Quantitative modeling of extrusion outcomes through QSPR-finite element approach

Overview

This research-based solution delivers predictive modeling capabilities for food extrusion processes, specifically targeting high-moisture and high-protein food matrices such as cereals and legumes. The approach combines quantitative structure-property relationships (QSPR) with finite element method (FEM) analysis to forecast key output variables including thermal properties, rheological behavior, texture, expansion index, chemical composition, starch and protein digestibility, and sensory attributes.

The platform addresses significant limitations of commercial extruders, which often struggle with high or low moisture content, varied particle sizes, and complex chemical compositions involving high fiber, protein, or carbohydrate content. By enabling more accurate prediction of extrusion outcomes, food manufacturers and product developers can optimize formulations, reduce trial-and-error experimentation, and accelerate product development cycles for nutritious snack products and other extruded food items.

Technical specifications

Key capabilities:

• Nonlinear system identification procedures to model complex relationships between physico-chemical inputs and extrusion outputs
• Finite element method (FEM) integration for thermal and rheological property prediction
• Quantitative structure-property relationship (QSPR) modeling linking molecular and compositional properties to product characteristics
• Prediction of physical attributes including texture and expansion index
• Modeling of biological activities such as starch and protein digestibility
• Coverage of diverse feedstocks including maize, sorghum, beans, and soy
• Custom-designed extruder prototypes capable of processing high-moisture, high-protein, high-fiber, and high-carbohydrate materials across different particle sizes
• Robust experimental design framework for controlled validation studies

Technology readiness level

The modeling approach is grounded in extensive prior research conducted at Universidad Autónoma de Querétaro, Instituto Politécnico Nacional - CICATA, and Tecnológico de Monterrey. The team has validated extrusion configurations with cereals and legumes to produce snack products with high nutritional properties.

Future validation is structured in four phases: determining target plant protein inputs and variables, performing controlled extrusion runs using both proprietary prototypes and potential commercial extruder access, building a quantitative database of input and output variables, and applying linear and nonlinear correlation methods to obtain and validate QSPR models. This staged approach supports progression toward a deployable predictive modeling capability for industrial food extrusion applications.