Investigating the applicability of a hybrid simulation methodology to motor bearing failure through incorporation of real system data (April 2026)

This paper will be included in An Open Access Journal of Production & Manufacturing Research, VOL. 14, NO. 1, 2661532, accessible via the following link: https://www.tandfonline.com/doi/full/10.1080/21693277.2026.2661532#abstract https://doi.org/10.1080/21693277.2026.2661532

Abstract

Predictive maintenance tools show promising results in the literature; however, theartificial intelligence and machine learning algorithms that underpin these approachesrequire large volumes of training data. This presents a challenge when sensor instru-mentation is recent or limited, restricting the availability of historical data. One poten-tial solution is the generation of simulated datasets capable of representing faultconditions. This paper investigates a hybrid simulation approach that incorporatesreal healthy operating data from the target system to improve simulation validitywithout explicitly modelling every component. The approach is evaluated usingmotor bearing defect experiments from Case Western Reserve University. The simula-tion successfully reproduces key macroscopic fault features and performs well inclassification tasks with real data. However, it fails to capture the full complexity ofreal signatures, as confirmed by a two-group t-test comparing simulated and experi-mental results, highlighting limitations of simulated data substitutes.