Advancements in AI-Driven Process Optimization and Quality Control for Edible Oils in Industry 4.0
Document Type : Review article
Authors
1 Quality Control and Quality Assurance Departments, Newsha Darian Agro-Industry Co., Tehran, Iran. Postal Code 1443955511
2 Department of Food Science and Engineering, College of Agriculture and Natural Resources, University of Tehran, Karaj Campus, Karaj, Iran
Abstract
Traditional quality control approaches are often reactive, labor-intensive, and limited in scalability, responsiveness, and precision. In contrast, AI and ML are transforming the edible oil manufacturing industry. They enable process optimization, real-time monitoring, and advanced quality control in line with Industry 4.0. This study reviews recent research and applications of AI and ML in edible oil extraction processes and quality control. It focuses on optimizing extraction parameters and yield, minimizing impurities, and ensuring safety to enable sustainable, intelligent production. Advanced algorithms such as ANNs and ANFIS offer superior accuracy for optimizing extraction, predicting antioxidant content, and controlling processes compared to conventional methods. For quality control, AI has enabled rapid, nondestructive assessments of oil authenticity and oxidation. Technologies such as LF-NMR, combined with CNNs, are used. AIoT sensor-based systems integrate intelligent sensors, cloud platforms, and deep learning models such as LSTM, ANNs, and CNNs. These systems enable real-time monitoring of rancidity, as well as volatile gas emissions and color changes during storage. Other advanced AI-driven innovations include image-based defect detection using DMEOI datasets and infrared cameras for real-time inspection. Emerging techniques such as HSI with ML, BME688 gas sensors, voltammetric electronic tongues, and visual array sensors detect adulteration in pure and blended oils. FPGAs are also used for real-time detection of gutter oils. Despite these advances, widespread industrial adoption faces challenges. Key issues include data quality, privacy, cybersecurity, workforce skills, and integration with legacy systems. Addressing these data issues is a major concern for industry and academia.
Keywords
Main Subjects
)