Software development for remaining useful life prediction of bearings

Bearings are crucial components in various industries, such as power generation, aerospace, and oil and gas. Predicting a bearing’s Remaining Useful Life (RUL) is essential for Condition-Based Maintenance. To achieve accurate RUL predictions, an accurate Health Indicator (HI) that represents degradation patterns is necessary. However, existing HIs can be affected by time varying working conditions and external interference, leading to decreased prediction accuracy. This project aims to develop reliable RUL prediction software for bearings by using a novel signal processing-based HI and data-driven prediction. The new HI leverages advanced signal processing techniques to identify degradation patterns. The software’s effectiveness will be validated using two public run-to-failure bearing datasets and one lab dataset created by the project team.

Faculty Supervisor:

Xihui Liang

Student:

Partner:

North Forge

Discipline:

Engineering

Sector:

Education; Management of companies and enterprises; Professional, scientific and technical services

University:

University of Manitoba

Program: