Response Surface Methodology Prediction Research Articles

Investigation of gold adsorption by ironbark biochar using response surface methodology and artificial neural network modelling

Due to the importance and economic value of gold in various applications, recovering gold from waste streams like tailings and industrial wastewater is essential. Biochar with a high surface area and porosity is considered a potential low-cost adsorbent for gold reclaiming from aqueous media. Therefore, this study aims to develop a high-quality biochar with excellent physicochemical properties to efficiently remove gold from aqueous media. To accomplish this, biochar was obtained from pyrolysis of ironbark (IB) biomass at 500 °C which has a surface area of 493.79 m2/g. Subsequently, the biochar was investigated for adsorption of gold from aqueous media, which exhibited a maximum adsorption capacity of 858 mg/g. Isotherm results showed that the adsorption of gold by biochar followed the Langmuir model, indicating monolayer adsorption. Further, response surface methodology (RSM) and an artificial neural network (ANN) combined with the Salp Swarm Algorithm (SSA) were used to analyze the generated results (ANN-SSA). The RSM prediction model fit was adequate (R2 = 0.99). However, comparing the statistical data revealed that ANN-SSA outperformed RSM in predicting experimental results. Overall, this study suggested that biochar derived from IB biomass could be used as a potential adsorbent to recover gold from an aqueous solution. This article presents a unique and innovative study on the utilization of ironbark biochar for the purpose of gold adsorption.

Data-driven approach for Cu recovery from hazardous e-waste

The Cu recovery from e-waste is beneficial since it contributes the most economic value to electronic scraps (∼322,000 tons Cu available with PCB). Data Science and Artificial Intelligence (AI) tools, experimental data-driven Adaptive Neuro-Fuzzy Inference System (ANFIS), and Response Surface Methodology (RSM) were adopted for the prediction of Cu recovery from e-waste. These models were developed using four variables (H2SO4, H2O2, Solid/Liquid ratio, and reaction time) and validated by validation experiments. The developed RSM-based model predicts the Cu recovery with an average error of ∼10.51%, whereas the ANFIS-based model predicts the Cu recovery with an error of ∼6.03%. Higher values of R2 (0.99), F (641.59), and a lower value of P (< 0.05) were observed in the ANOVA study indicating better suitability of the model. Based on findings from the comparison, the relevant data-driven ANFIS model is used for further analysis. The 3D interactive plots have been developed to find a suitable range of input parameters to improve the recovery of Cu from e-waste. The results and findings reported in this article will be valuable to Cu recovery, e-waste management, and hazardous waste management. The proposed ANFIS model is highly efficient for effective and accurate automation at the industrial scale. Environmental implicationScientific communities must pay attention to the challenges raised due to surplus e-waste. E-waste is reported to be the most dangerous anthropogenic material due to hazardous chemicals and toxic metals. This article explains how e-waste can be converted into economically beneficial products. Since Cu contributes the most economic value to e-waste, its recovery can minimize its hazardous effect and promote a circular economy. Modern data-driven tools are highly efficient for accurately predicting Cu recovery. In this article, RSM and ANFIS-based prediction models have been developed with high accuracy. It was observed that ANFIS is more efficient than RSM and better for the automation of copper recovery. The results and findings reported in this article will be valuable to Cu recovery, e-waste management, and hazardous management of e-waste. The proposed data-driven approach helps to minimize the load of electronic waste and reduces a significant proportion of hazardous effects of e-waste on the environment.

Unravelling adsorption potential of expired clindamycin as corrosion inhibitor at mild Steel/HCl interface via response surface methodology, artificial neural network, electrochemical, and computational strategies

