摘要
Abstract
The monitoring and life prediction methods of tunnel electromechanical equipment are studied,and compared with the traditional methods.LSTM-SAE network is used to predict equipment life,and intelligent monitoring work is carried out.First,the characteristic factors of equipment remaining life are determined,the noise is smoothed,and the data is normalized.Then,the equipment life is predicted based on LSTM.SAE sparse encoder is introduced to further improve the accuracy of network prediction,and LSTM-SAE network is established for prediction.A monitoring system with B/S structure is established,Ethernet is used to connect the server and browser,and a control system with status detection module,fault processing statistics module,life analysis and prediction module is established.Finally,the battery life prediction,BP neural network prediction,support vector machine prediction,Bayesian prediction and linear regression prediction are compared,proving that the system has high accuracy,MAE,RMSE and SF values are 8.68,9.23 and 39.12,respectively,with high accuracy.The fusion of LSTM and SAE is realized,which is more accurate than the traditional BP network and can meet the special needs of tunnel electromechanical equipment prediction.关键词
LSTM-SAE/隧道/机电设备/监控管理Key words
LSTM-SAE/tunnels/mechanical and electrical equipment/monitoring and management分类
交通工程
