基于改进多输出支持向量机的声发射源定位研究

为了提高根据声发射(AE)现象预报煤与瓦斯突出位置的精度,结合核主成分分析(KPCA),提出了一种改进的多输出最小二乘支持向量机(LSSVM)的目标定位方法.对于采集的声发射参数信号,采用核主成分分析提取重要定位特征;采用多输出最小二乘支持向量机建立定位模型,采用文化基因算法对多输出最小二乘支持向量机参数进行优化.试验测试定位性能,结果表明:算法提高了试验平台声发射定位的精度且定位时间少于其他定位算法,具有很高的实际应用价值.

Research on acoustic emission source localization based on improved multi-output SVM

In order to improve precision of predicting position of coal and gas outburst based on acoustic emission (AE)phenomena,combined with kernel principal component analysis(KPCA),a target positioning method based on improved multi-output least squares (LS)support vector machine (SVM),KPCA-LSSVM,is proposed. For collected acoustic emission parameter signals,KPCA is used to extract important localization features. Multi-output least squares support vector machine is used to establish positioning model. Memetic algorithm is used to optimize the parameters of the multi-output LSSVM. Simulation experiment is used to test positioning performance. The results show that the algorithm improves precision of acoustic emission localization of test platform and positioning time is less than other positioning algorithms,it has high practical application value.