一种模糊神经网络的采煤机滚筒温度实时故障预警方法

A real-time fault early warning method of shearer drum temperature based on fuzzy neural network

针对综采工作面采煤机设备割煤过程中滚筒温度过高导致的电气故障的问题,基于模糊神经网络设计了一种故障预警方法。模糊逻辑能对故障症状与故障原因之间的模糊关系进行准确描述,对模糊性信息具有较强的表达能力,而神经网络则具有强大的学习和获取知识的能力,能够降低模糊逻辑依赖模糊规则的局限性,结合两者的优点,设计出了一种模糊逻辑的概率神经网络。为了提高该算法模块的实用性和准确性,通过对采煤机的滚筒温度进行采集与模糊化处理,然后利用.NET提供的机器学习库进行模型的比较和选择,训练出最优模型,实时的预测出设备状态。结果表明,通过该方法进行实时地监测滚筒温度能够提前诊断出采煤机即将发生跳电故障,因此,设计的故障预警方法可以应用于采煤机设备的故障预警分析与决策,降低综采设备的维护成本。

Aiming at the electrical fault caused by the excessive temperature of the drum during the coal cutting process of the shearer equipment in a fully-mechanized mining face,a failure early warning method based on fuzzy neural network was designed.Fuzzy logic can describe the fuzzy relationship between failure symptoms and reasons accurately,and it has strong ability to express fuzzy information while neural networks have strong learning and knowledge acquisition capabilities,which can reduce the dependence of fuzzy logic on fuzzy rules.Combining the strengths of both,a fuzzy logical probabilistic neural network was designed.In order to improve the practicability and accuracy of the algorithm,the temperature of the shearer drum was collected and fuzzified,and then the machine learning library provided by.NET was used for model comparison and selection.In the end the optimal model was acquired and the real-time state was predicted.The results show thatthe real-time monitoring of drum temperature by this method can diagnose the impending power trip failure of the shearer in advance.Therefore,the designed fault early warning method can be applied to the shearer system fault warning analysis and decision-making,and reduce the maintenance cost of fully-mechanized mining equipment.