Supervision of industrial processes using artificial intelligence

The main objective of this book consists in the realization of different neural classifiers (types: MLP, RBF and RBF-BP) which allow to detect and diagnose the different modes of operation (Normal and Abnormal) in a column for the continuous automated distillation of methylcyclohexane from a binary toluene\ methylcyclohexane mixture. In this book, the author first proposed an overview of the different approaches to on-line and off-line fault monitoring and diagnosis developed by the different research communities. Then the author presented the results obtained during the application of the different approaches that allow to detect and diagnose the various malfunctions in a continuous automated distillation column, namely: ANN-RBF-BP, ANN-MLP and ANN-RBF which will allow the separation of two operating modes (normal and abnormal) by classification and compare them to other classification methods (KPPV, ACP, AFD).