Integrating System Failure Diagnostics Into Model-based System Engineering

Abstract

Ever-increasing system complexity is challenging for development engineers and service personnel troubleshooting system failures in the field. This paper presents a systematic, scalable approach to attain a diagnostic model. Automatic transformation into computational models is used 1) at design time to improve the diagnosability of the system, and 2) during operation for guided root cause analysis by calculating the most probable failures and suggesting diagnostic procedures based on available data and observations. The approach combines nicely with model-based systems engineering, showing the added value of using diagnostic models both during the design of a system and during operation when the system needs to be diagnosed.