Learning Automata-Based Methodology for Optimal Allocation of Renewable Distributed Generation Considering Network Reconfiguration

The inadequate capacity of distribution networks to consume renewable energy and the inappropriate allocation of renewable distributed generation (RDG) have become important issues. In this paper, a 3-level learning automata-based methodology in a master-slave structure is proposed for optimal RDG siting and sizing considering network reconfiguration. The RDG allocation optimization, i.e., the master problem, is proposed in the first level, with the objective of minimizing the annual investment cost and operation cost. Network reconfiguration is modeled as a slave problem in the second level to promote the consumption of RDG and decrease the operation cost. The RDG power control strategy, including active power curtailment and reactive power compensation, is introduced as a secondary slave problem in the third level. Considering the stochastic characteristics of renewable energy and loads, intelligent algorithms based on learning automata are proposed and embedded into the master-slave structure. The simulation results on the standard test systems demonstrate the feasibility and effectiveness of proposed method.