Corrosion has produced unprecedented disintegration of metals, constituting an imminent danger to mankind and triggering catastrophic global economic losses. The effectiveness of expired clindamycin (ECLI) as a low-cost corrosion control agent for mild steel was investigated utilising response surface methodology (RSM), artificial neural network (ANN), potentiodynamic polarization (PDP), electrochemical impedance spectroscopy (EIS), quantum chemical computation (QCC), and molecular dynamic simulation (MDS) studies in conjunction with thermometric and gasometric protocols at different HCl concentrations. The RSM model demonstrated an outstanding level of accuracy in predicting the mild steel corrosion inhibition efficiency (IE), the volume of hydrogen gas (VHG), reaction number (RN), and corrosion rate (CR). The model was significantly influenced by the operational parameters that were investigated, such as temperature (299–333 K), ECLI concentrations (100–500 mg/L), immersion time (1–6 h), and acid concentration (0.5–2.5 M). It was observed that as ECLI concentration increases, the VHG, RN, and CR decreased per time as well as % IE increased. The thermometric, gasometric, PDP, and EIS results showed percentage inhibition efficiency of 69.10, 69.49, 83.17, and 77.87 %, respectively. PDP revealed that ECLI operates as a mixed type of inhibitor, and EIS indicated that the inhibition process involves charge transfer. The Langmuir isotherm suits better and accurately describes the ECLI adsorption process on mild steel. The electron transfer propensity of the ECLI on the metal surface is measured by QCC using the DFT approach. MDS was implemented to establish the optimal adsorption orientation between ECLI and Fe (110). The inspection of surface morphology by SEM displayed the formation of a blanket-like layer on the steel by ECLI. To validate the experimental results, RSM and ANN prediction models were utilised, which were evaluated using a normal plot of residual, predicted versus actual, and residual versus run, and were found to be effective modelling tools. This study illustrates that ECLI can be utilised as a potent and affordable mild steel inhibitor, even at high acid concentrations.

Optimization of FSW parameters on bio-inspired jigsaw suture patterns to improve the tensile strength of dissimilar thermoplastics

This study aimed to enhance the Tensile Strength (TSFSW) of dissimilar thermoplastic joints by utilizing a bio-inspired jig saw suture and optimizing the Friction Stir Welding (FSW) limits are Traverse Speed (TS) and Plunge Depth (PD) and Rotational Speed (RS) at three varied levels. Statistical analysis, response surface methodology (RSM), and experimental validation were involved in achieving the research objectives. The outcomes showed that the TS and PD parameters had a higher significance on Tensile Strength compared to RS. The RSM prediction results were validated through experiments, achieving an extreme Tensile Strength of 11.1 MPa with a low error percentage. The best values of the FSW limits were found to be Rotational Speed (RS) of 1200 rpm, Plunge Depth (PD) of 0.37 mm, and Traverse Speed (TS) of 49.39 mm min−1. The formulated mathematical model with regression co-efficient R2 of 0.96 and RSM proved effective in predicting the optimal FSW parameters and achieving superior TSFSW. These findings prove that combination design can be reliably applied to optimise with a 95% confidence interval. The optical microscope and SEM morphological results in this study make congruently accurate predictions for the joint of the tensile fracture zone. These findings contribute to the advanced FSW techniques for dissimilar thermoplastic joints, providing insights for industrial applications requiring strong and reliable joints.

Optimization of the Biotreatment of GTL Process Water Using Pseudomonas aeruginosa Immobilized in PVA Hydrogel

The COD reduction in gas to liquid (GTL) process water was optimized using response surface methodology (RSM). The biodegradation process was carried out in a spouted bed bioreactor (SBBR) using Pseudomonas aeruginosa immobilized in polyvinyl alcohol (PVA) gel. Different factors affecting the biological treatment of GTL process water (PW) were investigated. Three variables including PVA volume fraction, initial COD, and pH were investigated in the batch experiments. The biodegradation experiments were carried out by varying the initial COD values from 1000 to 3000 mg/L, pH from 5 to 8, and PVA v% from 20 to 30%. The maximum COD reduction was estimated to occur at an initial COD of 2595 mg/L, PVA v% of 27%, and pH of 7.3. At optimum conditions, the bioreactor system was able to achieve a maximum COD reduction of 89%, which is quite close to the RSM prediction value of 90%. The optimum operating conditions were used to carry out continuous biodegradation, and the results indicated that the COD reduction increased from 60% to 62% with an increase in the air flow rate from 2 to 3.3 La/Lr.min. However, by increasing the liquid flow rate from 2.1 to 4.2 mL/min and back to 2.1 mL/min, the COD reduction decreased from 66% to 39%. The system responded quickly to the change in liquid flow rate and returned to the initial COD level. This indicates that the system is highly stable and can easily recover.

Evaluation of specific total work of fracture in polypropylene/ethylene‐propylene diene monomer nanocomposites based on various graphene derivatives: Optimization by response surface methodology technique and correlation with microstructure properties

AbstractThis study presents an experimental design approach for optimization and prediction of specific total work of fracture (wf) in polypropylene/ethylene‐propylene diene monomer (EPDM) nanocomposites based on disparate graphene derivatives including multilayer graphene, few‐layer graphene and graphene oxide. Box–Behnken design and response surface methodology (RSM) were exploited at three levels and four factors to construct a predictive mathematical model for the value of wf. The model predicted that the maximum value of wf (136.7 J/mm2) is achieved by applying dynamically vulcanized system at EPDM content and few‐layer graphene content of 20 wt% and of 0.5 wt%, respectively. In addition, by observing the fracture surfaces of specimens, a direct correlation between RSM predictions and morphology of nanocomposites was established. In this regard, transmission electron microscopy and wide‐angle X‐ray diffraction analyses showed that by using few‐layer graphene, thanks to formation of an exfoliated structure, more energy for the failure of nanocomposites is required which substantiates the model predictions. Scanning electron microscopy observations depicted that the droplets diameter of homogeneous EPDM particle size plays a pivotal role in number of micro‐voids and nano‐voids, which consequently promotes the creation of fibril structure and eventually heightens nanocomposite resistance against the applied load. The optimum status of this phenomenon happens at 20 wt% of EPDM and as the model showed, the value of wf reaches the highest point at this EPDM content. Likewise, by applying dynamic vulcanization, as truly predicted by the model, the formation of multi fibrillar structure led to dissipation of extra energy in nanocomposites inducing a remarkable improvement in the value of wf.

Surface Feature Prediction Modeling and Parameter Optimization for Turning TC17 Titanium Alloy

Surface integrity has a very significant effect on surface roughness and surface microhardness. These are the main characteristics of surface integrity. The present study investigated the influence of the cutting depth (ap), the cutting speed ( v c ), and the feed rate (f) on the surface roughness (Ra) and surface microhardness (HV) in turning TC17 titanium alloy. Data obtained from the Box-Behnken design experiments were used to develop the response surface methodology (RSM) and artificial neural network (ANN) models. Through analysis of variance (ANOVA), the relative effects of each cutting parameter on the responses have been determined. To examine the interaction effects of cutting parameters, 3D surface plots were generated. The desirability function approach (DFA) was used to optimize cutting parameters to achieve the lowest surface roughness and highest surface microhardness. The results show that ANN response prediction models have higher prediction accuracy and lower error than RSM prediction models. The optimization parameters are 60 m/min cutting speed, 0.06 mm/r feed rate, and 0.2 mm cutting depth for the minimum surface roughness and maximum surface microhardness with a maximum error of 2.83%.

Model prediction of coagulation by magnetised rice starch for wastewater treatment using response surface methodology (RSM) with artificial neural network (ANN)

This study synthesized magnetized rice starch (MRS) with recoverability benefits as an alternative coagulant for the coagulation treatment of industrial wastewater. The MRS was characterization by scanning electron microscopy and energy dispersive X-ray (SEM/EDX) and Brunauer–Emmett–Teller (BET). The engineered rice starch with BET surface area of 31.44 m2/g and cationic elementals showed good agglomeration behavior. Three process variables (coagulant dose, settling time, and mixing rate) were model and optimized for the coagulation removal efficiency of turbidity, color, and phosphate. Using the response surface methodology (RSM), Box-Behnken design (BBD) with 17 experimental runs was used to model and optimized the experimental parameters. At optimal conditions of coagulant dosage of 4 g, settling time of 15 mins, and mixing rate of 50 rpm resulted in removal efficiency of 72 %, 53.2 %, and 56.5 %, respectively for turbidity, color, and phosphate. This infers a desirability efficiency of 82.6% was attained via analysis of variance (ANOVA) at a 95% confidence level. The experimental results showed close agreement with the RSM and artificial neural network (ANN) model prediction. Results of the model's prediction significance (P <0.05) were compared in terms of coefficient of determination (R2 = 0.994) for the ANN model and RSM model (R2 = 0.97). Inclusively, the use of RSM and ANN for coagulation process optimization using magnetized rice starch is viable for wastewater management for reuse. Therefore, exploring the potency of the magnetized rice starch in real world application warrants the model adaptability to estimate its cost benefit.

