Abstract

Reliability modeling enables deriving reliability measurements and illustrating relevant fault-dependencies in manufacturing systems. Data-driven reliability modeling uses data generated in systems to either automate or at least support extraction of reliability models. To use these extracted models for decision support, we need to ensure models’ validity. In this extended abstract, we discuss our initial approach for validating data-driven reliability models. The challenge with validating data-driven models lies in the fact that these models are continuously generated and updated, implying that we need a new or updated validation approach to enable an ongoing validation of these models. The upside is that the systems of interest generate large amounts of data, which can significantly support the quantitative validation processes. Additionally, we briefly address the implications that could result from our proposed approach.