发变组失步保护原理与人工神经网络相结合的研究

Study on Artificial Neural Network Combined with Out-of-step Protection Principle for Large Generator Units

提出一种基于大型发电机失步保护原理的人工神经网络(ANN)技术,该技术将目前广泛应用的失步保护原理与人工神经网络有机结合起来,并充分利用两者的优点,对提高失步保护的可靠性做了一定的探索。整个判断过程既部分保留了目前广泛应用的阻抗元件原理又引入了ANN失步判断网络,从而提出一套完整的失步保护方案,大量的仿真计算表明了算法的优越性。

The paper put forward a new technique for applying artificial neural network （ANN） based out-of-step protection theory for large generator units. Extensive applied out-of-step protection principle is combined with ANN organically, which exert respective excellence fully .The new method seeks to improve the reliability of the out-of-step. The whole judgment of out-of-step not only reserves the extensive applied resistance principle,but advances the ANN to judge, accordingly put forward a set of integrated protection of out-of-step. The superiority of the technique is proved by lots of simulations.