We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
Model-based strategies for sensor fault accommodation in uncertain dynamic processes with multi-rate sampled measurements.
- Authors
Allen, James; Chen, Shaoru; El-Farra, Nael H.
- Abstract
Highlights • A methodological framework for sensor fault-tolerant control. • Active accommodation strategies for stability and performance degrading faults. • Rigorous closed-loop stability and performance assessments. • Insight into interplays of sensor errors, sampling rates and control parameters. • A simulated application to the control of non-isothermal chemical reactors. Abstract This work presents model-based strategies for the accommodation of sensor faults in dynamic processes with multi-rate sampled state measurements. The developed strategies account explicitly for both closed-loop stability and performance considerations. Initially, a model-based feedback control system, in which a model predictor compensates for the unavailability of state measurements between sampling times, is designed. The stability and performance characteristics of the multi-rate sampled-data closed-loop system are analyzed and explicitly characterized in terms of the state sampling rates, the fault parameters, and the various process, model and controller design parameters. The resulting characterizations provide insight into the robustness and margins of tolerable faults that can be accommodated, and are used to develop both stability-based and performance-based fault accommodation schemes that aim to maintain closed-loop stability and minimize the degradation of closed-loop performance in the presence of sensor faults. Two types of sensor faults are considered. These include faults that manifest themselves as improper sensor readings, as well as faults that cause drift in the sensor sampling rate. The first type of fault introduces errors in the model state updates at the sampling times, while the second type of fault alters the rate at which the sensors sample the process states. The developed methods are illustrated through a case study involving a cascade of two non-isothermal continuous-stirred tank reactors with plant-model mismatch and access to full state measurements that are sampled at different rates. The case study gives some insight into the effects of sensor faults on the stability and performance of the sampled-data state feedback control system, and how these effects can be mitigated through use of fault-tolerant control.
- Subjects
ISOTHERMAL processes; CHEMICAL reactors; CLOSED loop systems; ROBUST control; FEEDBACK control systems
- Publication
Chemical Engineering Research & Design: Transactions of the Institution of Chemical Engineers Part A, 2019, Vol 142, p204
- ISSN
0263-8762
- Publication type
Article
- DOI
10.1016/j.cherd.2018.12.003