Aerodynamic System Machine Learning Modeling with Gray Wolf Optimization Support Vector Regression and Instability...

Area of Science:

• Aerospace Engineering
• Computational Fluid Dynamics
• Machine Learning

Background:

• Stable aeroengine operation relies on predicting stall precursors in axial compressors.
• Early identification of system instability is crucial for safety and active control.

Purpose of the Study:

• To propose an advanced aerodynamic system modeling method for predicting stall precursors.
• To enhance the accuracy and reliability of instability prediction in aeroengines.

Main Methods:

• A hybrid model combining Wavelet Transform (WT) and Gray Wolf Algorithm optimized Support Vector Regression (GWO-SVR) was developed.
• The correlation-correlation (C-C) algorithm reconstructed phase space from chaotic data.
• A wavelet singular information entropy algorithm was used for stall inception detection.

Main Results:

• The WT-GWO-SVR model demonstrated superior prediction accuracy and reliability using wavelet reconstruction coefficients.
• The proposed method achieved early instability warnings approximately 102 revolutions in advance.
• The wavelet singular information entropy algorithm effectively identified stall inception.

Conclusions:

• The WT-GWO-SVR model offers a robust approach for predicting stall precursors in axial compressors.
• Early warning capabilities significantly aid in the active control and safety of aeroengines.
• This research advances intelligent identification of system instability in complex aerodynamic systems.