Simulation study on sensor fault diagnoses of the temperature of the boiler high-temperature part metal wall

摘要

A sensor fault diagnosis system is designed using SVR. SVR is trained off line, and used online. After being trained, SVR is used to simulate system dynamic characteristic. The simulation result is compared with actual output, and then fault error is drawn. Then use the categorized algorithm of fault based on the SVR of the decision tree, classify the faults. The simulation result shows that, SVR can simulate the system more accurately, thus the fault error is very precise and the classification to fault is correct. This assures the validity of this fault diagnosis system.