Modeling and Dynamic Parameterized Predictive Control of Dissolved Oxygen in Dual−Tank Bioreactor Systems

Uneven distribution and delayed system response of dissolved oxygen (DO) in dual−tank recirculating bioreactor systems pose significant challenges for oxygen supply. To address these issues, a dynamic parameterized predictive control (DPPC) approach is proposed and validated through simulation and bench−scale experiments. This method is underpinned by a mathematical model that integrates mass transfer kinetics and chemical thermodynamic principles, accurately capturing oxygen dissolution and transfer within a recirculating environment. By predicting future DO variations and continuously integrating real−time monitoring data, the controller adjusts oxygen injection parameters in real time, rapidly restoring DO levels to target values while minimizing overshoot and latency introduced by system circulation. Experimental results in dual−tank setups show an RMSE below 0.05 and an R² exceeding 0.99, affirming the model’s predictive accuracy under varying oxygen conditions. Compared with conventional feedback control strategies, the proposed method demonstrates improved stability, faster response, and lower overshoot, achieving a 47.8% reduction in ISE and a 41.4% reduction in IAE, thus highlighting its superior tracking accuracy. These findings suggest the DPPC method holds promise as a control framework for future application in oxygen−sensitive culture systems.