Thesis authored by Peter Shibu. Completed in 2019.
Thesis abstract: Due to the increase in demand for vehicles, the automotive industry ire constantly competing to launch new quality products into the market, which means the need for an efficient production line. However, companies are so focused on launching their products that they fail to address and update their maintenance systems. The following report aims to fill in this maintenance gap by developing a predictive maintenance design for an automotive assembly line. The purpose of this report is to reduce downtime caused by machine failure in an assembly line by predicting the maintenance of a machine before it actually breaks down. This is done by continuously monitoring the health of the machine, and collecting this raw data which is known as big data. The useful information is then extracted from this raw data by data mining techniques, which is then used to develop a predictive model algorithm. The predictive model algorithm is designed to make sense of the mined data by specifically pinpointing the root causes of machine failure and to raising an alarm when the machine is moving towards breakdown state. Due to the lack of research regarding ‘Big Data and Predictive Maintenance’ as it is a relatively new concept, this report seeks to promote the advantages of implementing these technologies into the automotive production systems. The report used a mix of qualitative and quantitative research approaches to find out the current maintenance practises followed in the automotive sector, and also to design a predictive maintenance model using relevant datasets and software environments.