Solving Customer Power Quality Problems Due to Voltage Magnification

Size: px

Start display at page:

Download "Solving Customer Power Quality Problems Due to Voltage Magnification"

Stephen McBride
6 years ago
Views:

1 PE-384-PWRD Solving Customer Power Quality Problems Due to Voltage Magnification R. A. Adams, Senior Member S. W. Middlekauff, Member Duke Power Company Charlotte, NC USA E. H. Camm, Member J. A. McGee S&C Electric Company Chicago, IL USA Duke Power Company Greenville, SC USA Abstract: Reports by a Duke Power customer concerning equipment malfunctioning coincident with the switching of a 100 kv shunt capacitor bank led to field measurements and EMTP simulations to identify the possible sources of the problems and to find an economical solution. Measurements and simulation revealed the occurrence of voltage magnification at the customer s site due to the effects of three different distribution capacitor banks on the feeder supplying the customer. A comparative evaluation of commercially-available capacitor-switching transient mitigation switching devices suggested the use of lowresistance pre-insertion inductors applied with circuit switchers to energize the 100 kv capacitor bank. Field measurements of capacitor-switching events after installation of pre-insertion inductors verified the elimination of voltage magnification and customer power quality problems. Keywords: transients. power quality, capacitor switching, switching 1. INTRODUCTION Voltage magnification occurring as a result of utility capacitor switching is now a well-know phenomenon. In particular, the effects of voltage magnification on adjustable-speed drives (ASDs) are well documented [l-3]. As utility customer power quality awareness increases, there is an increasing need to reduce or eliminate the effects of transients on customer electrical equipment. The devices that are commercially available to reduce or eliminate the effects of voltage magnification include high-resistance or low-resistance pre-insertion inductors used with circuit switchers, controlled closing circuit breakers or vacuum interrupters, and circuit breakers with pre-insertion resistors. Bellei, et al presented a comparative evaluation of capacitor-switching devices to prevent nuisance tripping of ASDs due to voltage magnification in [4]. PE-384-PWRD-O-1l-1997 A paper recommended and approved by the IEEE Transmission and Distribution Committee of the IEEE Power Engineering Society for publication in the IEEE Transactions on Power Delivery. Manuscript submitted August ; made available for printing November 7, II. CUSTOMER POWER QUALITY CONCERNS IDENTIFY VOLTAGE MAGNIFICATION A commercial customer called Duke Power Company with concern regarding the electrical service to their facility. The customer, a local TV station, was experiencing problems with high-voltage power supplies on a regular basis, resulting in loss of their audio signal, and partial power to their video signal. The problems experienced were reportedly very prevalent in fall and spring, but less during the winter and summer. The feeder supplying the customer had three installed distribution capacitor banks. A large underground subdivision was also served by the same feeder. Duke engineers suspected there may be a correlation between the high-voltage power supply breaker trips and the switching of the capacitors at the 100 kv tie station. To investigate, a pair of BMI 4800 power quality monitors were placed at the tie station, and at the electrical delivery of the TV station. Later, for a higher speed capture of the events, a Dranetz 658 was installed at the customer s electrical delivery. The customer was requested to keep a disturbance log of breaker trips, to attempt to correlate to system events. The disturbance logs indicated that the problems of the circuit breakers was coincident with the switching of a 100 kv, 57.6 Mvar capacitor bank at a 100 kv substation located approximately 14 miles from the customer s site. The capacitor bank was being switched on average 3 to 4 times a week, and resulted in tripping of circuit breakers at the customer s site in about 50% of the switching events. Recognizing the possibility of voltage magnification occurring due to the presence of the distribution capacitors, voltage measurements were performed at the customer s 480 V bus with all but one (a fixed 600 kvar bank) of the three distribution capacitor banks switched off. However, measurements indicated that the transient overvoltage was still sufficiently large (about 1.8 per unit of nominal line-toline voltage). See Fig IEEE. Reprinted, with permission, from IEEE/PES 1998 Winter Meeting, February 1-5, 1998, Tampa, Florida USA 71l-T66

