Fig 1 - uploaded by Majid Bagheri
Content may be subject to copyright.
Source publication
This study examined carbon, nitrogen and phosphorous removal from municipal wastewater in a sequencing batch reactor and biokinetic coefficients were evaluated according to results of BOD and COD. Furthermore, the MLVSS in the aeration reactor was modeled by using multilayer perceptron and radial basis function artificial neural networks (MLPANN an...
Similar publications
The terminal iterative learning control (TILC) is designed for nonlinear system based on neural network. A terminal output tracking error model is obtained by using a system input and output algebraic function as well as the differential mean value theorem. Radial basis function (RBF) neural network is utilized to construct the input for the system...
Sequencing batch reactor (SBR) is a bioreactor, based on activated sludge and operates on a sequence of fill and draws cycles, to treat the high strength wastewater. SBR process is complex and attains a high degree of nonlinearity due to time dependency of chemical and biochemical reactions and the presence of bioorganic constituents that are diffi...
Citations
... To comprehend and investigate the efficacy of biological treatment in RWTP, this research is devoted in deriving an understanding on performances in RWTP in terms of pollutant removal efficiency, pollutants loading rates and bio-kinetic performance. The Artificial Neural Network (ANN) is a cutting-edge black-box model with the capacity to predict outputs based on intricate inputs, resembling the structure and functionality of the human brain and nervous system [2]. Recently, the ANN has been widely applied in water and wastewater process [3]. ...
Uncontrolled development at the upstream of the river catchment have led to detrimental effect to the environment, including degradation of river water quality. River Water Treatment Plant (RWTP) technology was introduced to reduce the contamination loading into the river water system, worldwide. The technology offers the best biological treatment process including simplicity and stable removal efficiency. However, the plant performance plan is difficult task to predict, thus might have influence the operational control. Recently, artificial neural network (ANN) models have been widely applied in environmental engineering area due to the ability to skip the complexity process to assume of the unknown variables compare to conventional physical based model. In this study, the results of 3-yrs performance using ANN of RWTP were developed. Feed-forward back-propagation using Levenberg–Marquardt (trainlm) used as for this predictive approach. The ideal configuration involves utilizing the tangent sigmoid transfer function (Tansig) in the hidden layer and a linear transfer function (Purelin) in the output layer, with 25 neurons. This configuration yields an R ² value of 0.963 and the most least mean square error (MSE) of 30.39. From the comparison between two model (bio-kinetic and ANN), performance indicator for ANN model shows the best and the most optimum model. Ultimately, RWTP optimization using black-box model ANN is more reliable and timesaving as compared to conventional bio-kinetic model. The development of the proposed model can be implemented and used for various water quality improvement facilities and predict the effluent target parameter in RWTP with higher degree of accuracy.
... The MLSS and MLVSS concentrations were used to measure the activity of activated sludge. MLSS includes volatile and inert solids in the mixed liquor [18]. Fig. 2(a) shows the MLSS concentrations for R-C and R-PHA throughout the study. ...
Biofilm enhances the performances of biological wastewater treatment systems. This study aimed to investigate the feasibility of using biodegradable polyhydroxyalkanoate (PHA) pellets as novel biofilm carrier for phenol biodegradation. Two identical laboratory-scale reactors were operated with fill, react, settle, draw and idle periods in the ratio of 2:12:2:1:7 for a 24-h cycle. One reactor was supplemented with 2% (v/v) of PHA pellet and operated as sequencing batch biofilm reactor (SBBR), whereas the other reactor was operated as sequencing batch reactor (SBR) and used as the control reactor. The performances of SBBR and SBR in degrading phenol were studied at three phases with the introduction of 300, 500 and 1000 mg L ⁻¹ phenol, respectively. The removal of phenol was found best described using zero-order kinetics, with R ² > 0.97. At all phases, the phenol removal rate during react period for SBBR (7.30 ± 0.55 to 9.33 ± 1.06 mg L ⁻¹ min ⁻¹ ) was found higher compared to those for SBR (4.28 ± 0.66 to 8.35 ± 0.68 mg L ⁻¹ min ⁻¹ ), with significance difference observed at low phenol concentration. Whereas for chemical oxygen demand biodegradation kinetics, SBRR exhibited significantly higher rate compared to SBR at all phases. From the scanning electron microscopy image, the attachment of activated sludge onto PHA pellet was observed. The results indicated the potential of PHA serving as alternative biofilm carrier in biofilm process.
... Artificial neural network (ANN) is another computational approach inspired by the processing capability of the human nervous system. It is a mathematical structure that has a strong ability to approximate and predict complex non-linear processes [18,19]. Biodiesel production from various sources of lipid has been modeled using RSM [6,20] and ANN [17,18]. ...
