Diagnosis of Active Systems: Principles and Techniques (The Springer International Series in Engineering and Computer Science)

This book is about model-based diagnosis of a class of discrete-event systems called active systems. Roughly, model-based diagnosis is the task of finding out the faulty components of a physical system based on the observed behavior and the system model. An active system is the abstraction of a physical artefact that is modeled as a network of com­ municating automata. For example, the protection apparatus of a power transmission network can be conveniently modeled as an active system, where breakers, protection devices, and lines are naturally described by finite state machines. The asynchronous occurrence of a short circuit on a line or a bus-bar causes the reaction of the protection devices, which aims to isolate the shorted line. This reaction can be faulty and several lines might be eventually isolated, rather than the shorted line only. The diagnostic problem to be solved is uncovering the faulty devices based the visible part of the reaction. Once the diagnosis task has been on accomplished, the produced results are exploited to fix the apparatus (and also to localize the short circuit, in this sample case). Interestingly, the research presented in this book was triggered a decade ago by a project 011 short circuit localization, conducted by ENEL, the Italian electricity board, along with other industrial and academic European partners.

From the reviews:

"The book is an outstanding monograph of a recent research trend and reflects the experience of its authors in model-based diagnosis of a broad class of discrete-event systems, called active systems. ... The book is highly recommended to scientists and researchers working in the fields of artificial intelligence, control algorithms and data structures. The material can also be used for graduate courses in these areas. Engineers as well as software designers will find useful information to develop complex applications for supervising and monitoring industrial systems." (Octavian Pastravanu, Zentralblatt MATH, Vol. 1044 (19), 2004)