Tailoring interfacial properties of 3D-printed continuous natural fiber reinforced polypropylene composites through parameter optimization using machine learning methods

3D-printed continuous ramie fiber reinforced polypropylene composites (CRFRPP) are expected to ensure good mechanical properties while meeting the requirements of environmental friendliness and sustainability. To promote the wide industrial application of CRFRPP, this work investigated the effects of printing parameters (extrusion flow rate, printing temperature, layer thickness and printing speed) on the interfacial properties of CRFRPP. The interlayer and intralayer interfacial properties of CRFRPP with different printing parameters were studied using the design of experiment approach. Machine learning methods and response surface methodology prediction were also carried out based on the experimental results to bridge the printing parameters and interfacial properties. According to the prediction results, the printing parameters were optimized to improve the production efficiency while ensuring the desired interfacial performance. At last, the bending tests were conducted to investigate how the difference in interfacial properties can be translated to the mechanical performance. The results found that printed specimens with weak interfacial strength suffered interlaminar delamination failure when subjected to bending loads, greatly weakening the mechanical properties of the composites.

Production and Characterization of Rhamnolipids Produced by Pseudomonas aeruginosa DBM 3774: Response Surface Methodology Approach.

Rhamnolipids are extensively studied biosurfactants due to their potential in many industrial applications, eco-friendly production and properties. However, their availability for broader application is severely limited by their production costs, therefore the optimization of efficacy of their cultivation gains significance as well as the information regarding the physio-chemical properties of rhamnolipids resulting from various cultivation strategies. In this work, the bioprocess design focused on optimization of the rhamnolipid yield of Pseudomonas aeruginosa DBM 3774 utilizing the response surface methodology (RSM). Six carbon sources were investigated for their effect on the rhamnolipid production. The RSM prediction improved the total rhamnolipid yield from 2.2 to 13.5 g/L and the rhamnolipid productivity from 11.6 to 45.3 mg/L/h. A significant effect of the carbon source type, concentration and the C/N ratio on the composition of the rhamnolipid congeners has been demonstrated for cultivation of P. aeruginosa DBM 3774 in batch cultivation. Especially, changes in presence of saturated fatty acid in the rhamnolipid congeners, ranging from 18.8% of unsaturated fatty acids (carbon source glycerol; 40 g/L) to 0% (sodium citrate 20 g/L) were observed. This demonstrates possibilities of model based systems as basis in cultivation of industrially important compounds like biosurfactants rhamnolipids and the importance of detailed study of interconnection between cultivation conditions and rhamnolipid mixture composition and properties.

Modelling and optimization for methylene blue adsorption using graphene oxide/chitosan composites via artificial neural network-particle swarm optimization

Water pollution due to dyes from industrial effluents and domestic wastewater is a big environmental issue, so an effective adsorbent is needed. In this study, graphene oxide/chitosan (GO/CS) composites were synthesized and applied for methylene blue (MB) dye removal. Characterization was done on the GO and GO/CS composites using FTIR, EDX, SEM, and TGA. The adsorption studies were conducted to verify the effect of pH, adsorbent dosage and contact time. The interactive effects of the process variables were verified using response surface methodology (RSM), and optimal conditions for higher adsorption efficiency are evaluated by Artificial neural network (ANN)-Particle swarm optimization (PSO). ANN-PSO predictions are in good agreement with the experimental values and hence resulted in higher R2 (=0.998) compared to RSM predictions (R2 = 0.981). The MB adsorption process is found to be obeying the Langmuir isotherm and pseudo 1st order kinetic model. The maximum MB removal efficiency (90.34%) and adsorption amount (7.53 mg/g) can be obtained at an initial dye concentration of 10 mg/L and optimal values of pH (5), adsorbent dosage (0.143 g/L) and contact time (125 min). These results further confirm that the ANN-PSO-based approach is able to capture the inherent mechanisms of the MB adsorption process and can be used as a good modelling approach.

