Application LAMDA algorithm for Fault Detection and Isolation

For many complex technical devices the classic feed-back control does not satisfy. Therefore the Fault Tolerant Control (FTC) with progressive the Fault Detection & Isolation (FDI) is applied. The objective of this article is to describe the problem with application of data-based approach into the FDI block. The data-based methods (in our case we talk mainly about classification-based methods) provide the final information purely from the process data. The Learning Algorithm for Multivariate Data Analysis (LAMDA) is tested here for FDI task. The results of the recherche are demonstrated in a two-tank system utilized as a benchmark.