The paper focuses on the problem of robust fault detection using analytical methods and soft computing. Taking into account the model-based approach to Fault Detection and Isolation (FDI), possible applications of analytical models, and first of all observers with unknown inputs, are considered. The main objective is to show how to employ the bounded-error approach to determine the uncertainty of soft computing models (neural networks and neuro-fuzzy networks). It is shown that based on soft computing models uncertainty defined as a confidence range for the model output, adaptive thresholds can be described. The paper contains a numerical example that illustrates the effectiveness of the proposed approach for increasing the reliability of fault detection. A comprehensive simulation study regarding the DAMADICS benchmark problem is performed in the final part.
This paper proposes a practical tuning of closed loops with model based predictive control. The data assumed to be known from the process is the result of the bump test commonly applied in industry and known in engineering as step response data. A simplified context is assumed such that no prior know-how is required from the plant operator. The relevance of this assumption is very realistic in the context of first time users, both for industrial operators and as educational competence of first hand student training. A first order plus dead time is approximated and the controller parameters immediately follow by heuristic rules. Analysis has been performed in simulation on representative dynamics with guidelines for the various types of processes. Three single-input-single-output experimental setups have been used with no expert users available in different locations – both educational and industrial – these setups are representative for practical cases: a variable time delay dominant system, a non-minimum phase system and an open loop unstable system. Furthermore, in a multivariable control context, a train of separation columns has been tested for control in simulation, followed by experimental tests on a laboratory system with similar dynamics, i.e. a sextuple coupled water tank system. The results indicate the proposed methodology is suitable for hands-on tuning of predictive control loops with some limitations on performance and multivariable process control.