Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Control method for the stable operation of distributed inverters following the introduction of large-scale renewable energy to a remote island grid

Control method for the stable operation of distributed inverters following the introduction of... A control system that can operate multiple distributed inverters stably is proposed. The con- trol system is assumed to be applied to a remote island. From the simulation results, it was verified that the control method provides stable operation against various faults. Keywords: renewable energy; photovoltaic; battery energy storage system; inverter; remote island (Constant Voltage Constant Frequency) operation with one Introduction inverter or the master–slave method [2 1]. , However, if the The generation of CO and other greenhouse gases is a grid size is somewhat large, it is likely that inverters will be main contributor to global warming, which in turn af- installed in multiple locations. In this case, it is necessary fects different social and economic issues worldwide. As to coordinate the operation of the distributed inverters. a countermeasure, renewable energy is attracting atten- Various studies such as the VSG (Virtual Synchronous tion all over the world. However, since the output of re- Generator) [3, 6] have been carried out [7, 8] in an attempt to newables, specifically photovoltaic (PV) and wind power add an inertial property to an inverter, and recently it has generation, is heavily dependent on weather conditions, been developed into the concept of the GFM (Grid Forming) producing a stable output can be difficult. Output instabil- inverter [9, 10]. However, there are many variations in the ities can produce large frequency fluctuations in the grid. functions to be realized and the methods of realization, Moreover, the instances of output instability increase as and the control technologies for distributed inverters have the proportion of power contributed to the grid by renew- not yet been fully established. Therefore, it is difficult to able energy sources increases. In general, the introduction introduce large-scale renewable energy to an existing grid of large-scale renewable energy to the grid requires storage from a technical standpoint. Against this background, the batteries for supply-and-demand control and inverters are authors believe that it is necessary to consider a method essential to connect these to the power grid. In a system in that can control future grid configurations where renew- which a large number of inverter facilities are introduced, able energy sources predominate. it is difficult to acquire the inertial properties of the prime Many remote islands currently use diesel power as mover, because conventional inverter facilities could not their main source of energy and the unit cost of electri- have inertial properties. city generation is expensive due to the extra cost of trans- When the grid size is very small, such as in a remote porting fuel. In addition, there is a risk that the transport island or microgrid, it is possible to realize a grid centred of fuel itself will be severed in the event of a disaster. on the power supply by the inverter by applying CVCF Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 198 | Clean Energy, 2021, Vol. 5, No. 2 Therefore, the introduction of large-scale renewable en- supplies should be distributed. Therefore, it is considered ergy to remote islands is necessary. It is expected that re- appropriate for verifying the response of the system to newable energy, which does not consume fuel, will enable grid faults. Similar-scale remote islands include Oshima, the local production and local consumption of energy. In Tokyo and Kumejima, Okinawa Prefecture. In addition, the this regard, early studies include the design of remote is- results of this study can be applied to these similarly or land microgrids (0.5-MW demand) [11] and an empirical larger-scale remote islands. study of the renewable energy supply in small-scale grids (610-kW total demand) [12]. However, the small scale of 1.1 Inverter control these systems requires only one storage battery system for supply-and-demand control. Reference [13] presented and discussed the characteristics In this study, we examined a case in which a large of the BESS inverters proposed for load following and co- amount of renewable energy is introduced and multiple ordinated operation. In this study, since the multiple in- storage battery systems (hereinafter referred to as ‘BESS’ verters were distributed, the BESS for adjusting the output or ‘battery energy storage system’) are distributed. The au- has the same configuration as that for voltage control thors evaluated a method for adjusting the energy supply utilizing the ‘synchronization force’ required for coordin- and demand for remote islands that can be regarded as ated operation. Fig. 1 shows the main circuit configuration independent electric power systems by using multiple and control circuit. The features of this control system are BESSs that apply the control method proposed in Ref. [13]. briefly described below. Reference [13] presents a control and simulation for a • The inverter is operated by voltage control and the syn- large-scale system. References [14–16] and this paper are chronization force is used to maintain synchronization characterized by a plural number of BESSs in the system. between each inverter. The work described in this paper is a follow-up study to • The control system has properties equivalent to the in- Refs [14–16]. The stability of operation in the event of a ertia of the synchronous machine in order to avoid large fault in the proposed system under various conditions was fluctuations in frequency and to maintain stability. evaluated and investigated by instantaneous value-based • To improve the damping of power swings, phase com- simulations. pensation is provided. In Section 1, the model remote island grid system char - • The frequency control is equivalent to a governor. acteristics, the inverter controller configuration and the • The voltage-control part is equivalent to a STATCOM target system model are explained. In Section 2, the simu- (Static Compensator). lation conditions are assumed. The output of the PV and • Even if a system fault were to occur, the inverter would the BESS under steady-state conditions before reproducing continue to operate without stopping. the fault are mentioned. In Section 3, the performance of • The system reduces the overcurrent in short-circuit the control system under different conditions and with faults and limits the current to continue operation. two types of faults is presented systematically. In Section • AFC (automatic frequency control) is used. 4, the results for each case are considered. The changes in voltage, frequency and BESS output power in response to In addition, the PV inverter, whose output cannot be con- faults and fault removal are compared and discussed for trolled and is dependent on the weather, uses the conven- each case. In Section 5, the main results from the study tional current-control method. are recapitulated and the ability of the control system developed to operate the inverters stably is emphasized. Moreover, the scope for future work is also provided. 