Modeling and optimization of porous aerogel adsorbent for removal of cadmium from crab viscera homogenate using response surface method and artificial neural network

Abstract

The purpose of this study was to establish an efficient modern strategy for the reduction and control of cadmium (Cd) content in the crab viscera homogenate. To this end, the coupling of an artificial neural network (ANN) technique with the genetic algorithm (GA) method and compared the performance of GA-ANN with response surface methodology (RSM) were demonstrated in the prediction and optimization of the ultrasonic-assisted adsorption of Cd using a aerogel adsorbent. On the basis of single factor test, the data sample was established by Box–Behnken Design, and the effects of ultrasonic time, temperature, the addition amount of aerogel, and pH on the removal rate of Cd were researched. The results revealed that the RSM and GA-ANN models had relative errors and coefficients of determination R2 values of 3.90%, 0.8440, and 1.79%, 0.9283, respectively, suggesting that GA-ANN is more accurate than RSM. The maximum percentage of Cd extracted (79.13%) was obtained using GA-ANN model as ultrasonic time of 87 min, temperature of 81 °C, adsorbent dosage of 0.07 g, pH of 3.20. In conclusion, the present results provide an effective method for optimizing multi-technology coupling to reduce the toxic heavy metal content in aquatic products.