Predicting the capability of carboxylated cellulose nanowhiskers for the remediation of copper from water using response surface methodology (RSM) and artificial neural network (ANN) models

For the transesterification of biodiesel from Azolla oil, the safe and successful use of feed stocks is a very significant prerequisite. It is of high importance to determine the optimal reaction parameters to maximize the yield of low-cost biodiesel generated from Azolla oil. Ultrasonic energy was used in this work for the development of biodiesel from Azolla oil catalyzed by the KOH catalyst under different conditions. The effect on the transesterification of Azolla Oil to biodiesel of four reaction parameters, namely the methanol/Azolla oil molar ratio (A), KOH catalyst concentration (B), reaction time (C) and reaction temperature (D) were considered. In order to optimize the effects of reaction parameters for the transesterification of Azolla oil to biodiesel, response surface methodology (RSM) based on central composite rotatable design (CCRD) is applied. To obtain a good correlation between the input reaction parameters and the output response parameter (FAME yield) from Azolla oil to biodiesel, an artificial neural network (ANN) model with two feed-forward back-propagation neural-network architecture Multilayer Perceptron Network (MLP) and Radial Basis Function Network (RBFN) was developed. With the experimental information obtained from the RSM model, the built ANN models were trained and evaluated. Absolute Average Deviation (AAD), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and coefficient of determination were statistically compared with the predictive capacity of both RSM and ANN models (R2). The statistical analysis showed that the measured FAME yield from both the RSM and ANN models was able to predict the FAME yield, and the findings limited the ANN model to the much more reliable FAME yield prediction compared to the RSM model.

Application of response surface methodology and artificial neural network modeling to assess non-thermal plasma efficiency in simultaneous removal of BTEX from waste gases: Effect of operating parameters and prediction performance

This research investigated and optimized the separation of carbon dioxide (CO2) from natural gas in an adsorption column filled with grafted-beam nanofiber adsorbent. The main purpose of using ANN and RSM models in this manuscript is to compare these two methods in predicting the CO2 adsorption capacity. In other words, it was made to find a suitable model that has the highest agreement with the experimental data. Also, the other purpose of using the RSM model is to detect the optimized empirical conditions. Moreover, two common ANN models are applied in this work, including a multilayer perceptron (MLP) and radial basis function (RBF). The novelties of this work are explained as follows: (1) detecting the optimized synthesis factors of NF-PAN/PUGMA sorbent which possess the highest CO2 adsorption capacity with the help of response surface methodology (RSM), (2) studying the simultaneous interaction of synthesis parameters on the CO2 adsorption capacity with the help of both RSM and artificial neural networks (ANNs), (3) testing two types of ANN models including multilayer perceptron (MLP) and radial basis function (RBF) to predict the effect of monomer volume percentage and irradiation dose on the CO2 adsorption capacity. Indeed, finding the best model (ANN or RSM) can help engineers in practical applications predict CO2 adsorption capacity using NF-PAN/PUGMA under different conditions without incurring high-priced chemical materials, electricity, or human resources. The validation results were examined using correlation coefficients (R2) of RSM, RBF, and MLP models. The correlation coefficients for the RSM, RBF, and MLP models were 0.9910, 0.9949, and 0.9968, respectively. Additionally, the average absolute relative deviation (AARD) values for the RBF and MLP models were 0.00046512 and 0.00045511, respectively, indicating that the MLP model is better than the RBF model. To identify the optimal network structure, the trial-and-error method was conducted for MLP and RBF models. The number of neurons was found at 12 and 45 for MLP and RBF, respectively. The optimized effective parameters were obtained using RSM: 25.80% GMA, 66.45% amine, and an irradiation intensity of 28kGy.

Machine learning and response surface methods for predicting the compressive strength of high-strength concrete have not been adequately compared. Therefore, this research aimed to predict the compressive strength of high-strength concrete (HSC) using different methods. To achieve this purpose, neuro-fuzzy inference systems (ANFISs), artificial neural networks (ANNs), and response surface methodology (RSM) were used as ensemble methods. Using an ANN and ANFIS, high-strength concrete (HSC) output was modeled and optimized as a function of five independent variables. The RSM was designed with three input variables: cement, and fine and coarse aggregate. To facilitate data entry into Design Expert, the RSM model was divided into six groups, with p-values of responses 1 to 6 of 0.027, 0.010, 0.003, 0.023, 0.002, and 0.026. The following metrics were used to evaluate model compressive strength projection: R, R2, and MSE for ANN and ANFIS modeling; R2, Adj. R2, and Pred. R2 for RSM modeling. Based on the data, it can be concluded that the ANN model (R = 0.999, R2 = 0.998, and MSE = 0.417), RSM model (R = 0.981 and R2 = 0.963), and ANFIS model (R = 0.962, R2 = 0.926, and MSE = 0.655) have a good chance of accurately predicting the compressive strength of high-strength concrete (HSC). Furthermore, there is a strong correlation between the ANN, RSM, and ANFIS models and the experimental data. Nevertheless, the artificial neural network model demonstrates exceptional accuracy. The sensitivity analysis of the ANN model shows that cement and fine aggregate have the most significant effect on predicting compressive strength (45.29% and 35.87%, respectively), while superplasticizer has the least effect (0.227%). RSME values for cement and fine aggregate in the ANFIS model were 0.313 and 0.453 during the test process and 0.733 and 0.563 during the training process. Thus, it was found that both ANN and RSM models presented better results with higher accuracy and can be used for predicting the compressive strength of construction materials.