1.2 Modelling of the target system The PV systems and loads were assumed to be distributed in three locations. Assuming that the load in the island 1 The proposed system for evaluation (max. 11 MW and min. 5 MW) is dispersed in these three Hachijo Island, Tokyo was designated as an isolated is- locations with a ratio of 5:4:2 in heavy load and 1:2:2 in land. Hachijo Island is located 287  km south of Tokyo, light load, and the PV system and the BESS were placed with an area of 69  km and a population of ~8000. The in the vicinity of each load. In this case, the power-supply minimum electricity demand is 5  MW and the peak de- capacity does not need to be proportional to the load and mand is 11  MW. Since a geothermal power plant with a the total capacity was divided into three equal parts. Fig. rated capacity of 3.66 MVA was assumed to be in operation, 2a shows the distribution line connecting the loads, which the remaining demand was assumed to be supplied by PV. were 2 km apart, and the geothermal power plant, which The reason for targeting this island is mainly because the was 8 km away from the nearest load. Fig. 2b shows that scale of demand and the scale of area fit the purpose of the DC-side voltage is 850  V and the AC-side voltage is the study. That is, if the scale is small, CVCF operation by 6.6 kV. In addition, the PV capacity is 7.6 × 10  kW ÷ 3 = 2 one inverter or the master–slave method can be applied .53 ×  10   kW (25.3  MW) and the inverter output is 75% of but, on this scale, it is considered that inverter power the PV-panel capacity, or 19 MW. The capacity of the BESS Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 199 Inverter Transformer DC power AC bus of source Power system Gate pulse Output current Output voltage 1/Droop Frequency Frequency + 1 + sT control Frequency order Δf f AFC + + Phase 3Phase + 12 + π Output power order compen- voltage PWM + Ms + DS sation signal Δf Active power Inertia property AC-AVR Output voltage ⎥E⎥ Voltage Current Voltage order control limit + + Xslope Reactive power order 1 + sT Reactive power Output current Fig. 1: Detail configuration of inverter controller [13] geothermal power plant is 3.66  MVA and the output was A B 6.6kV (2km) 6.6kV (2km) 6.6kV (8km) adjusted to 3.66 × 0.9 = 3.3 MW at all times. P P P L LG L 2 Simulation conditions V V V o o o e . . a a a o This paper examines the transient stability of the proposed B B d d d model for faults during light loads. Simulations were per - E E E 3.3MW S S formed by first classifying the conditions into ‘daytime’ and Nighttime/Daytime S S S 1MW/5MW 2MW/4MW 2MW/2MW ‘night-time’. In the daytime, power generation by PV is high at 0.9 pu, with excess power charging the BESS. The maximum 6.6kV load was 11  MW. Simulation results for the three-phase line-to-ground faults (3LG) and the two-line short-circuit fault (2LS) in this condition show that stable operation is possible [15, 16]. The ‘night-time’ does not have PV power INV INV generation. Thus, power is supplied from the BESS and geo- DC850V DC850V thermal power. The night-time load (5 MW) was assumed to PV 19MW BESS 17MW be allocated following the ratio 1:2:2. Here, the BESS capacity Current control Proposal control (17  MW × 3 units) is sufficiently large compared to the de- mand (≤11 MW) in the island. BESS I indicates the use of one Fig. 2: System-model diagram BESS to meet all the demand on the island. In this case, the remaining BESSs were operated at output power order 0. The inverter was set to 17  MW. This appears to be an exces- pattern in which three BESS outputs were equal is indicated sive capacity for the apparent load, but was optimized as BESS III. The examples reported so far and the conditions to consider the economic efficiency, assuming that there in this study are shown in Table 1. is no diesel power generation throughout the year to ad- The fault was introduced 6  s after the simulation was just the supply and meet demand [14]. The capacity of the sufficiently stable and was removed after 0.1  s. For a Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 200 | Clean Energy, 2021, Vol. 5, No. 2 Table 1: Overview of simulation voltage decreases to 0 due to the fault and distribution is not carried out during the fault. Other BESSs increase Conditions Fault type their output to supply power to neighbouring loads and Time band Fault point 3LG 2LS cover losses caused by fault currents. Looking at point ‘P’ in Fig. 1, we find that the BESS active power output is used Daytime A and B Reference [5] Reference [6] to calculate the frequency deviation (∆f). Since the power- Night-time (BESS I) A Case 1 Case 2 order value is constant, the increase in deviation causes the power-detection value to decrease. Conversely, as the Night-time (BESS III) A Case 3 Case 4 detected value of the power increases, the frequency de- viation decreases. This operation explains the results shown in Fig. 3a. In Fig. 4c, distribution line impedance grid frequency of 50  Hz, 0.1  s corresponds to five cycles. exists between the fault point (point B) and the observa- Verification was performed on four different cases. tion point, so that the fault has little impact and produces little fluctuation. Fig. 3d shows the rotational speed of the (i) Case 1: 3LG at night-time (BESS I) geothermal generator. The rotational speed accelerates We assumed that one BESS is supplying power at ‘night- due to the impact of the fault and returns to the steady- time’. The system responses to 3LG at point A and point B state value with the removal of the fault. In Fig. 4d, the are evaluated. rotational speed slightly decreases immediately after the fault. It then accelerates and returns to steady state after (ii) Case 2: 2LS at night-time (BESS I) the fault removal. Since this is a fault in the vicinity of a We assumed that one BESS is supplying power at ‘night- geothermal generator, it is certain that angular backswing time’. The system responses to 2LS at point A and point B was produced at the moment of the fault. are evaluated. (iii) Case 3: 3LG at night-time (BESS III) 3.2 Case 2: 2LS at night-time (BESS I) We assumed that all BESSs are supplying power at The frequency deviation (a), the BESS active power output ‘night-time’. The system responses to 3LG at point A and (b) and the inverter terminal voltage (c) when 2LS occurs point B are evaluated. at point A are shown in Fig. 5. The corresponding response to 2LS at point B is shown in Fig. 6. From Figs 5 and 6, it (iv) Case 4: 2LS at night-time (BESS III) can be seen that the frequency deviation, the active power We assumed that all BESSs are supplying power at output and the inverter terminal voltage were all returned ‘night-time’. The system responses to 2LS at point A  and to steady-state values a short time after the elimination point B are evaluated. of the fault. The fluctuation of the frequency deviation is The simulation was carried out using the instantaneous slightly larger to that of Case 1 at point FA.  ig. 5b shows value-based software ATP (Alternative Transients Program) that the output power of BESS 1 was negative (about and its graphical user interface (GUI) tool, ATPDraw [17]. –0.2 