Difficulty in optimizing early fermentation operations
In pharmaceutical fermentation, where microorganisms produce target substances, stable operation for up to 400 hours is required. The initial 20–50 hours are especially critical, as they largely determine the final yield. However, manual operation often results in control deviations caused by human error, operator variability, and fatigue. Additionally, complex multivariate and nonlinear data, sensor discrepancies, and microbial variability further complicate PLC automation.
Developing optimal operation algorithms from PLC data and domain knowledge
PLC data were integrated with existing process rules and operation recipes to design standardized control algorithms. Parser programming was applied to formalize hidden operational rules and field logic based on domain expertise, enabling the development of optimal control algorithms that enhance productivity and enable full automation.
Achieving full automation and higher control accuracy
Two PLC-controlled fermentation units were fully automated, reducing control deviation from 4.3% to 0.2% compared to manual operation.The system also provides real-time recommendations for optimal operation timing and actions, eliminating human error and ensuring consistent, repeatable performance across operators.