The application of additive manufacturing technologies for producing parts from polymer composite materials has gained significant attention due to the ability to create fully functional components that leverage the advantages of both polymer matrices and fiber reinforcements while maintaining the benefits of additive technology. Polymer composites are among the most advanced and widely used composite materials, offering high strength and stiffness with low mass and variable resistance to different media. This study aims to experimentally investigate the impact of selected process parameters, namely, wall thickness, raster angle, printing temperature, and build plate temperature, on the flexural properties of carbon fiber reinforced polyamide (CFrPA) fused deposition modeling (FDM) printed samples, as per ISO 178 standards. Additionally, regression and artificial neural network (ANN) models have been developed to predict these flexural properties. ANN models are developed for both normal and augmented inputs, with the architecture and hyperparameters optimized using random search technique. Response surface methodology (RSM), which is based on face centered composite design, is employed to analyze the effects of process parameters. The RSM results indicate that the raster angle and build plate temperature have the greatest impact on the flexural properties, resulting in an increase of 51% in the flexural modulus. The performance metrics of the optimized RSM and ANN models, characterized by low MSE, RMSE, MAE, and MAPE values and high R2 values, suggest that these models provide highly accurate and reliable predictions of flexural strength and modulus for the CFrPA material. The study revealed that ANN models with augmented inputs outperform both RSM models and ANN models with normal inputs in predicting these properties.

This study provides a comparative evaluation of the ability of response surface methodology (RSM) and artificial neural network (ANN) to predict the performance of microbial fuel cell (MFC) driven by greywater-syrup substrate system as anolyte with respect to power generation and wastewater treatment. Fourier transform infrared instrumental analysis of the substrate shows the functional groups of compounds present. A 24 central composite design and a three-layered (4:n:1) feedforward ANN architecture trained by a backpropagation algorithm were used to study and predict the MFC process performance criteria. The ANN gave the best prediction with n = 10 neurons. The response variables (power density generation (mW/m2) and chemical oxygen demand (COD) removal efficiency (%)) were measured against four process input variables: mass of the clay component of the proton exchange membrane (PEM) (g), PEM preparation temperature (PPT), anolyte pH, and concentration. Optimal responses with respect to power density and COD removal of 88.3 mW/m2 and 95.2% were recorded at the values of 70 g, 300°C, 8.5, and 66.9 v/v for mass of clay, PPT, pH, and anolyte concentration, respectively. The power density and COD removal predictive abilities of the ANN and RSM models were evaluated in terms of error functions: root mean square error (RMSE) (0.512; 0.0557), chi-square (0.0510; 0.1240), model predictive error (MPE) (0.3326; 0.3526), and coefficient of determination (R2) (0.9954; 0.9051) and RMSE (0.0272; 0.0707), chi-square (0.0280; 0.181), MPE (0.08242; 0.1569), and R2 (0.9932; 0.9245), respectively. These results indicate the superiority of the ANN in predicting the performance of the MFC over the RSM.

Metal pollutants such as copper released into the aqueous environment have been increasing as a result of anthropogenic activities. Adsorption-based treatment technologies offer opportunities to remediate metal pollutants from municipal and industrial wastewater effluent. The aim of this work was to evaluate the capability of modified cellulose nanowhisker (CNW) adsorbents for the remediation of copper from water matrices under realistic conditions using response surface methodology (RSM) and artificial neural network (ANN) models. Considerations for design and application to remediate Cu(II) from wastewater by developing a continuous flow experiment are described in this study. However, the physical structure of modified CNW adsorbents renders them unsuitable for use in column operation. Therefore, a more detailed study of the mechanical properties of CNW adsorbents would be necessary in order to improve the strength and stability of the adsorbents. This work has demonstrated that modified CNW are promising adsorbents to remediate copper from water matrices under realistic conditions including wastewater complexity and variability. The use of models to predict the test parameter system and account for matrix variability when evaluating CNW adsorbents for remediating Cu from a real-world wastewater matrix may also provide the foundation for assessing other treatment technologies in the future.