pu) during the fault. This indicates that the BESS was charging during the fault. BESS 1 in Fig. 5c is near the fault point but, due to 2LS, the voltage was not 0 and decreased 3 Simulation results to ~0.5  pu. After the fault was eliminated, the output re- 3.1 Case 1: 3LG at night-time (BESS I) turned to 1 pu immediately. The common characteristic of the BESS active power output in Case 1 and Case 2 is that The frequency deviation (a), the BESS active power output only BESS 1 outputs ~0.15 pu, while BESS 2 and BESS 3 op- (b), the inverter terminal voltage (c) and the rotational erate at output 0 pu. Both the frequency deviation and the speed of the geothermal generator (d) when 3LG occurs at BESS active power output are detected values in the BESS point A  are shown in Fig. 3. The corresponding response block on the load side. Therefore, the response at point B is to 3LG at point B is shown in Fig. 4. These graphs of (a) to found to be small compared to that at point A. (c) show the values detected in the BESS inverter control block in Fig. 1. From Figs 3 and 4, it can be seen that the fre- quency deviation, the active power output and the inverter 3.3 Case 3: 3LG at night-time (BESS III) terminal voltage were all returned to steady-state values immediately after the elimination of the fault. Looking The system frequency deviation (a) and the BESS active at the frequency deviations in Fig. 3a, only the frequency power output (b) when 3LG occurs at point A are shown in deviation for BESS 1 in the vicinity of the fault point in- Fig. 7. The corresponding response at point B is shown in creased during the fault while the frequency deviations for Fig. 8. The fluctuation of the frequency deviation is similar BESS 2 and BESS 3 decreased. This response is explained to that for points A and B of Case 1. Looking at the initial by the active power output. In Fig. 3b , only BESS 1 showed value of the inverter output power of the BESS of Case 3 a decrease in output during the fault. This is because the (Figs 7 and 8), all three BESS units operate at an output Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 201 0.4 0.2 0.0 BESS 1 –0.2 BESS 2 BESS 3 –0.4 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 1.0 0.5 0.0 BESS 1 –0.5 BESS 2 BESS 3 –1.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 1.2 1.0 0.8 0.6 BESS 1 0.4 BESS 2 BESS 3 0.2 0.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 5.0 2.5 0.0 BESS 1 BESS 2 –2.5 BESS 3 –5.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] Fig. 3: Simulation results for Case 1: 3LG at point A of ~0.05  pu. In Fig. 7b, the BESS active power output de- in Fig. 10. The fluctuation of the frequency deviation is creased to 0 only for BESS 1 near fault point A  and the similar to that of Case 2 at both points A and B. All three power output of BESS 2 and BESS 3 increased as in Case BESS units operate at an initial value of ~0.05 pu. The BESS 1. The inverter terminal voltages at points A and B at 3LG active power output behaves like Figs 5b and 6b by sub- (not shown) were almost the same as those of Figs 3c and tracting bias. The inverter terminal voltages for 2LS at 4c. Thus, just as in Case 1, during the fault, only the BESS 1 points A and B (not shown) are similar to those of Figs 5c output became 0 and the other BESSs increased their out- and 6c. puts to supply power to the neighbouring load and cover the loss caused by fault currents. 4 Discussion The graphs of frequency, active power, voltage and geo- 3.4 Case 4: 2LS at night-time (BESS III) thermal generator rotational speed in Section 3 reveal that The system frequency deviation (a) and the BESS active they all converged to a sufficiently stable state 1–2 s after power output (b) when 2LS occurs at point A  are shown the elimination of the fault. From these results, it was veri- in Fig. 9. The corresponding response at point B is shown fied that, even if a short circuit or ground fault occurs in Rotational speed Output voltage [pu] Output power [pu] Frequency [Hz] [rad/s] Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 202 | Clean Energy, 2021, Vol. 5, No. 2 0.4 0.2 0.0 BESS 1 –0.2 BESS 2 BESS 3 –0.4 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 1.0 0.5 0.0 BESS 1 –0.5 BESS 2 BESS 3 –1.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 1.2 1.1 1.0 BESS 1 0.9 BESS 2 BESS 3 0.8 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 5.0 2.5 0.0 –2.5 –5.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] Fig. 4: Simulation results for Case 1: 3LG at point B the remote island system to which the proposed control inverter installation while point B is farther away. BESS 1 is method is applied, the system will return to the steady state most directly affected and the BESS 2 reaction is greater in immediately after the removal of the fault, thus enabling the case of point A faults. Conversely, the response of BESS stable operation. That is, a remote island power grid con- 3 is relatively greater at point B faults because it is closer. sisting of distributed inverters and geothermal power can This is a reasonable result considering the impedance of be operated stably by the proposed control system. Next, 6.6-kV distribution lines between loads 2 km apart. the results of each verification are compared. For the pur - pose of discussion, the maximum values of the frequency 4.2 Difference between types of faults (3LG deviation after the faults that were introduced, as shown and 2LS) in Figs 3–10, are summarized in Table 2. Looking at the frequency deviation in Table 2, 3LG  <  2LS for faults at point A.  For faults at point B, the frequency 4.1 Difference between fault points (points deviation is 3LG  >  2LS. Referring to the calculation for P A and B) in Fig. 1, the difference between the active power-output The frequency deviation due to the fault is greater at order and the detection value of the BESS is used to cal- point A. As can be seen from Fig. 2, point A is closer to the culate the frequency deviation. In the case of 3LG at point Rotational speed Output voltage [pu] Output power [pu] Frequency [Hz] [rad/s] Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 203 Fig. 5: Simulation results for Case 2: 2LS at point A A, the output of BESS 1 decreases from the initial value 4.4 Difference between load conditions (heavy to 0 pu, whereas the output of the BESS in 2LS decreases load and light load) from the initial value to a negative value (about –0.2 pu). As For the maximum daytime demand of 11  MW, three PVs the active power deviation increases, the frequency devi- (0.9 pu) and three BESSs (–0.8 pu) were assumed in Refs [15] ation also increases, resulting in 3LG < 2LS. In the case of a and [16], respectively. As in this study, 3LG and 2LS were fault at point B, the decrease in the voltage value and the simulated at points A and B. The maximum frequency de- change in the BESS active power output are small due to viation of both point A  and point B was more affected by the distance to the BESS installation location. Therefore, the 3LG fault. As described above, the change in the active the frequency deviation is 3LG > 2LS. power output of the BESS was found to vary from –0.8 to 0 pu in the case of 3LG, whereas the active power output was re- duced