【作者】 卢碧红；

【导师】 王启义； 赵亮；

【作者基本信息】 大连铁道学院 ， 机械制造及其自动化， 2003， 博士

【摘要】 本文采用虚拟制造的理论与方法，模拟真实加工条件，建立了面向虚拟数控车削加工环境的加工精度预测原型系统。该系统能够预测加工过程的稳定性、加工负荷状态和工件的加工精度。该研究成果为优化切削参数、提高生产效率与加工精度提供了技术支持，有助于工艺师在一个更高的水平上对工艺编排和数控程序加以验证，从而充分发挥数控机床的功能和效率，显著降低企业的生产准备时间和制造成本。因此，本论文研究成果对虚拟制造及其相关技术在我国的应用与开展具有重要的理论意义，对促进制造业技术水平和市场竞争能力的提高具有现实意义。论文着重研究并解决以下问题： 1．针对目前加工仿真尚未解决加工精度的预测问题，提出虚拟检测系统的概念；在分析系统功能需求的基础上，提出了建立精度预测系统的原则和构建基于边界模块与核心模块的虚拟检测系统的方法；用边界模块预测加工系统的稳定性和切削效率、核心模块预测相关精度参数，解决了现实检测到虚拟检测的映射问题。 2．针对车削加工的特点，在车削加工系统中定义了刀架坐标系、工艺基准坐标系和主轴坐标系，提出了虚拟工件成型的三瞬法，实现了对数控车削加工的建模。通过引入三个坐标系的误差及其映射关系，从而得到含有误差的虚拟工件，为加工精度预测奠定了基础。 3．采用人工神经网络建立了基于切削用量的切削力预测模型，其预测精度由经验公式的±2.5％提高到±0.25％；同时给出了切削功率预测以及对工艺系统加工能力评价的方法，实现了对加工性能的预测。 4．针对工艺系统受力变形的建模与求解，分别对解析法、基于反问题的有限元法进行了深入探讨，发现在有足够实测信息的前提下，利用反问题求出的虚拟结构可以代替原结构模型进行虚拟加工的实时计算与分析，并通过实验进行了验证。 5．探讨了机床运动误差的测量与辨识方法，通过标定工艺系统坐标系相关坐标轴上的误差分布实现了对机床几何误差与运动误差的辨识，简化了对众多误差源的处理，较好地解决了加工空间法标定难和综合法数学表达相对复杂的问题，为具体机床加工精度的预测奠定了基础。 6．针对加工精度预测的功能需求，分别从单一坐标系误差和整个工艺系统误差影响的两个层面，研究了车削加工工件的预测问题。在研究综合误差影响的预测时，就尺寸精度、形状精度、位置精度和表面粗糙度四个方面详细研究了精度预测模型的建立方法，推导出了基于刀尖圆弧半径、刀具副偏角和进给量预测加工表面粗糙度的大连铁道学院工学博士学位论文公式，通过相关模型的建立实现了虚拟车削加工工件的精度预测，同时为低成本优化加工过程提出了可行的技术路线。 7.提出用颤振作为预测加工状态稳定性的依据，通过分析力与加速度信号的时、频域特征，得出颤振特征提取的一般方法，借助不规则系数法以及颤振门限的确定实现了基于切削力的颤振预测。 8.基于上述理论与方法开发出面向虚拟车削加工环境的加工精度预测原型系统。将本预测系统嵌入虚拟加工环境，实现了面向虚拟加工环境的加工状态稳定性、机床加工能力和工件加工精度的预测，并结合具体加工工件对本文提出的主要理论方法进行了系统运行验证。更多还原

【Abstract】 Based on the theory of virtual manufacturing, a new prototype system of machining accuracy prediction in virtual NC turning has been developed, which is able to imitate the real machining conditions, and predict the stability of machining process, the state of machining load and the machining accuracy of parts. This system can provide the technical supports, on the one hand, for optimizing the machining parameters, increasing production efficiency and improving the machining accuracy, on the other hand, for validating the process planning and NC program at a high level so that the NC machine tools can be utilized to its full function and performance, and the preparation time of production and the production cost can be significantly reduced.The original contributions of this dissertation in both theory and application of virtual manufacturing are:1.A concept of virtual measuring system is proposed for solving the machining accuracy prediction problem. It follows that the principle of building up the accuracy prediction system is proposed, and a method of constructing the virtual measuring system based on boundary modules and core modules is presented. In this virtual measuring system, the boundary modules are used for predicting the stability and efficiency of the machining system, and the core modules are used for estimating the accuracy parameters; together they implement the mapping from real measuring to virtual measuring.2.A 3-Instantaneous-Points method is proposed for shaping the virtual parts, where the machining system is simplified to the tool and the part interconnected by the tool coordinate system, the processing reference coordinate system and the spindle coordinate system. Synthesizing the errors given in the three coordinate systems in a proper way, the virtual parts containing errors are built up, which lays the foundation for machining accuracy prediction.3. A neural-network-based cutting-force prediction (NNBCFP) model is developed, which can improve the predicted accuracy from ?.5% to ?0.25%. In the meantime, this NNBCFP model gives the methods of predicting cutting power and evaluating the machining capability of the processing system.4. Modeling and solving the forced deformation of machining system is explored based on both the analytical method and the finite element method for inverse problem. An important finding is that with enough real measurement information, the original structuremodel can be replaced by the virtual structure solved from the inverse problem for the real-time evaluation of the virtual machining. This finding is verified by the experiment.5. A method of measuring and identifying the lathe motion error is proposed, where the coordinate system is related with processing system and machining state, the identification of geometrical errors and motion errors of the machine tool is accomplished by calibrating the error distribution in the corresponding coordinate axes. This method simplifies processing multiple error sources and lays the foundation for machining accuracy prediction.6. The individual coordinate system error and the whole machining system error are studied respectively for predicting the machining accuracy. The accuracy prediction model which takes the dimension accuracy, form accuracy, position accuracy and surface roughness into account, is developed, and the surface roughness prediction formula is derived based on the nose radius, the minor edge angle of the tool and the feed. This accuracy prediction model not only implements the machining accuracy prediction of the virtual parts, but also provides an applied guideline to reduce cost and optimize machining process.7. Chatter is used to predict the stability of machining state in this dissertation. A new method to predict chatter based on cutting force measurement is proposed and implemented by the non-regular coefficient method and the chatter threshold determination, where the chatter features are abstracted from the features of force and acceleration signals in both更多还原