Intelligent diagnosis method for early faults of electric-hydraulic control system based on residual analysis

Early faults typically manifest as subtle changes on signals owing to its significant concealment and inherent randomness. The diagnosis of early fault holds significant importance for enhancing operational safety and production efficiency. To address the challenge of weak features and often high uncertainty associated with early fault characteristics, this study proposed an early fault diagnosis method for electric-hydraulic control system with features obtained by residual analysis. The residual features are extracted and analyses through residual signal extraction, residual processing, feature extraction, and residual feature sensitivity assessment. The new features obtained are applied to optimize the fault diagnostic model established based on Bayesian network. The incentive factor evaluation model based on residual feature analysis and the fault diagnosis result correction mechanism based on Bayesian network model are then established. The newly developed method is applied to a control system for subsea blowout preventer used as a case study to analyse the early fault evolution mechanism.