Fault Diagnosis Method for Check Valve Based on Stochastic Subspace Identification

Swing check valve plays a key role in preventing countercurrent in complex pipeline systems such as nuclear power plant. Aiming at the problem that swing check valve in loop devices is prone to stall and other faults during long-term operation and difficult to monitor, a fault diagnosis method based on stochastic subspace identification is proposed. The method can calculate and process the acceleration response data of check valve in real time. After calculating the modal parameters, the deviation from the standard modal parameters is calculated to realize the fault diagnosis and detection of the check valve. The feasibility of the proposed method is verified by simulation analysis from explicit dynamics and modal aspects respectively. Finally, the method is verified by experiments. The test results show that the check valve with stuck fault can be effectively monitored and identified by the method during the test, and the method has certain guiding significance for the fault detection of the valve without periodic characteristics in the loop.