Flight route EDR estimation using MHE to fuse three-dimensional wind information in the QAR data analysis

Area of Science:

• Aerospace Engineering
• Atmospheric Science
• Data Science

Background:

• Vertical wind data alone limits turbulence information, reducing accuracy in eddy dissipation rate (EDR) estimation.
• Current methods for EDR estimation using vertical wind (VWE) lack comprehensive turbulence data, leading to inaccuracies.

Purpose of the Study:

• To develop an improved flight route EDR estimation method by fusing three-dimensional wind information.
• To enhance the accuracy and applicability of EDR estimation for assessing turbulence intensity along flight routes.

Main Methods:

• Quality control of Quick Access Recorder (QAR) data to extract vertical and horizontal wind components.
• Utilizing a multi-head attention mechanism (MHE) to fuse 3D wind characteristics with vertical acceleration data.
• Employing maximum likelihood estimation (MLE) in the frequency domain for final EDR calculation.

Main Results:

• The proposed algorithm significantly outperforms existing methods (PCE, VWE) in EDR estimation accuracy.
• Estimated EDR values show reduced errors compared to true EDR values.
• Demonstrated optimal performance and practical applicability for flight path turbulence assessment.

Conclusions:

• The novel approach effectively fuses 3D wind data for highly accurate flight route EDR estimation.
• This method enhances the assessment of turbulence intensity, contributing to improved aircraft safety.
• The algorithm shows significant practical value for real-world aviation safety applications.