SASEL Lab
SASEL Lab
SASEL Lab
An NRC-funded digital twin platform for pilot-scale pea and faba bean protein fractionation, built to optimize quality, cost, and environmental performance in a single loop. Delivered an interactive decision-support dashboard for Canadian processors. Completed March 2026.
This project develops a digital twin simulation platform for dry fractionation of yellow pea and faba bean proteins. By combining multiphysics models with pilot-scale experimental data and machine learning, the platform enables multi-objective optimization across product quality, resource use, economic return, and environmental footprint. The outcome is an interactive decision support dashboard that helps Canadian processors select operating conditions that balance purity, safety, profitability, and sustainability.
Pulse proteins, particularly yellow peas and faba beans, have attracted significant investment across the Canadian Prairies, including the world's largest pea protein production plant in Manitoba with a processing capacity of 250,000 tonnes per year. Alongside this growth, the industry generates 60 to 70 percent of feedstock mass as co-products such as starch, fiber, micronutrients, phytochemicals, and lipids. Efficient and sustainable fractionation is therefore central to the Canadian bioeconomy.
Radio Frequency (RF) pre-treatment of pulses prior to protein extraction has been shown to improve fractionation efficiency and protein concentrate functionality. Full industrial adoption, however, requires process configurations that simultaneously maximize product quality, resource use, and economic return while minimizing environmental impact and addressing food safety. Experimental optimization alone is limited: it can only vary a small number of conditions at a time, it is time-consuming, and it produces models that rarely extrapolate beyond the tested range. Generic computational models, meanwhile, do not capture the physical and biochemical realities of pulse protein extraction.
This project addresses that gap by building a digital twin of pilot-scale pea and faba bean flour and protein fractionation. The model integrates physics-based simulation with real-time pilot data and machine learning so it can reflect, predict, and optimize process behavior. Stakeholders will be able to explore scenarios, identify optimal operating conditions, and quantify the trade-offs among product quality, cost, and environmental performance. The results will be delivered through an interactive decision support dashboard designed for industrial processors and research partners.
Use a digital twin approach to optimize Radio Frequency pre-treatment of pea and faba beans, improving flavour, reducing microbial load and anti-nutrient levels, and enhancing nutritional quality and functionality of the resulting flours.
Leverage digital twin simulation models to develop a product-quality based eco-efficiency assessment for pilot-scale air classification, supporting optimized production of pea and faba bean protein concentrates.
Build a digital twin based decision support dashboard that guides sustainable protein fractionation, informs operational choices, and helps resolve supply chain challenges.
Covers development of baseline multiphysics models in COMSOL, pilot-scale experimental trials in collaboration with NRC Saskatoon, training of the digital twin using regression neural networks, and characterization of physicochemical and techno-functional properties at optimized conditions.
Combines Discrete Element Method simulation of particle behaviour in the air classifier with pilot-scale sensor-enabled testing and machine learning to create a digital twin of the dry fractionation process. The model is then used for multi-objective optimization across protein yield, purity, cost, and environmental impact.
Structures data pipelines and builds a scalable ingestion, storage, and analytics stack using Apache Sedona, a NoSQL database, and Kibana. Delivers interactive dashboards built with modern web technologies. The platform is tested at pilot-scale and extended to pea protein supply chain scenarios.
Timeline Mar 2024 — Mar 2026
Research Areas
National Research Council of Canada, Saskatoon
Collaborator: Dr. Anusha Samaranayaka
Canadian pulse processing industry
Pilot validation and uptake
National Research Council of Canada (NRC), Canadian Sustainable Transformation Innovation Program (CSTIP)