2 Fig. 1. Field measurements of voltage magnification at customer s 480 V bus due to energizing a 100 kv, 57.6 Mvar, ungrounded-wye-connected shunt capacitor bank at Duke Power s 100 kv substation. Phase-to-phase voltages are shown. Only 600 kvar distribution capacitor bank is ON. III: EMTP SIMULATION OF CAPACITOR ENERGIZING TRANSIENTS An Electromagnetic Transients Program (EMTP) simulation model of the distribution system in the 100 kv substation area was developed to determine the effects of the three distribution capacitor banks and the capacitance of cables of a large subdivision in the area. Fig. 2 is a simplified one-line diagram of the 100 kv substation and the distribution system involved. The 450 kvar capacitors are located about 4800 ft (0.91 mile) from the customer s step-down transformer, the 600 kvar capacitors about 3.26 miles, and the 900 kvar capacitors about 4.02 miles. The large subdivision is located about 3.69 miles away from the customer s step-down transformer. Fig. 2. System one-line diagram showing 100 kv substation and distribution system to affected customer s utilization voltage bus. The EMTP equivalent circuit parameters were also used to calculate the approximate frequencies of oscillation of the capacitor-switching transient at the 100 kv substation and at the customer s 480 V bus. Fig. 3 shows the equivalent circuit for calculating the resonant frequencies at the two locations. The 3-phase EMTP simulation model included equivalent loads at the customer site, the subdivision, and the 100-kV system. The loads of the subdivision were modeled in detail, distributed along the lengths of equivalent cable models. This representation of the underground subdivision would show the effects of cable capacitance during simulations. The transient damping effected by long 100 kv transmission lines was represented at the 100 kv substation. The available 3-phase and phase-to-ground short-circuit currents were used to determine the parameters of the equivalent source at the 100 kv substation. The distribution feeder and lateral to the customer s site was represented by sections of overhead line and cable models. All capacitor banks were represented as lumped equivalent capacitances. 2 Fig. 3. Equivalent circuit illustrating circuit parameters determining transient frequencies at the switched capacitor bank and at remote capacitors. The frequency of the oscillation at the 100 kv substation is based on an equivalent source inductance of about 6.2 mh (i.e., about 24.7 ka 3-phase available short-circuit current) and the capacitance of the 100 kv, 57.6 Mvar capacitor bank of 15.3 uf. Frequencies of oscillation at the 480 V bus were calculated for each individual connected distribution capacitor bank to determine if the matching of the frequencies of the L-C circuit at the 100 kv substation and at the 480 V bus can be effectively changed to reduce the effects of voltage magnification. The resultant frequencies are summarized in Table 1. Actual frequencies will be slightly different from the calculated values due to the effects of damping from system loads and losses. When more than one distribution capacitor bank is connected, the

frequency of oscillation at the 480 V bus is determined by the interaction of the different L-C circuits.

It is clear from the tabulated frequencies in Table I that the frequencies of the L-C circuits at the two locations are sufficiently closely matched with each of the individual connected distribution

6 Mvar capacitor bank was simulated while only the 600 kvar distribution capacitors were connected. The closing instants of the three poles of the circuit switcher (without pre-insertion inductors).

