Detection and investigation of downtime causes

The use of machine learning and big data in production is still in its early stages as only a limited amount of data is used. This research aims to improve productivity by determining factors affecting downtime. Changeover’s time and performance rate improvement during production will be addressed considering not only equipment data and product data but also a variety of factors such as raw material consumption, environmental effect on material properties, production sequences and operators’ impact.
This project will include an evaluation of state-of-the-art techniques, the selection of data sources required for performance loss and changeover times analysis, the preprocessing of data, and the testing of different machine learning and big data techniques to identify the main factors impacting the downtime in production.

Faculty Supervisor:

Camélia Dadouchi;Bruno Agard;Robert Pellerin

Student:

Partner:

Bridgestone

Discipline:

Engineering

Sector:

Manufacturing

University:

Polytechnique Montréal

Program: