Selective Leaching of Arsenic from High-Arsenic Dust in the Alkaline System and its Prediction Model Using Artificial Neural Network - Mining, Metallurgy & Exploration (2021)

This study investigated the selective removal of arsenic from high-arsenic dust in alkaline systems and the effects of different leaching conditions. The results indicated that the liquid–solid ratio, NaOH concentration, and sulfur dosage had a significant influence on the process. The leaching efficiency of arsenic reached 99.57%, while that of lead was as low as 0.03% under appropriate conditions. In particular, the addition of sulfur can effectively promote the leaching of arsenic and reduce the dissolution of lead in the solutions. An artificial neural network was used to model the leaching process. It consisted of a back-propagation artificial neural network model with a “6–10–2” structure that could effectively simulate and predict the value with more than 99% accuracy. Based on the difference in weights of the different parameters in the neural network model, the relative importance of the parameters related to arsenic and lead leaching efficiency was obtained, which followed the order of NaOH concentration, liquid–solid ratio, sulfur dosage, temperature, time, and stirring speed.