Study the hydrotropic behaviour of butyl stearate using ANN tools

Area of Science:

• Physical Chemistry
• Materials Science
• Computational Chemistry

Background:

• Thermophysical properties are crucial for understanding solution behavior.
• Predicting these properties in complex mixtures can be challenging.
• Artificial Neural Networks (ANNs) offer a powerful tool for modeling complex relationships.

Purpose of the Study:

• To develop and validate an Artificial Neural Network (ANN) model for predicting thermophysical properties of butyl stearate in binary mixtures.
• To investigate the influence of hydrotropic concentration and temperature on these properties.
• To demonstrate the ANN model's accuracy and efficiency compared to traditional methods.

Main Methods:

• Experimental data collection for butyl stearate solutions with citric acid, urea, and nicotinamide at various temperatures (303-333 K).
• Training a committee of ANNs using iterative optimization to avoid overfitting.
• Utilizing visualizations (e.g., contour plots, 3D surface plots) for analysis and optimization.
• Implementing the ANN model in MATLAB for predictive tasks.

Main Results:

• The ANN model accurately predicted thermophysical properties of butyl stearate solutions.
• The model achieved a remarkable 2% margin of error in identifying optimal hydrotropic concentrations.
• The ANN approach proved versatile, showing success in predicting multi-pass turning operations.

Conclusions:

• ANNs provide a highly accurate and efficient method for predicting thermophysical properties in complex solutions.
• The developed model can predict properties at intermediate concentrations without further experimentation.
• This study highlights the broad applicability of ANNs in chemical and engineering applications.