Research on Fault Diagnosis of Secondary Processing Circuit of Temperature Measuring Instrument in Nuclear Power Plant

Aiming at the problem of fault diagnosis of secondary processing circuit of temperature measuring instrument in nuclear power plant. Based on the circuit model built by Multisim simulation software, the sensitive parameters of the temperature transmitter are analyzed. The degradation process is simulated by changing the resistance value of the temperature sensitive resistor, and the first-order voltage output of the temperature sensitive resistor is subjected to Monte Carlo sampling and temperature scanning to obtain the ideal state output sequence. Twenty classification algorithms are used as base classifiers, and the training data of each base classifier algorithm are traversed to obtain the diagnostic accuracy of each algorithm. The four algorithms with the highest diagnostic accuracy are selected, and the four base classifiers are combined by soft voter. In order to prevent the lower accuracy algorithms from “slowing down” the higher accuracy algorithms when combined with the higher accuracy algorithms applying the voter, the results of the soft voting are compared with the highest accuracy algorithms in the original base classifiers. If the accuracy of the algorithm with the highest accuracy in the original base classifier is higher than the soft voting result, the final output is the result of the algorithm with the highest accuracy in the base classifier and saved as a model, and vice versa, the final output is the result obtained by soft voting and saved as a model. Through simulation verification, the algorithm in this paper can adaptively select the optimal classification algorithm to get the optimal monitoring results according to different input data, realizing the real-time monitoring of the circuit state.