Mitigation of Transformer Inrush Current Associated with DER Facilities connected on the Distribution Grid

Transcription

1 Mitigation of Transformer Inrush Current Associated with DER Facilities connected on the Distribution Grid Marc Lacroix Eng, M. Eng, Senior Member IEEE Pierre Taillefer Eng, Member IEEE VIZIMAX Inc. Montréal, Canada André Mercier Eng IREQ Hydro-Québec Montréal, Canada Abstract Integration of Distributed Energy Resources (DER) to the distribution system brings technical challenges due to the inrush current resulting from the energization of power transformer. Besides impacting power quality, transients may cause undesired protection tripping. This paper compares different techniques used to mitigate these transients. It also presents results from research carried out on Hydro-Québec Research Institute s distribution test line (IREQ). Outcomes of additional tests using Gang Operated (GO) circuit breakers (CB) and Controlled Switching Device (CSD) are also presented. This approach solves the problem of the re-energization of a windmill on a live windfarm, in order to minimize transients and sympathetic interactions between the transformers in the DER facility. Since the CS solution virtually does not require any installation footprint, it has the advantage of being easily adapted to the existing switchgear designs integrated at the base of the wind towers. Index Terms--Distributed Energy Resources, Controlled switching, pre-insertion resistors, inrush current, Independent Pole Operated circuit breaker, Gang Operated circuit breaker. I. INTRODUCTION Integration of Distributed Energy Resources (DER) to the distribution grid involves major technical challenges for the DSO and utilities, due to the inrush current resulting from the energization of plant power transformer from the grid. Besides impacting power quality, transients may cause undesired protection tripping. Hydro-Québec has defined a list of requirements to guide the connection of DER to the distribution network [1]. Among these requirements, we may find design criteria and test procedures for harmonics content, load unbalance, transients and rapid voltage change of the installation of an industrial customer or energy producer. In these specifications, one of the requirements is that the installations must not cause rapid voltage variation higher than 4.75% on the distribution network. Consequently, energization of DER transformers from the grid is very critical since excessive inrush current is a main cause of disturbances. On a weak grid, it is practically impossible to meet these design criteria without using appropriate inrush current mitigation. In this respect, Hydro-Québec has started in 2000 the deployment of CS for power transformers in hydro-electric power plants and critical high voltage substations, including those involved in network restoration after a blackout. The interconnection of small DER facilities (such as wind farms) to the distribution network involves the same technical issues as in the transmission system. Since inrush current can reach up to 12 times the rated current, mitigation techniques should put in place to avoid feeder s breaker tripping. For example, after a feeder circuit breaker tripping, the wind farm is disconnected and should not reconnect immediately after the restoration of the power on the distribution line. According to the standards, the producer should wait for at least 5 minutes after the presence of a "normal" voltage. The DER facility unloaded power transformer must then be reenergized with a mitigation technique to limit inrush current and transient overvoltage on the distribution network. In anticipation of the addition of 151 MW of power on their 25kV distribution network in the next few years, Hydro- Québec has mandated its Research Institute (IREQ) to study different transformer energization techniques in order to mitigate the inrush current. Among possible solutions, three different methods were evaluated at the IREQ distribution test network facilities and the testing results are presented here. The complete results of the study are presented in the referenced document [2]. II. COMPARISON OF INRUSH CURRENT MITIGATION TECHNIQUES The first transformers energization tests conducted by IREQ were done to characterize the distribution of inrush currents while energizing the transformer directly from the networks without using any mitigation technique. Later, these results were used as a reference to compare the three different tested inrush current mitigation techniques approaches: 1. Pre insertion resistor technique: as shown in figure 1, a 20 Ohms resistor is temporarily connected in series with the load in order to limit the initial inrush current.

