Fault Detection and Control Systems

Fault detection and control systems research concerns how to reliably identify when an industrial process is behaving abnormally and respond before equipment fails or product quality degrades. The work draws on a combination of statistical analysis, data-driven modeling, and physical process knowledge to distinguish genuine faults from ordinary measurement noise — a distinction that grows harder as systems become larger and more interconnected. A central open question is how to build detection methods that generalize across operating conditions without requiring costly labeled fault data, which is rarely available in practice. Current active directions include integrating machine learning with model-based diagnosis, developing soft sensors that infer unmeasured process variables from available signals, and making these methods robust enough to handle the nonstationary, high-dimensional data streams common in real industrial environments.

Works

144,462

Total citations

1,553,346

Keywords

Process MonitoringFault DetectionData-Driven TechniquesStatistical AnalysisSoft SensorsModel-Based Diagnosis