Effects of the Firefly Optimization Algorithm Hyperparameters on the Optimal Placement Problem Results of Renewables-Based Power Plants

Integration of new capacities of renewable energy sources into existing electrical energy systems is one of the most relevant problems in the development of the power industry for most countries. The geographical location for construction should be accurately selected to achieve the high performance and efficiency of that generation type. The problem of proper placement of renewable energy sources generation could be solved with expert knowledge or with the help of mathematical optimization algorithms. This article is devoted to the consideration of renewable energy sources' optimal placement problem-solving with the use of the metaheuristic optimization algorithm named as the firefly algorithm. Values of the capacity factor for the renewable energy power station were used as the objective function to evaluate the proposed algorithm. Capacity factor values were calculated using a machine learning regression model based on the random forest algorithm. The proposed algorithm was tested in the case of placement optimization of a photovoltaic power plant with an installed capacity of 15 MW in Belarus. The article presents the influence of hyperparameters on the optimization results of the algorithm. Results are shown in the article in the form of boxplot diagrams of optimal capacity factor values, which were found while five hyperparameters were changed separately.