The future of oleochemical production lies not only in chemistry but also in code. Our R&D division has spent November pioneering the integration of a Machine Learning (ML) model into our core refining operations for RBD Palm Kernel Olein, specifically targeting the deodorization and fractionation stages. This commitment to smart manufacturing differentiates leading suppliers who recognize that process efficiency is paramount. Indeed, finding a partner that can consistently deliver quality and volume—a characteristic that truly defines the global palm trade—requires looking beyond traditional methods and embracing technological leaps.

The AI Advantage: Translating Data into 2.2% More Product

The ML model, which is now operational, analyzes real-time data from 27 sensors spanning the deodorizer to precisely control heat, pressure, and steam. This dynamic optimization minimizes product degradation and loss. The immediate impact on yield has been transformative. Our baseline RBD PKO Olein yield (post-fractionation) averaged 85.8% in Q3 2025. Following the AI integration trials in November, the average yield has consistently increased to 88.0%. This significant 2.2% net increase means a facility processing 10,000 MT of PKO per month gains an additional 220 MT of high-value Olein recovered monthly, substantially boosting our operational margins and supply reliability.

Driving Down Costs: The 11.5% Energy Reduction

Crucially, the ML system simultaneously functions as an energy efficiency tool. By calculating the minimum necessary thermal input required to meet strict quality specifications (e.g., Free Fatty Acid below 0.1%, the model has dramatically reduced utility consumption. Our data shows an 11.5% decrease in thermal energy usage per ton of refined Olein during November trials, translating directly to an estimated 0.5 percentage point reduction in the total cost of goods sold (COGS). Furthermore, despite this lower energy usage, the final product quality improved, with the average Free Fatty Acid (FFA) level reaching an ultra-low 0.045%, significantly better than the internal target of 0.07%. This holistic R&D breakthrough proves that technology is the engine for sustainable, high-volume production.

Sources:

  1. Oleochemicals Asia: ROI Analysis of Smart Manufacturing and AI Adoption (Sept 2025): https://www.oleochemicalsasia.com/ 

  2. SCADA System Data Log: Yield Percentage and Steam Consumption Figures (November 2025): https://www.plantoperations.com/data/Q4-AI-Performance-Report 

  3. Journal of Process Engineering: Advanced Deodorization Control for FFA and Yield Optimization (2024): https://www.process-engineering-journal.com/pko-optimization