Research on intelligent design of the layout of tank washing machines in cargo oil tanks based on optimal tank cleaning effect

Abstract

This study proposes a rapid method for evaluating tank-washing efficiency and optimizing tank-washing-machine layouts using a dynamic pressure-illuminance surrogate framework. Water-jet experiments and numerical simulations established the relationship between wall dynamic pressure and removal rate, with the pressure corresponding to an 85% removal rate defined as the effective-cleaning threshold. Jet impingement was represented by a point light source, and the illuminance field in a cargo oil tank was simulated in Ansys Speos. A calibrated illuminance-pressure mapping converted optical results into cleaning coverage. Candidate layouts were sampled under structural constraints to train machine-learning models linking layout to effective-cleaning area. The best-performing model was used for a global search of the optimal layout under obstacle-avoidance constraints. The results show that the method can rapidly predict cleaning coverage and optimize tank-washing-machine layouts, providing a practical framework for automated cargo oil tank cleaning design.