【作者】 贺军涛；

【导师】 宋玉泉；

【作者基本信息】 吉林大学 ， 材料加工工程， 2004， 硕士

【摘要】 拉深又称拉延，在板料成形领域，拉深是最常用、最重要的成形方法。目前拉深工艺被广泛应用于生产中成形筒形件、阶梯形件、锥形件、球形件、方盒形件、汽车覆盖件和其它不规则形状的薄壁零件。对于简单形状零件如筒形、圆锥形等的拉深，人们已经进行了大量的研究，并提出了工艺条件相关的极限拉延比(TLDR)来辅助确定拉延的次数。然而，由于传统的拉深工艺采用刚性模具，容易引起起皱、拉裂等缺陷。现在广泛采用在反复试验的基础上控制压边力，调节摩擦的相关输入参数，比如润滑剂、润滑量和改变板材金属的表面微结构，修改模具等方法来避免这些缺陷。这延长了产品的开发周期，增加了生产成本。1955年提出的液压拉深技术使用弹性体代替了刚性模具，在拉深成形过程中存在“摩擦保持效果”和“溢流润滑效果”，能有效的提高板材的拉延极限和拉深件的表面质量，相比传统拉深工艺获得的工件厚度分布也更均匀。此外，还可以避免因制造非对称的、几何形状复杂的模具带来的困难，降低模具生产成本、缩短产品的试制周期。目前此工艺在日本、美国、瑞士等工业发达国家已经获得了应用。对于大多数提出的液压拉深技术，人们只是简单介绍了其在成形筒形件上的优势，而对其在非轴对称件上的应用研究相对较少。而在实际工业生产中的拉深件多是盒形件、汽车覆盖件等非轴对称件。对盒形件和覆盖件的研究及成形分析，有助于指导生产、工艺的制定。目前人们对于盒形件拉深的变形特点、应力状态等成形机理有了一定的了解。然而，对于盒形件的拉深仍未能提出类似于轴对称壳形零件拉延极限比的确定公式。针对盒形件的传统拉深工艺过程，研究者已经提出了压边力大小对成形的影响的规律和新的压边力控制和装置方法；研究者还提出了采用神经网络、专家系统等来实现传统盒形件拉深工艺的参数选定。然而现在基本上还没有针对盒形件液压拉深技术的专门的研究。 <WP=100>从力学角度而言，板材成形是一个同时存在几何非线性（大位移、大应变）、物理非线性（弹塑性、弹粘塑性、各向异性）、接触和摩擦的非常复杂的力学过程。随着计算机的应用和发展和有限元方法的成熟，使得可以利用计算机的强大的计算能力进行板材成形的有限元分析。无凹模液压拉深技术，是由丹麦的Aalborg大学提出的一种新的液压拉深技术，他们已经针对筒形件的拉深作了实验和分析，证明此方法能提高盒形件的拉延极限比。另外，在此方法中由于没有下压边圈，所以不用考虑法兰部位的液体流动及压强变化。这种简单的边界条件使得可以对此工艺进行比较准确的数值模拟。本文使用动态显式有限元算法LS-DYNA对盒形件的传统拉深过程及无模液压拉深过程进行了模拟。为了证明本模拟中使用的材料参数及有限元程序的正确性。在传统拉深模拟中使用了NUMISHEET’93提出的盒形件及模具等，对压边力为19.6KN、10.4KN、5.3KN的模拟结果和Joachim Danckert的实验结果相吻合。通过液压拉深模拟结果和传统拉深工艺模拟结果的对比，证实了液压拉深应用于盒形件的拉深同样能获得更好的厚度分布的工件。然而，部分模拟结果也显示了使用无凹模液压拉深时，拐角处变形的抗力较大，如果液体压强变化曲线设置不当，在盒形件口部拐角处易出现褶皱。在传统拉深、充液拉深和液压机械拉深中的刚性凹模拐角能对拐角处产生很大的法向抗力，冲头的压力将促使其变形。无模液压拉深技术没有凹模，这种抗力是由液体压力来实现的，而过大的液体压强会导致直边处的材料不流动，造成侧壁拉裂。在液压拉深过程中，成形效果同样受多种工艺参数的影响。和传统拉深过程一样，其拉深效果受冲头速度、毛坯外形、压边力等参数的影响。这些参数对成形的效果的影响和传统拉深工艺中的规律是基本相同的。此外，在液压拉深过程还受液体压力、冲头与毛坯之间的距离大小、冲头与上压边圈内圈之间的侧隙大小的影响。论文作者做了大量的有限元模拟分析以探讨这些参数对成形效果影响的基本规律，并从模拟中得出以下的结论：当预胀形的高度较大（即初始时冲头与毛坯的距离较大）时，初始的预胀形阶段就相当于一次浅拉深，而随后的拉深则是反拉深阶段了。这样，在冲头下压拉深阶段金属材料的流动方向与预胀形阶段时的材料流动方向相反，有利于抵消拉深成形的残余应力。此外，预胀形后靠近冲头处的毛<WP=101>坯表面积增大，有利于在较小的液体压力下实现毛坯紧贴冲头加大盒形件内壁与冲头之间的摩擦力，减小已变形的盒形件的底部和角部的变形。成形过程中的液体压强对成形效果影响较大，如果液体压强太小会造成起皱，而压强过大则会造成侧壁拉裂。本论文，在我的导师宋玉泉教授的发明专利“精冲模架”的基础上，依据几种不同的液压拉深工艺，增设了盒形件液压拉深的工装和模具。此装置的特点是：无需液压和气压源，便能施加背压和压边力的气液结合弹力结构，而且弹力的大小及弹力变化的陡度或平稳度均可调。因此，具有结构简单、造价低和使用寿命长的优点，可在单动液压机上实现液压拉深，这也证实了“精冲模架”专利的创新性和科学性。更多还原

