Failure detection and diagnosis. Application to a grinding-classification circuit
Résumé
In this paper, we propose a hierarchical approach for failure detection and diagnosis on a plant. First we present local tests applied on sensors output. Secondly, at a global level, the use of models of the process allows to compute prediction errors and the statistical tests are applied to detect failures. With the aid of results of tests on identified parameters it is then possible to obtain a characterization of the failure. An example of failure diagnosis on a mineralurgical circuit is presented