Electric Load Forecast And Economic Load Dispatch Based On Neural Networks

Load forecasting is one important work of the power management system, it impacts directly on the power dispatch and safety, and provide the basis for power system generation, maintenance plans, and price quotes. Accuracy of the forecasting result is closely linked with the economic benefits of the power sectors. As the supply and demand of Chinese power market increase, social productive forces develop rapidly, and information develop acceleratly, the demand for accuracy of the current load forecast is rising, and therefore, it requires a reasonable and efficient prediction model to achieve the problem.The relationship between the load of power system and other influencing factors is strong coupling, multivariable, very non-linear, and dynamic. Traditional prediction cannot realize high accuracy, while dynamic recurrent neural network(Elman) can reflect the dynamic nature of the system in a more direct and effective way. A prediction model of power load was established based on EKF-Elman(a novel Elman neural network trained by extended Kalman filter) in this paper. By cases of prediction, the modeling effect of EKF-Elman network and Elman network and EKF-RBF network structure was compared. Simulation experiment shoes that EKF-Elman network has features such as dynamic, fast in network training and high accuracy, which proves that EKF-Elman prediction model is a fresh and reliable way of power load prediction.The economic load dispatch(ELD) problem is one of the important optimization problems in a power system. Traditionally, in the ELD problem, the cost function for each generator has been approximately represented by a single quadratic function. It is more realistic, however, to represent the generation cost function for fossil fired plants as a segmented piecewise quadratic function, as in the case of valve point loading. As fossil fuel costs increase, it becomes even more important to have a good model for the production cost of each generator.Hysteretic noisy chaotic neural network(HNCNN) has been proven to be a powerful tool in solving combinatorial optimization problems, which can increase the effective convergence toward optimal or near-optimal solutions by using both stochastic chaotic simulated annealing(SCSA) and hysteretic dynamics. Considering the excellent optimization performance of HNCNN, we apply HNCNN to better resolve economic load dispatch(ELD) of power system in this paper. In addition, the system loss and valve point effect are also involved in the simulation. Simulation results and analyses compared with other approaches are presented to illustrate efficiency of the HNCNN.