Abstract
Measurement channels of instruments used in nuclear power plants are designed on the principles of multiplicity, redundancy, and independence to obtain more reliable instrument measurements. Furthermore, a data acquisition system is set up for real-time monitoring of the state of a production process system. Based on the underlying principles and characteristics, cross validation was proposed to be used in fault diagnosis in our research, and a relevant model was built. According to the fault probability of instruments and the uncertainty of measurement channels, three sets of criteria were established for analyzing the degradation, and failure. These sets of criteria were also used as the basis for evaluating the operational state of the system or instrument. Improvements were seen in many plants; thus, proving that the model and method were highly effective in detecting anomalies and diagnosing faults, while being easy to operate. The developed model can act as a supplementary measure during maintenance, thereby reducing the frequency of interruptive maintenance and also reducing the scale of capital investment. • Cross validation model for fault diagnosis is proposed. • Establish three-level verification criteria according to the deviation. • A degradation criterion is proposed according to the uncertainty. • The failure criterion is proposed according to the failure probability analysis.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call