Reliability Analysis for Automobile Engines: Conditional Inference Trees
Shixuan Wang, Ying Liu, Carla Di Cairano‐Gilfedder, Scott Titmus, Mohamed Mohamed Naim, Aris A Syntetos · Procedia CIRP · 2018
The reliability model with covariates for machinery parts has been extensively studied by the proportional hazards model (PHM) and its variants. However, it is not straightforward to provide business recommendations based on the results of the PHM. We use a novel method, namely the Conditional Inference Tree, to conduct the reliability analysis for the automobile engines data, provided by a UK fleet company. We find that the reliability of automobile engines is significantly related to the vehicle age, early failure, and repair history. Our tree-structured model can be easily interpreted, and tangible business recommendations are provided for the fleet management and maintenance.