture performance. The ca< rupture curves, however, are probably not an adequs basis for predicting case rupture probability o f fa capacitors under a wide variety o f field conditions. FM experience with given fusing practices is the ulttnu measure ol case rupture probabilityC ertainty we'YKW not suggest using fuses rated higher than suggest by toe case rupture curves, but tower rated fuses m ay1 desirable to ^ ^ j^ y p e s of Installations. QUALITY Cun On distribution banks where system available fault curr exceeds about 5000 amperes, some form of current timii fusing is desirable to avoid excessive internal fault enei Also, on large substation banks with many units in pars current Smiting fuses are recommended to limit the en< discharged into a fault from paraHef units: HPC00026357 753694 ft' . ?y POOR QUALITY Relaying |f% K |A ia j A i and that nameplate ratings be verified. Some failures U i w S T I n L have occurred because the wrong rating of fuse or Inappropriate relay settings, defective relay systems, and capacitor was applied. absence of relay protection altogether have and can result in extensive capacitor bank damage in the event of arcing within the bank. During installation, follow the manufacturers recommenda tions in ensuring connection with adeauate bus or lead conductors with capability of at least 135* of rated Smaller group-fused banks {up to about 1200KVAC at nominal capacitor current. Normally the capacitor circuit 12 kV) usually involve no special relaying protection. Back' breaker' overload trip is set to trip between 135% and up protection is provided by system relaying. Relaying 200% of rated nominal capacitor current. AH bus joints should at least be considered for larger banks using and connections to terminals must be properly prepared individually fused capacitors. and tightened as a loosened, overheated, melted bus In considering relaying, at least three kinds of protection joint can induce a restriking bus fault with disastrous need to be considered: consequences to the capacitors. 1. Unbalance protection. Fuse operations on some units Depending on the method of protection against an internal may cause the voltage to increase on other units bank fault, it may be necessary to establish very close in the same group. Unbalance protection may give balance between the capacitances of the phases. Wher an alarm for a small amount of voltage rise and trip individual capacitor rack capacitance values are provided the bank off to avoid damage from excessive voltage arrange the racks in accordance with the manufacturen rise. recommendation. Check the capacitance value for esc! 2. Group arc-over protection. For some bank configura phase, independently, to ascertain that values are matche tions, an arc In one capacitor group will have limited fault current because of other groups in series. Transients associated with such an arc can damage capacitors, blow fuses, cause violent case ruptures from high voltages, etc. The bank should be dis connected quickly {certainly within one-half second) for a group arc-over. within tolerance. For banks connected wye-wye. compar the impedances of sections 1 and 2 of the same phase as well as the impedances of the three phases of th same section. Use these impedance checks to calcula' the effectiveness of the relay protection scheme. D withstand dielectric strength tests might be made on bai and spacing insulators to detect any cracks resulto 3. Phase-to-phase and phase-to-ground faults. These from shipment, handling, or installing. Upon initial energizati faults should be removed Quickly, and without re* of a large capacitor bank, the signal to any relays us dosing to avoid high transient voltages and currents. for detecting unbalance should be checked to reassi Sometimes this protection is provided by bus fault proper initial balance. or feeder protection. There has been some mention in the industry of the use of pressure switches in individual capacitor units to detect low current faults. White there is little doubt that pressure switches can reduce the probability of case- rupture from certain types of low current faults in switched capacitor equipment, there is some question ea to the economic viability o f wide scale application of pressure switches. Pressure switches are probably not a viable alternative on individually fused capacitors in large assemblies, particularly at transmission voltages. MAINTENANCE Regular maintenance and inspection can help reduce, probability of trouble and limit the possible consequent As a minimum, visual inspection for blown fuses, dama insulators, bushings, or containers (for leakage or swell is recommended. Polluted atmospheres or onvfronm1 may require periodic cleaning of porcelain insulators bushings. Weathered surfaces should be refinishec required. SHIPPING AND INSTALLATION In many cases transportation from the manufacturing Plant to the Installation site is the most severe duty required of a capacitor equipment. capacitor equipment should be thoroughly inspected Look especially for: (1) leaks, (2) loose J'JjttjW , terminals, or fuses, (3) broken porcelain, t ) dented containers, (5) broken frames or busses. Any "triage revealed requires restoring the parts to original condition or replacement of parts. ^ttong Instructions should be followed and proper lift Pi"0** used in unloading or moving. V j urth#r *Pflfloted that a check of Jhe material tvcriwed be made against the erection drawing requirements Blown fuses can result from improperly installed tu excessive harmonic or inrush current, external arc fault in the capacitor. Before replacing the fuse restoring service, inspect the capacitor for visible sij external damage and after shorting the capacitor, rec the capacitance. A small, hand held qapacttence r (such as the Doric Model 130 C-Meler) is conve and useful for this. Oo not re-fuse or reenergize a It or partly faulted capacitor as the damage suffered c the initial fuse