Diagnostic Row Reasoning Method Based on Multiple-Valued Evaluation of Residuals and Elementary Symptoms Sequence

The paper analyses the research problem of conducting diagnostic reasoning for dynamic objects to eliminate possibility formulating false diagnoses resulting from different delays symptoms related a particular fault while simultaneously striving obtain high distinguishability. aimed develop new inference method robust symptom and characterised by accuracy generated diagnosis. Known methods ensuring correctness in case but at cost reducing distinguishability faults have been characterised. A was developed, which uses three-valued residual evaluation knowledge regarding elementary sequences. formal description diagnosing system proposed are given. obtaining about order based on cause-and-effect graph method’s effectiveness presented simulation studies example set serially connected tanks. comparison obtained using other approaches illustrates advantages.