For Short-Term Optimum Thermal Planning, a Genetic Algorithm for Modeling Approach and Resolution Method

Abstract

The best weekly plan for power generation in a hydrothermal power plant is found using a genetic algorithm. This multi-reservoir transmitted hydroelectric system has an irregular relationship between power generation, net head, and water discharge rate. It also accounts for the water transport latency between linked reservoirs. A synopsis of the theory underlying the genetic algorithm approach and a discussion of the critical control parameters that impact the algorithm's effectiveness is provided. It is demonstrated that the best hourly packing of the scheme producers may be obtained using a manifold stage genetic algorithm exploration order. The short-term hydrothermal scheduling issue may be solved using the evolutionary algorithm technique, accounting for variations in the clear skull and water conveyance postponement issues. The problem may be solved efficiently under many operating circumstances if appropriate GA control settings are found. Evolutionary algorithms yield a straightforward formulation and solution strategy for the hydrothermal scheduling problem, which may be readily generalized to other complex operation and control issues that electrical companies face. Finally, as this study has shown, a simple modification to the GA, such as switching from one resolution to several resolutions, leads to significant gains in the quality of the solutions. Better solutions may also come from further fine-tuning the GA control parameters, which include population size, crossover, and mutation rates.