A comparative analysis of the performance of multiple meta-heuristic algorithms in sizing hybrid energy systems connected to an unreliable grid

Abstract

The availability of affordable and reliable power supply fosters social and economic growth and raises the standard of living. In most developing nations, there is a considerable gap between energy supply and demand, often resulting in load shedding and blackouts. Integrating two or more renewable power sources is a potential solution for the inconsistent nature of renewable energy, thereby supplying clean and sustainable electricity. However, proper component sizing and operation planning for different system components are necessary for a reliable and cost-effective system. This paper compares the performance of three widely used optimisation techniques (Artificial Bee Colony (ABC), Genetic Algorithm (GA), and Particle Swarm Optimisation (PSO)) in determining the size of a hybrid renewable energy system (HRES) with the lowest levelised cost of energy (LCOE) to meet the energy needs of a dairy farm in a rural settlement. PSO is observed to be the best-performed algorithm proposing a system with an LCOE of $0.162 per kWh, a net present cost (NPC) of 2.05 million dollars and a payback period of 5 years and 7 months when compared with the existing power system. The proposed HRES is determined to reduce annual diesel usage by 96%. Therefore, significantly decreasing greenhouse gas (GHG) emissions. The PSO algorithm performs satisfactorily in terms of results and convergence time compared to the results from commercially available hybrid optimisation software (HOMER Pro).

Keywords

• Hybrid renewable energy system
• HOMER Pro
• Meta-heuristic algorithms
• Unreliable grid
• Biogas generator
• Net present cost