from –0.8 to –0.2 pu in the case of 2LS. The deviation 4.3 Difference between BESS I and BESS III of the effective power output of the BESS at each fault was conditions caused by the magnitude of the steady-state power-output Table 2 shows that the frequency deviation is slightly lower value. In the case of a heavy load, the ratio of the output for the BESS III condition. In the BESS I condition, BESS 1 change due to the fault to the steady-state value is small, is in operation at 0.15 pu and BESS 2 and BESS 3 are in op- while, in the case of a light load, the ratio of the output eration at 0 pu. The fault reduced the output of BESS 1 to change due to the fault is very large. Although detailed fig- 0  pu at 3LG and to about –0.2  pu at 2LS, and the outputs ures are not given in this paper, the change in the BESS of the other BESSs also changed. In the BESS III condition, active power output between pre- and post-fault at point the output of BESS 1 decreased from 0.05 to 0 pu for 3LG A is 3LG > 2LS under heavy-load conditions. Therefore, the and –0.2 pu for 2LS due to the fault. The change in the ac- frequency deviation is also 3LG > 2LS, as shown in Table 2. tive power during the fault was smaller in BESS III. This From the above, in the light-load state, when 2LS occurs on resulted in a frequency-deviation characteristic of BESS the grid side as indicated by point A, the frequency devi- I > BESS III. ation may be slightly larger than that for 3LG. Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 204 | Clean Energy, 2021, Vol. 5, No. 2 Fig. 6: Simulation results for Case 2: 2LS at point B Fig. 7: Simulation results for Case 3: 3LG at point A Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 205 Fig. 8: Simulation results for Case 3: 3LG at point B Fig. 9: Simulation results for Case 4: 2LS at point A All the system properties (frequency, active power, 5 Conclusions voltage and geothermal generator rotational speed) were The proposed inverter control block was applied to a re- promptly returned to steady-state conditions after fault mote island-scale system with geothermal power gen- removal. In addition, 2LS produced a frequency deviation eration and multiple renewable energy sources and greater than that of 3LG at point A, which could be ex- distributed BESSs. In particular, this paper simulated faults plained by the operation of the inverter control block used (2LS and 3LG) in which the power was supplied only by in the BESS. From the results, the transient stability of the the BESS during light loads. In addition, the fault condition proposed model for the fault at the time of light load was was selected, with point A  on the system side and with evaluated. In reality, we are aiming to control supply and point B on the geothermal power plant side, and the tran- demand with 100% renewable energy and BESSs. sient stability of the power system was verified. Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 206 | Clean Energy, 2021, Vol. 5, No. 2 Fig. 10: Simulation results for Case 4: 2LS at point B Table 2: Absolute value of the maximum frequency deviation Simulation condition Maximum inverter frequency deviation BESS operation Fault point 3LG Value (Hz) 2LS Value (Hz) BESS I A Case 1 0.22 Case 2 0.34 B 0.17 0.10 BESS III A Case 3 0.19 Case 4 0.30 B 0.16 0.09 Daytime A Reference [15] 0.40 Reference [16] 0.28 B 0.27 0.14 In this paper, we simulated a remote island system [4] Torres  M, Lopes  LAC. Virtual synchronous generator control in autonomous wind-diesel power systems. In: IEEE Electrical powered by a geothermal power plant with synchronous Power & Energy Conference (EPEC), Montreal, QC, Canada, 22–23 generators and, for future studies, we also need to evaluate October 2009, 1–6. the existence of diesel generators in the power-supply mix [5] Yang  XZ, Su  JH, Ding  M, et  al. Control strategy for virtual and the stability of the generators when the output ratio is synchronous generator in microgrid. In: 4th International changed. In addition, in order to apply the proposed con- Conference on Electric Utility Deregulation and Restructuring and trol method to an actual system, it is necessary to carry Power Technologies (DRPT), Weihai, Shandong, China, 6–9 July 2011, 1633–1637. out sufficient verification via a test using a model system [6] Sakimoto  K, Miura  Y, Ise  T. Stabilization of a power system and a field test. including inverter type generators by the virtual synchronous generator. IEEJ Transactions on Power and Energy, 2012, 132:341–349. Conflict of Interest [7] Beck  HP, Hesse  R. Virtual synchronous machine. In: 9th International Conference on Electrical Power Quality and None declared. Utilization (EPQU), Barcelona, 9–11 October 2007, 1–6. [8] Zhong  QC, Weiss  G. Synchronverters: inverters that mimic References synchronous generators. IEEE Transactions on Industrial [1] Goda T, Shoyama M. Converter Principles and Design Methods for Electronics, 2011, 58:1259–1267. Renewable Energy. Kagakujyoho shuppan, Japan, 2016, 138–140. [9] Matevosyan  J, Badrzadeh  B, Prevost  T, et  al. Grid-forming in- [2] Naoi  S, Arai  J, Noro  Y. Control for inverter power sources in verters. IEEE Power and Energy Magazine, 2019, November/ a standalone microgrid. IEEJ Transactions on Power and Energy, December:89–98. 2017, 137:426–433. [10] Ramasubramanian  D, Farantatos  E, Ziaeinejad  S, et  al. [3] van Wesenbeeck MPN, de Haan SWH, Varela P, et al. Grid tied Operation paradigm of an all converter interfaced generation converter with virtual kinetic storage. In: IEEE Bucharest bulk power system. IET Generation, Transmission & Distribution, Power Tech, Bucharest, Romania, 28 June–2 July 2009, 1–7. 2018, 12:4240–4248. Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 207 [11] Toshikazu  Y. Demonstration test of microgrid system for ‘Power Systems Engineering’, IEEJ, PE-19–097, PSE-19–109, small remote islands. Journal of IEIE Japan, 2011, 31:907–911. 2019, 17–22. [12] Takano T, Yasuhiro K, Koji T, et  al. Isolated operation at [15] Shoichi S. Instantaneous value analysis in large introduction Hachinohe micro grid Project.IEEJ Transactions on Power and of renewable energy to remote island. In: Proceedings of the Energy, 2009, 129:499–505. 2019 IEIEJ Student Research Presentation, B-6, 44–45. [13] Noro Y. Proposal of inverter control method for electric power [16] Shoichi S, Noro Y. Instantaneous value analysis in large intro- system consisting of energy storage. IEEJ Transactions on Power duction of renewable energy to remote island part 2. In: 2020 and Energy, 2017, 138:854–861. IEEJ Annual Meeting, 2020, 6–227, 389–390. [14] Shoichi  S, Noro  Y. Inverter control method for the large [17] ATPDraw—The Graphical Preprocessor to ATP Electronic Transients introduction of renewable energy to Remote Island. In: The Program. https://www.atpdraw.net/ (10 February 2021, date Papers of Joint Technical Meeting on ‘Power Engineering’ and last accessed). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Clean Energy Oxford University Press

Control method for the stable operation of distributed inverters following the introduction of large-scale renewable energy to a remote island grid

Clean Energy , Volume 5 (2) – Jun 1, 2021

Loading next page...
 
/lp/oxford-university-press/control-method-for-the-stable-operation-of-distributed-inverters-4Rrd0v52Ki
Publisher
Oxford University Press
Copyright
Copyright © 2021 National Institute of Clean-and-Low-Carbon Energy
ISSN
2515-4230
eISSN
2515-396X
DOI
10.1093/ce/zkab004
Publisher site
See Article on Publisher Site

Abstract

A control system that can operate multiple distributed inverters stably is proposed. The con- trol system is assumed to be applied to a remote island. From the simulation results, it was verified that the control method provides stable operation against various faults. Keywords: renewable energy; photovoltaic; battery energy storage system; inverter; remote island (Constant Voltage Constant Frequency) operation with one Introduction inverter or the master–slave method [2 1]. , However, if the The generation of CO and other greenhouse gases is a grid size is somewhat large, it is likely that inverters will be main contributor to global warming, which in turn af- installed in multiple locations. In this case, it is necessary fects different social and economic issues worldwide. As to coordinate the operation of the distributed inverters. a countermeasure, renewable energy is attracting atten- Various studies such as the VSG (Virtual Synchronous tion all over the world. However, since the output of re- Generator) [3, 6] have been carried out [7, 8] in an attempt to newables, specifically photovoltaic (PV) and wind power add an inertial property to an inverter, and recently it has generation, is heavily dependent on weather conditions, been developed into the concept of the GFM (Grid Forming) producing a stable output can be difficult. Output instabil- inverter [9, 10]. However, there are many variations in the ities can produce large frequency fluctuations in the grid. functions to be realized and the methods of realization, Moreover, the instances of output instability increase as and the control technologies for distributed inverters have the proportion of power contributed to the grid by renew- not yet been fully established. Therefore, it is difficult to able energy sources increases. In general, the introduction introduce large-scale renewable energy to an existing grid of large-scale renewable energy to the grid requires storage from a technical standpoint. Against this background, the batteries for supply-and-demand control and inverters are authors believe that it is necessary to consider a method essential to connect these to the power grid. In a system in that can control future grid configurations where renew- which a large number of inverter facilities are introduced, able energy sources predominate. it is difficult to acquire the inertial properties of the prime Many remote islands currently use diesel power as mover, because conventional inverter facilities could not their main source of energy and the unit cost of electri- have inertial properties. city generation is expensive due to the extra cost of trans- When the grid size is very small, such as in a remote porting fuel. In addition, there is a risk that the transport island or microgrid, it is possible to realize a grid centred of fuel itself will be severed in the event of a disaster. on the power supply by the inverter by applying CVCF Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 198 | Clean Energy, 2021, Vol. 5, No. 2 Therefore, the introduction of large-scale renewable en- supplies should be distributed. Therefore, it is considered ergy to remote islands is necessary. It is expected that re- appropriate for verifying the response of the system to newable energy, which does not consume fuel, will enable grid faults. Similar-scale remote islands include Oshima, the local production and local consumption of energy. In Tokyo and Kumejima, Okinawa Prefecture. In addition, the this regard, early studies include the design of remote is- results of this study can be applied to these similarly or land microgrids (0.5-MW demand) [11] and an empirical larger-scale remote islands. study of the renewable energy supply in small-scale grids (610-kW total demand) [12]. However, the small scale of 1.1 Inverter control these systems requires only one storage battery system for supply-and-demand control. Reference [13] presented and discussed the characteristics In this study, we examined a case in which a large of the BESS inverters proposed for load following and co- amount of renewable energy is introduced and multiple ordinated operation. In this study, since the multiple in- storage battery systems (hereinafter referred to as ‘BESS’ verters were distributed, the BESS for adjusting the output or ‘battery energy storage system’) are distributed. The au- has the same configuration as that for voltage control thors evaluated a method for adjusting the energy supply utilizing the ‘synchronization force’ required for coordin- and demand for remote islands that can be regarded as ated operation. Fig. 1 shows the main circuit configuration independent electric power systems by using multiple and control circuit. The features of this control system are BESSs that apply the control method proposed in Ref. [13]. briefly described below. Reference [13] presents a control and simulation for a • The inverter is operated by voltage control and the syn- large-scale system. References [14–16] and this paper are chronization force is used to maintain synchronization characterized by a plural number of BESSs in the system. between each inverter. The work described in this paper is a follow-up study to • The control system has properties equivalent to the in- Refs [14–16]. The stability of operation in the event of a ertia of the synchronous machine in order to avoid large fault in the proposed system under various conditions was fluctuations in frequency and to maintain stability. evaluated and investigated by instantaneous value-based • To improve the damping of power swings, phase com- simulations. pensation is provided. In Section 1, the model remote island grid system char - • The frequency control is equivalent to a governor. acteristics, the inverter controller configuration and the • The voltage-control part is equivalent to a STATCOM target system model are explained. In Section 2, the simu- (Static Compensator). lation conditions are assumed. The output of the PV and • Even if a system fault were to occur, the inverter would the BESS under steady-state conditions before reproducing continue to operate without stopping. the fault are mentioned. In Section 3, the performance of • The system reduces the overcurrent in short-circuit the control system under different conditions and with faults and limits the current to continue operation. two types of faults is presented systematically. In Section • AFC (automatic frequency control) is used. 4, the results for each case are considered. The changes in voltage, frequency and BESS output power in response to In addition, the PV inverter, whose output cannot be con- faults and fault removal are compared and discussed for trolled and is dependent on the weather, uses the conven- each case. In Section 5, the main results from the study tional current-control method. are recapitulated and the ability of the control system developed to operate the inverters stably is emphasized. Moreover, the scope for future work is also provided. 