Fault Detection and Control Systems

Fault detection and control systems research investigates how to identify when an industrial process has deviated from normal behavior, diagnose the underlying cause, and intervene before failures propagate or cause harm. Methods range from physics-based models that encode known process dynamics to purely data-driven approaches—such as multivariate statistical analysis and machine learning—that infer normal behavior from historical sensor records, with soft sensors bridging the two by estimating variables that are difficult or expensive to measure directly. The practical stakes are high: undetected faults in chemical plants, power grids, or manufacturing lines can mean safety incidents, product defects, or costly downtime. Active research questions include how to build detection systems that remain reliable as processes drift over time or operate under novel conditions, and how to move from simply flagging an anomaly to providing operators with interpretable, actionable diagnoses.

Works

143,454

Total citations

1,529,274

Keywords

Process MonitoringFault DetectionData-Driven TechniquesStatistical AnalysisSoft SensorsModel-Based Diagnosis