Research on aero-engine physics-based model correction method based on mechanism fusion residual

Shu-bo Zhang, Qian-gang Zheng, Cheng Chen, Chang-peng Cai, Hai-bo Zhang, Yuan-wei Mou and Feng-ming Wang

Energy, 2025, vol. 332, issue C

Abstract: Physics-based model (PBM) accuracy plays an important role in the development of aircraft engine control system. However, the traditional physical modeling method affects the accuracy of PBM when dealing with uncertainties such as engine and model mismatch and actual complex flight conditions. In this paper, a new method of improving hybrid model and modifying model is proposed. Firstly, from the perspective of model mechanism, the mismatch problem of model gas path characteristics was solved, and the chunked cross-reliability domain method (CCD) was innovated to modify the characteristics of the PBM, which could effectively improve the repair parameter space of its local region. CCD can significantly improve the steady-state accuracy of the model, and the average steady-state error of the model is reduced from 1.13 % to 0.51 %, and the steady-state correction effect of the model is increased by 54 %. Secondly, based on CCD correction of PBM, residual fusion method is adopted to reduce the uncertainty error caused by PBM under actual complex flight conditions. Introduced the Local Linear Embedding (LLE) method to integrate the features of input data and reduce redundant features in the input data. On this basis, a residual fusion model (RFM) is constructed using a time-domain convolutional attention mechanism network (TCAM). Specifically, a series of residual data is used to establish RFM and PBM residual fusion for each parameter, and TCAM is independently selected for training. Finally, we validate the accuracy of the model using actual flight data. The average RSME value of RFM-TCAM is 0.22, with an R2 of 0.983. The residual learning performance of RFM-TCAM is better than that of traditional models. In summary, the mechanism fusion residual model correction method demonstrates stronger effectiveness and superiority.

Keywords: Aircraft engine; Physics-based model; Chunked cross-reliability domain method; Residual fusion model; Time-domain convolutional attention mechanism network (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0360544225025514
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:energy:v:332:y:2025:i:c:s0360544225025514

DOI: 10.1016/j.energy.2025.136909

Access Statistics for this article

Energy is currently edited by Henrik Lund and Mark J. Kaiser

More articles in Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().