Study Of Load-shedding Based On Neural Network For Power System Stability

With the development of power system, the important and difficulties of control in safety are enhanced in power system, especially about multi-machine system control problem.With the development of computer and communication technology, especially owing to the appearance of the artificial intelligent technology, they provide a new method to study the electrical power system control in safety and stability.This paper studies neural network emergent load-shedding and real-time calculation method for the electrical system stability controlled load-shedding. The paper applied neural network to the electrical system load-shedding stability control and approves its reasonableness. The paper also suggests that we should substitute faulty power fluctuated data for neural network character input to express fault models, and which can reduce input numbers and make it easy to implement. For GuanZhong power system emergent load-shedding control problem, i. e, the shedding out the both AnNan double lines in Northwest power system. Load-shedding simulation calculation is conducted by PSASP first, and obtained datum were used in neural network training. Then the load-shedding calculation method, which combine with the neural network and controlled strategy table, is produced. The simulation calculation results show that proposed method having a high calculation precision and computational efficiency, and which can meet the demand of stability control.