Allocation and sizing of reactive power compensators considering PV power and load demand uncertainty using beetle-antenna grey wolf optimization

Abstract

AbstractDistributed photovoltaic (PV) systems play an important role in supplying many recent microgrids. The absence of reactive power support for these small-scale PV plants increases total microgrid losses and voltage-instability threats. Reactive power compensations (RPCs) should be integrated to enhance both microgrid losses and voltage profiles. RPC planning is a non-linear, complicated problem. In this paper, a combined RPC allocation and sizing algorithm is proposed. The RPC-integrating buses are selected using a new adaptive approach of loss sensitivity analysis. In the sizing process, the uncertainties in PV power and load demand are modelled using proper probability density functions. Three simulation techniques for handling uncertainties are compared to define the accurate and fast accurate method as follows: Monte Carlo simulation (MCS), scenario tree construction and reduction method, and point estimation method (PEM). The load flow equations are solved using the forward–backward sweep method. RPCs are optimally sized using the beetle-antenna-based strategy with grey wolf optimization (BGWO) to overcome the local minima problem that appeared in the other pre-proposed methods. Results have been compared using particle swarm optimization and conventional GWO. The proposed model is verified using the IEEE 33 radial bus system. The expected power loss has been reduced by 22% and 31% using compensation of 26% and 44%, respectively. The results obtained prove that the BGWO optimal power flow and PEM to handle the uncertainty can significantly reduce the computation time with sufficient accuracy. Under the study conditions, PEM reduces the computation time to 4 minutes compared with 4 hours for MCS, with only a 3% error compared with MCS as an uncertainty benchmark method.