1. Contribution of Battery Energy Storage System (BESS) to power systems resilience
- Author
-
Liu, Haiyang, Levi, Victor, and Parisio, Alessandra
- Subjects
power system resilience ,optimization ,reliability ,multi-criteria decision-making ,battery energy storage system ,machine learning ,energy storage ,multi-criteria decision analysis - Abstract
Disastrous events, especially under extreme weather conditions, have severe impact on the resilience of electrical power systems. A power system should be reliable under normal circumstances and resilient to extreme events. To improve the resilience of a power system, infrastructural strategies ought to be devised to boost robustness, and operational strategies upgraded to improve the system's flexibility. The battery energy storage system (BESS) is one potential operational strategy to boost the resilience of a power system. This project aims to assess the contribution of BESS to the resilience of power systems under extreme weather events, especially via modelling of BESS utilization. In order to apply BESS analysis, a BESS modelling framework is proposed, taking reliability of battery cells and capacity fading of BESS into consideration. Case studies based on the modelling are performed to identify how BESS can improve reliability and resilience of a power system. Increasing numbers of BESS are planned to be installed in power systems around the world. Therefore, the locations to install BESS has become a prominent challenge to BESS utilization, given that an appropriate location of BESS installation will enable it to support not only the installation site, but surrounding network nodes. This project also develops assessments on location selection of BESS, as based on case studies. The performance of BESS to boost the resilience of a power system appears to vary with its installation locations. Therefore, in order to determine the appropriate locations to install BESS into a network, a 3-step dispatching procedure using an ELECTRE based multi-criteria decision analysis (MCDA) is proposed to effectively identify locations and amount of BESS to be installed into a network. This prediction is based on the capacity of the BESS in question and the features (topology, demand and fragility) of the network. Moreover, optimization of BESS utilization is taken into consideration. A better discharging profile improves the performance of BESS. Thus, this project develops case studies to analyse the optimization of BESS discharging, aiming to improve the resilience of a power system. Then, particle swarm optimization and Q-learning are tested for optimization of BESS control. A specific Q-learning algorithm is proposed to control BESS discharge. The proposed Q-learning optimization procedure is helpful with regards to reliability and resilience improvement of different power systems for BESS control. The final proposed 3-step BESS location selection procedure expands state-of-the-art research by introducing a new BESS dispatching method that provides an ELECTRE MCDA with an objective entropy weighting method. Conditional value at risk of energy not supplied (ENS) is also taken into consideration which enable the system operator to dispatch the BESS into a power system while considering both the amount of BESS and the power loss under extreme situation based on different confident level settings. The proposed Q-learning procedure is proved to be time efficient and able to improve the resilience of different power systems.
- Published
- 2022