基于高阶谱和支持向量机的溢流阀的故障诊断

摘要

By combining high order spectrum with support vector machine,the information of different state of the relief valve was collected and analyzed. By processing the power spectrum under different state as a support vector machine input,in the support vector machine (SVM)training,a vector parameter of different characteristics and fault type of mapping relations were established,so as to achieve the purpose of the fault diagnosis. The experimental results show that this method,which has high fault recognition abili-ty,is simple and effective.%　　将高阶谱和支持向量机相结合，提取不同工作状态下溢流阀的信息并进行分析，将不同状态下的功率谱经过处理作为支持向量机的输入，在SVM 中训练，建立不同特征的参数向量和故障类型的映射关系，从而达到故障诊断的目的。实验结果表明：该方法简单有效，并具有较高的故障识别能力。