Optimization numerically for five-pad tilting-pad journal bearings design based on particle swarm

Abstract

The design of five-pad tilting-pad journal bearings (TPJB) is complicated because of the uncertainty of design parameters which affects the performance of TPJB greatly. In this study, the numerical lubrication model is established to predict the performance of TPJB. Furthermore, a new optimal method combining the lubrication model and particle swarm method with a comprehensive target is brought out, which can obtain the optimal design parameters correctly and rapidly compared with enumeration method. The performance of TPJB is better after optimization, especially the temperature and frictional coefficient decreasing by about 31.1% and 35.8%, respectively. Moreover, different application scenarios of TPJB need different weight factor combinations in the comprehensive target, which is valuable for different design requirements.