Estimation based Fault Diagnosis and identification in sequential Industrial batch processes modeled as Hybrid Petri nets

Abstract

Fault Diagnosis and identification (FDI) in a process plant refers to the concept of detecting and locating faults occurring in a process plant. Faults mostly include sensor and actuator faults. Identification or location of a fault requires a special technique or algorithm and hence the concept of Hybrid Petri nets are proposed to achieve FDI since it has a strong mathematical background. In this paper the concept of estimation based FDI is proposed and for this purpose Hybrid Petri nets are used to model the process which is followed by detailed analysis in Petri net environment. Corresponding FDI algorithms are developed for a typical sequential Industrial batch process- a three tank hybrid system (bench mark system) which is equivalent to the sewage treatment process application considered in this paper for study. The algorithms are coded in MATLAB and implemented using a Graphical User interface and corresponding numerical results are obtained.