Demand‐side management using a distributed initialisation‐free optimisation in a smart grid

Abstract

Due to the integration of the renewable generation and the distributed load that inherently uncertain and unpredictable, developing an efficient distributed management structure of such a complex system remains a challenging issue. Most of the existing works on the demand-side management concentrate on the centralised methods or need a proper initialisation process. This study proposed a demand-side management strategy that can solve the optimisation problem in a distributed manner without initialisation. The objective of the designed demand management system is to maximise the social welfare of a smart grid by controlling the active power economically. The proposed optimisation strategy that generates the optimal power references uses the neighbouring information while considering the local feasible constraints by using a projection operation. Furthermore, the optimisation algorithm is initialisation free, which avoids any initialisation process when plugging-in new customers or plugging-out power units, such as demand loads, battery energy storage systems and distributed generators. The proposed strategy only uses the neighbouring information, so that the proposed approach is scalable and potentially applicable to large-scale smart grids. The effectiveness and scalability of the proposed algorithm are established and verified through case studies.