论文部分内容阅读
为了提高煤矿主扇风机故障诊断的准确性,将网格搜索法和支持向量机(SVM)应用到主扇风机的故障诊断中。首先,建立主扇风机运行故障的知识库,并将采集到的主扇风机振动信号进行小波消澡和归一化;然后,设计了网格搜索参数优化SVM的主扇风机故障诊断模型。最后,通过工程现场提取的数据进行实验验证,并与遗传算法和粒子群算法寻优的时间和诊断结果准确率进行比较。实验结果表明,网格搜索法SVM参数优化非常适合于煤矿主扇风机的故障系统中。
In order to improve the accuracy of fault diagnosis of main fan of coal mine, grid search method and support vector machine (SVM) are applied to fault diagnosis of main fan. Firstly, the knowledge base of the main fan running fault is established, and the main fan vibration signals collected are eliminated and normalized by wavelet. Secondly, the fault diagnosis model of main fan with grid search parameter optimization SVM is designed. Finally, the experimental data were extracted from the project site and compared with the accuracy of genetic algorithm and particle swarm optimization algorithm in time and diagnosis. The experimental results show that the SVM parameters optimization of grid search method is very suitable for the fault system of main fan of coal mine.