Loading...

Modeling and Predicting the Residual Life of Roller Bearings under Variable-Velocity Radial Loading Using Vibration Analysis and Neural Network

Golnary, Farshid | 2018

393 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 50899 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Behzad, Mehdi
  7. Abstract:
  8. The main objective of this research is to predict the life of a bearing that works in variable speed working condition. In this project, two series of experiments were conducted to investigate the vibration of the bearings. In the first series of experiments, multiple imperfections have been created on parts of the bearing that are susceptible including the inner ring, the outer ring and the balls. The results of this test have been used to study the importance of speed compared to other parameters in increasing the vibration levels of the bearing. In addition, the results of this experiment have been used as inputs to a neural network. This neural network is trained to classify bearing defects and is able to extract defective bearings from its vibrating characteristics. The trained neural network has been validated using new series of experimental data. In the second series of experiments, tests have been conducted to find the process of bearing failure (from the start-up to the bearing failure stage). The characteristics of these fault records are entered as inputs to the neural network and the neural network uses them to predict the remaining life of the bearings. In order to predict the remaining life under variable speed condition and to examine the relationship between speed and survival, a physical model that expresses the relationship between life and speed is used. Furthermore, the neural network has been evaluated and verified with some new data
  9. Keywords:
  10. Variable Speed ; Neural Networks ; Remaining Useful Life ; Vibrational Analysis ; Rolling Bearing

 Digital Object List

 Bookmark

No TOC