Artificial intelligence-driven control for enhancing carbon dioxide-based wastewater pH regulation in tubular reactor

Abstract

Alkaline wastewater treatment using carbon dioxide can reduce chemical costs and provide a safer alternative to traditional methods. However, complex gas-liquid reactions and narrow operating pH ranges present challenges. This research develops an artificial intelligence-driven control system for treating alkaline wastewater using carbon dioxide in a bench-scale tubular reactor. The proposed control system employs an inverse neural network to regulate the carbon dioxide gas based on the desired setpoint, along with a Smith predictor and a linear controller to compensate for natural delays, model mismatches, and pH disturbances. The inverse neural controller was trained using experimental data from a bench-scale reactor pH treatment of synthetic alkaline wastewater and verified on real influent from an electroplating wastewater treatment plant. The results show that the proposed method efficiently enforces the desired reactor outlet pH setpoint with up to 51.36% faster settling time than a proportional-integral controller while improving pH-adjusting efficiency by 72.24%.