Valve Stiction Detection Method Based on Dynamic Slow Feature Analysis and Hurst Exponent

Valve stiction is the most common root of oscillation faults in process control systems, and it can cause severe deterioration performance system instability, ultimately impacting product quality safety. A new method for detecting valve stiction, based on dynamic slow feature analysis (DSFA) Hurst exponent, proposed this paper. The first utilizes DSFA to extract features (SFs) from preprocessed reconstructed data controller output controlled variable; then, calculates exponent slowest SF quantify its long-term correlation; and, finally, defines a detection index identify stiction. results obtained simulations actual case studies demonstrate that method, DSFA–Hurst effectively detect loops.