基于CAS优化的LSTM循环神经网络的高压断路器故障预测方法研究

Fault Prediction Research of High Voltage Circuit Breakers Based on LSTM Recurrent Neural Network Optimized by CAS

在研究真空高压断路器运行历史数据的基础上,提出一种基于长短期记忆(LSTM)循环神经网络的断路器故障时间序列预测方法,采用混沌蚁群(chaotic ant swarm,CAS)优化算法训练LSTM模型,并在TensorFlow深度学习框架上搭建模型仿真,与其他常用的优化训练方法相比,基于CAS优化的LSTM模型具有更高的预测精度和更短的训练步数,且模型简单容易训练。该故障预测方法在基于时间序列的设备故障预测方面有较高的应用价值。

On the basis of studying the operation history data of vacuum high voltage circuit breaker,this paper proposes a fault time series prediction method based on Long Short-Term Memory( LSTM)Recurrent neural network. The LSTM model is trained by Chaotic Ant Swarm( CAS) optimization algorithm,and the model is built on the TensorFlow deep learning framework. Compared with other commonly used optimization training methods,LSTM model based on CAS optimization has higher prediction accuracy and shorter training steps,and the model is simple and easy to train. This method is an effective fault prediction method,and has high application value in equipment fault prediction based on time series.