Fault Diagnosis Models, Artificial Intelligence, Applications

All real systems in nature - physical, biological and engineering ones - can malfunction and fail due to faults in their components. Logically,the chances for malfunctions increase with the systems' complexity. The complexity of engineering systems is permanently growing due to their growing size and the degree of automation, and accordingly increasing is the danger of fail ing and aggravating their impact for man and the environment. Therefore, in the design and operation of engineering systems, increased attention has to be paid to reliability, safety and fault tolerance. But it is obvious that, compared to the high standard of perfection that nature has achieved with its self-healing and self-repairing capabilities in complex biological organisms, fault management in engineering systems is far behind the standards of their technological achievements; it is still in its infancy, and tremendous work is left to be done. In technical control systems, defects may happen in sensors, actuators, components of the controlled object - the plant, or in the hardware or soft ware of the control framework. Such defects in the components may develop into a failure of the whole system. This effect can easily be amplified by the closed loop, but the closed loop may also hide an incipient fault from be ing observed until a situation has occurred in which the failing of the whole system has become unavoidable.