2 This resistor is then shorted by the main CB to enable the full load current in the transformer. For power transformers applications, the CSD controls the closing moment of the CB according to the residual flux remaining in its core and resulting from previous deenergization. Using an IPO CB, this technique almost eliminates the inrush current and the associated voltage disturbances on the network. The residual flux in the transformer core is calculated by computing the mathematical integral of the voltage measured during the transformer deenergization on either the primary or secondary side of the transformer. III. Figure 1. Current inrush mitigation with pre insertion resistor 2. Energization transformers: The second approach also uses a series impedance current limiting technique, but this time two low cost 25kVA distribution transformers mounted back-to-back are used as smoothing inductors (figure 2). The equivalent impedance of these transformers was 330 Ohm. COMPARING THE MITIGATION TECHNIQUES The first transformers energization tests were done to characterize the distribution of inrush currents while energizing the transformer directly from the networks without using any mitigation technique (figure 4). In total, 51 energization tests were made with a 750-kVA distribution transformer, and 71 more tests were made with the addition of a 1MVA transformers connected in parallel for a total capacity of 1.75MVA. The purpose of adding a transformer in parallel was to verify if the inrush current would be proportional to the total capacity of the DER facility. Figure 4. Test bench for the transformer energization tests Figure 2. Current inrush mitigation using energization transformers 3. Controlled Switching Device integration (CSD): Since a few decades, this technique has been successfully used for switching of reactive devices, power transformers and transmission lines on HV power systems. Figure 3. Current inrush mitigation with CSD Comparing the average and maximum inrush current to the transformer capacity, it could be conservatively concluded from the tests that the maximum inrush current of an installation is proportional to the installed transformer capacity. During the subsequent tests, 25 to 40 energization tests were done on the 750kVA power transformer using the 3 inrush mitigation techniques previously described: 1. Gang operated CB equipped with pre-insertion resistor. 2. Gang operated CB with back-to-back energization transformers (equivalent to smoothing inductors). 3. Independent Pole Operated CB equipped with a Controlled Switching Device. Figure 5 summarizes the average value of the maximum inrush current expressed in Per Unit (PU) for the various mitigation techniques. During the tests, it has been observed that the best inrush current mitigation technique is using an independent pole operated CB equipped with a CSD that takes into account the residual flux. For the series impedance techniques (pre-insertion resistors and energization

3 transformer), the conclusion was that the higher the series impedance is, the better the inrush current mitigation is. windmills. An alternate solution was developed to overcome these problems. Figure 5. Comparison of the average of maximum inrush current (in pu) Figure 6 illustrates the cumulative probability distribution function of the inrush current for the various tests that were performed. It can be seen that only the CSD technique allows an almost complete elimination of the inrush current at all time. Using the other techniques, the inrush current was relatively high 50% of the time. Figure 6. Cumulative probability distribution function of the inrush current (comparison) The study conducted by IREQ did not take into account the situation where a windmill from a wind farm is returned to service after a period of maintenance (figure 7). Although the capacity of a single windmill transformer does not seem to be enough to cause a problem due to the limited capacity of short circuit, the rapid voltage variation at the utility interconnection point may exceed the limits established in the requirement document. The presence of sympathetic interactions between the transformers during the energization may result in further aggravation of the problem [5]. However, these phenomena can be controlled by the temporary de-energization of the wind farm in low wind speed conditions when returning the equipment in service. This technique may involve the residual flux equalization of each wind tower transformer before reenergizing the entire farm, which may result in a rather complex operation when there are a large number of Figure 7. Re-energization of a single transformer after maintenance IV. USING CSDS WITH GANG OPERATED CB The best solution for wind farms would be to mitigate the inrush current to an acceptable level at each windmill, solving both the re-energization of a single tower and the interconnection of a windmill at the grid voltage. However, due to the limited availability and higher cost of IPO CBs at medium voltage, an option would be to use a GO CB equipped with a CSD at each windmill of the facility. Since the CSD is compact, this solution has the advantage of being easily adapted to the existing switchgear designs integrated at the base of the windmill. Since there is an optimum GO CB closing moment that produces the minimum obtainable inrush current for each residual flux pattern in the transformer core [3][6], this CS technique requires the measurement of the residual flux in each phase of the windmill power transformer. It can be calculated by computing the integral of the transformer voltage during its de-energization [3] [4], either using an existing potential transformers or capacitive/resistive voltage divider sensor. Although that controlled opening of the CB can set a known residual flux pattern of +ϕ, 0, - ϕ that eliminates the transformer inrush current using a GO CB [1][3], it is almost impossible to control at all time the opening of the CB or the grid voltage loss on the distribution grid. Consequently, when a wind tower is de-energized, the residual flux pattern in the transformer core will always vary according to the transformer de-energization conditions. The transformer re-energization technique must therefore mitigate the inrush current at all time, no matter what is the residual flux pattern. In order to test this integrated CSD concept, additional testing was carried out on a 1 MVA power transformer (Dyn5) operated at 20kV using a GO vacuum CB, a configuration that is typically used at the base of small wind turbines. During these tests, the transformer was de-energized by controlling the CB opening with 30 degrees steps in order to produce different residual flux patterns in the transformer core. The CB was then closed back at the optimum electrical target corresponding to that flux pattern calculated by the CSD (figure 8).

