Método de estimativa pontual e algoritmo de Busca Cuco discreto adaptável para instalação ótima de geração fotovoltaica.

Abstract

A method for optimal location and sizing of photovoltaic generators in radial distribution systems, based on the 2m+1 point estimation method (PEM), where m represents the number of input random variables, and in the Adaptive Discrete Cuckoo Search (ADCS) is presented. The ultimate objective is to minimize the net present value cost, including annual energy losses. For that, the randomness of the generation and load demand is considered. To calculate the power flow, the backward-forward sweep method is used. For simplicity, the load is considered a Gaussian random variable. An improvement in the way of calculating energy losses by combining the PEM with Sobol sequence sampling is proposed. Therefore, Monte Carlo simulation is used as a comparison between the traditional point estimate method and the proposed method. The correlation found to exist between solar irradiance and ambient temperature is considered in the power flow calculations. Copula theory is used to incorporate the correlation during the performed Monte Carlo simulations. The proposed optimization algorithm, ADCS, is compared with the Cuckoo Search (CS), genetic algorithm (GA), and particle swarm optimization (PSO). From the initial tests, in which only one generator was inserted in the system, it was found that considering the correlation between irradiance and ambient temperature is irrelevant in the Monte Carlo simulations, however, it is important in the simulations using PEM. For validation, the proposed method is applied to the IEEE 69-bus test system. The feasibility of installing generators was verified, both by reducing total costs and the Levelized cost of energy.