薄壁件侧铣变形数值建模及实验研究

发布时间：2018-04-26 01:10

本文选题：薄壁件 + 热力耦合变形仿真　；参考：《哈尔滨工业大学》2017年硕士论文

【摘要】：薄壁类零件广泛应用于航空航天、汽车制造等领域。尤其是在航空航天领域,对薄壁类零件的加工精度有着极高的要求。由于薄壁类零件具有材料去除率高、结构刚性差等特点,在进行铣削加工时极易产生加工变形,这使得高精度薄壁类零件的加工制造成为了行业难点问题。本课题从减小薄壁件侧铣变形加工误差的角度出发,旨在通过对薄壁件侧铣变形规律和控制方法的研究,探索高精度薄壁类零件的制造方法,为实际加工生产和应用奠定基础。铣削力和铣削热是薄壁件侧铣变形的主要影响因素,针对这两种主要影响因素,本文通过分析对比几种经典的铣削力理论模型的优缺点,结合薄壁件侧铣加工方式的特点,建立了基于单位铣削力系数的力学模型。通过对比不同的温度场建模方式,分析了铣削过程中瞬时切屑厚度的变化规律,建立了基于Jeager正交切削移动热源带模型的薄壁件侧铣铣削温度预测模型,并利用Matlab模拟了铣削温度场的变化过程。通过铣削加工正交实验测定出不同铣削参数下的铣削力、铣削热、参考点变形量和已加工工件表面误差,根据实验数据计算出铣削力和铣削热模型中的相关系数,并将理论值与实验值进行对比分析,验证了理论模型的有效性。分析了不同铣削参数与铣削变形量之间内在联系,得出了较好的铣削参数组合。通过ANSYS参数化编程的方法编写了集几何建模、刀具走刀、材料去除于一体的薄壁件侧铣动态仿真程序,建立了侧铣变形分析模型。通过合理的假设和简化建立了薄壁件侧铣变形有限元分析模型,对比了只施加力载荷和热力耦合作用时薄壁件的变形规律,将仿真值与实验值进行对比,验证了薄壁件侧铣变形数值模拟的有效性。利用薄壁侧铣变形数值模拟的数据,建立了基于BP神经网络的变形预测模型和铣削参数优化模型。运用Matlab编写了BP神经网络薄壁件侧铣变形预测程序,用实验数据和已建立的BP神经网络变形预测模型,建立了基于遗传算法的铣削参数优化模型。对铣削参数进行了优化,通过铣削参数优化实验验证了优化模型的有效性。
[Abstract]:Thin-wall parts are widely used in aerospace, automobile manufacturing and other fields. Especially in the field of aeronautics and astronautics, the machining accuracy of thin-wall parts is very high. Due to the characteristics of high material removal rate and poor structural rigidity, thin-walled parts are easy to produce machining deformation in milling, which makes the manufacture of high-precision thin-walled parts become a difficult problem in the industry. From the point of view of reducing the machining error of side milling deformation of thin-walled parts, this paper aims at exploring the manufacturing methods of high-precision thin-walled parts by studying the law and control methods of side-milling deformation of thin-walled parts, which will lay a foundation for practical processing, production and application. Milling force and milling heat are the main influencing factors of side milling deformation of thin-walled parts. In view of these two main influencing factors, this paper analyzes and compares the advantages and disadvantages of several classical theoretical models of milling force, combined with the characteristics of side milling of thin-walled parts. A mechanical model based on unit milling force coefficient is established. By comparing different modeling methods of temperature field, the variation law of instantaneous chip thickness in milling process is analyzed, and the prediction model of side milling temperature of thin-walled parts based on Jeager orthogonal cutting moving heat source belt model is established. The change process of milling temperature field is simulated by Matlab. The milling force, milling heat, reference point deformation and the surface error of the machined workpiece are determined by orthogonal milling experiment. The milling force and the correlation coefficient in the milling heat model are calculated according to the experimental data. The validity of the theoretical model is verified by comparing the theoretical value with the experimental value. The internal relationship between different milling parameters and milling deformation is analyzed, and a better milling parameter combination is obtained. A dynamic simulation program for side milling of thin-walled parts is developed by means of ANSYS parameterized programming, which integrates geometric modeling, tool walking and material removal, and an analysis model of side milling deformation is established. Through reasonable assumption and simplification, the finite element analysis model of side milling deformation of thin-walled parts is established, and the deformation law of thin-walled parts under the coupling of force and heat is compared, and the simulation value is compared with the experimental value. The effectiveness of numerical simulation of side milling deformation of thin wall parts is verified. The deformation prediction model and milling parameter optimization model based on BP neural network are established based on the numerical simulation data of thin-walled side milling deformation. A BP neural network program for predicting the side milling deformation of thin-walled parts is compiled by using Matlab. Based on the experimental data and the established BP neural network deformation prediction model, the optimization model of milling parameters based on genetic algorithm is established. The milling parameters are optimized and the validity of the optimization model is verified by the experiment of milling parameters optimization.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG54

【相似文献】