Synergistic Application of Dimensional Analysis to Optimize Virgin Coconut Oil Press

Abstract

This research explores the optimization of virgin coconut oil (VCO) extraction by integrating dimensional analysis with experimental validation. It aims to create mathematical models that predict machine efficiency, throughput capacity, and oil yield quality based on key variables such as pressure, grind size, moisture level, viscosity, feed rate, and temperature. By applying Buckingham’s π theorem, dimensionless groups were formulated to define the functional relationships that regulate oil extraction behavior. The approach employed included constructing a dimensional matrix, developing parametric models, and deriving π-terms, followed by experimental verification. The created models were evaluated against the actual performance of the machine, resulting in regression coefficients (R²) of 0.0014 for efficiency, 0.0345 for throughput capacity, and 0.4009 for oil yield quality. These values indicate that, while the oil yield model had reasonable predictive strength, the efficiency and throughput models showed little correlation under the testing conditions, indicating opportunities for improvement. Nevertheless, the models captured key trends and exhibited promise for assisting in optimization efforts. This study highlights the effectiveness of dimensional analysis as an economical approach for improving VCO extraction, especially in low-resource settings. Suggested actions include refining the π-terms, applying machine learning techniques to account for non-linear effects, and performing field validations to enhance real-world relevance.