Sensitivity and uncertainty analyses of an extended ASM3-SMP model describing membrane bioreactor operation

Modelling of high-rate activated sludge process: Assessment of model parameters by sensitivity and uncertainty analyses

The objective of this study is to develop a mechanistic model to predict the long-term dynamic performance of High-Rate Activated Sludge (HRAS) process, including the removal of carbon (COD), nitrogen (N), and phosphorus (P). The model was formulated with inspiration from Activated Sludge Models No. 1 and 3 (ASM1 and ASM3) to incorporate essential mechanisms, such as adsorption and storage substrate, specific to HRAS systems. A stepwise protocol was followed for calibration with dynamic data from a pilot-scale HRAS plant. Sensitivity analysis identified influential model parameters, including maximum specific growth rate (μ), growth yield (YH), storage yield (YSTO), storage rate (kSTO), decay rate (b), and half saturation of the readily biodegradable substrate for growth (KS1). The calibrated model achieved prediction efficiencies above the normalized Mean Absolute Error (MAE) of 70 % for mixed liquor suspended solids (MLSS), total chemical oxygen demand (TCOD), soluble COD (SCOD), particulate COD (XCOD), total nitrogen (TN), ammonia nitrogen (SNH), total phosphorus (TP), soluble TP (STP), and particulate TP (XTP). Uncertainty analysis revealed that SCOD was underestimated. Based on the dynamic profiles of uncertainty bands and observed data, there is potential for improving the estimation of dynamic behavior in STP. The observed discrepancies may be attributed to variations in wastewater characteristics during the monitoring period, particularly concerning the phosphorus (P) fractions of the readily biodegradable substrate (SS) and soluble inerts (SI), which were not considered as dynamically changing parameters in the model.

Enhancing Real-Time Prediction of Effluent Water Quality of Wastewater Treatment Plant Based on Improved Feedforward Neural Network Coupled with Optimization Algorithm

To provide real-time prediction of wastewater treatment plant (WWTP) effluent water quality, a machine learning (ML) model was developed by combining an improved feedforward neural network (IFFNN) with an optimization algorithm. Data used as input variables of the IFFNN included hourly influent water quality parameters, influent flow rate and WWTP process monitoring and operational parameters. Additionally, input variables included historical effluent water quality parameters for future prediction. The model was demonstrated in a WWTP in Jiangsu Province, China, where prediction of effluent chemical oxygen demand (COD) and total nitrogen (TN) with large variations were tested. Relative to the traditional feedforward neural network (FFNN) model without considering historical effluent water quality parameter input, the IFFNN enhanced prediction performance by 52.3% (COD) and 72.6% (TN) based on the mean absolute percentage errors of test datasets, after its model structure was optimized with a genetic algorithm (GA). The problem of over-fitting could also be overcome through the use of the IFFNN, with the determination of coefficient increased from 0.20 to 0.76 for test datasets of effluent COD. The GA-IFFNN model, which was efficient in capturing complex non-linear relationships and extrapolation, could be a useful tool for real-time direction of regulatory changes in WWTP operations.

Utilization of random vector functional link integrated with manta ray foraging optimization for effluent prediction of wastewater treatment plant

An innovative predictive model was employed to predict the key performance indicators of a full-scale wastewater treatment plant (WWTP) operated with an activated sludge treatment process. The data-driven model was obtained using data gathered from Cairo, Egypt. The proposed model consists of Random Vector Functional Link (RVFL) Networks incorporated with Manta Ray Foraging Optimizer (MRFO). RVFL is used as an advanced Artificial Neural Network (ANN) that avoids the common conventional ANN problems such as overfitting. MRFO is employed to determine the best RVFL parameters to maximize the prediction accuracy of the model. The developed MRFO-RVFL is compared with conventional RVFL to figure out the role of MRFO as an optimization tool to enhance model performance. Both models were trained and tested using experimental data measured during a long period of 222 days. This study aims to provide an accurate prediction of the most widely treated effluent indicators of BOD₅ and TSS in the wastewater treatment plants. In this study, ten well-known influent wastewater parameters, BOD₅, TSS, and VSS, influent flow rate, pH, ambient temperature, F/M ratio, SRT, WAS, and RAS, the output BOD₅ and TSS were modeled and predicted using the integrated MRFO-RVFL algorithms and compared with the standalone RVFL model. The performance of the models was evaluated using different assessment measures such as R², RMSE, and others. The obtained results of R² and RMSE for the MRFO-RVFL model were 0.924 and 3.528 for BOD₅ and 0.917 and 6.153 for TSS, which were much better than the results of conventional RVFL with 0.840 and 6.207 for BOD₅ and 0.717 and 10.05 for TSS. Based on the obtained results, the selective model (MRFO-RVFL) exhibited a higher performance and validity to predict the TSS and optimal BOD₅.

