In an ordinary process industry facility, there are hundreds of control loops, making sure that all temperatures, pressures, speeds, flows and so on keep their preset levels. In order for the plant to deliver a final product with the right quality without excessive resource consumption, it is vital that the control loops perform satisfactory. Continuous monitoring of all these loops is, however, an unreasonable task for the operator personnel of the plant.
Consequently, some kind of automatic supervision of the control loops is required, in order to know when one of them operates poorly and needs tuning. This, however, requires knowledge on what is to be considered to be normal behavior of the control loop, which is time consuming to specify. In this project, the idea is to use the knowledge of the control loop that was gathered when it was tuned last, perhaps when the plant was commissioned. If this knowledge is formulated as a mathematical process model then the methods developed in the project ”Model-based fault detection” can be used in order to determine if the control loop has changed significantly. The project is carried out in cooperation with Optimation AB, a local consulting agency focusing on control solutions in the process industry.
Berndtsson, M & Johansson, A 2008, 'Design of a dynamic threshold generator for λ-tuned control loops', Control Engineering Practice.
Berndtsson, M 2006, Fault detection in lambda-tuned control loops, Licentiatsavhandling, (Luleå university of technology Licentiate thesis; 2006:47).
Berndtsson, M & Johansson, A 2006, 'Fault Detection in Lambda-Tuned Control Loops Using an Observer with Integral Action', I 45th IEEE Conference on Decision and Control, s. 538 - 543.