Journal of Northeastern University Natural Science ›› 2020, Vol. 41 ›› Issue (2): 223-228.DOI: 10.12068/j.issn.1005-3026.2020.02.013

• Mechanical Engineering • Previous Articles     Next Articles

Reliability Analysis Based on Active Learning for Complex Mechanical Structure

CAO Ru-nan, SUN Zhi-li, ZHANG Yi-bo, WANG Jian   

  1. School of Mechanical Engineering & Automation, Northeastern University, Shenyang 110819, China.
  • Received:2019-06-21 Revised:2019-06-21 Online:2020-02-15 Published:2020-03-06
  • Contact: CAO Ru-nan
  • About author:-
  • Supported by:
    -

Abstract: In order to advance the efficiency for analyzing the reliability of complex mechanical structure, a new adaptive learning method(AK-MCS-K), which combines adaptive learning function VF with k-means clustering, is proposed. The AK-MCS-K method balances the precision of failure probability estimation and computational efficiency. It can realize parallel computation, that is, several samples can be simulated simultaneously at every iteration on a few computers. When evaluating the reliability of complex structures with time-consuming and implicit function, reducing iterations can effectively save time and improve computational efficiency. Compared with others, the AK-MCS-K method can reduce the number of iterations significantly with higher computational efficiency under the condition of satisfying the precision. Finally, the new method is used for analyzing the reliability of the motion accuracy of a simplified model of a rigid-flexible coupling artillery coordinator.

Key words: AK-MCS method, learning function VF, k-means algorithm, complex mechanical structure, reliability analysis

CLC Number: