Development of a Digital Image Processing- and Machine Learning-Based Approach to Predict the Morphology and Thermal...

Area of Science:

• Materials Science
• Computational Chemistry
• Polymer Science

Background:

• Polyurethane foams are vital for thermal insulation due to their porous structure.
• The synthesis of these foams involves blowing agents that create insulating pores.
• Understanding pore stability and neighboring effects is crucial for optimizing thermal performance.

Purpose of the Study:

• To computationally analyze pore stability and neighboring effects in polyurethane foams synthesized with water and cyclohexane.
• To develop a digital image processing and machine learning algorithm for predicting foam properties.
• To establish a correlation between pore structure analysis and thermal conductivity.

Main Methods:

• Synthesis of five distinct polyurethane foam samples using water and cyclohexane blowing agents.
• Computational analysis of pore stability and neighboring effects using MP2/aug-cc-pVDZ.
• Development of a machine learning algorithm based on Voronoi tessellation for analyzing pore neighboring effect distances from digital images.

Main Results:

• A strong correlation was identified between calculated Voronoi neighboring effect distances and the thermal conductivity coefficients of the polyurethane foams.
• The developed algorithm successfully predicted the mean neighboring effect distances.
• The mobile phone-based imaging and Python algorithm offered a faster, non-destructive analysis compared to Scanning Electron Microscopy (SEM).

Conclusions:

• The Voronoi neighboring effect distance is a reliable indicator for predicting the thermal conductivity of polyurethane foams.
• This novel, accessible method provides a rapid and non-destructive approach to foam characterization.
• The findings enable the prediction of thermal conductivity for similar polyurethane foams using standard mobile phone imaging.