Prediction of contact angle of coke-pitch system from raw material properties using artificial neural network

Carbon anodes used during the electrolytic aluminum production are made of aggregate material (petroleum coke, butts, and recycled green and baked anodes) and coal tar pitch. A clear understanding of chemical and physical interactions taking place during mixing would facilitate the selection of raw materials and optimization of mixing parameters that improve anode properties. It is well-known that good interaction between coke and pitch is essential for the creation of a satisfactory bond between them, and contact angle is a measure of this interaction. To optimize and predict the contact angle for a given coke/pitch pair artificial neural network (ANN) model is employed to predict the contact angles at 80 s and 1500 s of contact time. A quantitative relationship between raw material chemical properties and contact angle is established. It was found that oxygen containing functional groups are the most important factor impacting the wettability of coke by pitch. The obtained results demonstrated that the developed models are highly effective in estimating the contact angle of coke/pitch pair. The analysis provided an insight into the effect of different parameters on contact angle. In turn, this might help improve the quality of bonding between coke-pitch, consequently, anode properties.