基于薄板件装夹变形控制的夹紧点位置优化方法研究

摘要

对薄板件装夹变形控制问题进行了研究,分析了薄板件装夹变形控制的不足,提出了一种基于工件变形控制的装夹夹紧点位置优化方法--耦合BP神经网络与遗传算法.采用均匀试验设计表设计仿真方案,基于ABAQUS仿真数据建立了BP神经网络预测模型,结合该预测模型在夹紧点设定区域内构建优化数学模型,并用遗传算法进行优化.对优化后的结果进行了仿真验证.结果表明装夹变形明显减小.说明耦合BP神经网络与遗传算法的优化方法是可行的.%The paper studied the problem on control of deformation in fixing thin-walled workpiece, analyzed the deficiencies on control of deformation causing by fixing.A clamping-point optimization method-by coupling neural network and genetic algorithm (GA) was proposed.The simulation program was designed based on uniform experimental design technique.A prediction model for deformation was created using artificial neural network exploiting simulating data in ABAQUS.Neural network model and initialed range were employed in the construction of optimization problem.The optimization process was solved by genetic algorithm.An Additional simulation was conducted to validate the optimal results.This indicates that the neural network model coupled with GA can effectively utilized in fixation of clamping points.