1.2 Modelling of the target system The PV systems and loads were assumed to be distributed in three locations. Assuming that the load in the island 1 The proposed system for evaluation (max. 11 MW and min. 5 MW) is dispersed in these three Hachijo Island, Tokyo was designated as an isolated is- locations with a ratio of 5:4:2 in heavy load and 1:2:2 in land. Hachijo Island is located 287  km south of Tokyo, light load, and the PV system and the BESS were placed with an area of 69  km and a population of ~8000. The in the vicinity of each load. In this case, the power-supply minimum electricity demand is 5  MW and the peak de- capacity does not need to be proportional to the load and mand is 11  MW. Since a geothermal power plant with a the total capacity was divided into three equal parts. Fig. rated capacity of 3.66 MVA was assumed to be in operation, 2a shows the distribution line connecting the loads, which the remaining demand was assumed to be supplied by PV. were 2 km apart, and the geothermal power plant, which The reason for targeting this island is mainly because the was 8 km away from the nearest load. Fig. 2b shows that scale of demand and the scale of area fit the purpose of the DC-side voltage is 850  V and the AC-side voltage is the study. That is, if the scale is small, CVCF operation by 6.6 kV. In addition, the PV capacity is 7.6 × 10  kW ÷ 3 = 2 one inverter or the master–slave method can be applied .53 ×  10   kW (25.3  MW) and the inverter output is 75% of but, on this scale, it is considered that inverter power the PV-panel capacity, or 19 MW. The capacity of the BESS Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 199 Inverter Transformer DC power AC bus of source Power system Gate pulse Output current Output voltage 1/Droop Frequency Frequency + 1 + sT control Frequency order Δf f AFC + + Phase 3Phase + 12 + π Output power order compen- voltage PWM + Ms + DS sation signal Δf Active power Inertia property AC-AVR Output voltage ⎥E⎥ Voltage Current Voltage order control limit + + Xslope Reactive power order 1 + sT Reactive power Output current Fig. 1: Detail configuration of inverter controller [13] geothermal power plant is 3.66  MVA and the output was A B 6.6kV (2km) 6.6kV (2km) 6.6kV (8km) adjusted to 3.66 × 0.9 = 3.3 MW at all times. P P P L LG L 2 Simulation conditions V V V o o o e . . a a a o This paper examines the transient stability of the proposed B B d d d model for faults during light loads. Simulations were per - E E E 3.3MW S S formed by first classifying the conditions into ‘daytime’ and Nighttime/Daytime S S S 1MW/5MW 2MW/4MW 2MW/2MW ‘night-time’. In the daytime, power generation by PV is high at 0.9 pu, with excess power charging the BESS. The maximum 6.6kV load was 11  MW. Simulation results for the three-phase line-to-ground faults (3LG) and the two-line short-circuit fault (2LS) in this condition show that stable operation is possible [15, 16]. The ‘night-time’ does not have PV power INV INV generation. Thus, power is supplied from the BESS and geo- DC850V DC850V thermal power. The night-time load (5 MW) was assumed to PV 19MW BESS 17MW be allocated following the ratio 1:2:2. Here, the BESS capacity Current control Proposal control (17  MW × 3 units) is sufficiently large compared to the de- mand (≤11 MW) in the island. BESS I indicates the use of one Fig. 2: System-model diagram BESS to meet all the demand on the island. In this case, the remaining BESSs were operated at output power order 0. The inverter was set to 17  MW. This appears to be an exces- pattern in which three BESS outputs were equal is indicated sive capacity for the apparent load, but was optimized as BESS III. The examples reported so far and the conditions to consider the economic efficiency, assuming that there in this study are shown in Table 1. is no diesel power generation throughout the year to ad- The fault was introduced 6  s after the simulation was just the supply and meet demand [14]. The capacity of the sufficiently stable and was removed after 0.1  s. For a Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 200 | Clean Energy, 2021, Vol. 5, No. 2 Table 1: Overview of simulation voltage decreases to 0 due to the fault and distribution is not carried out during the fault. Other BESSs increase Conditions Fault type their output to supply power to neighbouring loads and Time band Fault point 3LG 2LS cover losses caused by fault currents. Looking at point ‘P’ in Fig. 1, we find that the BESS active power output is used Daytime A and B Reference [5] Reference [6] to calculate the frequency deviation (∆f). Since the power- Night-time (BESS I) A Case 1 Case 2 order value is constant, the increase in deviation causes the power-detection value to decrease. Conversely, as the Night-time (BESS III) A Case 3 Case 4 detected value of the power increases, the frequency de- viation decreases. This operation explains the results shown in Fig. 3a. In Fig. 4c, distribution line impedance grid frequency of 50  Hz, 0.1  s corresponds to five cycles. exists between the fault point (point B) and the observa- Verification was performed on four different cases. tion point, so that the fault has little impact and produces little fluctuation. Fig. 3d shows the rotational speed of the (i) Case 1: 3LG at night-time (BESS I) geothermal generator. The rotational speed accelerates We assumed that one BESS is supplying power at ‘night- due to the impact of the fault and returns to the steady- time’. The system responses to 3LG at point A and point B state value with the removal of the fault. In Fig. 4d, the are evaluated. rotational speed slightly decreases immediately after the fault. It then accelerates and returns to steady state after (ii) Case 2: 2LS at night-time (BESS I) the fault removal. Since this is a fault in the vicinity of a We assumed that one BESS is supplying power at ‘night- geothermal generator, it is certain that angular backswing time’. The system responses to 2LS at point A and point B was produced at the moment of the fault. are evaluated. (iii) Case 3: 3LG at night-time (BESS III) 3.2 Case 2: 2LS at night-time (BESS I) We assumed that all BESSs are supplying power at The frequency deviation (a), the BESS active power output ‘night-time’. The system responses to 3LG at point A and (b) and the inverter terminal voltage (c) when 2LS occurs point B are evaluated. at point A are shown in Fig. 5. The corresponding response to 2LS at point B is shown in Fig. 6. From Figs 5 and 6, it (iv) Case 4: 2LS at night-time (BESS III) can be seen that the frequency deviation, the active power We assumed that all BESSs are supplying power at output and the inverter terminal voltage were all returned ‘night-time’. The system responses to 2LS at point A  and to steady-state values a short time after the elimination point B are evaluated. of the fault. The fluctuation of the frequency deviation is The simulation was carried out using the instantaneous slightly larger to that of Case 1 at point FA.  ig. 5b shows value-based software ATP (Alternative Transients Program) that the output