Optimal Control of Li-Ion Battery Energy Storage System for Frequency Regulation

Abstract

The usage of renewable energy to supply electricity demands is expected to increase drastically in the coming years. This is attribute to the encouragement of the government through policies such as RPS(renewable energy portfolio standards) and FIT(feed-in-tariffs). However, the performance of power system is sometimes critically constrained by stability characteristics. It is very challenging for system operators to manage a high penetration of renewable energy resources in a bulk power system, as they are known to deteriorate the dynamic performance of power systems.Therefore, it has been proposed that the large-scale Energy Storage System (ESS) of which power output can be controlled accurately and quickly can be applied to a bulk power system for securing the stability of the system despite a high penetration of the renewable energy resources. By increasing deployment of renewable energy, importance of power grid stability has been increased. For example, Wind Power has characteristics of difficult prediction and rapid variation depending on the regional climate characteristics. Energy Storage System aims to transform an uncontrollably variable and partially unpredictable renewable energy into a controlled and predictable one. For these purpose, a quickly responsive and highly efficient control-algorithm is essential. It is for using such a high performance of ESS when a disturbance occurs on a electrical grid system.KEPCO implemented a demonstration project on large-scale grid connected 4 MW / 8 MWh Battery ESS (BESS) using Lithium-Ion Battery (LIB). Fig. 1 shows the interconnection of ESS at the demonstration site. The 4 MW / 8 MWh BESS is connected to both the 22.9kV substation bus and the distribution line in Jocheon, Jeju Island for offshore wind turbines. In this study, Operation strategies for the BESS were developed for frequency regulation of the grid-system. Fig. 1 Interconnection of ESS at Jocheon Substation For power system frequency regulation, There are two kinds of control strategies through ESS here. One is for the normal status that the ratio of frequency change is not bigger than reference ratio of frequency change(ξ). The reference ratio is frequency changes per second when the smallest power supply was failed from the power system. The other is for the dynamic status that the ratio of frequency change is bigger than the reference ratio. In case of South Korea, it is about -0.306 Hz/sec. It could be variable every year because the total amount of power supply in the grid-system would be different. In this paper, we researched briefly about the normal status.At normal control mode, the power output(P) of the ESS is proportional to the change in frequency(f). kd is a proportional coefficient set by the droop characteristics of a battery generation.P=kd(60-f) In order to distribute output power of each battery, this study established a SOC(State Of Charge) weighted control strategy that can consider the SOC of each battery. The following factors were considered to provide frequency regulation service while considering the SOC :1) Set proportional coefficient (kd) for calculating total output demand2) Set priority at output distribution according to SOC3) Consider rating output of each battery when distributing outputAt normal status, the proportional coefficient(kd) was set by applying the droop of a general generator. This is to control the ESS output with a performance similar to the droop of the general generator. When allocating output demand determined by each battery SOC through this strategy, the battery with a higher SOC have priority in distribution. This is more efficient way to fulfill the total output demand cost-effectively. The SOC weighted control result is shown in Figure 2.Fig. 2 Result of battery output during SOC weighted control (simulation)It can be found that the distributed quantity of output was provided depending on the SOC level of each battery. It's likely to be seen that a battery with a higher SOC has priority in the distribution of output.SOC management effect of the control algorithm was analyzed through a 5-hour too. The result obtained from a simulation frequency of five hours before and after a credible accident. We can find that, in the range within dead band, the control of battery was conducted in order to reach a SOC sustainment section. When the frequency is in the dead band, ESS is not working for frequency regulation. In the outside range of the frequency dead band, it was observed that control capacity was distributed according to SOC in order to provide frequency regulation.