论文部分内容阅读
在机械故障诊断中,通常不具备有大量的故障样本,因此,制约了故障诊断技术向智能化方向发展。而基于统计学习理论(SLT)的支持向量机(SVM)方法正好克服了这方面的不足。统计学习理论是专门研究少样本情况下的统计规律及学习方法的理论。SLT理论和SVM方法为故障诊断技术向智能化发展提供了新的途径。该文讨论了支持向量机在故障诊断领域中应用的分类算法。并以滚动轴承的振动信号为例进行了试验论证。试验表明:SVM方法对具有少样本的故障诊断领域具有很强的适应性。
In the mechanical fault diagnosis, usually do not have a large number of fault samples, thus restricting the fault diagnosis technology to the intelligent direction. Support Vector Machine (SVM) based on statistical learning theory (SLT) just overcomes this problem. Statistical learning theory is a case study of statistical law and learning methods in the case of few samples. The SLT theory and the SVM method provide a new way for the intelligent development of the fault diagnosis technology. This paper discusses the classification algorithm used by SVM in fault diagnosis. The vibration signal of rolling bearing is taken as an example to demonstrate the experiment. Experiments show that SVM has strong adaptability to the field of fault diagnosis with few samples.