Optimization of crankshaft main bearing lubrication performance considering bearing profiles

It is the aim of this work to reduce friction power loss of main bearings by optimization. To this purpose, elastohydrodynamic (EHD) model is used for EHD calculations for different main bearings. BP neural network is implemented to establish the approximation model for bearings. Then, multi-objective optimization of bearings using genetic algorithm is formulated and conducted. It is found that a more compliant bearing profile can provide hydrodynamic lift during film lubrication while bearing profiles have more significant impact on lubrication performance in comparison to other key parameters. The results of the BP network model using the genetic algorithm agree closely with the calculated value based on EHD-MBD model. The presented approach allows reliably to conduct the optimization of bearings. After optimization, the friction power loss is significantly reduced while the minimum oil film thickness increases and the total pressure drops.