Measurement of performance and emission distinctiveness of Aegle marmelos seed cake pyrolysis oil/diesel/TBHQ opus powered in a DI diesel engine using ANN and RSM

In this study, poly ethyleneimine (PEI)/Titania (TiO2) multilayer film on quartz tubes have been successfully fabricated via a layer-by-layer (LbL) self-assembly method. Scanning electron microscopy (SEM) and Brunauer-Emmett-Teller (BET) surface area analysis were carried out for characterization of the layers on quartz tube. The SEM pictures showed that the film surface is smooth and uniform. The BET characterization verified the formation of multilayer thin film. The photocatalytic activity of the PEI/TiO2 multilayer deposited on the quartz tubes was evaluated in the treatment of raw petroleum refinery wastewater (PRW) under UV light irradiation in three annular photocatalytic reactors. This study examined the impact of initial chemical oxygen demand (COD) concentration, H2O2 concentration, pH and reaction time on the PRW treatment and the results were used to generate both a response surface methodology (RSM) model and an artificial neural network (ANN) model. Maximum COD removal (98 %) was achieved at the optimum conditions (initial COD concentration of 300 mg/l, hydrogen peroxide concentration of 8.8 mM, pH of 5 and reaction time of 120 min). A comparison between the model results and experimental data gave a high correlation coefficient (R ANN 2 = 0.9632, R RSM 2 = 0.943) and showed that two models were able to predict COD removal from PRW by PEI/TiO2/UV process. However, ANN model was superior to RSM model with higher value of coefficient of determination (0.9632ANN > 0.94RSM) and the lower root mean square error (RMSE) (3.377AAN < 3.569RSM). The average percentage error for ANN and RSM models was 0.18 and 0.73, respectively, indicating the superiority of ANN in capturing the nonlinear behavior of the system. It was clear that the best networks were able to predict the experimental responses more accurately than the multiple regression analysis.

Hybrid optimisation studies on the microstructural properties and wear resistance of maraging steel 1.2709 parts produced by laser powder bed fusion

Metal matrix composites (MMCs) have found many successful industrial applications in recent past as high-technology materials due to their properties. Wire electric discharge cutting (WEDC) process is considered to be one of the most suitable processes for machining MMCs. Lot of research work has been done on WEDC, but very few investigations have been done on WEDC of MMCs. This paper reports work on the analysis of material removal rate (MRR) and cutting width (kerf) during WEDC of 6061 Al MMC reinforced with silicon carbide particles (i.e. SiCp/6061 Al). Four WEDC parameters namely servo voltage (SV), pulse-on time (TON), pulse-off time (TOFF) and wire feed rate (WF) were chosen as machining process parameters. Artificial neural network (ANN) models and response surface methodology (RSM) models were developed to predict the MRR and kerf using Box-Behnken design (BBD) to generate the input/output database. It was observed that prediction of responses from both models closely agree with the experimental values. The ANN models and RSM models for WEDC of MMC were compared with each other on the basis of prediction accuracy which shows that ANN models are more accurate than RSM models for MRR and kerf because the values of percentage absolute errors are higher for RSM models than ANN models.

In the current report, both response surface methodology (RSM) and artificial neural network (ANN) were employed to develop an innovative way for removing crystal violet (CV) from aqueous media using Haloxylon salicornicum (HS) as a cost‐effective, eco‐friendly adsorbent. HS was characterized using scanning electron microscopy (SEM) and Fourier‐transform infrared (FTIR) spectroscopy. The effects of operational parameters such as adsorbent dosage, initial dye concentration, and pH on HS were studied using a central composite design (CCD). A comparative analysis of the model findings and experimental measurements revealed high correlation coefficients (R2ANN = 0.994, R2RSM = 0.971), indicating both models accurately predicted HS. The predictive performance of the ANN and RSM models was evaluated using metrics such as mean absolute deviation (MAD), mean absolute percentage error (MAPE), root mean square error (RMSE), mean square error (MSE), and the correlation coefficient (R2). The results indicate that the ANN model provides greater accuracy compared to the RSM model. The experimental data were analyzed using both linear and nonlinear forms of pseudo‐first and pseudo‐second order kinetic models (LPFO, NLPFO, LPSO, and NLPSO). Statistical error analysis was conducted to identify the best‐fitting kinetic or isotherm models for the adsorption data. The adsorption process of CV/HS was best described by NLPSO and LPSO. Additionally, the adsorption isotherms were analyzed using linear and nonlinear regression methods. The findings indicated that the linear Langmuir and Freundlich isotherms provided a more accurate fit compared to the nonlinear models, demonstrating greater effectiveness in accounting for the adsorption parameters. Thermodynamic investigations clearly demonstrate that the biosorption of CV is spontaneous and exothermic. This cost‐effective adsorbent is highly promising for treating textile wastewater.