Optimization of ingredients level for developing yellow passion (passiflora edulis) fruit jam

ABSTRACT This study was carried out to investigate the optimum level of ingredients to develop an acceptable jam product from yellow passion fruit. The jam was developed from fresh yellow passion fruit using four ingredients: sugar, pectin and citric acid of different levels. The different ingredient levels were optimized using response surface methodology (RSM). A uniform precision type Box-Behnken design consisting of three variables in a three-level pattern with 17 runs (with five center points) was employed to optimize the jam to obtain a product with adequate sensory acceptability, especially, overall acceptability. Sensory analysis for appearance, color, flavor, sweetness, taste, texture and overall acceptability was performed. Jam produced under the optimum conditions was subjected to sensory evaluation, and the results were compared with the RSM predictions. Data obtained from RSM were subjected to the analysis of variance (ANOVA) using a second-order polynomial equation. Results showed that the model fit was significant (p < .05) for appearance, flavor, taste, sweetness and even highly significant (p < .01) for texture, color and overall acceptability. There was also a satisfactory correlation between actual and predicted values. Depending on the best sensory score obtained, the optimum ingredient levels were 400 g/kg, 5.24 g/kg and 1 g/kg for sugar, pectin and citric acid, respectively.

Optimization and prediction of engine block vibration using micro-electro-mechanical systems capacitive accelerometer, fueled with diesel-bioethanol (water-hyacinth) blends by response surface methodology and artificial neural network

The current investigation is focused on the vibration signals analysis for health status diagnosis of the single-cylinder diesel engine fueled with bioethanol diesel mixture. The water hyacinth (WH) plants (Eichhornia crassipes) are used as raw materials for bioethanol production. The bio-ethanol obtained from WH has been mixed with diesel fuel (WBED) to various extent. Systematically designed experiments were conducted with different working parameters like load, fuel injection pressure (FIP), and compression ratio (CR) in a diesel engine. The Micro-Electro-Mechanical Systems (MEMS) capacitive accelerometer was used to get vibration signals from the engine while operating with blended fuels. The obtained experimental vibrations data have been used to predict the engine vibration by using Response Surface Methodology (RSM) technique and Artificial Neural Network (ANN). The experimental results have been compared with RSM and ANN prediction results. From results, it is elicited that the acceleration declines with the increase in load and CR. At all tested blends, FIP produces a significant effect on the engine block vibration. Among all blends, WBED 5 and WBED 10 produce less vibration as compared to other diesel bioethanol blends. At optimized operating condition the engine block vibration for WBED 5; the experimental acceleration is 0.016962 m/s2 and the predicted acceleration by RSM and ANN is 0.016182 m/s2 and 0.0166 m/s2, respectively. For WBED 10, the acceleration is 0.0172604 m/s2 and the predicted acceleration by RSM and ANN 0.016207 m/s2 0.017 m/s2, respectively, has been found.

Lead (Pb) removal from contaminated water using constructed wetland planted with Scirpus grossus: Optimization using response surface methodology (RSM) and assessment of rhizobacterial addition

Lead (Pb) is one of the toxic heavy metals that pollute the environment as a result of industrial activities. This study aims to optimize Pb removal from water by using horizontal free surface flow constructed wetland (HFSFCW) planted with Scirpus grossus. Optimization was conducted using response surface methodology (RSM) under Box–Behnken design with the operational parameters of initial Pb concentration, retention time, and aeration. Optimization results showed that 37 mg/L of initial Pb concentration, 32 days of retention time, and no aeration were the optimum conditions for Pb removal by using the systems. Validation test was run under two different conditions, namely, non-bioaugmented and bioaugmented with rhizobacteria (Bacillus cereus, B. pumilus, B. subtilis, Brevibacillus choshinensis, and Rhodococcus rhodochrous). Results of the validation test showed that Pb removal in water achieved 99.99% efficiency with 0.2% error from the RSM prediction, while the adsorption of Pb by plants reached 5160.18 mg/kg with 10.6% error from the RSM prediction. The bioaugmentation of the five rhizobacterial species showed a slight improvement in Pb removal from water and Pb adsorption by plants. However, no significant improvement was achieved (p < 0.05). Overall results suggested that operating the HFSFCW under optimum conditions with no bioaugmentation might be a feasible choice for the treatment of Pb-contaminated water.

