Real-Time Analysis of an Active Distribution Network - Coordinated Frequency Control for Islanding Operation

Abstract

The increasing penetration of distributed generation (DG) and distributed energy re-sources (DERs), and the consequential requirement to accommodate and integrate them within distribution networks brings both challenges and opportunities to the distribution system operator (DSO). This will enable and require a transition from today’s passive distribution networks to future active distribution networks (ADNs) which utilizes ad-vanced operation and control strategies in order to improve power supply reliability, and realize the potential of DG to provide system support. The presence of DERs within distribution networks makes it possible to operate the distribution networks independently which is called islanding operation. However, it is a challenge to ensure secure and reliable operation of the islanded system due to a num-ber of reasons, e.g. low inertia in the islanded system, intermittency of some of the DERs, etc. Particularly during islanding operation, with relatively few DG units, the frequency and voltage control of the islanded system is not straightforward. DG units, specially based on renewable energy sources (RESs), i.e. wind and solar, have an inter-mittent nature and intrinsic characteristics, they can’t ensure the constant power supply required by loads. Furthermore, the DG units with relatively slow response have insuffi-cient dynamic performance in terms of load following. In order to meet the challenges, coordinated control strategies are needed to ensure smooth transition to the islanding operation and reliable operation of the islanded sys-tem. The goal of this Ph.D project is to develop effective frequency control strategies for the islanding operation of ADNs. The developed control strategies are comprised of a primary frequency control scenario with a battery energy storage system (BESS) and two secondary frequency control scenarios with BESS and DG units. During the island-ing transition, the frequency is regulated by the fast-acting primary control of the BESS. The secondary control of the main management system (MMS) detects the status of the BESS and tries to return the power output of the BESS to reference value by assigning the total power difference to the dispatchable DG units. Hence, the dispatchable DG units can be coordinated to share the load following burden of the BESS. To that end, firstly, a reliable real-time model of the Bornholm distribution system is constructed using the real-time digital simulator (RTDS). The resulting model is capable of performing dynamic simulations of the islanded Bornholm distribution system to investigate the frequency regulation performance. In addition, a generic model of Born-holm distribution system is constructed, which can be used as a benchmark model for smart grid testing purposes. In both cases, the simulation results are compared and pro-vided a desirable performance with very high degree of accuracy. Secondly, the simplified battery model is adopted and has been modeled in the RTDS in order to investigate the role of the BESS as a primary frequency regulator during island-ing transition. The effectiveness of proposed primary frequency control strategy is illus-trated by using two test cases (i.e. IEEE 9-bus and Bornholm). In both cases, the fre-quency regulation performance is highly improved without degrading the proposed con-trol performance. Thirdly, a new fuzzy logic based secondary frequency control strategy between a BESS and dispatchable DG units is proposed for further improving the system frequency per-formance as well as reducing output power fluctuations. The simulation results show that the frequency regulation performance is highly improved with fuzzy logic control (FLC) when the system enters into islanding operation. Lastly, an intelligent multi-agent based secondary frequency control strategy for the islanding operation of ADN is proposed. A complete software-in-the-loop (SIL) simula-tion is carried out and optimization of the parameters of the secondary controller is achieved in a simple manner through the effective application of particle swarm optimi-zation (PSO) technique. Simulation results show that the proposed multi-agent based secondary frequency control strategy performs well, in comparison to the performance of proportional integral (PI) control design.