Establishment of Aero-Engine Improved On-Board Adaptive Model with Contracted Kalman Filter Estimation

Zhidan Liu, Linfeng Gou, Ding Fan, Chujia Sun

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

2 Scopus citations

Abstract

Due to the limited number of sensors in the aeroengine, the estimation of performance degradation of the engine is inaccurate. In this paper, an improved on-board adaptive model with contracted Kalman filter is proposed. This is accomplished by constructing a transformation matrix to reduce the dimension of the health parameter vector, and taking the estimated deviation of the simplified Kalman filter as the goal of minimization, and using genetic algorithm and deep learning models to build more accurate health parameters to reflect the engine performance. Use the inverse transform to obtain the original health parameters. The simulation results show that the improved on-board adaptive model established in this paper can still estimate the health parameters with high accuracy when the measured parameter dimension is less than the health parameter dimension, and the accuracy of improved on-board adaptive model is very high.

Original languageEnglish
Title of host publicationProceedings of the 40th Chinese Control Conference, CCC 2021
EditorsChen Peng, Jian Sun
PublisherIEEE Computer Society
Pages1379-1383
Number of pages5
ISBN (Electronic)9789881563804
DOIs
StatePublished - 26 Jul 2021
Event40th Chinese Control Conference, CCC 2021 - Shanghai, China
Duration: 26 Jul 202128 Jul 2021

Publication series

NameChinese Control Conference, CCC
Volume2021-July
ISSN (Print)1934-1768
ISSN (Electronic)2161-2927

Conference

Conference40th Chinese Control Conference, CCC 2021
Country/TerritoryChina
CityShanghai
Period26/07/2128/07/21

Keywords

  • aeroengine
  • contracted Kalman filter
  • genetic algorithm
  • performance degradation
  • transformation matrix

Fingerprint

Dive into the research topics of 'Establishment of Aero-Engine Improved On-Board Adaptive Model with Contracted Kalman Filter Estimation'. Together they form a unique fingerprint.

Cite this