FOSS-Based Method for Thin-Walled Structure Deformation Perception and Shape Reconstruction.

Area of Science:

• Engineering
• Materials Science
• Computational Mechanics

Background:

• Flexible thin-walled structures are crucial in aerospace and renewable energy.
• Accurate deformation perception and shape reconstruction are vital for their performance and safety.
• Existing methods often face challenges in precision and real-time monitoring.

Purpose of the Study:

• To develop an accurate method for deformation perception and shape reconstruction of flexible thin-walled structures.
• To integrate fiber optic sensor systems (FOSS) with machine learning techniques.
• To provide a new approach for real-time monitoring and analysis.

Main Methods:

• Utilized ANSYS finite element analysis for strain measurement and deformation data collection.
• Employed the one-class support vector machine (OCSVM) model to identify and remove outliers.
• Developed a neural network model to establish a mapping relationship between strain values and deformation variables (x, y, z axes).

Main Results:

• The proposed method achieved maximum errors of 2.01% (x-axis), 29.49% (y-axis), and 15.52% (z-axis).
• Despite higher errors in y and z directions for small deformations, the reconstructed shape showed good consistency with the actual deformation state.
• The integration of FOSS and machine learning proved effective for deformation analysis.

Conclusions:

• The combined FOSS and machine learning method offers a highly accurate solution for real-time monitoring and shape reconstruction of flexible thin-walled structures.
• This approach provides a valuable tool for analyzing structures like wings, helicopter blades, and solar panels.
• The study demonstrates the potential of advanced computational and sensing techniques in structural health monitoring.