SVM和改进粒子群算法在冲压成形优化中的应用

Applications of SVM and Improved Particle Swarm Algorithm to Sheet Metal Forming Optimization

针对薄板冲压成形工艺优化中的复杂变量关系,提出了一种新的优化方法。将非线性动态改进惯性权值、非线性动态调整加速因子和引入自适应粒子变异3种策略应用于标准粒子群算法的改进中,基于支持向量机构建了工艺参数与成形质量之间精确的回归模型,并与改进粒子群优化算法相结合,对板料冲压成形工艺参数进行了优化,优化结果有效地控制了起皱和开裂缺陷,提高了板料成形质量。

In view of the complex variable relation in the optimization of sheet metal forming process, a novel optimization method is proposed. By applying three strategies （nonlinear dynamic improvement of inertial weights, nonlinear dynamic adjustment of acceleration factor and introduction of adaptive particle mutation） to the improvement of standard particle swarm optimization algorithm, an accurate regression model between process pa- rameters and forming quality is constructed based on support vector machine, and combined with improved particle swarm optimization algorithm, an optimization on the parameters of sheet metal forming is conducted. The results of optimization effectively control the cracking and wrinkling defects and improve the quality of sheet metal forming.