金属露天矿旋回式破碎机在线故障诊断模型构建与实时分析

矿石破碎是露天矿采矿生产工艺的重要环节,旋回式破碎机作为露天矿进行破碎的大型设备,其故障的诊断通常较为复杂,难以利用主观经验进行精确及快速诊断。针对旋回式破碎设备的故障诊断问题,利用传感器实时采集数据,基于改进后BP神经网络构建故障诊断模型,以偏心套故障、轴承磨损、平行轴缺油、平行轴油温异常作为故障类型,以回流油温、润滑油油压、轴承振动频率、轴承转速作为故障特征参数,利用已知故障类型和故障样本数据对BP神经网络故障诊断模型进行了训练和优化,最后通过测试数据对优化后的旋回式破碎机故障诊断模型进行了验证。结果表明:基于BP神经网络的故障诊断模型能够对旋回式破碎机故障状态实时地做出有效判断,实现以预防为主的故障诊断方式,满足了露天矿大型旋回式破碎机的故障诊断的需求。

On-line Fault Diagnosis Model Construction and Real-time Analysis of Rotary Crusher in Metal Open-pit Mine

Ore crushing is an important link in mining production process of open-pit mines. As the large-scale device for the crushing of open-pit mines,in general,gyratory crusher′s failure diagnosis is relatively complicated and it′s hard to make accurate and rapid diagnosis based on subjective experience. Aiming at the failure diagnosis problem of gyratory crusher,real-time data collection is carried out through sensor and the failure diagnosis model is built up based on the improved BP neural network. Taking eccentric bushing failure,bearing wear,oil storage of parallel axis and abnormal oil temperature of parallel axis as the failure types and taking reflux oil temperature,lubricating oil pressure,bearing vibration frequency and revolving speed of bearing as the failure feature parameters,the training and optimization for the failure diagnosis model based on BP neural network are made by utilizing the known failure types and failure sample data. Finally,the optimized failure diagnosis model of gyratory crusher was verified according to the test data. The result indicated that the failure diagnosis model based on BP neural network could effectively judge the real-time failure state of gyratory crusher and the prevention-oriented failure diagnosis model could be realized,which satisfied the failure diagnosis demand of the large-scale gyratory crusher for open-pit mines.