摘要
在分类诊断煤矿用破碎装载卧底机运行故障时,若以大量原始故障特征为依据易导致检测结果的均方误差较大。为此设计了新的运行故障检测方法。在采集卧底机运行输出信号后,利用测量矩阵明确采集信号所包含信息。然后根据小波变换原理建立信号特征提取机制,对原始故障特征进行多层处理,得到时域特征和频域特征相结合的特征向量。采用动态模糊函数建立自学习模式,根据节点能量实时更新有效特征数据。最后通过构建以BP神经网络为核心的故障检测模型快速诊断、分类运行故障。实验结果表明,与灰色关联检测法、粗糙集检测法相比,该方法故障检测结果的均方误差降低了53.85%、65.38%。
If a large number of original fault features are taken as the basis for the classification and diagnosis of crushing and loading undercover machine operation faults in coal mine, the mean square error of detection results will be large. Therefore, a new operation fault detection method was designed.After the output signal of the undercover machine was collected, the information contained in the signal was defined by the measurement matrix. Then a signal feature extraction mechanism was established based on the principle of wavelet transform, the original fault features were processed by multiple layers, and the feature vectors of time domain feature and frequency domain feature were obtained. A self-learning model was established by using dynamic fuzzy function, and effective feature data were updated in real time according to node energy. Finally a fault detection model based on BP neural network was constructed to diagnose and classify operating faults quickly. Experimental results show that compared with grey correlation detection method and rough set detection method, the mean square error of fault detection results of this method is reduced by 53.85% and 65.38%.
作者
鞠晨
Ju Chen(Shendong Coal Group,CHN Energy Group,Yulin 719315,China)
出处
《煤矿机械》
2022年第11期179-183,共5页
Coal Mine Machinery
关键词
煤矿用破碎装载卧底机
故障检测
信号特征
动态模糊
小波神经网络
推理模型
crushing and loading undercover machine for coal mine
fault detection
signal characteristic
dynamic fuzzy
wavelet neural network
reasoning model
作者简介
鞠晨(1988-),陕西韩城人,工程师,本科,研究方向:机电信息。