【Abstract】 Deep drawing is one of the most used and a very important sheet metal forming process. Deep drawing is widely used in the forming of cylinder cups, ladder-shaped cups, conical cups, rectangular cups, auto-body panel and the other irregular shapes thin-walled works, etc. Now, many researches have been done to study the deep drawing process of cylinder cups and conical cups etc. TLDR(Technological Limiting Drawing Ratio) has been developed to assistant the determination of stages needed in traditional deep drawing process. However, wrinkle and crack often occurr during the traditional process for the using of rigid tools. Trial and error is used to avoid the defaults. Controling the blank holder force, adjusting the input parameters such as lubricate and changing the microstructure on the sheet metal surface, reworking the tools, all these are needed during the tial and error process, which prolonging the lead time and increasing the cost.Hydroforming deep drawing (HDD) was first brought forward in 1955. In this process the rigid tool is replaced by flexible tools. There are friction-holding effect and friction-reducing effect during the hydroforming deep drawing process. These effects can improve the LDR, improve the quality of the surfaces. we can get more uniform thickness through this process than the traditional deep drawing process. Furthermore, it can reduce the cost of the manufacture of the mold, decrease the lead time, for that we can avoid the difficulties brought by the manufacture of anisomerous and complex-shaped mould.To most of the hydroforming deep drawing process, researchers only introduce the advantages in the forming of cylinder cups. But there are fewer <WP=103>researches on the non-axisymmetric parts. Those most used in industry product are rectangular cups, auto-body panel and other non-axisymmetric parts. The research on the deep drawing process and the analysis on forming of rectangular box and auto-body panel will help to instructing the production and the working-out of work process. Now, researchers have some knowleges about the mechanism of deformation such as the characteristic of deformation, stress state etc. Howerver, there is no formula like that of axisymmetric shell has been proposed for the LDR of rectangular box. To the forming of rectangular box by traditional deep drawing process, researchers have found the rule for the effect of BHF on forming product, and proposed the new control methods for BHF and new equipments. Researchers proposed to apply neural network and expert system in the choice of parameters in the deep drawing process. But until now, there is no special research on the forming of rectangular box using hydroforming deep drawing process. In terms of mechanics, sheet forming process is a complicate process including geometrical non-linear (large displacement, large deformation), physical non-linear (elastoplasticity, elastic viscous plasticity, anisotropy), contact problem, and friction problem. Along with the application and development of the computer science and the development of FEM simulation, sheet metal forming process can be simulated by FEM code. Hydromechanical deep drawing without draw die, which was proposed by Aalborg University in Danmark, is a new HDD process. The researchers of Aalborg University have done some experiments and analysis about this process on cylinder cups, which show that LDR can be increased by using this process. There is no lower ring in this method, so it is not needed to consider the flow of the liquid and the pressure changing at the flange area. For the simple boundary condition, exact FEM simulation on this process can be carried out. In this thesis, explicit dynamic code LS-DYNA is used to simulate the process of both traditional deep drawing and HDD. To verify the correctness of the material and other parameters, the balnk and tools used in this thesis are NUMISHEET’93 <WP=104>benchmark geometries. The simulation results under the BHF of 19.6KN, 10.4KN and 5.3KN is consistent with the experiments directed by Jo更多还原