Identification for fault detection in an industrial condenser

Abstract The condenser of a batch distillation column, equipped with temperature and flow sensors, is considered. The aim of the paper is to build a system which is able to detect flow sensor failures occurring in a batch. Three methods are considered For all of them, the first step consists of the parameter identification of a grey box model. For the first two tests, the classical least squares approach is used The first fault detection test is based on the value of the sum of the squares of the prediction errors obtained with the current batch measurements. The second one is a classical hypothesis test relying on the log-likelihood ratio. Both methods are shown to lack robustness with respect to the process variability from batch to batch. A third approach is then investigated, in which during the identification phase, the optimised cost function is chosen in order to reflect the sensitivity of the model to the flow sensor fruits. It yields significantly better results than the other two methods for the considered data.