Auxiliary power unit failure prediction using quantified generalized renewal process

Yujie Zhang, Lulu Wang, Shaonian Wang, Peng Wang, Haitao Liao, Yu Peng

Research output: Contribution to journalArticle

9 Scopus citations

Abstract

Auxiliary Power Unit (APU) is an essential component utilized in modern civil aircraft. In order to meet APU availability requirements, failure prediction should be performed in an effective way. To this end, APU performance deterioration can be modeled by a Generalized Renewal Process (GRP). In this paper, aircraft APU failure prediction is implemented using a Weibull-based GRP (WGRP). However, the effect of maintenance activities on aircraft APU required in WGRP is difficult to be quantified. To solve this problem, a new model named Quantified Generalized Renewal Process (QGRP) is developed in this paper. In this model, APU performance-related test parameters after repairs are utilized to quantify the maintenance effect. Based on the proposed QGRP model, the conditional failure rate and hazard rate of each aircraft APU at a future point in time can be calculated based on the APU's virtual age and be combined to predict the number of failures of a fleet of APUs. The performance of the proposed QGRP model is validated using a three-year data set provided by China Southern Airlines. The results show that the QGRP model is effective in aircraft APU failure prediction.

Original languageEnglish (US)
Pages (from-to)215-225
Number of pages11
JournalMicroelectronics Reliability
Volume84
DOIs
StatePublished - May 1 2018
Externally publishedYes

Keywords

  • APU
  • Condition-based maintenance
  • Deterioration
  • Failure prediction
  • QGRP
  • Repairable systems
  • Virtual age

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Atomic and Molecular Physics, and Optics
  • Safety, Risk, Reliability and Quality
  • Condensed Matter Physics
  • Surfaces, Coatings and Films
  • Electrical and Electronic Engineering

Fingerprint Dive into the research topics of 'Auxiliary power unit failure prediction using quantified generalized renewal process'. Together they form a unique fingerprint.

  • Cite this