Improved Genetic Algorithm-Based Unit Commitment Considering Uncertainty Integration Method

Abstract

In light of the dissemination of renewable energy connected to the power grid, it has become necessary to consider the uncertainty in the generation of renewable energy as a unit commitment (UC) problem. A methodology for solving the UC problem is presented by considering various uncertainties, which are assumed to have a normal distribution, by using a Monte Carlo simulation. Based on the constructed scenarios for load, wind, solar, and generator outages, a combination of scenarios is found that meets the reserve requirement to secure the power balance of the power grid. In those scenarios, the uncertainty integration method (UIM) identifies the best combination by minimizing the additional reserve requirements caused by the uncertainty of power sources. An integration process for uncertainties is formulated for stochastic unit commitment (SUC) problems and optimized by the improved genetic algorithm (IGA). The IGA is composed of five procedures and finds the optimal combination of unit status at the scheduled time, based on the determined source data. According to the number of unit systems, the IGA demonstrates better performance than the other optimization methods by applying reserve repairing and an approximation process. To account for the result of the proposed method, various UC strategies are tested with a modified 24-h UC test system and compared.