3 frequency of oscillation at the 480 V bus is determined by the interaction of the different L-C circuits. TABLE I APPROXIMATE FREQUENCIES OF OSCILLATION OF CAPACITOR Lumped capacitance of subdivision cables. It is clear from the tabulated frequencies in Table I that the frequencies of the L-C circuits at the two locations are sufficiently closely matched with each of the individual connected distribution capacitors to result in voltage magnification. In order to validate the EMTP simulation model, the energizing of the 100 kv, 57.6 Mvar capacitor bank was simulated while only the 600 kvar distribution capacitors were connected. The closing instants of the three poles of the circuit switcher (without pre-insertion inductors). was selected to coincide with the closing instants which yielded the voltage measurements at the customer s 480 V bus shown in Fig. 1. This also served to confirm the apparent voltage magnification caused by the 600 kvar distribution capacitors. The resulting phase-to-phase voltages at the 100 kv substation and at the customer s 480 V bus are shown in Fig. 4. Note that the effects of voltage magnification due to the 600 kvar capacitors are clearly shown. The peak phase-to-phase transient overvoltage at the 100 kv substation is only kv (or 1.52 per unit), while the corresponding peak overvoltage at the customer s 480 V bus is about 1383 V (or 2.04 per unit). The frequency of the simulated transient overvoltage at the 100 kv substation is about 565 Hz (compared to the calculated frequency of 518 Hz), while the corresponding frequency at the 480 V bus is about 627 Hz (compared to the calculated frequency of 681 Hz). The magnitude of the simulated transient overvoltage at the 480 V bus is somewhat higher than the measured overvoltage shown in Fig. 1, but the simulated transient frequency agrees very closely with the 625 Hz of the measured transient. The difference in transient overvoltage magnitude is due to the increased damping in the field, particularly since the effects of the harmonics at the customer s site were not represented in the simulation model. connected to the feeder individually, and also in combination with one or both of the other banks. For comparison, the same closing instants of the three poles of the circuit switcher as during the field test were used throughout the simulations. The resulting peak overvoltages are summarized in Table II. As expected, voltage magnification is apparent with all different combinations of connected distribution capacitor banks. The most severe voltage magnification at the 480 V bus (1645 V or 2.42 per unit of nominal peak line-to-line voltage) occurs with the 900 and 450 kvar distribution capacitor banks switched on. See Fig. 5. Note also the slight magnification due to the capacitance of the large subdivision cables when all three distribution capacitor banks are switched off. Fig. 4. Simulated phase-to-phase voltages at (a) 100 kv bus and at (b) customer s 480 V bus when energizing the 100 kv, 57.6 Mvar capacitor bank at Duke Power s 100 kv substation. Only 600 kvar distribution capacitor bank is ON. TABLE Il. PEAK TRANSIENT OVERVOLTAGES WHEN ENERGIZING THE 100 kv, 57.6 MVAR CAPACITOR BANK. The EMTP simulation model was then used to determine the severity of voltage magnification due to the three distribution capacitor banks. Simulations were performed without any of the three distribution capacitor. banks connected, and with each of the three capacitor banks 3

Fig. 5. Simulated most severe phase-to-phase voltage at customer s 480 V bus when energizing the 100 kv, 57.6 Mvar capacitor bank with the 900 and 450 kvar distribution capacitor banks switched ON.

either high-resistance or low-resistance pre-insertion inductors to energize the 100 kv, 57.6 Mvar capacitor bank at the 100 kv substation.

4 Fig. 5. Simulated most severe phase-to-phase voltage at customer s 480 V bus when energizing the 100 kv, 57.6 Mvar capacitor bank with the 900 and 450 kvar distribution capacitor banks switched ON. IV. EMTP SIMULATION OF TRANSIENT MITIGATION USING PRE-INSERTION INDUCTORS Alternatives for minimizing the effects of voltage magnification due to the distribution capacitor banks included applying either high-resistance or low-resistance pre-insertion inductors to energize the 100 kv, 57.6 Mvar capacitor bank at the 100 kv substation. These were the preferred transient mitigation options since the capacitor bank at the 100 kv substation was being switched with a circuit switcher to which pre-insertion inductors could easily be retrofitted. Pre-insertion inductors furnish an impedance, which is frequency dependent, in series with the bank capacitance during the initial energization of the capacitor bank. This impedance reduces the collapse in bus voltage by the amount of voltage developed across the inductor during the inrush of current into the bank. The pre-insertion inductor also limits the magnitude of the initial inrush current. Since the impedance of the pre-insertion inductor is frequencydependent, its value appears to be quite large during initial inrush current into the bank when the frequency is quite high. Thereafter, the effective impedance of the preinsertion inductor is reduced when the steady state, 60 Hz, current value of the bank is obtained. The pre-insertion inductor - like any other pre-insertion impedance - gives rise to a second transient when the inductor is bypassed (after the 60-Hz current is obtained). This transient, referred to as the bypass transient, is generally much smaller than the initial transient, but can be larger depending on the size of the bank and the impedance of the inductor. The pre-insertion inductor is typically comprised of a number of close-coupled layers of stainless-steel (high resistance) wire wound to form a hollow glass-reinforced tube; low-resistance aluminum wire is also sometimes used. The pre-insertion inductor is typically applied with a circuit switcher having a high-speed disconnect blade, which inserts the pre-insertion inductor for 7 to 12 cycles (depending on system voltage) during closing, to energize the bank. See Fig. 6. Fig. 6: Pre-insertion inductors mounted on a Mark V circuit switcher at the Duke Power 100 kv tie station. High-resistance pre-insertion inductors are usually recommended for mitigating capacitor-switching transients due to voltage magnification. However, for larger capacitor banks where the bypass transient may be a concern, the use of low-resistance pre-insertion inductors may be more suitable. With high-resistance pre-insertion inductors, the bypass transient, which is proportional to the voltage developed across the inductor due to the capacitor bank current flowing through it, increases as the capacitor bank size increases. For the 57.6 Mvar capacitor bank the peak bypass transient magnitude when using 40 mh-81 ohm high-resistance pre-insertion inductors would be approximately 0.47 per unit of nominal peak phase-toground voltage. Realizing that this magnitude bypass transient could result in magnified overvoltages at the customer s 480 V bus which may be unacceptably high, the use of 40 mh-5.5 ohm low-resistance pre-insertion inductors were considered instead. The low-resistance preinsertion inductors are somewhat less effective than the high-resistance pre-insertion inductors during the initial insertion transient since the resistance thereof is too low to effectively damp the oscillatory transient. However, the inductance of the low-resistance pre-insertion inductors is as effective as that of the high-resistance pre-insertion inductors in reducing the initial collapse in the bus voltage, which reduces the excitation of the remote L-C circuit. Furthermore, since the impedance of the low-resistance preinsertion inductors is small compared to that of the highresistance pre-insertion inductors, the bypass transient is always small compared to the insertion transient. To determine the effectiveness of the low-resistance preinsertion inductors, the performance of 40 mh-5.5 ohm pre-insertion inductors was determined using the EMTP simulation model with all combinations of connected distribution capacitor banks. For comparison purposes, the closing instant of the three poles of the circuit switcher was considered to be the same as for the previously simulated cases without pre-insertion inductors. Resulting peak phase-to-phase transient overvoltages are summarized in 4

