Modeling Xanthan Gum Foam's Material Properties Using Machine Learning Methods.

Area of Science:

• Materials Science and Engineering
• Biopolymer Applications
• Sustainable Materials Development

Background:

• Xanthan gum, a natural biopolymer, is widely used in pharmaceuticals, cosmetics, and food industries.
• Limited research exists on its application as a foam material for insulation and packaging.
• Artificial neural network (ANN) modeling has not been applied to xanthan gum-based foams.

Purpose of the Study:

• To investigate the production of novel foam materials using xanthan gum and cellulose fiber.
• To evaluate the physical and mechanical properties of these bio-based foams.
• To model the foam properties using various machine learning techniques, including ANNs.

Main Methods:

• Foam materials were produced using varying ratios of xanthan gum and cellulose fiber in a citric acid medium.
• Physical and mechanical properties (density, compressive modulus, flexural modulus) were experimentally determined.
• Five machine learning models (MLR, SVM, ANN, LS, GRNN) were employed for data analysis and prediction.

Main Results:

• Xanthan gum significantly influenced foam properties more than cellulose.
• Foam densities ranged from 49.42 to 172.2 kg/m³.
• Compressive moduli varied between 235.25-1257.52 KPa, and flexural moduli between 1939.76-12,736.39 KPa.
• Machine learning models accurately predicted experimental results, enhancing process efficiency.
• The generalized regression neural network (GRNN) model achieved R² values > 0.97 for predicting key properties.

Conclusions:

• Xanthan gum is a promising biopolymer for developing sustainable foam materials for insulation and packaging.
• Machine learning, particularly GRNN, effectively models and predicts the properties of these novel bio-foams.
• This approach allows for efficient acquisition of quantitative data, reducing experimental costs and time.