Predictive Maintenance – Analysis of Seasonal Dependence of Vehicle Engine Faults

Abstract

This paper presents methods and results for the analysis of interrelationships between the occurrence of specific engine faults according to seasons. The issue of maintenance is substantial for the automotive industry and improvements are requested due to the enhancement of profitability. Findings of this paper are based on logged-vehicle data from 760.976 vehicles provided by the company Geotab. Utilization of such data gains importance for the automotive service sector with special regard of increasing importance of predictive maintenance. The visualization of the data of three different engine faults was realized with the free graphic and statistics program “Tableau Desktop 2018.1” as well as “IBM SPSS Statistics Subscription Trial for Mac OS”. The result is that the tested interrelations are significant, leading to the conclusion that the engine faults of “vehicle battery has low voltage”, “low priority warning light on” and “general warning light on” are dependent of seasons. This finding can be used to help car manufacturers and car service providers to reduce maintenance costs.

Keywords: automotive, big data, predictive maintenance, seasonal engine faults, vehicle error codes