Generic Framework for Hybrid Fault Diagnosis and Health Monitoring of the Tennessee Eastman Process

Fault Diagnosis and Health Monitoring (FD-HM) based on hybrid approaches have been an active field of research and a key challenge over the last few years. In many applications, generic and unified approaches are usually required for designing a complete robust FD-HM system. The main contribution of this article is to develop a hybrid approach properly tailored for such challenge, by bridging the Bond-Graph (BG), the Signed Bond-Graph (SBG) and the Bayesian Network (BN) methods, through a probabilistic common framework for decision-making. This new hybrid methodology benefits from different types of information emanating from the quantitative model, the qualitative reasoning and the available data, in order to increase the overall confidence in the diagnosis performances. The effectiveness of the proposed hybrid approach is validated by the well-known Tennessee Eastman Process.