【摘要】 针对运行状态中的电梯制动器进行物理参数监测和故障预警,设计了制动器试验平台模拟制动器真实工作状态,结合传感器监测装置,对制动器全生命周期过程中制动器间隙、摩擦噪声、制动器电压、制动器电流、线圈温度以及微动开关等关键物理参数进行采集。基于长短期记忆网络自编码器对关键参数数据进行重构,生成工作状态时间序列,结合时间序列回归模型对制动器剩余生命周期进行预测。实验结果表明,监测系统可以采集到制动器失效过程中的关键物理参数,结合采集到的数据进行故障预警,得到剩余生命周期预测误差低于5.5%,证明了该方法可以实现制动器运行状态智能监测并进行故障预警。更多还原

【Abstract】 In order to monitor the physical parameters of the elevator brakes in the running state and provide early warning of faults,this article uses an independently designed brake test platform to simulate the true working state of the brakes. Sensor monitoring devices was used to collect key physical parameters such as brake clearance,friction noise,brake voltage,brake current,coil temperature,and micro switch during the entire life cycle of the brake. Based on the long short-term memory network encoder-decoder( LSTM-ED),key parameter data was reconstructed to generate a time series of working states,and the remaining useful life( RUL) of the brake was predicted in conjunction with the time series regression model. The experimental results show that the platform can collect key physical parameters in the brake failure process in a short time.Combined the collected data for fault warning,the RUL prediction error is less than 5.5% which proves that this method can realize intelligent monitoring of brake operating status and realize the fault early warning.更多还原