基于SOM神经网络的煤矿用防爆柴油机故障诊断

Fault Diagnosis of Mine Explosion-proof Diesel Engine Based on SOM Neural Network

以某矿用防爆柴油机为研究对象,采集其在实际工况下的实时数据。基于SOM神经网络建立矿用防爆柴油机故障诊断模型,运用MATLAB专业模块对该模型进行数值仿真。仿真结果表明:矿用防爆柴油机存在发动机冒烟、发动机异响、发动机冬季启动困难、曲轴不转动等故障,且不同故障在不同因素下呈现一定的规律性及聚集性。该研究为矿用防爆柴油机故障诊断、故障处理时间缩短等方面提供理论依据及技术支持。

A mine explosion-proof diesel engine has been taken as the research object, the real-time data under actual working conditions have been collected. The fault diagnosis model of mine explosionproof diesel engine has been established based on SOM neural network, and the model has been numerically simulated by using MATLAB professional module. The simulation results show that the mine explosion-proof diesel engine has faults such as engine smoke, engine abnormal noise, difficulty in starting the engine in winter and crankshaft non-rotation, and the different fault shows certain regularity and clustering under different factors. The research provides theoretical basis and technical support for fault diagnosis and shortening of fault processing time of mine explosion-proof diesel engine.