A novel framework of multivariate modeling of water distribution network through 33 factorial design and artificial neural network.

The water distribution network is largely affected by the change in the influencing factors, such as input pressure, demand and supply duration. The change in each parameter requires the extensive design of the network and the interactive effect of the influencing parameters are hardly explored. The main hurdles for the water providers lie in the absence of a prediction model, which can be used as a decision tool to assess the effect of the change in parameter and estimating the cost for the changed scenario. The present study developed a novel framework based on the artificial neural network for multivariate prediction modeling taking the response as the cost of the pipe network. The application of the 33 factorial design was used for the selection of the influencing parameters and outcome was taken as the input to the neural network model. The adequacy of the model was tested through error functions and analysis of variance. The low values of the error functions (0.0004-0.228) and high F value (162,442) and R2 (0.999) established the significance of the model. The model can be used for predicting the cost of the changed scenarios and assessment of the optimal solution for the system variables.

Sorptive equilibrium profile of fluoride onto aluminum olivine [(FexMg1-x)2SiO4] composite (AOC): Physicochemical insights and isotherm modeling by non-linear least squares regression and a novel neural-network-based method.

A novel aluminum/olivine composite (AOC) was prepared by wet impregnation followed by calcination and was introduced as an efficient adsorbent for defluoridation. The adsorption of fluoride was modeled with one-, two- and three-parameter isotherm equations by non-linear regression to demonstrate the adsorption equilibrium. The FI was the best-fitted model among the two-parameter isotherms with a R2 value of 0.995. The three-parameter models were found to have better performance with low values of the error functions and high F values. The neural-network-based model was applied for the first time in the isotherm study. The optimized model was framed with eight neurons in hidden layer with a mean square of error of 0.0481 and correlation coefficient greater than 0.999. The neural-based model has the better predictability with a higher F value of 9484 and R2 value of 0.998 compared to regression models, exhibiting the F value and the R2 in the range of 86-3572 and 0.835-0.995, respectively. The material characterization established the formation of the aluminum oxide, silicate, etc. onto the olivine which is conducive of the removal of fluoride by the formation of aluminum fluoride compounds, such as AlF3 in the spent material after defluoridation.

Adsorptive removal of arsenic by novel iron/olivine composite: Insights into preparation and adsorption process by response surface methodology and artificial neural network.

Olivine, a low-cost natural material, impregnated with iron is introduced in the adsorptive removal of arsenic. A wet impregnation method and subsequent calcination were employed for the preparation of iron/olivine composite. The major preparation process parameter, viz., iron loading and calcination temperature were optimized through the response surface methodology coupled with a factorial design. A significant variation of adsorption capacity of arsenic (measured as total arsenic), i.e., 63.15 to 310.85 mg/kg for arsenite [As(III)T] and 76.46 to 329.72 mg/kg for arsenate [As(V)T] was observed, which exhibited the significant effect of the preparation process parameters on the adsorption potential. The iron loading delineated the optima at central points, whereas a monotonous decreasing trend of adsorption capacity for both the As(III)T and As(V)T was observed with the increasing calcination temperature. The variation of adsorption capacity with the increased iron loading is more at lower calcination temperature showing the interactive effect between the factors. The adsorbent prepared at the optimized condition of iron loading and calcination temperature, i.e., 10% and 200 °C, effectively removed the As(III)T and As(V)T by more than 96 and 99%, respectively. The material characterization of the adsorbent showed the formation of the iron compound in the olivine and increase in specific surface area to the tune of 10 multifold compared to the base material, which is conducive to the enhancement of the adsorption capacity. An artificial neural network was applied for the multivariate optimization of the adsorption process from the experimental data of the univariate optimization study and the optimized model showed low values of error functions and high R2 values of more than 0.99 for As(III)T and As(V)T. The adsorption isotherm and kinetics followed Langmuir model and pseudo second order model, respectively demonstrating the chemisorption in this study.

Multivariate optimization of process parameters in the synthesis of calcined Ca‒Al (NO3) LDH for defluoridation using 3(3) factorial, central composite and Box-Behnken design.

Response surface methodology was applied for the first time in the optimization of the preparation of layered double hydroxide (LDH) for defluoridation. The influence of three vital process parameters (viz. pH, molar ratio and calcination temperature) in the synthesis of the adsorbent 'Calcined Ca‒Al (NO3) LDH' was thoroughly examined to maximize its fluoride scavenging potential. The process parameters were optimized using the 3(3) factorial, face centered central composite and Box-Behnken designs and a comparative assessment of the methods was conducted. The maximum fluoride removal efficiency was achieved at a calcination temperature of approximately 500ºC; however, the efficiency decreased with increasing pH and molar ratio. The outcome of the comparative assessment clearly delineates the case specific nature of the models. A better predictability over the entire experimental domain was obtained with the 3(3) factorial method, whereas the Box-Behnken design was found to be the most efficient model with lesser number of experimental runs. The desirability function technique was performed for optimizing the response, wherein face centered central composite design exhibited a maximum desirability. The calcined Ca‒Al (NO3) LDH, synthesized under the optimum conditions, demonstrated the removal efficiencies of 95% and 99% for the doses of 3 g L(-1) and 5 g L(-1), respectively.