A noise annealing neural network for hydroelectric generation scheduling with pumped-storage units

A new approach based on neural networks is proposed for the hydroelectric generation scheduling with pumped-storage units in the Taiwan power system. The purpose of hydroelectric generation scheduling is to determine the optimal amounts of generated powers for the hydro units in the system. To achieve an economical dispatching schedule for the hydro units including two large pumped-storage plants, a neural network is employed to reach a schedule in which total fuel cost of the thermal units over the study period is minimized. The neural network model presented can solve nonlinear constrained optimization problems with continuous decision variables. Incorporating the noise annealing concepts, the model is able to produce such a solution which is the global optimum of the original problem with probability close to 1. The proposed approach is applied to hydroelectric generation scheduling of Taiwan power system. It is concluded from the results that the proposed approach is very effective in reaching proper hydro generation schedules.