Estimation of optimal physico-chemical characteristics of nano-sized inorganic blue pigment by combined artificial neural network and response surface methodology

Abstract

The application of response surface methodology combined with artificial neural network was used to estimate the optimal characteristics of nano-sized cobalt based pigment. Firstly, the nano-sized blue powder was synthesized by stoichiometric contents of cobalt and aluminum nitrates using autoignition technique. The study was conducted over a wide range of operating conditions, designed by response surface methodology, in terms of pH, fuel ratio and calcination temperature. The crystallite size, specific surface area, color behavior and crystallinity of powders were determined according to the standard methods. Secondly, several artificial neural networks were designed and then examined for prediction of pigment characteristics. The appropriate model was obtained to achieve better prediction and then the response surface methodology was applied to screen the artificial neural network output data for optimizing synthesis condition. It was concluded that the trained artificial neural network combined with response surface methodology can provide the synergetic pigment synthesis conditions. The additional validations were performed and the results showed acceptable error between the predicted and experimental data. The application of presented algorithm can be important tool for reliable synthesis nano-sized blue powder.