Optimization of FDM 3D Printing Process Parameters on ABS based Bone Hammer using RSM Technique

Rapid prototyping (RP) uses a cycle where a real model is made by explicitly adding material as thin cross-sectional layers. Fused deposition modelling (FDM) 3D printer is being use for synthesis of ABS based bone hammer. Response surface methodology (RSM) based L27 design of experiment were adopted to perform the experiment using four influencing parameters such as layer thickness, infill percentage, orientation and nozzle temperature for the three responses deflection, hardness and weight. Response surface methodology was used for modelling and optimization of considered process parameters. In present investigation, it is evident that bone hammer fabrication process parameters have been optimized on data such as bone hammer weight 19.8091g, hardness 104.5921 BHN, and force of 15 degree deflection 36.0681 N has been produced with RSM prediction with influence of process parameters such as layer thickness 0.250 mm, infill percentage 63.3333, orientation 60 degree, nozzle temperature 240°C.

Modeling and combined application of MOEA/D and TOPSIS to optimize WEDM performances of A286 superalloy

Superalloys are categorized as difficult to process materials with a broad spectrum of applications in industries. Process modeling and optimization of WEDM performances on nickel- and titanium-based superalloys are widely investigated. However, such investigations on iron-based superalloy are still lacking and hence probed in the present article. Thus, the paper targets modeling the correlation between the performance parameters and the control parameters with two popular techniques: response surface methodology (RSM) and artificial neural network (ANN) for WEDM of a typical iron-based superalloy, i.e., A286 superalloy. A comparison between the model estimates and the experimental values is made to check ANN and RSM's prediction accuracy. The estimates by the ANN model are exact and consistent with the experimental results. An analysis of variance (ANOVA) test is performed to perceive the degree of statistical significance of parameters. Moreover, a novel two-stage procedure, i.e., MOEA/D in collaboration with TOPSIS method, is implemented to search the optimal condition for process performances. The quality of Pareto-optimal solutions acquired using MOEA/D is compared to that of Pareto-optimal solutions obtained using NSGA II, PESA II, and MMOPSO through the use of a hypervolume (HV) parameter. Wilcoxon’s test is performed to identify the statistical difference between MOEA/D and competing algorithms. The optimal parametric combination recommended by the proposed optimization approach is Ton = 130 µs, Toff = 52 µs, Ipeak = 12 A, Wf = 5 m/min and SV = 30 V. The proposed optimization technique can also be exploited in other manufacturing processes.

Optimization and modeling of efficient photocatalytic TiO2-ZnO composite preparation parameters by response surface methodology

The major focus of this work is the design and optimization of the preparation conditions of binary titanium dioxide-zinc oxide (TiO2-ZnO) composite using response surface methodology (RSM). Calcination temperature and ZnO weight percent (%) with respect to TiO2 were chosen as the key parameters for optimizing the preparation conditions. The evaluation of photocatalytic nitrogen oxides (NOx) removal efficiency was selected for modeling the RSM. Thirteen factorial design runs were first generated with one replicate and 5 center points in order to perform all the experiments. The experimental results on NOx removal efficiency for the composites calcined at 747, 602, and 558 °C containing 5.48%, 3.48%, and 2.1% of ZnO with respect to the weight of TiO2, designated as MaxT(Z%), MedT(Z%) and MinT(Z%) respectively, agree well with RSM predictions. The response coefficient difference between experimental and RSM results of MaxT(Z%), MedT(Z%), and MinT(Z%) were approximately 0.019, 0.044, and 0.11, respectively. The microstructural and optical properties as well the surface composition of MaxT(Z%), MedT(Z%), and MinT(Z%) were evaluated by X-ray diffraction (XRD), diffuse reflectance spectroscopy (DRS), and X-ray photoelectron spectroscopy (XPS), respectively. The results from the photodecomposition of NOx gas under illuminated UV-light confirmed that the TiO2-ZnO prepared with a 2.1 wt% of ZnO and calcined at 558 °C witnessed the best photocatalytic activity and the reasons for the excellent performance are corroborated with XRD, XPS, and DRS results. This study provides a simple and systematic approach for the optimization and modeling of the binary composite preparation for achieving enhanced photoactivity.

