Optimization of Surface Texture in Double Rectangular Cavity Hydrostatic Thrust Bearing Through GA genetic algorithm

Abstract

Abstract Optimization of surface texture in liquid hydrostatic thrust bearing is particularly important in order to improve quality of processed products. There are excellent nonlinear ability and quite flexible network structure in BP neural networks, which can be used to achieve optimization of surface texture in all aspects of thrust bearing. The model of surface texture size parameters and the oil cavity pressure are established by BP neural network, and the experiment is designed based on orthogonal experimental samples. The optimal parameters of the texture size were optimized using the GA genetic algorithm, yielding a distance L = 1.2323 between the texture and the oil cavity, a width B = 0.99547, a depth H = 1.4714, and a corresponding mean pressure of the oil cavity P = 0.11882MPa. In particular, the sensitivity simulation method is able to find the optimal number of "type 1" surface textures on the oil sealing edge.