多点柔性工装装夹布局优化研究

Research on clamping distribution optimization of multi-point flexible tooling system

薄壁件在多点柔性工装系统定位和支撑作用下,支撑单元布局对其加工变形的影响尤为明显。为此,结合有限元技术、正交实验理论和多元非线性回归的方法,提出以支撑单元的布局为设计参数,建立以薄壁件的最大变形和平均变形为指标的加工变形预测模型。为了简化计算,采用加权的方法将多目标预测模型转化为单目标预测模型,进一步以该单目标预测模型为目标函数,采用全局寻优能力较强的遗传算法进行优化计算,得到理论最优的装夹布局。实验结果表明:多点柔性工装系统的装夹布局与薄壁件的加工变形确实有非线性回归关系,且通过遗传算法得到的最优装夹布局,经过加工分析其最大变形量和平均变形量分别比均匀分布布局降低了53.0%和54.4%,比正交实验最低降低了19.0%和8.9%。

During the multi - point flexible processing of thin - walled components, support unit distribution has an obvious influence on machine deformation. Therefore, combining the theory of finite element analysis, the orthogonal experiment and the method of multivariate nonlinear regression, put forward to regard the support unit distribute as design parameters, and establish the prediction model putting the maximum deformation and the average deformation as a target. In order to simplify the calculation, a weighted way is used to make the multi -objective prediction model into single objective prediction model, further to treat the single objective prediction model as the objective function, by applying the genetic algorithm to optimize the support unit distribution, and gain the optimal layout. The result shows that support unit distribution and milling deformation do have a nonlinear regressing relationship, and compared to the maximum deformation and the average deformation of uniform distribution, the milling deformation of the optimal layout decrease by 53.0% and 54.4% ; and compared to the minimum deformation of orthogonal experiment decrease by 19.0% and 8.9%.