Wastewater sludge from a poultry slaughterhouse treatment plant was recovered through conversion into biodiesel by ultrasound-assisted in situ transesterification. The main effects of the process parameters were investigated at three levels, and their empirical relationship was modeled using artificial neural network (ANN) and response surface methodology (RSM). The developed models predicted the process behavior with excellent accuracy. Although both models had similar prediction performances, ANN marginally outperformed RSM. The capability of the genetic algorithm (GA) combined with the RSM (RSM-GA) and ANN (ANN-GA) models was evaluated for optimizing the process variables. The maximum biodiesel yield (21.45% w/w) was obtained using the ANN-GA model under optimized conditions, i.e., at the reaction time of 39.69 min, H2SO4 concentration of 3.34% (v/v), methanol-to-sludge relative content of 14.91:1 (mL/g), and ultrasound power of 104.87 W. Consequently, a combination of ANN and GA was proposed to model and optimize the transesterification process. The biodiesel yield obtained in this study was higher than the previously reported values from tannery (10.98%), dairy (13.46%), and municipal (18.58%) sewage sludge. This study specified biodiesel with a fatty acid methyl esters content of 96.86% using gas chromatography-mass spectrometry, Fourier transform infrared, and nuclear magnetic resonance spectroscopy.
... In recent years, artificial neural networks (ANNs) have been employed in diverse scientific areas, such as water resources and environmental engineering [21,22]. Besides, neural networks have been efficiently applied for controlling and monitoring water and wastewater treatment systems [23][24][25]. Neural networks fundamentally compute relationships among input values and output values via some internal calculations. Hence, a multi-layer perceptron neural network (MLPNN) was used to predict wastewater characteristics at the effluent of the wastewater treatment system in our study. ...
This study assesses the performance of a hybrid municipal wastewater treatment system. The proposed system attempted to improve the performance of an activated sludge system by using fully immersed vertical rotating biological contactors in the aeration basin of the system. Besides, a multi-layer perceptron neural network (MLPNN) was used to predict pollutants in the effluent. Overall treatment efficiencies of chemical oxygen demand (COD), total suspended solids (TSS), total phosphorus (TP), turbidity, and NH3 removal were 94%, 95%, 74%, 94%, and 86%, respectively. Via the training procedure of the MLPNN model, an almost perfect match was achieved between predicted values and experimental values. The correlation coefficient (R) was higher than 0.95 for all models predicting COD, TSS, TP, turbidity, and NH3, and mean squared error was satisfactory. The results were verified and demonstrated the effectiveness of the MLPNN model to predict effluent values. Also, the findings confirmed the effectiveness of the system in municipal wastewater treatment. Consequently, this system is recommended for treating municipal wastewater.
... On the whole, modeling can be used as a practical approach to monitor the changes over time of water and wastewater treatment systems and predict effluent quality parameters. Recently, artificial neural network (ANN) methods have been used for various areas of environmental issues such as wastewater and water treatment [18][19][20]. While wastewater treatment processes are pretty complicated, the improvements in intelligent methods make them possible to employ in the modeling of complex systems [21]. ...
... The results of the different data almost showed a perfect match between experimental values and predicted values for the effluent COD, turbidity, NH 3 , and TSS. The results of this study confirm the high generalization capability of the MLP-NN algorithm, and this has been reported in some studies [19,56]. MLP-NN predicting effluent TSS were 0.353 and 0.718 based on the train and all data sets. ...
This research was an effort to improve the performance of an activated sludge system by using biofilm carriers in the aeration basin of the system for treating petroleum refinery wastewater. Eventually, a granular activated carbon column was used in the last part of the treatment process. A neural network was employed to predict pollutants in the effluent and analyze the operating parameters. Overall treatment efficiencies of COD, turbidity, NH3, and TSS removal were 93%, 94%, 94%, and 92%, respectively. The results indicated that the removal efficiencies of pollutants in our hybrid system were superior to conventional activated sludge systems. The training procedure of the NN model was promising, and virtually an acceptable match was achieved between predicted values and experimental values. For all models predicting effluent COD, turbidity, NH3, and TSS, the correlation coefficient was higher than 0.9, and the mean squared error approached zero. According to the analysis of input parameters, the influent concentration is the essential factor in the modelling of effluent characteristics.
Keywords: Petroleum refinery wastewater treatment, Activated sludge process, Biofilm carriers, Activated carbon, Neural networks, Prediction.
... The primary treatment of the tannery effluents consists in the removal of a substantial part of the sedimentary or floating material, which represents significant percentages of the contaminating solids. The two types of primary treatment used are coagulation and flocculation (Aguinaga, 1996;Emmer & Del Campo, 2014;Mirbagheri et al., 2015). ...