4 During the tests conducted with the CSD, it was observed that for any transformer residual flux pattern, the maximum inrush current was 2.2 PU with an average of 1.5 PU. It could also be observed that the maximum inrush current in the 3 phases is in the same magnitude range regardless of the residual flux pattern. Figure 8. Inrush current as a function of the CB opening angle To establish a comparison, uncontrolled energization tests were also conducted using the same transformer by closing the CB at random time. During the tests, the maximum and minimum recorded inrushes current were 5.8 and 4.6 PU respectively, compared to 2.2 and 1.1 PU using the CSD mitigation technique (Figure 9). Figure 10. Inrush current as a function of the CB opening angle The main objective of the additional tests was not only to establish a comparison of the inrush current with/without using a CSD, but also to measure the effect of the closing angle deviation from the theoretical optimum target point. This target deviation is mainly caused by the mechanical repeatability of the CB, and the variation of the pre-arcing time when switching the 3 phases simultaneously with the CB. During these tests, the CB was repeatedly opened at an angle of 30 degrees to produce a known residual flux pattern in the transformer core, and both random and controlled closing of the CB were done to measure the resulting power transformer inrush current (figure 11). From these tests, it can be observed that a CB timing deviation of ±20 (1ms at 60Hz) would not produce a significant change in the resulting maximum inrush current. As an additional conclusion, the CSD technique offers stable inrush current mitigation and the mechanical repeatability of the CB to a certain extent would not produce excessive inrush current. Figure 9. Inrush current as a function of the CB opening angle As an initial conclusion from these tests, the use of a CSD with a GO CB offers a comparable inrush current reduction as provided by the energization transformer technique (Figure 2). Since it virtually does not require any installation footprint, this solution has the advantage of being easily adapted to the existing windmill switchgear designs. In order to verify the initial results, additional tests were carried out on a 21 MVA/66kV dry type power transformer (Yyn0) using a CSD controlling a GO SF 6 circuit breaker. During the tests, the maximum measured inrush current with random energization of the power transformer was 4.2 PU, compared with 1.3 PU with the CS technique. Figure 10. Inrush current as a function of the CB target angle deviation

5 V. CONCLUSION The tests conducted at IREQ research institute led to the conclusion that the use of series impedance mitigation techniques is simple but more expensive due to extra equipment and related installation, while reducing inrush current by a relatively small factor. However, the CS technique accounting for the transformer residual flux offers the best possible inrush current reduction, allowing DER facilities to meet stringent rapid voltage changes design criteria imposed by the grid code. Combined with an IPO CB, the CSD solution will almost eliminate the power transformer energization inrush current, making it the ideal switching solution at the HV grid interconnection point. When involving the interconnection of a wind farm on the MV distribution grid, the optimum solution is to mitigate the inrush current to an acceptable level at each of each windmill. This approach solves the re-energization of a single tower on a live farm, both minimizing the inrush current impacts on the grid and the sympathetic interactions between the transformers in the DER facility. However, due to the limited availability and higher cost of IPO CBs at medium voltage, this solution cannot be economically implemented in existing designs of wind tower switchgears. In this paper, an inrush current mitigation technique solution based on the use of a simultaneous pole operated VCB equipped with a CSD has been successfully tested on a 1MVA@20kV Dyn5 power transformer. During these tests, the maximum inrush current has been reduced by 2.6 times compared to random energization of the power transformer. It was also observed that the maximum inrush current in the 3 phases is in the same magnitude range regardless of the residual flux pattern. Additional testing has also been conducted with an SF6 GO CB equipped with a CSD on a 21 MVA@66kV Yyn0 dry type power transformer. These tests have not only shown similar inrush current mitigation (1.3 PU compared to 4.2 PU with random energization), but also demonstrated that the CS technique offer stable inrush current mitigation despite the CB closing time variation due to its mechanical repeatability. These testing results show an inrush mitigation that is in the same range as those obtained using a gang operated CB with back-to-back energization transformers initially tested at IREQ and presented here. Since it is compact, the CS solution can be easily adapted to the existing switchgear designs integrated at the base of the windmill. REFERENCES [1] Hydro-Québec, Vice-présidence réseau, "Exigences techniques relatives à la protection et à l émission de perturbations des installations de clients raccordées au réseau de distribution d'hydro-québec", March 2004 [2] Abbey, C. and Taillefer, P., "Mitigation of Transformer Inrush Current Associated with DER Facilities" PAC World Conference, Raleigh, USA, 2014 [3] Brunke, JH and Frohlich KJ, "Elimination of inrush currents by controlled switching. I. Theoretical considerations," IEEE Transactions on Power Delivery, vol.16, no.2, pp , April [4] A. Mercier, E. Portales, Y. Filion, and A. Salibi, Transformer Controlled Switching taking into account the core residual flux, a real case study, CIGRÉ Conference 2002, # [5] Rioual, M., & Sow, M. (2008). Study of the Sympathetic Interactions When Energizing Transformers for Wind-Farms: Description of the Phenomena Involved and Determination of the Stresses during their Energization. In 7th International Workshop on Large-Scale Integration of Wind Power into Power Systems and Transmission Networks for Offshore Wind Farms. Madrid, Spain. [6] Nishiwaki, S., Koshizuka, T., Saito, M. and Sato, Y. Residual Magnetic Fluxes in the Iron Core of Neutral Ungrounded Transformer That Are Required to Be Known for Controlled Switching, The International Conference on Electrical Engineering (2008)