23-05-2014, 03:48 PM
Multi-fault classification based on support vector machine trained by chaos particle swarm optimization
Multi-fault classification based on support.pdf (Size: 263.34 KB / Downloads: 15)
ABSTRACT
A novel method of training support vector machine (SVM) by using chaos particle swarm optimization
(CPSO) is proposed. A multi-fault classification model based on the SVM trained by CPSO is established
and applied to the fault diagnosis of rotating machines. The results show that the method of training
SVM using CPSO is feasible, the proposed fault classification model outperforms the neural network
trained by chaos particle swarm optimization and least squares support vector machine, the precision
and reliability of the fault classification results can meet the requirement of practical application.
Introduction
The fault of machinery reduces the production rate and in-
creases the costs of production and maintenance, so knowledge
of what, where and how faults occur is very important [1].
In order to improve the veracity and reliability of the fault diag-
nosis results, many artificial intelligence methods such as neural
networks have been widely used [2,3]. These methods are based
on an empirical risk minimization principle and have some disad-
vantages such as local optimal solution; low convergence rate;
obvious ‘‘over-fitting” and especially poor generalization when
the number of fault samples is limited. Shen et a1. [4] proposed
a rough set theory based method that can diagnose more than
one category of faults in a generic manner. However, one disadvan-
tage of this method is that the rough set theory cannot be used to
deal with continuous attributes. To apply this method the discret-
ization method has to be used. Because a prior knowledge about
the attribute is difficult to obtain, it is hard to choose an appropri-
ate discretization method. This disadvantage may adversely affect
the robustness and accuracy of fault diagnosis.
Experimental results and analysis
In this section, we use several standard data sets to illustrate
the differences between canonical SVM and CPSO–SVM firstly,
and then establish a multi-fault classification model based
on CPSO–SVM and apply it to the fault diagnosis of rotating
machines.
Conclusion and future work
In this study, SVM trained by CPSO is applied to the fault diag-
nosis of rotating machines. Results show that CPSO–SVM can serve
as a promising method for training SVM, the PSO algorithm is simple to implement, and does not require any background of numer-
ical methods. It can be seen from the experiment that the fault
classification model overcomes the main shortage of artificial neu-
ral network without defining network structure and trapping in
the local optimum. Compared with LS–SVM and CPSO–NN,
CPSO–SVM has better performance.