Abstract:

A fault detection method based on fuzzy sets is presented to overcome the shortcomings of the limitchecking fault detection method. The method pays attention to the uncertainty of the measured value and the fuzzy attribute of the thresholds and divides the set of possible values of the sensor data into several fuzzy subsets, the difference degree between the measured value and the expected value is expressed by the membership function and the alarm weights of fuzzy subsets, taking the effect of time on the checking result into consideration, a method of window function is developed to calculate the sum of alarm weights as the rule of fuzzy checking decision. The simulation results demonstrate the method based on fuzzy sets outperforms the method based on PersistenceCounter.