Model-based diagnostic decision-support system for satellites

We propose a novel framework for Model-Based Diagnosis (MBD) that uses active testing to de-crease the diagnostic uncertainty. This frame-work is called LYDIA-NG and combines sev-eral diagnostic, simulation, and active-testing al-gorithms. We have illustrated the workings of LYDIA-NG by building a LYDIA-NG-based de-cision support system for the Gravity field and steady-state Ocean Circulation Explorer (GOCE) satellite. This paper discusses a model of the GOCE Electrical Power System (EPS), the al-gorithms for diagnosis and disambiguation, and the experiments performed with a number of di-agnostic scenarios. Our experiments produced no false positive scenarios, no false negative sce-narios, the average number of classification er-rors per scenario is 1.25, and the fault detection time is equal to the computation time. We have further computed an average fault uncertainty of 2.06×10−3 which can be automatically reduced to 9.5×10−4 by sending a single, automatically com-puted, telecommand, thus dramatically reducing the fault isolation time.