Learning Strategies for Volt-VAR Optimization with PV Inverters and Battery Energy Storage Systems in Distribution Network

As the integration of PV inverters and battery energy storage systems (BESSs) gradually increases in the distribution network, the rapid fluctuation and random nature of these distributed generators (DGs) put forwards an urgent demand for real-time Volt-VAR optimization (VVO). To address this issue, this paper proposes a VVO strategy learning method via machine learning framework. First, the VVO problem with PV and BESS is formulated as a mixed integer programming (MIP). Second, a strategy learning framework is proposed, which maps from uncertain parameters to the optimal value of integer variables and tight constraints. The solution finding of the MIP can be accelerated by applying the strategy. Then, a decision tree model is created to tackle the classification task. Finally, case studies are conducted on the modified IEEE 33-bus system, which demonstrates the feasibility and reliability of the proposed method. The result shows that the proposed method can reach near-optimal performance compared with commercial solvers.