Implementation Of Kriging Model Based Sequential Design On The Optimization Of Sliding Bearing

Liquid sliding bearings play a vital role in high-speed rotating machinery due to their excellent stability and low friction power consumption.As the modern industry puts higher requirements on bearing performance,and the optimization method is applied in sliding bearings to adapt to the continuous progress and development of rotating machinery.Although the conventional optimization method can get a better design plan,when faced with complex multi-objective and nonlinear functions and evaluating the optimal parameters,it is easy to fall into the local optimal and low efficiency.A built-in prime benefit of Kriging surrogate model resides in its unbiased prediction and the associated confidence intervals.Therefore Kriging model effectively guide the subsequent optimization algorithm toward the global optimal direction.The optimization method of the Kriging model has high optimization efficiency,and has been widely used in many fields such as geological engineering,aerospace and many other fields.However,there are few applications in sliding bearing design,introducing the method into the Kriging model based sequential design in sliding bearing design can greatly improve the optimization efficiency and accuracy.This paper analyzes and discusses the influence of the traditional experimental design method and the sequential method based on Kriging model.The Kriging model is used to provide the advantages of the estimated value of the objective function and the corresponding theoretical error.By studying the global accuracy of the model and the local accuracy of the interval near the optimal solution,introducing the parallel plus-point criterion and the corresponding convergence conditions,a proxy model with high accuracy and good efficiency is obtained.Based on the sequential addition method of Kriging,the model has higher global accuracy and can converge to the real optimal solution of the optimization problem more quickly.Taking radial sliding bearings of different structural forms as the optimal design object,Kriging mode based sequential design has the most significant effect on reducing the friction power consumption per unit bearing capacity under the limited iteration steps,and the rapid convergence of the method is verified.In multi-objective agents,the self-organizing mapping methods take the advantages of hidden high-dimensional multi-attribute data features in the low-dimensional visual space.The self-organizing mapping method is used to optimize the multi-objective Pareto obtained by using NSGA-II.The clustering operation is performed on the optimal solution data to obtain the characteristics distribution and mapping relationship of the target and parameter in the high-dimensional data.By analyzing the mapping relationship between the goals and parameters in the multi-objective optimization results,and their correlation characteristics is extracted to visualize the optimization design results.The optimization results show that the SOM diagram accurately expresses the mapping relationship between the actual design space and the target space,and can be used to guide the designer to choose the optimal solution.After the optimization of this method,the friction power consumption and temperature rise of the sliding bearing's unit bearing capacity are reduced,and the unstable speed is obviously increased,which is highly consistent with the current research results.