[...]... industrial processes Preface xiii How to Use this Book Successful grey−box identification of industrial processes requires knowledge of two kinds, i) how the process works, and ii) how the software works Since the knowledge is normally not resident within the same person, two must contribute Call them process engineer” and “model designer” The latter should preferably have taken a course in Process. .. knowledge of the process and it would be a waste of information not to use it After the first session on grey−box identification at the 4th IFAC Symposium on Adaptive Systems in Control and Signal Processing in 1992, and the first special issue in Int J Adaptive Control and Signal Processing in 1994, the approach has now been reasonably well accepted as a paradigm for how to address the practical problems... describe the process as a “black box” A linear model is the most popular first choice, but if one would suspect that the process is nonlinear, and also take into account some rudimentary prior knowledge (that a hole at the bottom tends to reduce the level), the following heuristic form would also be conceivable: dz dt = p 1 z + p 2 p + p 3 f − [p 4 z + p 5 p + p 6 f] α (1.9) 8 Practical Grey−box Process. .. in practical grey−box process identification The style of parts II and III deviates somewhat from what is customary in text books, namely to use sentences in passive form, free of an explicit subject The idea of the customary practice is that science and engineering statements should be valid irrespective of the subject Unfortunately, the custom is devastating for the understanding, when describing processes... Bengt Nilsson, who contributed process knowledge to the Cardboard case study Jan Erik Gustavsson, who contributed process knowledge to the Recovery Boiler case study Alf Isaksson, who participated in the Pulp Refiner and Drive Train cases, and headed the MoCaVa project between 1998 and 2001 Linus Loquist, who designed the MoCaVa home page Contents Part I Theory of Grey−box Process Identification 1 Prospects... What one can generally do in order to take prior knowledge into account is to start with a versatile class of models, for which there are general tools available for analysis and identification, and try 4 Practical Grey−box Process Identification and adapt its freedom, its ‘design parameters’, i.e., the specifications one has to enter into the identification program, to the prior knowledge This means... doing grey− box identification properly ii) A software tool MoCaVa (Process Model Calibrator & Validator) based on one of the procedures (Bohlin and Isaksson, 2003) iii) A number of case studies of grey−box identification of industrial processes They were carried out in order to see whether the theoretical procedure would also be a practical one, and to test the software being developed in parallel... scenario in mind: Suppose a production process is to be described by a dynamic model for simulation or other purposes A number of submodels (or first principles or heuristic relations) for parts of the process are available as prior information, developed under more or less well controlled conditions However, when the submodels are assembled into a model for the integrated process, all their input and output... is downloadable from www .springer. com/1−84628−402−3 together with all material needed for running the case studies (The package also contains the HTML−manual as well as on−line help facilities.) This offers a possibility to get more direct experience of the model−design session It would therefore be possible to use Parts II and III as study material for a course in grey−box process identification Acknowledgements... it would be impractical to try and solve all problems of grey−box identification by relying on intuition and reasoning alone, however clever Therefore, the mathematics is interpreted in intuitive terms, and necessary approximations motivated in the same way, whenever the mathematical problems become unsurmountable, or an exact solution would take prohibitively long for a computer to process The following . Data Bohlin, Torsten, 1931- Practical grey-box process identification : theory and applications. - (Advances in industrial control) 1 .Process control - Mathematical models 2 .Process control - Mathematical. in a rolling-mill process) , continuous pulp digestion, cement milling, an industrial recovery boiler process (pulp production process unit) and cardboard manufacturing. The practical experience. of Tokamak Plasmas MarcoAriolaandAlfredoPironti Publication due May 2007 Torsten Bohlin Practical Grey-box Process Identification Theory and Applications With 186 Figures 123 Torsten Bohlin Automa