论文作者 武云霞
论文导师 王成勇,论文学位 硕士,论文专业 机械制造及其自动化
论文单位 广东工业大学,点击次数 39,论文页数 77页File Size4863k
论文网 http://www.lw23.com/lunwen_126703557/ 石墨电极;电火花加工;BP神经网络
Graphite electrode;EDM;BP neural network
    石墨电极具有电极损耗小、加工速度快、耐高温、加工精度高等优点，是模具电火花加工理想的电极材料。开展石墨电极电火花加工工艺研究，对于推动石墨电极的应用，提高模具的加工精度具有理论指导意义。 本论文通过对石墨电极电火花加工特性及电加工工艺与电加工机床的适应性研究，建立石墨电极电火花加工参数优化模型，为石墨电极电火花加工提供依据。 1、通过大量的电加工实验和多种测试手段，研究了电极极性、脉冲宽度、脉冲间隔、开路电压、峰值电流、石墨颗粒尺寸和工件材料对石墨电极电火花加工的影响规律。 2、分别在4种不同的电火花加工机床上对石墨电极电加工特性进行了研究，分析了石墨电极电火花加工工艺与电火花机床的适应性。 3、初步利用BP神经网络技术建立了石墨电极电火花加工参数优化模型。 通过系列实验研究和理论分析，获得以下结论： 1、脉冲宽度和峰值电流对石墨电极电火花加工特性(电极损耗、加工速度和表面粗糙度)影响比较显著，脉冲宽度越大，电极损耗越小，存在负损耗。开路电压和脉冲间隔存在最优值，石墨颗粒尺寸和工件材料对石墨电极电火花加工特性影响也比较显著。 2、Charmilles ROBOFORM 35机床的加工速度较快，加工精度高，但电极损耗较大，适合精密加工；Sodick A35R机床加工精度良好，电极损耗较小，但加工速度慢，效率低；GOLD SAN机床和AGIE机床放电状态不稳定，容易发生电弧放电和烧蚀现象，加工精度低，不太适合石墨电极电火花加工。 3、石墨电极电火花加工工艺模型可以有效地预测加工效果，该模型对加工速度、表面粗糙度和电极损耗比的平均预测误差分别为3.62％、2％、15.7％，真实反映了机床的加工工艺规律。

    Because of the advantages such as lower wear ratio, rapid machining speed, high heat-resistant and machining precision, graphite electrode is a kind of perfect electrode material in the EDM process of die and mould. In order to improve its use in the industry, it is valuable and important to research machining technology of graphite electrode for EDM.In this thesis, parameters optimized selection model of graphite electrode for EDM is established based on the research of EDM characteristics of graphite electrode and adaptability of machining technology to different EDM machine tools.Effects of electrode polarity, pulse duration, pulse interval, open-circuit voltage, peak current, grain size of graphite electrode and workpiece material on EDM characteristics of graphite electrode were studied through a lot of experiments and analysis. EDM characteristics of graphite electrode were measured on four EDM machine tools in order to analyze adaptability of machining technology to EDM machine tools. EDM machining technology model with graphite electrode established by BP neural network were carried out.The results indicate that effects of pulse duration and peak current on EDM characteristics, including the electrode wear, material removal rate and surface roughness, are significant. There are an optimum pulse-off time and open-circuit voltage in this experiment. Grain size of graphite, workpiece material and it was also found in this thesis that EDM machine tools affect the results of EDM characteristics. It was proved that that EDM machining technology model with graphite electrode established by BP neural network can effectively predict machining effect, and really reflect machining technology rule of machine tools. The mean predicted errors were 3.62%, 2%, 15.7% for material removal rate, surface roughness and electrode wear, respectively,