Identification for fault detection in an industrial condenser

The condenser of a batch distillation column, equipped with temperature and flow sensors, is considered here. The aim of the paper is to build a syste m that is able to detect flow sensor failures occurring in a batch. Three m ethods are considered. For all of them, the first step consists of the para meter identification of a grey-box model. For the first two tests, the clas sical least-squares approach is used. The first fault-detection test is bas ed 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 hypothe sis test, relying on the log-likelihood ratio. Both methods are shown to la ck robustness with respect to the process variability from batch to batch. A third approach is then investigated, in which the optimised cost function of the identification phase is chosen in order to reflect the sensitivity of the model to the how sensor faults. II yields significantly better resul ts than the other two methods for the data being considered. (C) 1998 Elsev ier Science Ltd. AII rights reserved.