Practical grey-box process identification. Theory and applications (Q2500519)

The book is devoted to the investigation of process identification problems by the grey-box method. The latter is being implemented while modelling different objects such as industrial processes. Usually, two sources of information are available when building a model: some prior knowledge on the model or an object of investigation and experimental data. White-box identification uses mainly the first source of data, and black-box identification substantially uses the second one. Each of these methods has both advantages and drawbacks. The idea of grey-box identification consists in combining both approaches, i.e. in using both available sources of data in order to combine the strengths of the methods. The monography consists of three parts. The first part describes the theoretical fundamentals of white-box, black-box, and grey-box identification. The main emphasis is made on those elements of the theory that are important from the practical standpoint and necessary for understanding principles underlying the software described in Part II. Second part of the book is a tutorial introduction and a user's guide to MoCaVa, a software for calibrating and validating tentative model structures. MoCaVa is a grey-box tool which means that it takes into account the need for using both the prior information about process if available and the impact of random disturbances. The software is available for free download from the Springer website. It is distributed as a package for MATLAB, and since MoCaVa is an open source software, it allows one to expand its capabilities to needs of an end user. In the conclusive part of the book, two case studies of grey-box identification are described. The examples are drawn from the paper and steel industries. The main attention is paid to the two latter parts of the book which describe the MoCaVa software and the case studies.