Fault Detection and Accommodation in Particulate Processes with Delayed, Sampled Measurements

Measurement sampling and delays, when ignored in the monitoring and control system design, may erode the diagnostic capabilities of fault-tolerant control systems. This study, therefore, considered these in a proposed framework for fault detection and accommodation in particulate processes. An observer-based output feedback controller is designed based on a reduced-order model and includes an inter-sample model predictor and a propagation unit that compensates for measurement intermittency and delays, respectively. Rules for fault detection and accommodation are derived from the fault-free stability properties in terms of the sampling rate and delay size. The methodology is illustrated using a continuous crystallizer.