A three-volume work bringing together papers presented at 'SAFEPROCESS 2003', including four plenary papers on statistical, physical-model-based and logical-model-based approaches to fault detection and diagnosis, as well as 178 regular papers.

The safe and reliable operation of technical systems is of great significance for the protection of human life and health, the environment, and of the vested economic value. The correct functioning of those systems has a profound impact also on production cost and product quality. The early detection of faults is critical in avoiding performance degradation and damage to the machinery or human life. Accurate diagnosis then helps to make the right decisions on emergency actions and repairs. Fault detection and diagnosis (FDD) has developed into a major area of research, at the intersection of systems and control engineering, artificial intelligence, applied mathematics and statistics, and such application fields as chemical, electrical, mechanical and aerospace engineering. IFAC has recognized the significance of FDD by launching a triennial symposium series dedicated to the subject. The SAFEPROCESS Symposium is organized every three years since the first symposium held in Baden-Baden in 1991. SAFEPROCESS 2006, the 6th IFAC Symposium on Fault Detection, Supervision and Safety of Technical Processes was held in Beijing, PR China. The program included three plenary papers, two semi-plenary papers, two industrial talks by internationally recognized experts and 258 regular papers, which have been selected out of a total of 387 regular and invited papers submitted. * Discusses the developments and future challenges in all aspects of fault diagnosis and fault tolerant control * 8 invited and 36 contributed sessions included with a special session on the demonstration of process monitoring and diagnostic software tools

Bayesian Networks, the result of the convergence of artificial intelligence with statistics, are growing in popularity. Their versatility and modelling power is now employed across a variety of fields for the purposes of analysis, simulation, prediction and diagnosis. This book provides a general introduction to Bayesian networks, defining and illustrating the basic concepts with pedagogical examples and twenty real-life case studies drawn from a range of fields including medicine, computing, natural sciences and engineering. Designed to help analysts, engineers, scientists and professionals taking part in complex decision processes to successfully implement Bayesian networks, this book equips readers with proven methods to generate, calibrate, evaluate and validate Bayesian networks. The book: Provides the tools to overcome common practical challenges such as the treatment of missing input data, interaction with experts and decision makers, determination of the optimal granularity and size of the model. Highlights the strengths of Bayesian networks whilst also presenting a discussion of their limitations. Compares Bayesian networks with other modelling techniques such as neural networks, fuzzy logic and fault trees. Describes, for ease of comparison, the main features of the major Bayesian network software packages: Netica, Hugin, Elvira and Discoverer, from the point of view of the user. Offers a historical perspective on the subject and analyses future directions for research. Written by leading experts with practical experience of applying Bayesian networks in finance, banking, medicine, robotics, civil engineering, geology, geography, genetics, forensic science, ecology, and industry, the book has much to offer both practitioners and researchers involved in statistical analysis or modelling in any of these fields.

Nonlinear problems in flight control have stimulated cooperation among engineers and scientists from a range of disciplines. Developments in computer technology allowed for numerical solutions of nonlinear control problems, while industrial recognition and applications of nonlinear mathematical models in solving technological problems is increasing. The aim of the book Advances in Flight Control Systems is to bring together reputable researchers from different countries in order to provide a comprehensive coverage of advanced and modern topics in flight control not yet reflected by other books. This product comprises 14 contributions submitted by 38 authors from 11 different countries and areas. It covers most of the currents main streams of flight control researches, ranging from adaptive flight control mechanism, fault tolerant flight control, acceleration based flight control, helicopter flight control, comparison of flight control systems and fundamentals. According to these themes the contributions are grouped in six categories, corresponding to six parts of the book.

This work considers the problem of identifying the fault in a faulty dynamical system on the basis of the system's input and output signals only. For this purpose, a model-based method for the design of diagnostic tests which consist of specific input signals and appropriate residual generators is developed. The method extends the structure graph of dynamical systems in order to represent the couplings in a system which has been brought to a specific operating region. The resulting local structure graph is used to determine specific residual generators which can distinguish between faults on the basis of the system's input and output signals in the corresponding operating region. Algorithms to determine advantageous operating regions and input signals which drive the system into such operating regions are given. The application of the method to determine diagnostic tests is demonstrated using a typical automotive system, a throttle valve.

Distributed control systems offer the advantages of control local to the process being controlled while retaining the ease of control at a single centralised location. Typically such a system has involved a great deal of hard-wiring and has been of most use only in situations where flexibility is not essential. Now, however, distributed control systems are being applied more often in process control, autonomous systems and safety-critical systems where control needs to change to cope with fault appearance or other (possibly intentional) process disturbance. Reconfigurable Distributed Control helps meet the challenge of applying distributed control to dynamical systems, integrating different approaches to the problem. It presents an holistic view based on the appropriate use of stochastic, formal and robust control paradigms. The use of smart peripheral elements means that the degree of effort required for the reconfiguration of a networked control system can now be reduced, particular emphasis being placed on the reduction of time delays. Case studies are employed to demonstrate the real applications of the theory. While being of most interest to academic researchers and graduate students grappling with the problem of making distributed control systems more responsive to changes in process and plant, Reconfigurable Distributed Control will also be informative for readers with a background in more general distributed computing.

With increasing demands for efficiency and product quality plus progress in the integration of automatic control systems in high-cost mechatronic and safety-critical processes, the field of supervision (or monitoring), fault detection and fault diagnosis plays an important role. The book gives an introduction into advanced methods of fault detection and diagnosis (FDD). After definitions of important terms, it considers the reliability, availability, safety and systems integrity of technical processes. Then fault-detection methods for single signals without models such as limit and trend checking and with harmonic and stochastic models, such as Fourier analysis, correlation and wavelets are treated. This is followed by fault detection with process models using the relationships between signals such as parameter estimation, parity equations, observers and principal component analysis. The treated fault-diagnosis methods include classification methods from Bayes classification to neural networks with decision trees and inference methods from approximate reasoning with fuzzy logic to hybrid fuzzy-neuro systems. Several practical examples for fault detection and diagnosis of DC motor drives, a centrifugal pump, automotive suspension and tire demonstrate applications.

Thomas Hubauer addresses the challenge of providing reasonable interpretations of incomplete observational data in the context of imperfect domain models – a situation often encountered in the context of industrial diagnostics. To tackle this problem, the author proposes a novel approach called Relaxed Abduction, which is able to derive pragmatic interpretations in situations where existing methods either fail or provide overly complex solutions. To strengthen the link to applications in industrial diagnostics, he develops a methodology to structure diagnosis problems according to ISO 13379 and express it using multiple description logic knowledge bases.

Rob Milne was a remarkable man. He died of a heart attack on the 5th of June 2005 while climbing Mount Everest in Nepal. Milne (48) lived an active life: combining his three ‘careers’ seemingly effortlessly. He was a hi-tech entrepreneur, an AI researcher and a passionate mountaineer. Mount Everest was last on his list of the highest summits on each continent. He was only 400 meters from the top when he died. This publication commemorates and celebrates the life of Rob Milne. It covers all facets of Rob Milne's life and contains contributions by the people who have known him well and pay tribute to his life and his legacy. Rob Milne is survived by his wife Val and his two children Alex and Rosemary. After he died, his wife said in a radio interview: “Rob died at the top, doing what he loved.”