摘要
针对发动机的故障分类问题,提出了一种基于粒子群(PSO)优化支持向量机(SVM)的发动机故障诊断方法,采用粒子群算法优化支持向量机的惩罚系数C和核宽度系数σ,并在MATLAB环境下对发动机进行故障类别诊断,通过对发动机典型故障的诊断研究表明,采用PSO-SVM算法模型的故障诊断的精确度和效率都得到了提高,该方法与BP神经网络、PSO-BP、标准SVM相比,有较高的分类准确率,准确率可高达100%;与GA-SVM方法相比,诊断效率有所提高,从而验证了该方法在发动机故障诊断中的有效性。
Concerned with the problem of engine fault classification, this paper proposes an algorithm based on particle swarm optimiza tion--support vector machine （PSO--SVM） for engine fault diagnosis. The method uses the MATLAB for engine fault diagnosis and optimi zes the penalty factor C and kernel width coefficient a of the support vector machine （SVM） by the PSO algorithm. A typical engine fault di- agnosis shows that the proposed algorithm can improve the fault diagnosis accuracy and efficiency. Compared with BP neural network, PSO--BP, the standard SVM, the proposed method has a high classification accuracy which can reach 100%. Compared with GA--SVM, the diagnosis efficiency of the proposed method has greatly improved. Thus verify the effectiveness of the method to engine fault diagnosis.
出处
《计算机测量与控制》
北大核心
2014年第2期355-357,360,共4页
Computer Measurement &Control
基金
重庆市教委科学技术研究项目(KJ120511)
关键词
粒子群优化算法
支持向量机
发动机
故障诊断
particle swarm optimization
support vector machine~ engine
fault diagnosis
