Orthotopic‐filtering‐based hierarchical fault diagnosis algorithm for linear recursive models

Abstract

An orthotopic-filtering-based hierarchical fault diagnosis algorithm is proposed for complex systems with multiple fault types. The given algorithm uses an orthotopic method to describe a feasible parameter set and detects whether a fault occurs by determining whether the feasible parameter set is empty. Then the hierarchical clustering method is applied to analyse the fault library. The clustering result is used as the prior knowledge for fault diagnosis analysis, and a discriminant analysis is conducted layer by layer. Finally, a model-matching method is applied to realise the fault identification. However, if the fault type is not included in the fault library, the fault type is then added to the fault library. Therefore, when the fault diagnosis is performed again, the fault library is re-hierarchically clustered. The analysis on the false alarm rate and the missing detection rate of the fault diagnosis algorithm are also studied. Finally, the fault diagnosis of a buck circuit is taken as an example to demonstrate the effectiveness and feasibility of the proposed method by analysing the fault diagnosis results.