Multi-objective Shape Optimization of Hydrodynamic Journal Bearings Using Non-dorminated Sorting Genetic Algorithm Ⅱ

In order to overcome the shortcoming caused by taking the eccentricity as the initial design variables in the shape optimization model, i.e., the optimized shape was localized in the original shape, a new mathematic model of multi-objective shape optimization for hydrodynamic journal bearings was built through the general film thickness equation based on the Fourier series. In the model, the objective function is minimization of oil leakage and power loss, which is subjected to the minimum oil film thickness and the minimum bearing capacity, and the coefficients of general oil film thickness are design variables. The multi-objective optimization for the profile design of a hydrodynamic journal bearing is accomplished using a modified non-dominated sorting genetic algorithm II (NSGA-Ⅱ). The governing equations based on the general film thickness were solved by Matlab PDE (partial differential equation) toolbox. The results of a case study show that the bearing shape obtained by multi-objective optimization based on the general film thickness equation is not limited by the original shape. In addition,a Pareto-optimal set is obtained, where, for one of the optimized non-circle bearings with the maximum load capacity, the power loss and leakage rate are decreased by 80.8% and three orders of magnitude, respectively, compared with that just taking the load capacity as the objective function.