Abstract A novel semi-mechanistic data-driven technique based on an evolutionary polynomial regression (EPR) was developed along with a kinetic analysis to evaluate the performance of an external loop upflow partially-aerated fixed-film bioreactor (UP/ASFF). Modelling of experimental data was carried out under different organic loading rates (OLR), solid retention time (SRT), and food to microorganisms’ ratio (F/M ratio). The results showed that the EPR model was the best approach for the performance evaluation of the UP/ASFF system with R2 of 0.93 and normalized MSE of 0.02 although the Stover-Kincannon’s model showed also a promising predication with high accuracy and R2 of 0.92.

中文翻译：

---

部分曝气浸没固定膜生物反应器的性能建模：机械分析与半数据驱动方法

摘要 开发了一种基于进化多项式回归 (EPR) 的新型半机械数据驱动技术以及动力学分析，以评估外部循环上流式部分曝气固定膜生物反应器 (UP/ASFF) 的性能。在不同的有机负载率 (OLR)、固体保留时间 (SRT) 和食物与微生物的比率 (F/M 比率) 下对实验数据进行建模。结果表明，EPR 模型是 UP/ASFF 系统性能评估的最佳方法，R2 为 0.93，归一化 MSE 为 0.02，尽管 Stover-Kincannon 模型也显示出具有高准确度和 R2 为 0.92 的有前景的预测。