Fault Diagnosis Based on Neural Networks and Decision Trees: Application to Damadics

This paper presents a new technique of fault detection and isolation (FDI). This technique is based on Neural Networks fault-free and Faulty behaviors Models (NNF Ms). NNFMs are used for residual generation, while decision trees are introduced for residual selection and evaluation. Each part of the tree corresponds to specific residuals and with the decision tree it becomes possible to take the appropriate decision regarding the actual process behavior by evaluating few residuals. In comparison to the systematic evaluation of all residuals, the proposed technique requires less computational effort and is suitable for on line diagnosis implementation. An application to DAMADICS (Development and Application of Methods for Actuator Diagnosis in Industrial Control Systems) Actuator system is presented to illustrate and confirm the effectiveness and the accuracy of the proposed approach.