Ensemble deep learning and enhanced metaheuristic methods for wind power integrated generation scheduling problem in a smart grid framework

Abstract

The major task of traditional electric power system is to optimally schedule the generators to meet out the consumer demands while minimizing the fuel cost despite emission caused. With increase in environment concern and alternative strategies, environmental pollution has to be reduced from electric power plants. Thus economic emission dispatch handles the objective of minimizing fuel cost and emission which are conflicting in nature and are simultaneously optimized subjected to various system constraints. Integration of renewable energy resources such as wind farms with electric power plants has been aiding in meeting out the demand. Since a part of the load is supplied by wind farms the fuel cost has also been reduced and also the emission level of electric power plants is reduced. Due to stochastic nature of wind and uncertainty of wind speed, integrating of renewable energy resources with thermal power plants is a tedious process. Accurate prediction of wind speed is necessary in order to integrate the wind farm output with generating units. Fluctuation in the wind power may influence in the violation of generating units ramping constraints. Thus the problem to incorporate the varying nature of wind power output has to be investigated in scheduling with thermal plants. Minimizing the wind power curtailment for the scheduling period can result in maximum usage of wind power and this is considered as an objective in this research work newline