An integrated deep learning-based approach for automobile maintenance prediction with GIS data

Chen, Chong, Liu, Ying ORCID: https://orcid.org/0000-0001-9319-5940, Sun, Xianfang, Cairano-Gilfedder, Carla Di and Titmus, Scott 2021. An integrated deep learning-based approach for automobile maintenance prediction with GIS data. Reliability Engineering and System Safety 216 , 107919. 10.1016/j.ress.2021.107919

PDF - Accepted Post-Print Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (1MB)

Abstract

Predictive maintenance (PdM) can be beneficial to the industry in terms of lowering maintenance cost and improve productivity. Remaining useful life (RUL) prediction is an important task in PdM. The RUL of an automobile can be impacted by various surrounding factors such as weather, traffic and terrain, which can be captured by the geographical information system (GIS). Recently, most researchers have conducted studies of RUL modelling based on sensor data. Owing to the fact that the collection of sensor data is expensive, while maintenance data is relatively easy to obtain. This study aims to establish an automobile RUL prediction model with GIS data through a data-driven approach. In this approach, firstly, due to the data type and sampling rate of the maintenance data and GIS data are different, a data integration scheme was researched. Secondly, the Cox proportional hazard model (Cox PHM) was introduced to construct the health index (HI) for the integrated data. Then, a deep learning structure called M-LSTM (Merged-long-short term memory) network was designed for HI modelling based on the integrated data which contains both sequential data and ordinary numeric data. Finally, the RUL was mapped by predicted HI and the Cox PHM. An experimental study using a sizable real-world fleet maintenance dataset provided by a UK fleet company revealed the effectiveness of the proposed approach and the impact of the GIS factors on the automobiles under investigation.

Item Type:	Article
Date Type:	Publication
Status:	Published
Schools:	Engineering
Publisher:	Elsevier
ISSN:	0951-8320
Date of First Compliant Deposit:	20 July 2021
Date of Acceptance:	11 July 2021
Last Modified:	06 Nov 2023 18:57
URI:	https://orca.cardiff.ac.uk/id/eprint/142731

Citation Data

Cited 7 times in Scopus. View in Scopus. Powered By Scopus® Data

Actions (repository staff only)

Edit Item