Operation was a Success, but the Patient Died – Call for More Advanced Process Diagnostics
Authors: | Hietanen Ville, Metso Automation, Finland Happonen Harri, Metso Automation, Finland Friman Mats, Metso Automation, Finland Huhtelin Taisto, Metso Automation, Finland Kaunonen Antti, Metso Automation, Finland |
---|
Topic: | 6.1 Chemical Process Control |
---|
Session: | Advances in Automation in Paper Making Industry |
---|
Keywords: | diagnostic, modelling, learning algorithms, multivariable systems |
---|
Abstract
The old cliché in the title of this paper is irritating, but describes all too well end-users' common experience of process diagnostics. Expensive tools are not in systematic use, exhaustive process models become outdated and information produced is scattered and detached from operational routines. At the same time, development of computational capability, information and communications technology and user interfaces, among other things, enable reliable easy-to-use hybrid diagnostic solutions to be made. In this paper, the topical need for modular, easy-to-use, situational and contextual process diagnostic approaches is discussed. Examples are given in the field of pulp and paper industry. Insight into process knowledge and its user-friendly representation in form of models and diagnostics tools, in particular, is provided.