白月香,黄昭明,王利,等.基于GA–BPNN的圆筒形件多步冲压成形工艺参数优化[J].精密成形工程,2022,14(9):79-85.
BAI Yue-xiang,HUANG Zhao-ming,WANG Li,et al.Optimization of Multi-step Stamping Process Parameters of Cylindrical Parts Based on GA-BPNN[J].Journal of Netshape Forming Engineering,2022,14(9):79-85. |
| 基于GA–BPNN的圆筒形件多步冲压成形工艺参数优化 |
| Optimization of Multi-step Stamping Process Parameters of Cylindrical Parts Based on GA-BPNN |
| |
| DOI：10.3969/j.issn.1674-6457.2022.09.011 |
| 中文关键词: 正交试验 BP神经网络 遗传算法 工艺参数 减薄率 |
| 英文关键词: orthogonal experimental BP neural network genetic algorithm process parameters thinning rate |
| 基金项目:江西省教育厅科学技术研究项目(204606);马鞍山工程技术研究中心开放基金(QMSG202102，QMSG202104) |
|
| 摘要点击次数: 1072 |
| 全文下载次数: 430 |
| 中文摘要: |
| 目的 研究精密多步冲压成形过程中圆筒形件出现的成形性问题。方法 基于正交试验设计和极差分析方法、BP神经网络建模以及遗传算法得到最优工艺参数，研究制件成形过程中相关工艺参数对减薄率的影响。结果 通过正交试验设计和极差分析，获得各参数对极大减薄率的影响主次顺序为拉延摩擦因数>拉延压边力>反拉延摩擦因数>反拉延压边力，相对最优工艺参数如下:拉延摩擦因数为0.200、反拉延摩擦因数为0.100、拉延压边力为50 kN、反拉延压边力为30 kN。极大减薄率的仿真极值为0.144 4、极小减薄率的仿真极值为−0.127 7;以极大减薄率为成形质量评价指标，经BP神经网络建模并结合遗传算法寻优，获得最优工艺参数如下:拉延摩擦因数为0.200、反拉延摩擦因数为0.159、拉延压边力为55 kN、拉延压边力为40 kN，极大减薄率预测值(0.134 9)与仿真值(0.140 1)的相对误差仅为3.7%，优化后的制件成形质量良好。结论 所提出的方法对量化调整制件的成形工艺具有良好的工程应用价值。 |
| 英文摘要: |
| The work aims to study the formability of cylindrical parts in precision multi-step stamping. Aiming at the formability problems in multi-step stamping of cylindrical parts, the optimal process parameters were obtained based on orthogonal experimental design, range analysis method, BP neural network modeling and genetic algorithm, and the effect of relevant process parameters on the thinning rate in the forming process was studied. Through orthogonal experimental design and range analysis, the primary and secondary order of the effect of each parameter on the maximum thinning rate was drawn friction factor > drawn blank holder force > reverse drawn friction factor > reverse drawn blank holder force. The relative optimal process parameters were as follows:the drawn friction factor was 0.200, the reverse drawn friction factor was 0.100, the drawn blank holder force was 50 kN, and the reverse drawn blank holder force was 30 kN. The extreme simulation value of the maximum thinning rate was 0.144 4 and the extreme simulation value of the minimum thinning rate was −0.127 7. Taking the maximum thinning rate as the forming quality evaluation index, through BP neural network modeling and genetic algorithm optimization, the optimal process parameters were obtained as follows:the drawn friction factor was 0.200, the reverse drawn friction factor was 0.159, the drawn blank holder force was 55 kN, and the drawn blank holder force was 40 kN. The relative error between the predicted value of the maximum thinning rate (0.134 9) and the simulation value (0.140 1) was only 3.7%, and the forming quality of the parts after optimization was good. The proposed method has good engineering application value for quantificationally adjusting the forming process of parts. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
渝公网安备 50010702501719号