The co-production of immiscible oil and water phases frequently results in the formation of stable emulsions, posing a persistent challenge in the petroleum industry. These emulsions are metastable colloidal systems, typically stabilized by the presence of surface-active agents, amphiphilic polymers, or fine solid particles. Their occurrence can lead to substantial operational and economic drawbacks, making the ability to accurately predict emulsion stability critical for effective mitigation strategies. In this study, synthetic emulsions were systematically prepared based on experimental conditions prescribed by the Stratigraphic Centurion VII design of experiments. Emulsion stability was evaluated by measuring the volume of water separated at 60 °C over a duration of 30 days (or 15 days in selected cases). The experimental design followed the Box-Behnken Design (BBD) framework, and the resulting data were used to develop predictive models using Response Surface Methodology (RSM), Artificial Neural Networks (ANN), and Gene Expression Programming (GEP). Model performance was assessed through Analysis of Variance (ANOVA), alongside statistical metrics such as the coefficient of determination (R²) and root mean square error (RMSE). A Pareto chart was also employed to quantify the relative importance of input variables on emulsion stability. Comparative analysis between model predictions and experimental results revealed that the ANN model achieved superior accuracy, with an R² of 0.9828 and minimal deviation. In contrast, the RSM and GEP models exhibited lower predictive performance, with R² values of 0.8682 and 0.8723, respectively. Despite its higher accuracy, the ANN model's black-box nature limits its interpretability. Conversely, the RSM and GEP models provide explicit mathematical expressions, making them valuable tools for understanding system behaviour and guiding process optimization.

Camelina [Camelina sativa (L.) Crantz] is an annual versatile oilseed crop of the Brassicaceae family, with an increasingly cultivated area. Predicting camelina seed yield response to fertilization and planting density is of great importance in understanding production potential and management planning. Therefore, the current study aimed to estimate the seed yield of camelina by response surface methodology (RSM) and artificial neural network (ANN) as affected by different levels of planting row spacing and nitrogen (N), sulfur (S), and cow manure (CM) fertilization. The experiment was conducted in two growing years of 2019–2020 and 2020–2021, based on a central composite design with four factors including row spacing (15–35 cm), N (0–200 kg ha−1), S (0–100 kg ha−1), and CM (0–40 t ha−1). The RSM models for seed yield versus fertilization and row spacing factors in both years were statistically significant and had an acceptable predictive ability. Camelina seed yield decreased with increasing row spacing but showed a positive response to increasing the amount of N, S, and CM fertilizers. Comparing the performance of the models showed that, although the RSM models were significant and had the necessary efficiency in predicting camelina seed yield, the ANN models were more accurate. The performance criteria of coefficient of determination (R2), root mean square error (RMSE), standard error of prediction (SEP), mean absolute error (MAE), and Akaike information criterion (AICc) averaged over the two years for the RSM model were 0.924, 51.60, 5.51, 41.14, and 394.05, respectively, and for the ANN model were 0.968, 32.62, 3.54, 19.55, and 351.33, respectively. Based on the results, the ANN modeling can be used in predicting camelina seed yield in field conditions with more confidence than the RSM technique.

In this study, Response Surface Methodology (RSM) and Artificial Neural Networks (ANN) was used to predict the mechanical properties of self-compacting concrete (SCC) with silica fume as partial cement replacement and Polyethylene terephthalate (PET) solid waste as partial sand replacement. PET plastic was varied between 0 and 20 wt% while the silica fume was varied between 0 and 40 wt%. The parameters investigated were the compressive strength, tensile strength and impact strength of SCC. The RSM model was fairly accurate (R2 ≥ 0.92) in predicting the mechanical properties. The model was statistically significant (p-value < 0.5) and did not possess any prediction bias. The ANN model was able to capture the variability of the data as evidenced by the good R2 threshold (R2 > 0.93) for training, testing and validation. Parity plots revealed that both the ANN and RSM models do not have any prediction bias. However, the ANN model is superior because of its higher accuracy and the use of admixtures enhanced the workability suitability for dataset. The 3D microstructural analysis showed that the interfacial adhesion between the aggregates and the cementitious materials reduced at increased partial replacement leading to a decrease in the strength.