Large-scale monitoring applications in process industry

Abstract

Modern industrial processes are increasingly complex and managed with limited personnel resources. In order to ease the task of process management new networked communication and monitoring technologies have been developed and implemented in production plants. To a great extent this has been enabled by technological advances in electronics, communications and software techniques. The technological advances have resulted in an increasingly decentralized periphery and improved the possibilities in embedded data processing. On the other hand new integration techniques have resulted in improved integration of previously separate process assets. The integrated process assets and added intelligence form more manageable and compact units which improves the accessibility and usefulness of process data.

The theoretical part of the thesis discusses process monitoring methods and the trend of decentralization in the process industry. Several related methods and technologies are introduced and the evolution of process automation discussed. In the applied part two subprojects and their results are presented. The first subproject concentrates on the development of a large scale process monitoring application. The application was designed to be generic so that the chosen methods don’t limit the implementation area to any specific processes and the deployment costs would be reasonable. The application was then deployed in a pulp drying process. The second subproject studies the possibilities of exploiting the features of intelligent field devices in process monitoring by incorporating field device level diagnostics into the developed monitoring concept.

The results of these projects were mixed. The results of the first subproject were very promising and all the participants were pleased with the developed application. The monitoring application learned the typical behaviour of the process and is a good tool for pointing out problem areas in the process. The results of the second subproject were not as good due to lack of interesting events in the process providing excitation for the monitoring system. The problem was a result of “too well” functioning machinery and thus nothing could be done about it. However it could be established that intelligent field devices can be used as a part of developed monitoring concept. The results of the subprojects suggest that such monitoring and performance measurement systems are useful tools but further work is needed, especially in the integration of separate information sources.