power of BESS 1 was negative (about and its graphical user interface (GUI) tool, ATPDraw [17]. –0.2 pu) during the fault. This indicates that the BESS was charging during the fault. BESS 1 in Fig. 5c is near the fault point but, due to 2LS, the voltage was not 0 and decreased 3 Simulation results to ~0.5  pu. After the fault was eliminated, the output re- 3.1 Case 1: 3LG at night-time (BESS I) turned to 1 pu immediately. The common characteristic of the BESS active power output in Case 1 and Case 2 is that The frequency deviation (a), the BESS active power output only BESS 1 outputs ~0.15 pu, while BESS 2 and BESS 3 op- (b), the inverter terminal voltage (c) and the rotational erate at output 0 pu. Both the frequency deviation and the speed of the geothermal generator (d) when 3LG occurs at BESS active power output are detected values in the BESS point A  are shown in Fig. 3. The corresponding response block on the load side. Therefore, the response at point B is to 3LG at point B is shown in Fig. 4. These graphs of (a) to found to be small compared to that at point A. (c) show the values detected in the BESS inverter control block in Fig. 1. From Figs 3 and 4, it can be seen that the fre- quency deviation, the active power output and the inverter 3.3 Case 3: 3LG at night-time (BESS III) terminal voltage were all returned to steady-state values immediately after the elimination of the fault. Looking The system frequency deviation (a) and the BESS active at the frequency deviations in Fig. 3a, only the frequency power output (b) when 3LG occurs at point A are shown in deviation for BESS 1 in the vicinity of the fault point in- Fig. 7. The corresponding response at point B is shown in creased during the fault while the frequency deviations for Fig. 8. The fluctuation of the frequency deviation is similar BESS 2 and BESS 3 decreased. This response is explained to that for points A and B of Case 1. Looking at the initial by the active power output. In Fig. 3b , only BESS 1 showed value of the inverter output power of the BESS of Case 3 a decrease in output during the fault. This is because the (Figs 7 and 8), all three BESS units operate at an output Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 201 0.4 0.2 0.0 BESS 1 –0.2 BESS 2 BESS 3 –0.4 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 1.0 0.5 0.0 BESS 1 –0.5 BESS 2 BESS 3 –1.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 1.2 1.0 0.8 0.6 BESS 1 0.4 BESS 2 BESS 3 0.2 0.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 5.0 2.5 0.0 BESS 1 BESS 2 –2.5 BESS 3 –5.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] Fig. 3: Simulation results for Case 1: 3LG at point A of ~0.05  pu. In Fig. 7b, the BESS active power output de- in Fig. 10. The fluctuation of the frequency deviation is creased to 0 only for BESS 1 near fault point A  and the similar to that of Case 2 at both points A and B. All three power output of BESS 2 and BESS 3 increased as in Case BESS units operate at an initial value of ~0.05 pu. The BESS 1. The inverter terminal voltages at points A and B at 3LG active power output behaves like Figs 5b and 6b by sub- (not shown) were almost the same as those of Figs 3c and tracting bias. The inverter terminal voltages for 2LS at 4c. Thus, just as in Case 1, during the fault, only the BESS 1 points A and B (not shown) are similar to those of Figs 5c output became 0 and the other BESSs increased their out- and 6c. puts to supply power to the neighbouring load and cover the loss caused by fault currents. 4 Discussion The graphs of frequency, active power, voltage and geo- 3.4 Case 4: 2LS at night-time (BESS III) thermal generator rotational speed in Section 3 reveal that The system frequency deviation (a) and the BESS active they all converged to a sufficiently stable state 1–2 s after power output (b) when 2LS occurs at point A  are shown the elimination of the fault. From these results, it was veri- in Fig. 9. The corresponding response at point B is shown fied that, even if a short circuit or ground fault occurs in Rotational speed Output voltage [pu] Output power [pu] Frequency [Hz] [rad/s] Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 202 | Clean Energy, 2021, Vol. 5, No. 2 0.4 0.2 0.0 BESS 1 –0.2 BESS 2 BESS 3 –0.4 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 1.0 0.5 0.0 BESS 1 –0.5 BESS 2 BESS 3 –1.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 1.2 1.1 1.0 BESS 1 0.9 BESS 2 BESS 3 0.8 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] 5.0 2.5 0.0 –2.5 –5.0 5.5 6.0 6.5 7.0 7.5 8.0 Time [s] Fig. 4: Simulation results for Case 1: 3LG at point B the remote island system to which the proposed control inverter installation while point B is farther away. BESS 1 is method is applied, the system will return to the steady state most directly affected and the BESS 2 reaction is greater in immediately after the removal of the fault, thus enabling the case of point A faults. Conversely, the response of BESS stable operation. That is, a remote island power grid con- 3 is relatively greater at point B faults because it is closer. sisting of distributed inverters and geothermal power can This is a reasonable result considering the impedance of be operated stably by the proposed control system. Next, 6.6-kV distribution lines between loads 2 km apart. the results of each verification are compared. For the pur - pose of discussion, the maximum values of the frequency 4.2 Difference between types of faults (3LG deviation after the faults that were introduced, as shown and 2LS) in Figs 3–10, are summarized in Table 2. Looking at the frequency deviation in Table 2, 3LG  <  2LS for faults at point A.  For faults at point B, the frequency 4.1 Difference between fault points (points deviation is 3LG  >  2LS. Referring to the calculation for P A and B) in Fig. 1, the difference between the active power-output The frequency deviation due to the fault is greater at order and the detection value of the BESS is used to cal- point A. As can be seen from Fig. 2, point A is closer to the culate the frequency deviation. In the case of 3LG at point Rotational speed Output voltage [pu] Output power [pu] Frequency [Hz] [rad/s] Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 203 Fig. 5: Simulation results for Case 2: 2LS at point A A, the output of BESS 1 decreases from the initial value 4.4 Difference between load conditions (heavy to 0 pu, whereas the output of the BESS in 2LS decreases load and light load) from the initial value to a negative value (about –0.2 pu). As For the maximum daytime demand of 11  MW, three PVs the active power deviation increases, the frequency devi- (0.9 pu) and three BESSs (–0.8 pu) were assumed in Refs [15] ation also increases, resulting in 3LG < 2LS. In the case of a and [16], respectively. As in this study, 3LG and 2LS were fault at point B, the decrease in the voltage value and the simulated at points A and B. The maximum frequency de- change in the BESS active power output are small due to viation of both point A  and point B was more affected by the distance to the BESS installation location. Therefore, the 3LG fault. As described above, the change in the active the frequency deviation is 3LG > 2LS. power