Dongqu Mine, Shanxi Coking Coal Xishan Coal Electricity

针对东曲矿采煤机电设备运转时间长,工作环境恶劣,导致机器设备故障频发,传统方法通过实时温度监测的数据来反映设备工作状态,但鉴于温度为滞后指标,设备在参数异常前就已经故障,无法达到预防目的,提出使用长短期记忆神经网络(LSTM)模型对温度时间序列进行建模,最终通过实际温度与计算温度的差距来判断设备的工作状态,达到预防设备故障的目的。经实验表明:通过LSTM模型对机电设备进行温度预测的方法是可行的,且准确的预测模型应当选取前24h的温度数据作为自变量,当理论计算值与实际值出现显著的差距时,应当对设备的截割部位、冷却系统以及润滑系统进行及时的监控检查。

Aiming at the long operation time and harsh working environment of coal mining electromechanical equipments in Dongqu Mine, which lead to frequent machine equipment failures, traditional methods reflect the working status of equipment through real-time temperature monitoring data. However, considering that temperature is a lagging indicator, the equipment has already malfunctioned before parameter abnormalities, which cannot achieve the preventive purpose. A long short-term memory neural network model (LSTM) model is proposed and used to model the temperature time series, and ultimately determine the working status of equipment by the difference between actual temperature and calculated temperature, in order to prevent equipment failure. Numerical experiments show that the method of using LSTM model for temperature prediction of electromechanical equipment is feasible, and the accurate prediction model should select the temperature data of the first 24 hours as the independent variable. When there is a significant difference between the theoretical calculation value and the actual value, the cutting part, cooling system, and lubrication system of the equipment should be monitored and inspected in a timely manner.