Two-layer volt/var/total harmonic distortion control in distribution network based on PVs output and load forecast errors
A two-layer control method is proposed for voltage and reactive power control in harmonic polluted distribution network with penetration of photovoltaic (PV) systems. Optimal scheduling of load tap changer and shunt capacitors for minimising energy losses and improving the power quality simultaneously are performed using P-particle swarm optimisation (PSO) optimisation method. Here, the minimising cost of real power losses and improving the power quality criteria have been pursued as the goals of an optimisation problem. Considering load and PVs output forecast, the first-layer control determines the optimal reactive power and control settings for all mechanical controllers. Hourly errors of load and power forecasts and mechanical control setting of the first layer are used to estimate optimised reactive power of PV in order to achieve maximum voltage regulation, reduce network losses and total harmonic distortion (THD). These data are trained a neural network (NN) to estimate optimised PV reactive power. This NN in the second layer is used to optimise the online reactive power setting based on online PV power. For more practical applications of the proposed method, simulation is carried out in a large distorted 37-bus distribution system. The algorithm will increase use of renewable energies, reduce voltage fluctuations, THD, and wear and tear of mechanical equipment.