基于改进粒子群与人工鱼群混合SVM故障诊断方法研究

Research on Fault Diagnosis Method Based on Improved Particle Swarm Optimization and Artificial Fish Swarm Hybrid SVM

针对传统支持向量机(Support Vector Machine,SVM)在工业过程故障诊断应用中存在参数优化准确率低、识别率低等问题,提出一种基于改进粒子群(particle swarm optimization,PSO)与人工鱼群(attificial fish swarms algorithm,AFSA)混合SVM智能算法进行参数优化,以提高SVM故障诊断的效果。由于标准PSO算法易陷入局部最优,基本AFSA算法后期收敛速度慢,这都会导致优化模型收敛速度以及收敛精度的下降,影响最终结果的准确性。首先利用改进后的惯性权重以及学习因子对PSO进行优化,并将改进后的PSO算法与AFSA算法相结合,综合利用PSO算法的局部收敛性以及AFSA算法的全局收敛性,提高混合算法的收敛速度以及收敛精度;其次将SVM惩罚因子和核参数作为共同优化对象,获得最优参数;最后,通过仿真模拟实验验证,结果表明这种混合算法对于SVM中的参数优化收敛速度更快、精度更强,从而使得过程故障诊断的效果更佳。

Aiming at the problems of low accuracy and low recognition rate of parameter optimization in the application of tradi-tional support vector machine(SVM)in industrial process fault diagnosis,this paper proposes a hybrid SVM intelligent algorithm based on Improved Particle Swarm Optimization(PSO)and artificial fish swarm algorithm(AFSA)to optimize the parameters,so as to improve the effect of SVM fault diagnosis.Because the standard PSO algorithm is easy to fall into local optimization and the later convergence speed of the basic AFSA algorithm is slow,which will lead to the decline of the convergence speed and convergence accu-racy of the optimization model and affect the accuracy of the final result.Firstly,PSO is optimized by using the improved inertia weight and learning factor,and the improved PSO algorithm is combined with AFSA algorithm to improve the convergence speed and conver-gence accuracy of the hybrid algorithm by comprehensively using the local convergence of PSO algorithm and the global convergence of AFSA algorithm;Secondly,the SVM penalty factor and kernel parameters are taken as the common optimization object to obtain the optimal parameters;Finally,through simulation experiments,the results show that this hybrid algorithm has faster convergence speed and higher accuracy for parameter optimization in SVM,which makes the effect of process fault diagnosis better.