基于振动和声发射信号的截齿磨损程度识别研究

Study on Wear Degree Recognition of Picks Based on Vibration and Acoustic Emission Signals

为实现采煤机截齿截割过程中磨损程度的在线监测和识别，提出一种基于RBF（RadicalBasisFunction）神经网络的截齿磨损程度多特征信号监测方法。提取截割过程中不同磨损程度截齿的振动和声发射特征信号，分别分析振动和声发射信号的峰值、时域图方差和频域图均方根值这6个特征参数，获取振动信号、声发射信号与不同磨损程度截齿的变化规律。建立5种不同磨损程度截齿的多特征信号样本数据库，采用多特征信号样本对RBF神经网络进行学习和训练，建立截齿磨损程度的识别模型，实现截齿磨损程度在线监测与精确识别。实验结果表明：随着截齿磨损程度的加剧，截齿振动和声发射时域信号中信号峰值和方差均呈增大的趋势；振动和声发射的频域信号中频谱图均方根也呈现逐渐增大的趋势；基于RBF神经网络的截齿磨损程度监测系统的网络判别结果和测试样本的实际磨损程度类别相符，该RBF神经网络系统能够对截齿磨损程度类型进行准确的监测和识别。

In order to realize on-line monitoring and identification of pick wear degree in the shearer cutting process , a multi-feature signals detection pick wear method based on RBF （ Radial Basis Function） neural network was proposed. The vibration and acoustic emission characteristic signals of different picks wear degree were extracted in the cutting process, the six characteristic parameters of vibration and acoustic emission signal＇ s the peak value, variance of the time domain and mean square value of the spectrum were analyzed respectively. The variation law of the vibration acceleration signal, acoustic emission signal and different pick wear degrees were obtained. A multi-feature signal sample database with five different pick wear degrees were established. The multi-feature signal samples were used to study and train the RBF neural network, the recognition model of pick wear degree was established to realize the on-line monitoring and accurate identification of pick wear degree. The experimental results show that with the increase of pick wear degree, the peak value and variance of the signal are increasingas well. The frequency spectrum of vibration and acoustic emission＇ s root mean square value also shows an increasing trend. The monitoring system of pick wear degree based on RBF neural network, the results of network are consistent with the actual wear degree of the test samples, theRBF neural network system can accurately monitor and identify the types of pick wear degree.