Loading...
Search
Search in this resource
sort by
Send to Friend
 Digital Object List
 Bookmark
 Review List
  1. Simple search
  2. Search result
  3. پیش آگاهی خرابی یاتاقان های غلتشی تحت شرایط تغییر بار با استفاده از ترکیب روش‌های داده محور و فیزیکی

Bearing Prognostics under the Load-Varying Conditions, Using Combination of Data-Drive and Physic-based Methods

Arghand, Hesamoddin | 2019

1369 Viewed
  1. Type of Document: Ph.D. Dissertation
  2. Language: Farsi
  3. Document No: 52454 (08)
  4. University: Sharif University of Technology
  5. Department: Mechanical Engineering
  6. Advisor(s): Behzad, Mehdi; Rohani, Abbas; Ming J. Zuo
  7. Abstract:
  8. In this research, an artificial intelligence (AI) based model and a physics-based model have individually employed for prognostics of rolling element bearing (REB) under constant operating condition. Then a combination of physics-based models and data-driven model (Hybrid model) has been developed for remaining useful life (RUL) prediction of rolling element bearing (REB) under the load de-rating condition. The well-known crack propagation model Paris’ law has been employed to develop a physics-based model which can describe the degradation rate of the REB when the applying load is changed. To this aim, the elastic solution of the contact mechanics and Hertz stress between curved surfaces is used to explain the relationship between the radial load of REB and maximum shear stress. On the other hand, a feedforward neural network (FFNN) is employed for estimating the degradation rate parameter in the physics-based model. The proposed hybrid framework for combining data-driven model and physics-based models explains how long RUL can be extended if the applying load of REB is decreased to a new lower level. The results of a simulated degradation process under variable loading condition shows that the proposed algorithm can predict the RUL for each given operating condition. In addition, the performance of proposed models tested with bearing accelerated life test data in constant operating condition as well as in de-rated loading condition. There is an excellent agreement between the simulated prediction results and the actual life of REBs in both constant loading condition and de-rated loading condition
  9. Keywords:
  10. Condition Monitoring ; Prognostics Failure ; Hybrid Methods ; Neural Network ; Rolling Bearing ; Remaining Useful Life ; Paris Law

 Digital Object List

 Bookmark

No TOC