operation may be such as to cat high probability of case rupture during the second ooeratton. In event that no damage to the capacito be detected, re-fuse with the correct fuse eteme all cases when energizing a capacitor after re-t observe safety precautions against potential failur case rupture. aSfcMWKfcfltft&S NPC00026358 753695 t LITY At least several capacitor failures have resulted in A L and recresents Ihe loss of a targe number cf damage to capacitor installation because of apparent lack units. of maintenance of the protective system Regular checks should be made that the correct current through a CT (or Filter Banks voltage on a PT) will cause proper relay, alarm, and While overvoltage and overcurrent guidelines were given breaker operation. Protective gaps on CT's should be in an earlier part of this presentation, the design of filter checked for proper clearance id breakers should be bank3 should almost always be done in cooperation to-potted on e reasonably regular basis. with the manufacturer. Mounting of reactors, connection If there has been a major failure resulting in external arcing in a capacitor Installation, the entire Installation should be thoroughly inspected before hying to re-energlze it: and protection of capacitors, selection of ratings, etc. can almost always be aided by the manufacturer's experience with previous banks. 1. AII capacitors should be checked for proper capaci tance; installations with several series groups should SUMMARY have the capacitance balanced so the the voltage Capacitor case rupture is the result of many, varied divides equally among the series groups. phenomena. Most important in minimizing the probability of 2. All capacitor units should be cheeked for excessive case rupture b the correct application of reliable capacitor corrosion or bushing contamination and any indication units. In addition, proper protection can help minimize the of leaks. Units should be painted and bushings cleaned probability of rupture In the event of failure. as required. Leaking units should be eliminated. 3. AB fuses should be checked for continuity, expulsion fuses cheeked fo r tightness. 4. All insulators should inspected and any badly chipped ones replaced. 5. In addition, the relaying check suggested earlier should be done before re-energizing the bank. Many of the factors involved in capacitor reliability application, and protection have been discussed Satisfactory performance b not a matter of fusing alone, or of capacito construction alone, or of protective relaying alone, or c application criteria alone....but requires balanced attentio to aHof the factors as outlined in the paper. Based on past experience, a thorough examination of the bank wfll likely reveal some unexpected Hems. References -- OTHER CONSIDERATIONS 1. American National Standard for Shunt Power Capacito S w itch in g (ANSI C 55.1, IEEE No. 18} Inappropriate selection of switches or breakers for capaci tor switching has and can cause problems for both the 2. NEMA Standards Publication tor Shunt Capacit NEMA CPI-1976) capacitor bank and the system. Switches used tor capacitor switching should be rated for capacitor switching; the con* Recent A rticles Relating to Cepaeltor Failure .1 tacts should not bounce on closing or restrike on opening. 1. Protective Characteristics of Current lim iting Capacl ii It is also desirable to test the breaker or switch in place after Installation but before energizing the capacitor Fuses, W.J. Huber and R.H. Amundsen. IEEE Pai F 75 538-9 bank. Interuptera should be topotted in place. It Is also 2. Safe Shunt Capacitor Operation, M.B. Delaney ( it! desirable to check the timing of interrupters where there sented to Pacific Coast Electric Association. Ma Jj are several Interrupters In series per phase. Shipping 2 0 *2 1 ,1 9 7 5 damage, which may not be obvious by visual examination 3. Shunt Capacitor Tank Rupture Considerate 'i * o f the breaker, has caused breaker malfunction end resulting capacitor bank damage. Lawrence M. Burrange, IEEE Paper A 76 04 4. Optimum Shunt Capacitor Group Fusing, John *i i/ Bank Configuration Grounded vs. Ungrounded. Grounded banks result in Harder, IEEE Power Apparatus and Rvtm* Ma April 1977 5. Capacitor Case Rupture Withstand Capability and f f* I I lower transient voltages on the capacitor switch and may Protection Considerations; R.A. Pratt. W.W. Olive, result ki leas unit overvoltage for individual fuse operation. Ungrounded banks have no ground path for certain harmonic B.D. Whitman, R.W. Brown; presehted to EC Electee Institute T & D Subccmmittee'May 5*6,1 f currents, minimize ground currents during switching, and 6. Relay Prevents Cascading Capacitor Failures. may Rmit fault current for some fault conditions. For transmission voltage banks, the base Insulation b reduced or eliminated by a grounded wye connection; most banks 230 kV and higher are connected grounded wye. Johanson and R J. Haas, Electrical World. July 1. 1 - 7. Derivation and Application of Capacitor Case Rui Curves. L.E. Bock, presented to Power Dislrib Conference at University of Texas at Austin. Oc Series end parallel'groups. Generally a bank with the 2 6 .1 9 7 7 . fewest number of series groups and the maximum number 6. Superfuse - Group Fusing Can Prevent Cap of parallel units b easiest to protect. In thb cassia certain Tank Ruptures. Ken Christianson, Electric Ugh par cent change in capacitance for one group b easiest Power, October 1977. ^m ssm B W B O ^ssm m NPC00026359 753696 9. Capacitor Fusing to Overcome TenK Rupture, Peter Marsion, Transmission and Distribution, December * 1977 10. Use Tank Pressure to Protect Capacitors, Larry E. Bock and George R. Newcomb, Transmission & Distribution. February 1978 11. Duai Fusing to Prevent Capacitor Case Ruptures, 6.J. Pulaski, Transmission and Distnbutfon, March 1978 12. Capacitor Case Ruptures Beg Solution, Ken Owens. Transmission and Distribution, March 1978. & NPC00026360 753697