Research On Approximate Model And Multi-objective Optimization For Static And Dynamic Performance Of CNC Lathe Bed

As a key load-bearing component of lathe,the lathe bed's static and dynamic performance directly affects the processing quality of workpiece.With the improvement of lathe speed and machining accuracy,higher requirements are put forward for all aspects of the lathe bed performance.Research shows that optimizing the lathe bed structure and its parameters is one of the effective ways to improve the performance of lathe.In view of this,this paper chooses CNC horizontal lathe of an enterprise as the research object,and carries out research work on static and dynamic performance of the lathe bed.Specific research contents are as follows:Firstly,the structural characteristics of the original CNC lathe bed are analyzed.The load distribution of the lathe bed joint surface is discussed,and the mechanical model of the bed is established accordingly.By using ANSYS software,the static and modal analysis of the bed was carried out,the performance parameters of the bed were extracted,and the weak parts of the lathe bed structure were found.Secondly,the methods of improving the static and dynamic performance of the lathe bed are summarized,and the weak parts of the lathe bed are improved.The results show that the static deformation is reduced by 15.24%,the first order mode frequency is increased by 3.43%,and the mass is increased by 99.3Kg for structurally improved lathe bed.Indicating that the lathe bed still has an optimization space and needs to be optimized twice.In view of the improved lathe bed,17 structural parameters were selected as the variables to be optimized.Based on Morris sensitivity analysis results,13 parameters which have great influence on the performance of the lathe bed were selected.And Sample information was obtained by the the optimal Latin hypercube design method accordingly.Thirdly,five commonly used approximate model parameter selection methods,precision influencing factors,and the construction method in MATLAB are analyzed,and cross validation,cyclic structure and genetic algorithm are applied to optimize model parameters,which improves the generalization ability of approximate model.Various approximate models are used to express the non-linear relationship between lathe bed structure parameters and performance indicators,and the prediction accuracy of the models are compared and analyzed.The results show that the second-order polynomial response surface model,Kriging model with first-order polynomial regression term and Kriging model with constant regression term have the highest prediction accuracy for mass,static total deformation and first-order modal frequency of lathe bed respectively.Finally,in order to realize the lightweight design that takes into account the static dynamic performance of the lathe bed,an optimization design method aiming at mass,static total deformation and first-order modal frequency is proposed.Based on approximate model technology,a multi-objective optimization mathematical model of lathe bed is established,and the optimal solution set of Pareto is obtained by using the gamultiobj function based on NSGA-II,from which a set of optimal design schemes are selected by TOPSIS method.Comparing the lathe bed optimization results before and after,the results show that three objectives of the optimized lathe bed have been improved.In summary,a multi-objective optimization method of lathe bed structure parameters combining approximate model with intelligent algorithm is proposed in this paper,which can guide the design and optimization of bed body.