output of the BESS was found to vary from –0.8 to 0 pu in the case of 3LG, whereas the active power output was re- duced from –0.8 to –0.2 pu in the case of 2LS. The deviation 4.3 Difference between BESS I and BESS III of the effective power output of the BESS at each fault was conditions caused by the magnitude of the steady-state power-output Table 2 shows that the frequency deviation is slightly lower value. In the case of a heavy load, the ratio of the output for the BESS III condition. In the BESS I condition, BESS 1 change due to the fault to the steady-state value is small, is in operation at 0.15 pu and BESS 2 and BESS 3 are in op- while, in the case of a light load, the ratio of the output eration at 0 pu. The fault reduced the output of BESS 1 to change due to the fault is very large. Although detailed fig- 0  pu at 3LG and to about –0.2  pu at 2LS, and the outputs ures are not given in this paper, the change in the BESS of the other BESSs also changed. In the BESS III condition, active power output between pre- and post-fault at point the output of BESS 1 decreased from 0.05 to 0 pu for 3LG A is 3LG > 2LS under heavy-load conditions. Therefore, the and –0.2 pu for 2LS due to the fault. The change in the ac- frequency deviation is also 3LG > 2LS, as shown in Table 2. tive power during the fault was smaller in BESS III. This From the above, in the light-load state, when 2LS occurs on resulted in a frequency-deviation characteristic of BESS the grid side as indicated by point A, the frequency devi- I > BESS III. ation may be slightly larger than that for 3LG. Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 204 | Clean Energy, 2021, Vol. 5, No. 2 Fig. 6: Simulation results for Case 2: 2LS at point B Fig. 7: Simulation results for Case 3: 3LG at point A Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 205 Fig. 8: Simulation results for Case 3: 3LG at point B Fig. 9: Simulation results for Case 4: 2LS at point A All the system properties (frequency, active power, 5 Conclusions voltage and geothermal generator rotational speed) were The proposed inverter control block was applied to a re- promptly returned to steady-state conditions after fault mote island-scale system with geothermal power gen- removal. In addition, 2LS produced a frequency deviation eration and multiple renewable energy sources and greater than that of 3LG at point A, which could be ex- distributed BESSs. In particular, this paper simulated faults plained by the operation of the inverter control block used (2LS and 3LG) in which the power was supplied only by in the BESS. From the results, the transient stability of the the BESS during light loads. In addition, the fault condition proposed model for the fault at the time of light load was was selected, with point A  on the system side and with evaluated. In reality, we are aiming to control supply and point B on the geothermal power plant side, and the tran- demand with 100% renewable energy and BESSs. sient stability of the power system was verified. Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 206 | Clean Energy, 2021, Vol. 5, No. 2 Fig. 10: Simulation results for Case 4: 2LS at point B Table 2: Absolute value of the maximum frequency deviation Simulation condition Maximum inverter frequency deviation BESS operation Fault point 3LG Value (Hz) 2LS Value (Hz) BESS I A Case 1 0.22 Case 2 0.34 B 0.17 0.10 BESS III A Case 3 0.19 Case 4 0.30 B 0.16 0.09 Daytime A Reference [15] 0.40 Reference [16] 0.28 B 0.27 0.14 In this paper, we simulated a remote island system [4] Torres  M, Lopes  LAC. Virtual synchronous generator control in autonomous wind-diesel power systems. In: IEEE Electrical powered by a geothermal power plant with synchronous Power & Energy Conference (EPEC), Montreal, QC, Canada, 22–23 generators and, for future studies, we also need to evaluate October 2009, 1–6. the existence of diesel generators in the power-supply mix [5] Yang  XZ, Su  JH, Ding  M, et  al. Control strategy for virtual and the stability of the generators when the output ratio is synchronous generator in microgrid. In: 4th International changed. In addition, in order to apply the proposed con- Conference on Electric Utility Deregulation and Restructuring and trol method to an actual system, it is necessary to carry Power Technologies (DRPT), Weihai, Shandong, China, 6–9 July 2011, 1633–1637. out sufficient verification via a test using a model system [6] Sakimoto  K, Miura  Y, Ise  T. Stabilization of a power system and a field test. including inverter type generators by the virtual synchronous generator. IEEJ Transactions on Power and Energy, 2012, 132:341–349. Conflict of Interest [7] Beck  HP, Hesse  R. Virtual synchronous machine. In: 9th International Conference on Electrical Power Quality and None declared. Utilization (EPQU), Barcelona, 9–11 October 2007, 1–6. [8] Zhong  QC, Weiss  G. Synchronverters: inverters that mimic References synchronous generators. IEEE Transactions on Industrial [1] Goda T, Shoyama M. Converter Principles and Design Methods for Electronics, 2011, 58:1259–1267. Renewable Energy. Kagakujyoho shuppan, Japan, 2016, 138–140. [9] Matevosyan  J, Badrzadeh  B, Prevost  T, et  al. Grid-forming in- [2] Naoi  S, Arai  J, Noro  Y. Control for inverter power sources in verters. IEEE Power and Energy Magazine, 2019, November/ a standalone microgrid. IEEJ Transactions on Power and Energy, December:89–98. 2017, 137:426–433. [10] Ramasubramanian  D, Farantatos  E, Ziaeinejad  S, et  al. [3] van Wesenbeeck MPN, de Haan SWH, Varela P, et al. Grid tied Operation paradigm of an all converter interfaced generation converter with virtual kinetic storage. In: IEEE Bucharest bulk power system. IET Generation, Transmission & Distribution, Power Tech, Bucharest, Romania, 28 June–2 July 2009, 1–7. 2018, 12:4240–4248. Downloaded from https://academic.oup.com/ce/article/5/2/196/6271253 by DeepDyve user on 11 May 2021 Sato and Noro | 207 [11] Toshikazu  Y. Demonstration test of microgrid system for ‘Power Systems Engineering’, IEEJ, PE-19–097, PSE-19–109, small remote islands. Journal of IEIE Japan, 2011, 31:907–911. 2019, 17–22. [12] Takano T, Yasuhiro K, Koji T, et  al. Isolated operation at [15] Shoichi S. Instantaneous value analysis in large introduction Hachinohe micro grid Project.IEEJ Transactions on Power and of renewable energy to remote island. In: Proceedings of the Energy, 2009, 129:499–505. 2019 IEIEJ Student Research Presentation, B-6, 44–45. [13] Noro Y. Proposal of inverter control method for electric power [16] Shoichi S, Noro Y. Instantaneous value analysis in large intro- system consisting of energy storage. IEEJ Transactions on Power duction of renewable energy to remote island part 2. In: 2020 and Energy, 2017, 138:854–861. IEEJ Annual Meeting, 2020, 6–227, 389–390. [14] Shoichi  S, Noro  Y. Inverter control method for the large [17] ATPDraw—The Graphical Preprocessor to ATP Electronic Transients introduction of renewable energy to Remote Island. In: The Program. https://www.atpdraw.net/ (10 February 2021, date Papers of Joint Technical Meeting on ‘Power Engineering’ and last accessed).

Journal

Clean EnergyOxford University Press

Published: Jun 1, 2021

References