A Three-Dimensional Microstructure Reconstruction Framework for Permeable Pavement Analysis Based on 3D-IWGAN with...

Area of Science:

• Civil Engineering
• Materials Science
• Computational Modeling

Background:

• Permeable pavements are increasingly used, but clogging reduces their functionality.
• Traditional clogging analysis methods are costly and time-consuming.
• There's a need for efficient methods to study permeable pavement design and durability.

Purpose of the Study:

• To propose an image-based computational system for analyzing permeable pavement clogging.
• To overcome the limitations of traditional physical experiments.
• To improve the design and durability of permeable pavements.

Main Methods:

• Developed a 3D microstructure reconstruction framework using 3D-IDWGAN with an enhanced gradient penalty.
• Input: two-dimensional (2D) image of pavement microstructure.
• Output: 3D microstructure image generated via a generative adversarial network (GAN).

Main Results:

• The proposed system effectively generates 3D microstructure images.
• Achieved improved 3D image generation compared to other GAN methods.
• Extracted physical properties from the reconstructed 3D model closely matched those from real pavement samples.

Conclusions:

• The 3D-IDWGAN framework offers a cost-effective and time-efficient alternative for permeable pavement clogging analysis.
• This computational approach aids in understanding and improving permeable pavement performance.
• The method shows promise for enhancing the design and durability of permeable pavement systems.