Minor fault detection of thermocouple sensor in nuclear power plants using time series analysis

his paper addresses a scheme for detection a minor deflection of thermocouple signal reading in nuclear power plants (NPPs). These minor deflection signal does not exceed the system threshold value, therefore they are neither separated by the data classification methods nor detected by the data approximation methods. The proposed approach is based on time series analysis methods. In this approach, the minor failure signals are detected by computing the distance between the empirical cumulative distribution function of test signal and the standard signal. The moving average method is applied to smooth out the small fluctuation and noise of the signals before computing the empirical distribution. The time series mean shift detection technique is employed to the faulty signal for surety that the signal has a fault and find out its change point. The proposed method is validated by the real data from the Fast Breeder Test Reactor (FBTR).