Book Review

Journal of Sound and Vibration, 238(5), December 2000, 907.

Models and Reality in Systems Dynamics, by Hans Gunther Natke. UNSER Verlag, Hannover, Germany, 2000, ISBN 3-934208-03-7.

The aim of this book is to put dynamics within a framework of systems theory and engineering. Central to this aim is a discussion of mathematical modelling and the inherent uncertainties in models and measurements. In essence this work is a summary of the modelling philosophy of the author and his significant contributions over many years. The book starts by defining systems and their models within a framework of set theory, and emphasises that models must be 'goal-orientated'. Measured data is considered the best information on the actual system, and its use in the model building processes is considered. System identification is interpreted as test supported model building, and the need for validated and verified models is highlighted. The sources of model uncertainties and methods of their reduction are discussed. The author then moves onto methods of model correction (or model updating) and reviews the available residual types, identification methods and regularisation. The book ends with a summary approaches to model based diagnosis, and finally some comments on the future, concentrating on holistic models, nonstationary signals and active systems.

This is a short book, with only 68 pages of text. With such a large scope most topics can only be mentioned briefly. However, there are a large number of references, and the reader is directed to these for more detailed information. Furthermore, space has not allowed the examples to be discussed in depth. Thus this is not a book for the engineer learning modelling and identification techniques. More it is for the scientist or engineer who is already experienced in these techniques, who would appreciate a more philosophical and fundamental view of the modelling process.