Abnormality detection of a continuous pan in a sugar mill

This paper reports an industrial application of principle component analysis to process abnormality detection in a sugar mill. The process under investigation is a continuous pan which is one of the most crucial processes in sugar production. Two sets of experimental results are obtained in this work: one is related to an artificially induced fault by a stuck valve; the other captures a naturally occurred fault due to equipment failure. The principle component analysis algorithm successfully detects both faults. Only the latter case is reported in this paper.