Supercritical CO2 extraction plays a significant role in the cannabis industry. However, there are many factors that affect extraction outcomes, such as extractor temperature and pressure, flow-rate, and separator conditions. Previously, DELIC Labs developed an extraction optimization process and have accumulated a large and diverse database of cannabis extraction runs. This database has allowed the data scientists’ team at DELIC Labs to save clients time, money, and resources by leveraging their proprietary ML algorithm to create various computational models to mathematically and statistically identify optimal extraction parameters, a process that is unmatched by other licensed cannabis R&D labs and undergoing constant enhancement.
The animation above illustrates one of our computational models, using total CBD recovery as our prediction surface. For this particular model, three parameters were considered: residence time, extraction pressure, and extraction run-time. To keep the three parameters measurable and consistent, the values were normalized from -1 to 1, allowing DELIC Labs to track the trends and explore ways to optimize extraction parameters.
Through our industry-leading data team, DELIC Labs hopes to bring our computational models and ML-related technical skills to better serve not only our customers but the overall industry as well in this fast-growing and competitive market.
Written by: Dr. Tom Dupree and Duo Lu