Process modelling is an effective tool for describing and predicting the performance of an aerobic membrane bioreactor (AeMBR) for fish canning wastewater (FCWW) treatment under different operating conditions. Three machine learning (ML) algorithms were developed, random forest (RF), decision tree regressor (DTR) and adaptive boosting regression (AdaBoost-R), based on various physico-chemical characteristics of the influent and operating conditions, including hydraulic retention times (HRT), organic loading rates (OLR), total dissolved solids (TDS), aeration rate and permeate volumetric rates. Predicted values for chemical oxygen demand (COD), biochemical oxygen demand (BOD5), total Kjeldahl nitrogen (TKN), and nitrate (NO3−) are compared with those reported from the experiment. As regards the quantitative assessment of the three predictive models, the DTR model demonstrated a modest determination coefficient (R2) value of 0.654, the AdaBoost-R model achieved 0.739, whereas the RF model showed the highest performance at 0.98. Due to its robustness and accuracy, the RF model was chosen for its superior ability to predict the performance of the AeMBR. Based on OLR of 4.27 (kg COD)∙(m3∙d)−1, a HRT of 24 h, a TDS of 3 g∙dm−3, an aeration rate of 1,300 Ndm3∙h−1 and a permeate volumetric rate of 15 dm3∙h−1, the average effluent characteristics comply with discharge and reuse limits.

This study examined the impact of adding bentonite clay (concentration of 1.5 to 10 g/L) to a pilot-scale aerobic ceramic membrane bioreactor (AeCMBR) for treating pharmaceutical wastewater (PhWW). The hydraulic retention time (HRT) was maintained at 24 h; the dissolved oxygen was between 2 mg/L (on) and 4 mg/L (off) throughout operation. Organic and nitrogen pollution removal rates and heavy metal (Cu, Ni, Pb, Zn) reduction rates were assessed. The chemical oxygen demand (COD) removal efficiency exceeded 82%. Adsorption improved ammonia (NH4+) removal to 78%; the addition of 5 g of bentonite resulted in a 38% improvement compared with the process without bentonite. The average nitrate concentration decreased from 169.69 mg/L to 43.72 mg/L. The average removal efficiencies for Cu, Ni, Pb and Zn were 86%, 68.52%, 46.90% and 56.76%, respectively. Bentonite at 5 g/L significantly reduced membrane fouling. The cost–benefit analysis enabled us to predict that the process will meet the multiple objectives of durability, treatment performance and economic viability. The combination of an AeCMBR and bentonite adsorption has proven to be a valuable solution for treating highly polluted wastewater.

Aerobic ceramic membrane bioreactors (AeCeMBR) have shown great potential in treating wastewater (WW) from the textile industry; however, their operation faces challenges such as process variability, membrane contamination, and the need for accurate prediction of treated water quality under varying conditions. In this study, chemical oxygen demand (COD) and turbidity were selected as key indicators, as they directly reflect organic load removal and solids separation efficiency in MBR systems. The effect of four operational parameters: hydraulic retention time (HRT), organic loading rate (OLR), mixed liquor suspended solids (MLSS), and transmembrane pressure (TMP), was investigated using a response surface methodology (RSM) based on a Box–Behnken design. A random forest (RF) model coupled with Monte Carlo simulation (MC) was also developed using 174 experimental data points to enhance predictive power and quantify uncertainty. The RSM model showed strong agreement with experimental results (coefficient of determination (R²) > 0.95), achieving approximately 96% removal for both COD and turbidity, with validation errors of less than 2%. MC simulation (10,000 iterations) was applied to assess the effect of ±10% variance under operating conditions, providing a probabilistic view of system performance. The RF-MC framework demonstrated high predictive accuracy, with strong correlations between predicted and observed values (R² = 0.92 for COD and 0.97 for turbidity) and low uncertainty. Overall, this study proposes an integrated RSM, RF–MC approach for AeCeMBR systems, providing a robust and uncertainty-aware framework for process optimization and performance prediction under changing operating conditions.