Table III. The case resulting in the most severe overvoltage at the customer s 480 V bus is illustrated in Fig. 7.

6 MVAR CAPACITOR BANK WITH A CIRCUIT SWITCHER WITH LOW-RESISTANCE PRE-INSERTION INDUCTORS. V.

substation and at the customer s 480 V bus were again performed. At the tie station, a 20 khz digital oscilloscope was used to capture switching events at a high sampling rate.

5 Table III. The case resulting in the most severe overvoltage at the customer s 480 V bus is illustrated in Fig. 7. Note that the pre-insertion inductors practically eliminated the transient overvoltages at the customer s 480 V bus. TABLE III PEAK TRANSIENT OVERVOLTAGES WHEN ENERGIZING THE 100 kv, 57.6 MVAR CAPACITOR BANK WITH A CIRCUIT SWITCHER WITH LOW-RESISTANCE PRE-INSERTION INDUCTORS. V. FIELD MEASUREMENTS OF TRANSIENT MITIGATION USING PRE-INSERTION INDUCTORS To verify the performance of the low-resistance preinsertion inductors, field measurements of transient voltages at the 100 kv substation and at the customer s 480 V bus were again performed. At the tie station, a 20 khz digital oscilloscope was used to capture switching events at a high sampling rate. A Dranetz 658 was again used to capture the transients seen at the customer s 480 V delivery. Note: Since the 100 kv capacitor bank is ungrounded, the peak overvoltages at the two locations do not necessarily occur on the same phases due to the transient reduction by the pre-insertion inductors. Fig. 8. Measured phase-to-phase voltages at (a) 100 kv substation and at (b) customer s 480 V bus when energizing the 100 kv, 57.6 Mvar capacitor bank with a circuit switcher with low-resistance pre-insertion inductors. Note that the capacitor switching transient is hardly visible due to the harmonic voltage distortion at the 480 V bus. (b) Fig. 7. Simulated phase-to-phase voltages at (a) 100 kv substation and at (b) customer s 480 V bus when energizing the 100 kv, 57.6 Mvar capacitor bank with a circuit switcher with low-resistance pre-insertion inductors. Illustration of most severe overvoltage at 480 V bus. All distribution capacitor banks are ON. Fig. 8 shows the resulting waveform from the switching of the capacitor bank at the 100 kv tie station. Fig. 8(a) was measured using the 20 khz oscilloscope at the tie station. The transient, occurring at the trigger point (indicated by the arrow), is very subtle, a vast improvement from before. Fig. 8(b) was measured using the Dranetz, at the customer 480 V electrical delivery. The transient is almost non-existent at that point, and is masked further by the voltage distortion from the electronics of the transmitter.

