Process parametric optimization toward augmentation of silica yield using Taguchi technique and artificial neural network approach

Abstract

This study was attempted towards the retrieval of silica from rice husk ash to annihilate the local problems of disposal from the rice milling industries for enhancement of silica purity. Optimization of process factors using the Taguchi technique involved variation in sodium hydroxide concentration (NaOH), alkali impregnation volume per unit weight of the rice husk ash, and reaction time for designing the experimental matrix utilizing L16 orthogonal array at four different levels. The maximum silica extraction was 98.26% obtained with 4 N of NaOH, 20 ml/g of alkali volume, and treatment time 60 min. The identical experimental data set was also applied to an artificial neural network model (ANN) with the LM algorithm for predicting the feasibility of the extraction process. Both Taguchi and neural networks suggested a high coefficient of determination and a satisfactory correlation between experimental and predicted silica recovery values. The detailed characterization of the synthesized silica powder and residual rice husk ash was executed using field emission scanning electron microscopy (energy-dispersive spectroscopy), Fourier transform infrared spectroscopy, thermogravimetric, Brunauer Emmett Tellet surface area, and particle size analysis. The simultaneous reuse of residual ash and silicate was performed to ensure the best possible reclamation of silica and reusability of rice husk ash. The detailed cost estimation of the synthesized silica powder further suggested the effectiveness of the optimized process. Thus, a comprehensive approach for enhancement of the silica yield and purity by adopting Taguchi and ANN optimization proved to be useful in this study.