Aeroengine surge prediction method based on multi-branch feature fusion network

Through the method based on the multi-branch feature fusion network, multiple features in the operation data of the aircraft engine are extracted and fused, which solves the problem of insufficient surge prediction accuracy in the prior art, and achieves more accurate surge prediction and early warning.

CN112580267BActive Publication Date: 2025-02-28NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
View PDF 4 Cites 1 Cited by

Patent Information

Application Number
CN202110040087.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-01-13
Publication Date
2025-02-28
Estimated Expiration
2041-01-13

Smart Images

  • Figure CN112580267B_ABST
    Figure CN112580267B_ABST
Patent Text Reader

Abstract

The present invention proposes an aircraft engine surge prediction method based on a multi-branch feature fusion network, which belongs to the field of active control of aircraft engine stability. The steps are: (1) collecting the operating data of the engine surge test; (2) pre-processing the engine operating data and using a sliding window to divide the samples; (3) using short-time Fourier transform to monitor the power change of the specific frequency signal component, determine the surge state of the sample and set the label to make a data set; (4) preliminarily extracting the time domain statistical features, time-frequency spectrum features and whole machine parameter change features of the sample; (5) constructing a multi-branch feature fusion network based on FC-VGG16-TCN and training it; (6) inputting the real-time operating parameters of the engine into the trained network to detect the precursors of instability online. Based on deep learning and feature fusion algorithms, the present invention accurately detects the precursors of instability through deep feature extraction and classification of the whole engine operating data, makes surge warnings in advance, and provides sufficient response time for active control.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Early warning method for rotating stall of axial-flow air compressor based on frequency domain hump identification

    CN107165850A

  • Online recognition method of aero-engine surge precursor

    CN110610026A

  • Axial flow compressor stall surge prediction method based on deep learning

    CN111737910A

  • Axial flow compressor stall surge prediction method based on deep learning

    CN112001128A

Cited By

  • Compressor stall time-space intelligent early warning method and system based on graph deviation neural network

    CN122527506A