6 VI. CONCLUSIONS Voltage magnification occurring as a result of switching a 100 kv, 57.6 MVAR capacitor bank on Duke Power Company s system caused problems with high-voltage power supplies at a local TV station. Field measurements and EMTP simulations resulted in identification of the problem and suggested the use of low-resistance preinsertion inductors for transient mitigation. Pre-insertion inductors offer a cost-effective solution for reducing capacitor switching transients, especially in existing installations where circuit switchers are used. Capacitor switching transients often cause problems for customer equipment, however the transient does not need to be completely eliminated. The pre-insertion inductor provides a means for controlling the transient to an acceptable level. For most capacitor-switching transient mitigation applications where voltage magnification is an issue, the high-resistance pre-insertion inductor is recommended. However, for larger capacitor bank sizes where the bypass transient becomes significant, low-resistance pre-insertion inductors may provide better performance. This is particularly true for applications where the source impedance is low. EMTP simulation was used effectively to confirm this theory. Low-resistance pre-insertion inductors retrofitted on a circuit switcher worked as expected, eliminating power quality problems at the TV station. Technical Services within the Electric Transmission Business unit. Ron is a past Chairman of the IEEE Power Engineering Society, Charlotte Chapter and been active with the IEEE Working Group for Recommended Practices and Requirements for Harmonic Control in Electrical Power Systems (IEEE 519). Ron is registered as a professional engineer in both North and South Carolina. Stephen W. Middlekauff received his BSEE and MSEE from Clemson University, Clemson, SC, in 1993 and 1996 respectively. He has been employed by Duke Power in the System Power Quality department since graduation in May As a graduate student and Duke employee, Stephen served as a principle investigator on the Westinghouse Dynamic Voltage Restorer (DVR) project. He is a member of IEEE and the Power Engineering Society, and serves as a chapter chairman for the Custom Power Task Force Application Guide. His primary research interests include power quality, custom power, and harmonics. Ernst H. Camm received his BSc(Eng) degree in Electrical and Electronic Engineering from the University of Cape Town, South Africa in 1984, and his MSEE degree from the Ohio State University in From 1984 to 1990, he held various positions in Plant and Project Engineering. He is currently a Project Engineer in the Engineering Services Department at S&C Electric Company. Ernst has had extensive involvement in capacitor switching transient analysis at S&C, including analysis in the development of optimally sized pre-insertion inductors for capacitor switching transient mitigation. He is the author of and copresenter of S&C s Seminar on Capacitor Switching Transients and Their Impact on Your System. He is a member of the Switching Transients Task Force of the IEEE s Modeling and Analysis of System Transients Working Group and the Shunt Capacitor Application Guide Working Group. James A. McGee is a 1976 graduate of North Carolina State University with a B.S. degree in Electrical Engineering. Since that time he has been employed by Duke Energy in a variety of assignments, including protective relay engineering and transmission systems operations. Currently, Jim serves as a power quality specialist. In 1981 he was awarded an M.E. degree in Electrical Engineering from Clemson University. He is a registered Professional Engineer in North and South Carolina. VII. REFERENCES [I] M. F. McGranaghan, et al, Impact of Utility Switched Capacitors on Customer Systems: Part II - Adjustable Speed Drive Concerns, Presented at the 1991 IEEE-PES Winter Meeting, New York, New York, February 3-7, [2]. J. A. Oliver, and R. A. Ferraro, The Myths of ASD Power Quality, in Proceedings of the Second International Conference on Power Quality: End-Use Applications and Perspectives, September 28-30, 1992, Atlanta, Georgia. [3]. H. G. Murphy, Power Quality Issues with Adjustable-Frequency Drives: Coping with Power Loss and Voltage Transients, Power Quality Assurance Magazine, May/June [4] T. A. Bellei, R.P. O Leary, and E. H. Camm, Evaluating Capacitor- Switching Devices for Preventing Nuisance Tripping of Adjustable- Speed Drives Due to Voltage Magnification, in IEEE Transactions on Power Delivery, Vol. II, NO. 3, July VIII. BIOGRAPHIES Ron A. Adams received his B.S. degree in Electrical Engineering from Clemson University in May, He has been employed with Duke Energy for 12 years and has held various engineering positions within Substation Project Engineering, Industrial Marketing and System Power Quality. Currently he holds the position of Subprocess Owner of Provide

Do Capacitor Switching Transients Still Cause Problems?

Do Capacitor Switching Transients Still Cause Problems? Mark McGranaghan We have been evaluating problems related to capacitor switching transients for many years. Capacitor banks have been used on distribution