Chemical dosage optimization in kitchen wastewater treatment via an integrated bio-chemical model.

To address the challenges of achieving organic compliance in kitchen wastewater treatment and the high cost of chemical dosing, this study established a two-stage reactor incorporating both biological and chemical processes and proposed an optimal regulation system. The reactor achieves the effluent COD standard compliance. Additionally, both exceeding-standard and standard-compliant data were collected from the reactor. This avoids the limitation that relying solely on standard-compliant data would prevent chemical dosing optimization from reaching the optimal solution. By adopting the XGBoost algorithm, biological and chemical prediction models for effluent COD are constructed. The R2 values of the biological model and the chemical model for accuracy are 0.879 and 0.904, respectively. Subsequently, a chemical dosage optimization method for wastewater treatment processes integrating the Bio-Chemical model and particle swarm optimization (BCPSO) is proposed. After optimization, all the original standard-exceeding COD samples meet the discharge standards. The average effluent COD concentration is reduced by 31.17%, and the average total removal rate is increased by 4.82%. On the premise of ensuring effluent standard compliance, the optimization of originally standard-compliant samples reduces the chemical cost by 22.45%. The BCPSO proposed in this study can be extended to the treatment processes of other refractory organic wastewater, reducing the operating costs without posing a threat to the environment.