Zhu Zhengwei

Model study of transformer fault diagnosis based on principal component analysis and neural network

Models of transformer fault diagnosis were developed by using on-line data to improve the conventional testing method and physical law methods. The operation data of 7 variables that affect transformer fault had been studied by using principal component analysis method, 5 principal components had been obtained and the contributions of the principal components had been computed. Based on the factors, a three-layer RBF neural network is designed. It is proved by MATLAB experiment that RBF neural network is a strong classifier which can be used to diagnose transformer fault effectively.

Study on ultrasonic measuring component of binary gas

In this paper, a new method is proposed for analyzing components of binary gas. An equation for measuring binary gas is deduced theoretically. It is found that the velocity of the ultrasonic wave in the binary gas has a relation with the gas concentration. Although the rate of speed change is very small, change in the phase of velocity can be obvious. An implementation of the research project is introduced from the points of theoretical analysis, system collectivity design and functional modules. Dual channel and measuring phase are used to measure the ultrasonic wave propagation time. Difference phase comparison is used to realize the accurate measurement for gas mixture. The shortcoming in conventional sound-time measurement is overcomed. In order to increase the devices precision, measures such as temperature compensation are adopted. The relation between the temperature and the dual-ultrasonic phase difference and the relation between the gas thickness and the dual-ultrasonic phase difference are found by experiments. Reasonable compensation has been made to those relations by software.