Optimal Hydro-Thermal Generation Scheduling Using an Efficient Feedback Neural Network Optimization Model

This study demonstrates the use of a high-performance feedback neural network optimizer based on a new idea of successive approximation for finding the hourly optimal release schedules of interconnected multi-reservoir power system in such a way to minimize the overall cost of thermal generations spanned over the planning period. The main advantages of the proposed neural network optimizer over the existing neural network optimization models are that no dual variables, penalty parameters or lagrange multipliers are required. This network uses a simple structure with the least number of state variables and has better asymptotic stability. For an arbitrarily chosen initial point, the trajectory of the network converges to an optimal solution of the convex nonlinear programming problem. The proposed optimizer has been tested on a nonlinear practical system consisting of a multi-chain cascade of four linked reservoir type hydro-plants and a number of thermal units represented by a single equivalent thermal power plant and so obtained results have been validated using conventional conjugate gradient method and genetic algorithm based approach.