Predictive capability evaluation and optimization of Pb(II) removal by reduced graphene oxide-based inverse spinel nickel ferrite nanocomposite

Pb(II) is a heavy metal that is a prominent contaminant in water contamination. Among the different pollution removal strategies, adsorption was determined to be the most effective. The adsorbent and its type determine the adsorption process's efficiency. As part of this effort, a magnetic reduced graphene oxide-based inverse spinel nickel ferrite (rGNF) nanocomposite for Pb(II) removal is synthesized, and the optimal values of the independent process variables (like initial concentration, pH, residence time, temperature, and adsorbent dosage) to achieve maximum removal efficiency are investigated using conventional response surface methodology (RSM) and artificial neural networks (ANN). The results indicate that the initial concentration, adsorbent dose, residence time, pH, and process temperature are set to 15 mg/L, 0.55 g/L, 100 min, 5, and 30 °C, respectively, the maximum removal efficiency (99.8%) can be obtained. Using the interactive effects of process variables findings, the adsorption surface mechanism was examined in relation to process factors. A data-driven quadratic equation is derived based on the ANOVA, and its predictions are compared with ANN predictions to evaluate the predictive capabilities of both approaches. The R2 values of RSM and ANN predictions are 0.979 and 0.991 respectively and confirm the superiority of the ANN approach.

ANFIS modeling of biodiesels' physical and engine characteristics: A review

Abstract Population increase has resulted in an increase in the worldwide demand for alternative fuels due to depleting resources. There is a periodic increase in concern about the engine performance, pollutant emissions, and their predictions, from an engine using biodiesels. The use of intelligent algorithms in modeling and forecasting alternative fuels characteristics and their performance in engines are critically reviewed in this study. The paper aims at demonstrating with artificial intelligence methodologies the main conclusions of the recent research done for the above topic from 2012 to 2020. This article attempted to demonstrate an exploratory examination of the adaptive neuro‐fuzzy inference system (ANFIS) soft computing technique used for the exact measurement and analysis of engine performance, emissions of exhaust engines when biodiesel is used as an alternative fuel. Additionally, the yield of biodiesel and their different characteristics predicted using ANFIS are also reviewed. Integration of particle swarm optimization (PSO), genetic algorithm (GA), and response surface methodology (RSM), either for comparison or optimization with ANFIS is presented. The summary of all studies is provided in tabular form. For the demonstration purpose, the ANFIS studies predicting different biodiesel and engine characters are provided with illustrative figures. The ANFIS prediction related to biodiesel used engine and biodiesel self‐characteristics is found to be excellent. The ANFIS accuracy reported is better than the artificial neural network (ANN) accuracy. A minimum of 0.9R2value is generally obtained which is around 5% greater than the ANN modeling results reported. However, the ANFIS predictions are much more fitter than the RSM predictions. The integration of ANFIS‐PSO and ANFIS‐GA provided much more optimized results.

Evaluation of surface phenomena of magnetic biomass for dye removal via surface modeling

Methylene blue (MB) dye is used in various industries; however, it is detrimental to both humans and the environment. This study reports on the adsorptive removal of methylene blue (MB) using a magnetically modified agro-waste: iron oxide loaded Lonicera japonica flower powder (IOLJF). The interaction effects of operating variables such as the initial MB concentration, temperature, dosage of IOLJF, and pH on the IOLJF removal efficiency for MB were investigated systematically and statistically. The results showed that the adsorptive removal of MB was explained well by the pseudo-second-order (PSO) kinetic model and the Langmuir and Freundlich isotherms. The maximum uptake capacity was 403.44 mg/g at pH 6.0 and 293 K. The MB adsorption process was complex on the surface of the IOLJF. To comprehend the essential characteristics and envision the removal efficiency in the batch adsorption process, two approaches were used, namely the response surface methodology (RSM) and the data-driven adaptive neuro-fuzzy inference system (ANFIS). The RSM (R2 = 0.9782) and ANFIS (R2 = 0.9964) model predictions were both good and comparable with the experimental data. Further, it was found that the magnetic biomass could be re-used for up to seven cycles without losing its efficacy of ~95%. Overall, the results indicated that the IOLJF had sufficient potential for wastewater treatment.