改进粒子群算法优化支持向量机在故障诊断中的应用研究

支持向量机(SVM)作为当前新型的机器学习方式,凭借解决小样本问题、高维问题和局部极值问题等方面的优越性,在当前故障诊断方面有突出的表现;文章根据对支持向量机的研究,发现其在分类模型参数选择上存在困难,为此,提出利用改进粒子群算法优化的办法,解决粒子群前期收敛速度过快导致后期容易优化不均的现象;通过粒子群算法优化与支持向量机分类模型结合,以轴承故障检测和诊断为例,分析次方法的优越性和提高支持向量机在故障诊断过程中的精准度;通过实际检测得出,这种算法优化的方法改进的支持向量机对于聚类性较差的故障分类具有很好的诊断功能。

Application of Improved Particle Swarm Optimization Support Vector Machine in Fault Diagnosis

Support vector machine (SVM) as the new machine learning method, with the advantages of solving the problem of small sample, high dimension and local extremum problems, with outstanding performance in the current fault diagnosis. According to the research on support vector machine and found that it has difficulty in parameter selection of classification model, this paper proposes an improved particle swarm optimization algorithm to solve the particle swarm pre convergence speed causes the latter easily optimization inequality.The particle swarm optimization algorithm combined with support vector machine classification model used for the detection and diagnosis of bearing fault case analysis method, the superiority and improvement of support vector machine in fault diagnosis process accuracy. Through the actual detection, the improved SVM method has a good diagnosis function for clustering fault classification.