Multi-Timescale Model Predictive Control of Battery Energy Storage System Using Conic Relaxation in Smart Distribution Grids

This paper proposes a multi-timescale volt/var optimization for the optimal dispatch of battery energy storage system in smart distribution grids. It aims to coordinate the substation on-load tap changer operation on slow-timescale (hourly basis) with the photovoltaic inverters and battery storage operations on fast-timescale (15 min basis). This coordination is achieved by using two-stage stochastic programming and implemented via model predictive control. The power loss and energy purchase cost are reduced while maintaining voltages within limits. The forecasting uncertainties are modeled by generating a large number of random scenarios and then subsequently reducing scenario numbers to establish a tradeoff between computational burden and accuracy of the solution. The mixed-integer second-order cone program (MISOCP) is formulated with reduced scenarios, which achieves global optimum. Simulation results demonstrate the effectiveness of proposed MISOCP model in keeping the voltages within limits under forecasting uncertainties.