Auxiliary signal design in fault detection and diagnosis (Q1187695)

This book with rapid fault-detection in an industrial process, detecting something like the Three-Mile accident. It is concerned with the best statistical tests, based on a list of types of faults and various interval measurements, to determine whether a particular fault is actually occurring. Although other methods are discussed, the method used in the book is the sequential probability ratio test (SPRT) of Wald which has an optimality property of requiring on the average a minimum number of measurements. The author discusses standard topics like robustness, generalizes the theory to auxiliary signals, which are controls by which the quantity measured can be adjusted and to more than 2 hypotheses. He works out in detail an example for a caustic soda plant and gives other examples.