Abstract:
The paper is concerned with search for faulty components in unidimensional linear {continuous and sampled-data) systems by introduction of test points into the system. For multidimensional linear systems with unidimensional components the faulty components are detected by disturbing certain system inputs. The study is performed for the case of no exogenous or internal noises in the system. For such conditions the extreme diagnosis depth is determined within the framework of the proposed approach.