Monitoring of complex industrial processes based on self-organizing maps and watershed transformations

An efficient operation of complex industrial processes requires the continuous diagnosis of the asset functionality. The early detection of potential failures and malfunctions, the identification and localization of present or impending component failures and, in particular, the monitoring of the underlying physical process behaviour is of crucial importance for a cost-effective operation of complex industry assets. With respect to these suppositions a monitoring concept based on machine learning methods has been developed, which allows an integrated and continuous diagnosis of the physical process behavior and phases. The present paper outlines briefly the architecture of the developed distributed diagnostic concept and presents in detail the developed approach for the identification of intrinsic process-phases and the monitoring functionality of the unknown process behaviour based on self-organizing-maps and watershed transformations.