Modelling wax deposition of diesel in sequential transportation of product oil pipeline using optimized back propagation neural network

Contamination of gasoline by wax deposit of diesel is a severe problem in sequential transportation of product oil pipelines in cold areas. However, most works on wax deposition are focused on crude oil. In response, this paper aims to investigate wax deposition from a unique perspective of diesel oil in sequential transportation. To this end, a cold finger apparatus was designed and constructed. It is found that the wax deposition rate of diesel oil increases with oil temperature and wax content, and decreases with cold finger temperature. A non-monotonic variation trend is observed against shear stress. To predict diesel wax deposition rate, a back propagation (BP) neural network optimized by bald eagle search (BES) algorithm is proposed. Grey relational analysis (GRA) is employed to get the highly relevant factors as input parameters of the developed model. Prediction accuracy and generalization ability of the BES-BP model is experimentally verified. This work will be helpful to schedule the transportation program of product oil to avoid contamination of gasoline by diesel wax deposit.