Fault diagnosis based on pole learning and MCS-LSSVM

In this paper, a fault diagnosis method using the poles of dynamical system as the learning sample of least squares support vector machine(LSSVM) optimized by modified cuckoo search(MCS) is presented. The system poles can reflect the failure state of the system more intuitively and effectively with less quantity, so they are selected as fault feature information to solve the problems that fault diagnosis results caused by information redundancy are inaccurate and the system design becomes more difficult. Moreover, in order to enhance the efficiency of LSSVMu0027s pole learning, the CS is improved from the aspect of discovery probability, which can ameliorate the conditions of the commonly used optimization algorithms. A benchmark example of adaptive reconfiguration control shows, the MCS-LSSVM fault diagnosis method with the pole learning has better modeling effect and higher classification accuracy, which verifies the effectiveness, progressiveness and accuracy of the method.