FAULT DETECTION AND DIAGNOSIS WITH THE HELP OF FUZZY-LOGIC AND WITH APPLICATION TO A LABORATORY TURBOGENERATOR

摘要

In this contribution a new approach to the fault detection and fault diagnosis problems is presented. The main efforts are concentrated to obtain, through the utilization of some concepts of the fuzzy-logic, a better diagnosis about the causes of a fault. The solution consists basically of the following three levels: the residual generation, the detection and the decision. The residual generation is carried out through the utilization of m-MISO-ARX models of the process, where m is the number of outputs. The faults are detected with the help of the Hotelling's and Scheffee's statistics, as well as with the help of some additional computations for the operating point, which after a fuzzification procedure are given to the decision level. The decision level consists of some logic rules, which have been classified into two stages, and of a final decision block, which can give a crisp evaluation about the state of the plant, especially about the faults. The proposed solution is applied to a laboratory turbogenerator and good experimental results have been obtained.