【作者】 梅自元；

【导师】 周新建；

【作者基本信息】 华东交通大学 ， 载运工具运用工程， 2007， 硕士

【摘要】 车身覆盖件由于结构复杂、精度和表面质量要求严格以及影响其冲压成形的因素之多，决定了其冲压过程中变形规律的复杂性，分析起来十分困难。汽车整车产品中，覆盖件特别是车身覆盖件的市场生命周期最短，变化最频繁。随着汽车工业的发展，覆盖件冲压成形中回弹问题变得越来越棘手，这对提高车身覆盖件制造精度、降低模具成本、缩短车身覆盖件模具开发周期具有十分重要的现实意义。目前回弹预测与控制的问题已成为该领域的研究热点与难点之一。前围板是一种基本的车身覆盖件。论文按照不同的冲压参数，对前围板冲压成形过程进行了仿真试验。结合仿真试验结果，对影响冲压成形性能的部分冲压参数进行了研究。然后，依据成形仿真试验，对前围板进行了回弹仿真试验，并利用正交试验法对影响回弹的冲压参数进行了定性的分析。在前围板回弹仿真正交试验的基础上，本文应用BP神经网络方法建立了基于前围板冲压成形仿真的BP神经网络回弹预测模型，并预测了不同冲压参数(压边力、摩擦系数、板料厚度)下前围板的回弹量，得到了较好的结果。然后，利用BP神经网络预测的回弹量，并与仿真试验结果进行了对比分析，验证了采用仿真试验和BP神经网络对回弹预测的准确性和可靠性。最后，对前围板进行回弹补偿，并且进行了拉深模具设计。最后，利用BP神经网络预测的回弹量，对前围板进行回弹补偿，进行了拉深模具设计。并对实际与理论冲压结果进行了对比分析，验证了采用仿真试验和BP神经网络对回弹预测的准确性和可靠性。更多还原

【Abstract】 In the autobody panel stamping forming process, the rule of blank deforming is complicated and difficult to be analyzed because of the complex autobody panel structure and exigent precision and surface quality and various forming parameters affecting the stamping forming process of autobody panel. Compared with other parts of the automobile, the panels, especially autobody panels change most frequently and their market time are shortest. With the development of the automobile industry, the springback defect of the formed autobody panel becomes more and more intractable, it has much important realistic meaning to improve the accuracy of manufacture, reduce the cost of die and shorten the time cycle of die development. Consequently, the prediction and control of the springback become one of the hottest and most difficult subjects in autobody panel manufacturing.The front panel is a basic type of autobody panel. The simulation of stamping forming process for the front panel is experimentized according to various stamping parameters. With the result of simulation experiment, partial parameters which effect on the stamping forming performance are studied. And then, depending on the simulation experiment, the springback simulation experiment of front panel is finished, and using orthogonal experiment qualitatively analyzes the stamping parameters which effect on springback. Based on orthogonal experiment, BP neural network is used to establish the prediction model of springback, and predicts the springback magnitude of front panel with various stamping parameters (blank holder pressure, frictional coefficient, thickness of blank), the result of prediction is satisfactory. And then, the springback magnitude which is predicted by BP neural network is adopted to compare with the result of simulation experiment, the veracity and reliability of using simulation experiment and BP neural network to predict springback are validated. At last, the springback is compensated, and the cupping die of front panel is designed.更多还原