誤差逆伝搬法による電力系統安定化装置のパラメータ設計

抄録

Case studies of artificial intelligence (AI) technologies have been reported. Applications of Deep Learning as known as one of Machine Learning technologies have expanded in some fields such an engineering, medical, and game. Back Propagation (BP) method is a basic technology of Deep Learning, is also one of optimization approaches. This paper has tried to applying the AI technology to the system control field. In enhancement of transient stability in a power system, a power system stabilizer (PSS) equipped in a generator with an automatic voltage regulator (AVR) has been useful. The PSS has some Lead Lag blocks, and has been designed by tuning of time constants in the blocks. These time constants have been determined with trial and error in a damping and synchronous torque diagrams of frequency domain by design engineers. In this paper, Back Propagation method has been applied to design of PSS. The BP method determines time constants in a Lead Lag block in PSS. This design approach realizes a controller design graphically without trial and error.