Support Vector Regression Modelling of an Aerobic Granular Sludge in Sequential Batch Reactor

Support vector regression (SVR) models have been designed to predict the concentration of chemical oxygen demand in sequential batch reactors under high temperatures. The complex internal interaction between the sludge characteristics and their influent were used to develop the models. The prediction becomes harder when dealing with a limited dataset due to the limitation of the experimental works. A radial basis function algorithm with selected kernel parameters of cost and gamma was used to developed SVR models. The kernel parameters were selected by using a grid search method and were further optimized by using particle swarm optimization and genetic algorithm. The SVR models were then compared with an artificial neural network. The prediction results R² were within >90% for all predicted concentration of COD. The results showed the potential of SVR for simulating the complex aerobic granulation process and providing an excellent tool to help predict the behaviour in aerobic granular reactors of wastewater treatment.

Uncertainty analysis of rising sewer models with respect to input parameters and model structure using Monte Carlo simulations and computational fluid dynamics

Modelling conversion processes in sewers can help minimize odour and pipe corrosion issues, but model uncertainties and errors must be understood. In this study, the Wastewater Aerobic/Anaerobic Transformation in Sewers (WATS) model is implemented in two different frameworks; 1-D (CSTR-in-series) and computational fluid dynamics (CFD) to study the uncertainties due to model parameters and its mathematical form. The 1-D model is used to conduct uncertainty/sensitivity analysis using Monte Carlo simulations. Time-averaged outputs were represented using a general linearized model to quantify the importance of specific parameters. The sulfide formation rate per unit area of the biofilm is the most influential parameter. Parameters controlling anaerobic hydrolysis and fermentation are also significant. Uncertainty due to model structure is studied using CFD to explore the influences of non-homogeneous surface reactions and solids settling. These showed that the 1-D model provides a reasonable characterisation of the process for simple flows in pressure mains.

Operational performance and biofoulants in a dynamic membrane bioreactor

In this study, a mathematical model was developed to have a better understanding of the process and be used in future reactor scale models to predict its process performance. This model utilizes the Activated Sludge Model NO.1 (ASM1) framework and incorporates bioprocesses of formation and degradation of soluble microbial products (SMP) and extracellular polymeric substances (EPS). Simulation result shows the model could very well predict the bioreactor performance. The average error of COD, BOD and NH₃-N removal efficiency was 0.48, 0.28 and 1.18%, respectively.

Performance evaluation and modeling of a submerged membrane bioreactor treating combined municipal and industrial wastewater using radial basis function artificial neural networks

Treatment process models are efficient tools to assure proper operation and better control of wastewater treatment systems. The current research was an effort to evaluate performance of a submerged membrane bioreactor (SMBR) treating combined municipal and industrial wastewater and to simulate effluent quality parameters of the SMBR using a radial basis function artificial neural network (RBFANN). The results showed that the treatment efficiencies increase and hydraulic retention time (HRT) decreases for combined wastewater compared with municipal and industrial wastewaters. The BOD, COD, [Formula: see text] and total phosphorous (TP) removal efficiencies for combined wastewater at HRT of 7 hours were 96.9%, 96%, 96.7% and 92%, respectively. As desirable criteria for treating wastewater, the TBOD/TP ratio increased, the BOD and COD concentrations decreased to 700 and 1000 mg/L, respectively and the BOD/COD ratio was about 0.5 for combined wastewater. The training procedures of the RBFANN models were successful for all predicted components. The train and test models showed an almost perfect match between the experimental and predicted values of effluent BOD, COD, [Formula: see text] and TP. The coefficient of determination (R(2)) values were higher than 0.98 and root mean squared error (RMSE) values did not exceed 7% for train and test models.

Sensitivity analyses and simulations of a full-scale experimental membrane bioreactor system using the activated sludge model No. 3 (ASM3)

An ASM3-based model was implemented in the numerical software MATHEMATICA where sensitivity analyses and simulations of a membrane bioreactor (MBR) system were carried out. These results were compared with those obtained using the commercial simulator WEST. Predicted values did not show significant variations between both software and simulations showed that the most influential operational conditions were influent flow rate and concentrations and bioreactor volumes. On the other hand, sensitivity analyses were carried out with both software programs for the same five outputs: COD, ammonium and nitrate concentrations in the effluent, total suspended solids concentration and oxygen uptake rate in the aerobic bioreactor. Similar results were in general obtained in both cases and according to these analyses, the most significant inputs over the model predictions were growth and storage heterotrophic biomass yields and decay coefficient. Other parameters related to the hydrolysis process or to the autotrophic biomass also significantly influenced model outputs.

Variance-based sensitivity analysis for wastewater treatment plant modelling

Global sensitivity analysis (GSA) is a valuable tool to support the use of mathematical models that characterise technical or natural systems. In the field of wastewater modelling, most of the recent applications of GSA use either regression-based methods, which require close to linear relationships between the model outputs and model factors, or screening methods, which only yield qualitative results. However, due to the characteristics of membrane bioreactors (MBR) (non-linear kinetics, complexity, etc.) there is an interest to adequately quantify the effects of non-linearity and interactions. This can be achieved with variance-based sensitivity analysis methods. In this paper, the Extended Fourier Amplitude Sensitivity Testing (Extended-FAST) method is applied to an integrated activated sludge model (ASM2d) for an MBR system including microbial product formation and physical separation processes. Twenty-one model outputs located throughout the different sections of the bioreactor and 79 model factors are considered. Significant interactions among the model factors are found. Contrary to previous GSA studies for ASM models, we find the relationship between variables and factors to be non-linear and non-additive. By analysing the pattern of the variance decomposition along the plant, the model factors having the highest variance contributions were identified. This study demonstrates the usefulness of variance-based methods in membrane bioreactor modelling where, due to the presence of membranes and different operating conditions than those typically found in conventional activated sludge systems, several highly non-linear effects are present. Further, the obtained results highlight the relevant role played by the modelling approach for MBR taking into account simultaneously biological and physical processes.

Soluble microbial products (SMPs) release in activated sludge systems: a review

This review discusses the characterization, production and implications of soluble microbial products (SMPs) in biological wastewater treatment. The precise definition of SMPs is open to talk about, but is currently regarded as “the pool of organic compounds that are released into solution from substrate metabolism and biomass decay”'. Some of the SMPs have been identified as humic acids, polysaccharides, proteins, amino acids, antibiotics, extracellular enzymes and structural components of cells and products of energy metabolism. They adversely affect the kinetic activity, flocculating and settling properties of sludge. This review outlines some important findings with regard to biodegradability and treatability of SMPs and also the effect of process parameters on their production. As SMPs are produced during biological treatment process, their trace amounts normally remain in the effluent that defines the highest COD removal efficiency. Their presence in effluent represents a high potential risk of toxic by-product formation during chlorine disinfection. Studies have indicated that among all wastewater post-treatment processes, the adsorption by granular activated carbon combined with biologically induced degradation is the most effective method for removal of SMPs. However, it may be concludes that the knowledge regarding SMPs is still under progress and more work is required to fully understand their contribution to the treatment process.