A Predictive Rolling Optimization Method of Load Balancing in Low-voltage Active Distribution Network

As the penetration of distributed generators connected to the low-voltage distribution network (LVDN) gradually increases, the three-phase imbalance problem becomes more and more serious. The traditional governance model suffers from the lack of timely calculation results and incompatibility with distributed generators (DG) connected to the network. Therefore, this paper proposes a novel three-phase unbalance governance method that combines load connection phase pre-decision and rolling optimization. First, the pre-decision module uses a wavelet denoising regression tree for short-term load forecasting based on historical data and a Chaos Particle Swarm Optimization (CPSO) for feeder current balance optimization. Secondly, the rolling optimization module uses a hierarchical optimization method to realize real-time fine optimization based on real-time data. The simulation results in the actual LVDN show that the proposed method can not only obtain a better balancing effect in the LVDN with distributed generator access but also the independent work of the pre-decision module and the rolling optimization module can still have the three-phase unbalance suppression capability in the absence of historical data or real-time data.