Multi-agent-based approach for generation expansion planning in isolated micro-grid with renewable energy sources and battery storage

Generation expansion planning (GEP) is a widely studied problem in the literature. However, with increasing participation of renewable energy sources (RES), the problem has to be looked upon considering intermittency associated with these sources. Integration of storage devices is seen as a viable option for counteracting the intermittency, particularly in isolated micro-grid (IMG). This paper puts forth a generalized mathematical structure for GEP centred on techno-economic criteria which duly addresses uncertainties associated with RES and correspondingly proposes a plan for addition of RES and storage units. As GEP problem is a highly constrained, large-scale problem involving assessment of enormous combinations, a decentralized multi-agent system (MAS) is used along with butterfly particle swarm optimization (BFPSO) technique for parallel processing. Hybridization of features of MAS with BFPSO called as multi-agent-based butterfly PSO algorithm (MABFPSO) presents a time effective and more accurate alternative to existing optimization techniques for addressing GEP problem. The formulation has been applied on IMG comprising of RES (solar and wind) and battery storage system. Results obtained from MABFPSO are compared with PSO with the purpose of establishing superiority of MABFPSO. It has been found that MABFPSO provides a cost benefit of 3.5% in comparison with classic PSO. For the same reliability standards, the capacity addition required for solar and wind generators is similar for both the techniques. However, the storage capacity requirement with PSO is 100% higher in comparison with that of MABFPSO. The comparison of convergence characteristics of two techniques clearly demonstrates that on the iteration scale, MABFPSO presents 29.6% faster convergence in comparison with that of PSO.