Innovative approach for predicting biogas production from large-scale anaerobic digester using long-short term memory (LSTM) coupled with genetic algorithm (GA).

An artificial neural network (ANN) model called long-short term memory (LSTM), coupled with a genetic algorithm (GA) for feature selection, was used to predict biogas production of large-scale anaerobic digesters (ADs) of Tehran South Wastewater Treatment Plant (Iran), with a biogas production of approximately 30,000 Nm3/d. In order to employ the real conditions, the hydraulic retention time (HRT) of the ADs (21 days) was considered as the LSTM look-back window. To evaluate the model predictions, three different scenarios were defined. In the first scenario, the model predicted the produced biogas by using raw wastewater characteristics and reached the coefficient of determination of R2 = 0.84. The GA selected four out of eleven parameters of raw wastewater, including loads of BOD5, COD, TSS, and TN (kg/d), as the most informative data for the model. In the second scenario, the model predicted the produced biogas by employing the data of the thickened sludge streams entering the ADs and yielded a higher accuracy (R2 = 0.89). In this scenario, GA selected two out of six parameters of the sludge streams, including total flow rate (m3/d) and average solids content (w/w%). Finally, in the third scenario, by putting the parameters of the two previous scenarios together, the model's prediction accuracy increased slightly (R2 = 0.90). The results demonstrated that the GA-LSTM modeling technique could achieve reliable performance in predicting biogas production of large-scale ADs by including HRT in modeling procedure. It was also found that the raw wastewater characteristics severely affect AD behavior and can be successfully used as the input data of the AD models.

Fouling mechanisms in anoxic-aerobic sequencing batch membrane bioreactor based on adapted Hermia models and main foulant characteristics.

Various derivatives of Hermia models (complete pore blocking, intermediate pore blocking, cake layer formation, and standard pore blocking) and different assessments of foulant characteristics have long been used to determine the membrane fouling mechanisms. Accordingly, this study aims to adapt Hermia models and their combination according to the operating conditions of an anoxic-aerobic sequencing batch membrane bioreactor (A/O-SBMBR). In addition, fouling mechanisms of the A/O-SBMBR were assessed using these models along with the main foulant characteristics. Models fitting with the transmembrane pressure (TMP) data indicated that the intermediate-standard model was accounting for the increased fouling during the whole regular operating period, with the residual sum of squares (RSS) of 58.3. A more detailed study on the distinct stages of TMP curve showed that the intermediate-standard model had the best fit in stages of 2 and 3, with the RSS equal to 2.6 and 2.8, respectively. Also, the complete-standard model provided the best description of the fouling mechanism in stage 4, with the RSS of 12.5. Different analyzes revealed how the main foulant characteristics affect the occurrence of intermediate, complete and standard fouling mechanisms in the A/O-SBMBR, which is consistent with the fitting results of the adapted Hermia models. The modeling and experimental methods used in the presented study provided a valuable basis to prevent and control membrane fouling in membrane bioreactors.

Predicting coagulation-flocculation process for turbidity removal from water using graphene oxide: a comparative study on ANN, SVR, ANFIS, and RSM models.

Three artificial intelligence (AI) data-driven techniques, including artificial neural network (ANN), support vector regression (SVR), and adaptive neuro-fuzzy inference system (ANFIS), were applied for modeling and predicting turbidity removal from water using graphene oxide (GO). Based on partial mutual information (PIM) algorithm, pH, GO dosage, and initial turbidity were selected as the input variables for developing the models. The prediction performance of the AI-based models was compared with each other and with the response surface methodology (RSM) model, previously reported by the authors, as well. The models' estimation accuracy was assessed through statistical measures, including mean-squared error (MSE), root-mean-square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2). Among the evaluated models, ANN had the highest estimation accuracy as it showed the highest R2 for the validation data (0.949) and the lowest MSE, RMSE, and MAE values. Furthermore, ANN predicted 76.1% of data points with relative errors (RE) less than 10%. In contrast, the weakest prediction performance belonged to the SVR model with the lowest R2 for both calibration (0.712) and validation (0.864) data. Besides, only 57.1% of the SVR's predictions were characterized by RE < 10%. The ANFIS and RSM models exhibited a more or less similar performance in terms of R2 for the validation data (0.877 and 0.871, respectively) and other statistical parameters. According to the results, the ANN technique is proposed as the best option for modeling the process. Nevertheless, as the RSM technique provides valuable information about the contribution of the independent operational parameters and their complex interaction effects using the least number of experiments, simulating the process by this technique before modeling by ANN is inevitable.

Performance evaluation of Fe-based water treatment sludge for dewatering of iron ore tailings slurry using coagulation-flocculation process: Optimization through response surface methodology.

This research attempted to investigate the feasibility of using drinking water treatment sludge (ferric chloride sludge, FCS) as a coagulant for turbidity removal from iron ore tailings slurry. The evaluation was performed in two phases. In the first phase, the one factor at a time (OFAT) approach was used to study the effects of FCS dosage, initial pH, and initial turbidity on turbidity removal efficiency (TR%) and the volume of the sediment produced at the end of the process (SV). In the second phase, response surface methodology (RSM) was employed to assess the individual and interaction effects of the parameters on TR% and SV. Numerical multiple-response optimization was carried out using RSM to maximize TR% and minimize SV simultaneously. At optimum condition (FCS dose of 0.13 g dried FCS/L, initial pH of 10, and initial turbidity of 538 NTU), the removal of all particles in the range of 0.25-1 µm and 2-55 µm from slurry led to the TR% of 78.80% and SV of 0.74 mL (per 250 mL of tailings). Characterization tests indicated that at alkaline pH values, the higher presence of hydroxide compounds intensified the enmeshment in a precipitate or sweep-floc mechanism, which was the predominant removal mechanism in this work. This study demonstrated the remarkable performance of FCS as a coagulant in water reclamation from iron beneficiation wastewater.

Coagulation-flocculation of turbid water using graphene oxide: simulation through response surface methodology and process characterization.

It was aimed to precisely investigate the coagulation properties of graphene oxide (GO) as a novel coagulant for turbidity removal from water. For this purpose, the process was simulated through response surface methodology (RSM) to determine the effect of the preselected independent factors (pH, GO dosage, and initial turbidity) and their interaction effects on the process. Based on the results, increased turbidity removal efficiencies were obtained as pH decreased from 10 to 3. Besides, increase of GO dosage within the test range (2.5-30 mg/L) was highly beneficial for enhancing the process performance. However, a slight overdosing of GO was observed for dosages of more than 20 mg/L under pH values of less than about 4. For initial turbidity with test range of 25-300 NTU, there was an optimum range (approximately 120-200 NTU) out of which the removal efficiency declined. According to the results of the analysis of variance (ANOVA), pH and GO dosage, orderly, had the strongest individual effect on the process performance. The most significant interaction effect was also observed between pH and GO dosage. The optimal coagulation conditions with GO dosage of 4.0 mg/L, pH of 3.0, and initial turbidity of 193.34 NTU led to a turbidity removal efficiency of about 98.3%, which was in good agreement with RSM results. Under basic pH levels, the sweeping effect was recognized as the main coagulation mechanism occurred between the negatively surface charged particles of GO and soil. However, according to zeta potential (ZP) analysis results, under acidic pH conditions in addition to the sweep coagulation, the electric double layer compression, and the subsequent ZP reduction also contributed significantly to the process. Scanning electron microscopy (SEM) images showed that the layered structure of GO particles provided an appropriate platform on which the flocs were formed.