Fault detection and isolation in a FED-batch penicillin fermentation process

This paper deals with fault detection and isolation (FDI) in a fed-batch penicillin fermentation process. The method is based on an empirical model developed using the multi-way partial least squares technique. The paper begins by demonstrating how this model can be used to provide FDI capabilities and then shows how this FDI scheme can be integrated within a model predictive controller to provide accurate control in the presence of fault conditions within the process.