基于BP和Elman神经网络的智能变电站录波启动判据算法

针对传统故障录波启动判据算法的局限性,提出一种基于BP神经网络和Elman神经网络的算法。以A、B两相电流越限为例进行了算法的研究,通过选取启动判据样本来训练BP和Elman神经网络,将启动判据信息输入到训练好的两种模型中,由输出结果就可以判断是否需要启动录波。Matlab输出表明:基于BP神经网络的故障录波启动判据算法能有效地完成录波启动,误差较小,但是速度相对较慢;而基于Elman神经网络的故障录波启动判据算法也可以完成录波启动,但是误差稍大,由于带有反馈环节,所以速度较平稳,易于工程实现。较之两种算法,可针对故障录波数据量的大小进行择优选择。

Criteria algorithm for smart substation recorder starting based on BP & Elman neural network

As to the limitation of traditional starting criteria for fault recorder algorithm, this paper proposes an algorithm based on BP neural network and Elman neural network. An example of phase A and phase B current out-of-limit is studied with the algorithm. By choosing starting criteria samples to train BP and Elman neural network, then inputting the starting criteria information to the two trained models, whether to start recording can be judged from the output results. The outcome of MATLAB simulation shows that the starting criteria for fault recorder algorithm based on BP neural network can effectively complete the recording start with minor error, but the pace is comparatively slower. The starting criteria for fault recorder algorithm based on Elman neural network can also complete the recording start, but the error is bigger. Thanks to the part of feedback, the pace is smooth and steady and easy to accomplish in engineering project. Comparing two algorithms, the suitable one can be selected according to the amount of recorded data.