Predicting failures in networking equipment using machine learning

This project is in collaboration with Ciena, an international company specialized in the supply of telecommunications networking equipment and software services. Electrical components produced by Ciena are important in many networking equipment such as routers or switches. Every year, a large number of components is produced. However, the production process may have some flaws, causing the production of faulty components. Detecting which components are faulty and what are the causes of the failures is a difficult and costly process and has been identified as a major challenge by Ciena in order to maintain a very low failure rate. Besides, failures can arise following a normal long usage of the component. Predicting when wear failures will occur in order to replace the equipment before it fails is also a challenge. This research project is within this context. The goal is to use state-of-the-art machine learning tools in order to improve the reliability of Ciena production process by detecting automatically faulty components as they are produced.

Faculty Supervisor:

Kim Khoa Nguyen;Quentin Cappart;Brigitte Jaumard;Daniel Aloise;Brigitte Jaumard

Student:

Partner:

Ciena Canada (Saint-Laurent, QC)

Discipline:

Computer science

Sector:

Information and cultural industries; Manufacturing

University:

École de technologie supérieure; Polytechnique Montréal

Program: