WNN模型在预测声发射信号方面的应用

小波神经网络具有预测精度高、结构简单以及收敛快等众多优点，因此，试图将这一优势模型用于声发射的预测方面，进而为矿业领域完善一种新的预测方法，并根据实验室岩石加载实验过程中采集的大量声发射数据，建立了一种与之相适应的预测模型。首先，针对实验室实验过程中监测得到的声发射数据建立了小波神经网络模型（WNN模型），然后对声发射监测得到的声发射事件率进行网络自主学习，得到预测结果，最后与实际值相比并计算其误差。结果表明：WNN模型预测精度较高，与实际监测得到的结果基本吻合，证明WNN可以用于声发射信号方面的预测。

Application of the WNN Model to Prediction of Acoustic Emission Signal

Wavelet neural network has advantages of high precision,simple structure and fast convergence,etc. Therefore,attempts with this advantage model is applied to forecast aspect of acoustic emission,and a new predicting method in mining field would be further improved.Based on a huge number of data in the process of coustic emission at laboratory rock loading experiment,a corresponding prediction model can be set up.Firstly,acoustic emission data obtained in the process of monitoring of laboratory experiments can be used to establish Wavelet Neural Network Model,and the acoustic emission monitoring of acoustic emission events rate for network autonomous learning can be further conducted.Finally,the obtained prediction results were compared with the actual value in order to calculate the error.The results demonstrated that the prediction accuracy was higher,and basically comparable with the actual monitoring results,which suggested that the Wavelet Neural Network Model can be employed to predict the acoustic emission signal in future.