Search

In view of artificial neural network, there are some deficiencies in condenser fault diagnosis. The BP neural network used for condenser fault diagnosis is highly nonlinear pattern recognition and high precision of fault diagnosis. The PSO-BP neural network can effectively solve the problem of BP neural network that training time is long and training process is easy to fall into the local minimum. The training results of PSO-BP network in convergence speed and convergence effect are significantly improved. Under the condition of small samples, the calculation results of SVM method are better than the calculation result of the other two methods. Although the recognition ratio of improved PSO-BP(2) and SVM is the same, training time of improved PSO-BP(2) is longer than training time of SVM. The generalization ability of SVM is stronger, and the efficiency of SVM is higher than the neural network. With MATLAB programming, three different algorithms, which are BP neural network, PSO-BP neural network and SVM, are studied and compared for the performance of condenser fault diagnosis. In the models of this study, the research results show that condenser fault diagnosis based on SVM has the fastest convergence speed and the best accuracy.

For the reason that the application of the neural network in the condenser fault diagnosis has some restrictions, the result of the fault diagnosis based on neural network is unsatisfied. On the basis of the fault diagnosis of condenser based on neural network, the thought of data fusion is introduced in this paper and a method of condenser integrated fault diagnosis based on multiple neural networks and D-S evidence theory is proposed. According to BP neural network and CPN network, the respective diagnosis results regarded as the primary evidences of D-S theory evidence in decision layer are obtained first, and then these results are fused by using of the evidence theory to obtain the final diagnosis result. The result of simulation shows that: Comparing with the result from the single network, this method has a smaller error and higher diagnosis reliability.

An approach of image segmentation was used in the application of particle holography analysis for the diameter measurement. The whole process of image segmentation based on boundary detection was delivered. For improving the accuracy, the segmented particle image has been obtained by the bi-threshold method which decreases influence of background noise and inhomogeneous gray distribution of holography. The experiment results showed the most relative error is less than 6%. The study has some meanings on the data processing in particle field holography.

The high quality of the information source can improve the accuracy of target recognition when data is fused via multi-information sources. In this paper, it proposes a new method to determine the reliability coefficient of the information source by information criterion. With calculating the reliabilities of information sources, the more reliable source would be used in the combination process to improve the recognition rate. The experiment results prove that the better results will be obtained with the reliability coefficient to choose the source in information fusion.

In the application of particle holography analysis, the diameter measurement is basic and important. A novel approach of segmentation is proposed. For improving the accuracy of measurement, the method of bi-threshold is used firstly and then the edge linking method is adopted to get the more accuracy segmented particle image. The experiment tests showed the less relative error is in the diameter measurement with the algorithm, compared with the bi-threshold algorithm. The study has further meanings on the processing of particles field holography.