Machinery Health Monitoring using Multiple Sensor Fusion and Deep Learning

Manufacturing system reliability is significant in modern industry that requires high production speed, low maintenance cost, and enhanced operation safety. The proposed work is aiming to establish a systematic and applicable approach for condition monitoring of key machinery (e.g. bearings, gearboxes, motors) through multiple sensor fusion and deep learning. With the capability of working under high industry electromagnetic noise environment, optical-fiber-based sensors are incorporated to collect various signals of the machinery. Sensory information from various types of sensors will then be fused by sensor fusion algorithms. Deep learning, as an emerging technique successfully applied in many areas will be utilized to automatically extract the knowledge inside the massive sensory data. An intelligent machine fault diagnosis and prognosis system will be developed to achieve increased accuracy and reliability.

Faculty Supervisor:

Clarence W de Silva

Student:

Partner:

The University of Tokyo

Discipline:

Engineering

Sector:

Education

University:

The University of British Columbia

Program: