A Non-convex Dispatch Problem with Generator Constraints Using Neural Network and Particle Swarm Optimization

Abstract

In this paper, we consider an economic emission dispatch problem consisting of the thermal generators with valve point effect and the wind turbines. Due to the uncertainty of wind power generation, the cost of reserve capacity penalty and the cost of wind abandoning are introduced into the wind generation model. Moreover, the aforementioned problem is subject to the capacity constraint and supply–demand balance constraint which include the power loss of thermal power generation and wind power generation. In modeling the wind generation, the wavelet neural network is introduced to predict the wind power in order to get more accurate power data in a day. Next, the penalty function method is used to deal with the supply–demand balance constraint; then, the local optima is found using projection neural network, and particle swarm optimization is utilized to update initial points to obtain the global optima. Finally, the power distribution of ten units with different load demands within 24 h is described.