Statistical Modeling And Optimization Of Process Parameters For 2,4-Dichlorophenoxyacetic Acid Removal By Using Ac Pdmaema Hydrogel Adsorbent Comparison Of Different Rsm Designs And Ann Training Methods

• Main Author: Azhar, Emillia Eizleen Md
• Format: Monograph
• Language: English
• Published: Universiti Sains Malaysia 2022
• Subjects: T Technology; TP155-156 Chemical engineering
• Online Access: http://eprints.usm.my/55169/1/Statistical%20Modeling%20And%20Optimization%20Of%20Process%20Parameters%20For%202%2C4-Dichlorophenoxyacetic%20Acid%20Removal%20By%20Using%20Ac%20Pdmaema%20Hydrogel%20Adsorbent%20Comparison%20Of%20Different%20Rsm%20Designs%20And%20Ann%20Training%20Methods.pdf; http://eprints.usm.my/55169/

The effect of pH, initial concentration of 2,4-D and the activated carbon content toward the adsorption process of 2,4-D by the modified hydrogel (AC/PDMAEMA hydrogel) were analysed. The experimental data taken from previous study was used to predict the removal of 2,4-D and the adsorption capacity. The simulation was done with Design Expert V12.0 and Matlab R2021a, where different design of response surface methodology (RSM) and training methods of Artificial Neural Network (ANN) were used. RSM was used to analyse the effect of the process parameter toward the adsorption process of 2,4-D as well as to build an empirical model which display the relationship between the factors and the responses. The analysis of the empirical model build by the two level factorial, face centred composite and custom designs were done with the analysis of variance (ANOVA) and compared. Apart from performance of empirical model, the optimum condition for the maximum removal of 2, 4-D and adsorption capacity was also obtained with the RSM simulation. It was found that among these three design, the optimal design has the highest accuracy in predicting the responses. The maximum removal of 2, 4-D and adsorption capacity at 65.01 % and 65.29 mg/g respectively were obtained at pH of 3, initial concentration of 2,4-D of 94.52 mg/L and 2.5 wt% of activated carbon. Apart from optimization of process parameter, the neural network architecture was also optimized by trial and error with different number of hidden neurons in the layers to obtain the best performance of the response. The optimization of the neural network was done with different training methods and compared. It was found that among the three training methods of ANN model, Bayesian Regularization methods had the highest R2 and lowest MSE with optimum network architecture of 3:9:2. The optimum condition obtained from RSM was also simulated with the optimized neural network architecture to validate the responses and adequacy of the RSM model.