Coloured Petri net diagnosers for lumped process systems

A model-based method is proposed in this paper for algorithmic generation of diagnosers in coloured Petri net (CPN) form from characteristic input-output traces obtained from a qualitative model of lumped process systems. The qualitative model contains the description of the considered persistent faults in the form of fault indicators, and it is transformed into a CPN. The diagnosers are constructed from a CPN obtained by the process mining methodology using the generated inputoutput traces for identically constant inputs. The concepts and methods are illustrated using a simple case study consisting of an industrial storage tank system with additive and multiplicative failures on sensors, and on the behaviour of a pump