An Enhanced Network Learning Method for Dynamic Probabilistic LCF Evaluation of Turbine Blisk

Cheng Wei Fei, Cheng Lu, Wen Zhong Tang

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

To reasonably predict the blisk LCF life, an enhanced network learning method (ENLM) was proposed by integrating neural network of regression with extremum thought. The developed ENLM was mathematically modeled and the method of reliability analysis with the ENLM were investigated. Then the LCF life reliability of a turbine blisk were evaluated utilizing the proposed ENLM by considering thermal-structural interaction and the stochastic characteristics of parameters. It is illustrated in the analysis to the reliability degree of blisk (0.998, 5) and blisk LCF life (9, 419 cycles) under the allowable value 6, 000 cycles, which ensures some life margin for the blisk relative to the deterministic analysis. The comparison of methods reveals that the proposed ENLM has highly-computing efficiency and precision. The advantages of this method will more outstand with the increasing simulations. The outcome of this paper is to develop a promising tool for the reliability-based design optimization of gas turbine blisk LCF life in future work.

Original languageEnglish
Title of host publication2019 Prognostics and System Health Management Conference, PHAI-Qingdao 2019
EditorsWei Guo, Steven Li, Qiang Miao
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728108612
DOIs
StatePublished - Oct 2019
Event10th Prognostics and System Health Management Conference, PHM-Qingdao 2019 - Qingdao, China
Duration: 25 Oct 201927 Oct 2019

Publication series

Name2019 Prognostics and System Health Management Conference, PHM-Qingdao 2019

Conference

Conference10th Prognostics and System Health Management Conference, PHM-Qingdao 2019
Country/TerritoryChina
CityQingdao
Period25/10/1927/10/19

Keywords

  • ENLM
  • extremum response
  • LCF life
  • probabilistic analysis
  • turbine blisk

Fingerprint

Dive into the research topics of 'An Enhanced Network Learning Method for Dynamic Probabilistic LCF Evaluation of Turbine Blisk'. Together they form a unique fingerprint.

Cite this