Due to the characteristics of their processes, tanneries generate liquid effluents at different acidity or alkalinity conditions, with high turbidity, suspended solids and color. These effluents must be treated to avoid environmental contamination. The objective of the present work was to study, at a laboratory scale, coagulation processes that allow the reduction of turbidity and color of tannery effluents. The work focused on the use of collagen hydrolyzate (CH) as a coagulant, obtained from the shavings that the leather process discards. Tests were made with pure CH, in combination with a commercial coagulant (aluminum sulfate) or a commercial flocculant (cationic quaternary polyamine), in order to compare the efficiency of the process. Effluents were generated with a generic tanning formula, which were treated by aeration followed by sedimentation. The results obtained show that the CH alone does not show sufficient effectiveness, but combined with both the aluminum sulphate and the cationic quaternary polyamine gives very good results, involving smaller amounts of these commercial additives. The fact that the CH is obtained from a residue of the industry itself and it can successfully replace other chemicals, lowers the costs of treating the liquid effluents while avoiding having to manage that solid waste.
... However, the high concentration of MLSS in reactor 1 was due to more precipitation of inorganic (sediments) in the system which was suspected to have caused by the higher evaporation rate from the treatment unit. Considering the relevance of MLSS in wastewater treatment, as too low concentration will lead to poor treatment and very high values will amount to self-degradation of microorganisms due to insufficient food [25]. The concentration used in the present study was maintained relative to the recommended range (1500-4000 mg/L) for completely mixed suspended growth biological process for organic carbon and ammonium removal [7]. ...
The unrestricted discharge of domestic and industrial wastewaters along with agricultural runoff water into the environment as mixed-wastewater pose serious threat to freshwater resources in many countries. Mixed-wastewater pollution is a common phenomenon in the developing countries as the technologies to treat the individual waste streams at source are lacking due to high operational and maintenance costs. Therefore, the need to explore the potential of suspended growth process (SGP) which is a well-established process technology for biological wastewater treatment, is the focus of this paper. Different wastewater constituents: representing domestic, pharmaceutical, textile, petroleum, and agricultural runoff were synthesized as a representative of mixed-wastewater and treated in two semi-continuous bioreactors (R1 & R2) operated at constant operating conditions vis., MLSS (mg/L): 4640-R1, 4440-R2, SRT: 21-d, HRT: 48 - 72-h, and uncontrolled pH. The system attained stable condition in day 97, with average COD, BOD and TSS reduction as 84.5, 86.2, and 72.2% for R1; and 85.1, 87.9, and 75.1% for R2 respectively. Phosphate removal on average was by 74.3% in R1 and 76.6% in R2, while average nitrification achieved in systems 1 and 2 were 56.8and 54.7% respectively. The biological treatment system has shown potential for improving the quality of mixed-wastewater to the state where reuse may be considered and tertiary treatment can be employed to polish the effluent quality.
This study evaluates and models the impacts of employing biofilm carriers in sequencing batch reactors (SBR). A neural network (NN) was used to predict contaminants in the effluent and analyze importance of operating parameters. With a hydraulic retention time of 7 h, removal efficiency of chemical oxygen demand (COD), total phosphorous (TP), and total suspended solids (TSS) was 85%, 82%, and 98.9%, respectively. The removal efficiency of COD, TP, and TSS in our hybrid system was superior to regular single SBR systems. The training procedure of the NN model was successful and almost a perfect match was achieved between predicted values and experimental values. For all models predicting effluent COD, TP, and TSS, the correlation coefficient was higher than 0.99, and mean squared error approached zero. The analysis of input parameters demonstrated that influent concentration is the significant factor in the modelling of effluent characteristics.
This paper critically reviews all artificial intelligence (AI) and machine learning (ML) techniques for the better control of membrane fouling in filtration processes, with the focus on water and wastewater treatment systems. Artificial neural networks (ANNs), fuzzy logic, genetic programming and model trees were found to be four successfully employed modeling techniques. The results show that well-known ANNs such as multilayer perceptron and radial basis function can predict membrane fouling with an R² equal to 0.99 and an error approaching zero. Genetic algorithm (GA) and particle swarm optimization (PSO) are optimization methods successfully applied to optimize parameters related to membrane fouling. These optimization techniques indicated high capabilities in tuning various parameters such as transmembrane pressure, crossflow velocity, feed temperature, and feed pH. The results of this survey demonstrate that hybrid intelligent models utilizing intelligent optimization methods such as GA and PSO for adjusting their weights and functions perform better than single models. Clustering analysis, image recognition, and feature selection are other employed intelligent techniques with positive role in the control of membrane fouling. The application of AI and ML techniques in an advanced control system can reduce the costs of treatment by monitoring of membrane fouling, and taking the best action when necessary.
