多轴数控机床准静态空间误差建模及误差辨识方法研究

发布时间：2018-08-12 18:07

【摘要】：随着我国制造业转型升级的深入发展,各行业对精密机械产品的需求量不断攀升。数控机床作为制造业的基础装备,其加工精度直接影响工件的品质。因此,数控机床加工精度一直是学术界和工程领域的研究热点。然而,精度和误差是一对相互矛盾关系,揭示数控机床各种误差产生的机理、研究误差建模及控制方法是提高数控机床加工精度亟待解决的关键科学问题。为了解决上述问题,本论文深入研究多轴数控机床准静态空间误差建模及误差项检测辨识方法,并结合机床实验,验证所提出方法的正确性和可行性。本论文的主要研究工作和成果如下:(1)提出一种基于增量矩阵的空间误差建模方法:采用增量的形式探讨机床相邻体间的实际坐标变换,通过叠加相邻体间位姿引起的误差求解机床总空间误差,为直观识别多轴数控机床关键误差源提供依据。基于所提出的建模方法,利用微分法建立了数控机床误差灵敏度分析模型,为数控机床精度分配、局部误差补偿提供理论依据。采用所提出的建模方法,建立了三轴数控机床和五轴RTTTR型数控机床空间综合误差模型,并对两种数控机床进行误差灵敏度分析。(2)设计多轴数控机床移动轴和旋转轴误差检测实验,建立了误差辨识模型:利用球杆仪将旋转轴位移误差和转角误差进行分开测量辨识,以RTTTR型五轴联动机床的A、C轴为研究对象,采用轴向和径向测量法测量辨识位移误差;结合A、C轴的具体特点,分别设计了两轴转角误差检测实验,建立了转角误差辨识模型。针对激光干涉仪九线测量法对测点要求严格、误差项计算值含累积误差、直线度测量值含垂直度问题,提出改进措施并建立新的移动轴误差辨识模型。对五轴龙门机床QLM27100-5X开展移动轴误差检测辨识,探讨三轴移动引起的空间几何误差和热误差分布,并对空间几何误差进行灵敏度分析,验证了所提出的误差辨识方法的可行性。(3)建立了多轴数控机床进给轴几何误差支持向量机模型和基于热变形修正系数的位移热误差模型:基于多轴数控机床进给轴几何误差测量数据样本少、非线性强的特点,建立移动轴和旋转轴几何误差支持向量机模型;针对当前研究忽略移动轴轴向和径向热变形系数不一致、对热误差特性把握不全面的问题,从分析移动轴热变形机理着手,考虑晶体材料线膨胀系数和体膨胀系数的关联性,从理论上修正丝杠轴向和径向热变形公式,根据移动进给系统温度场非均匀性特点,建立基于热变形修正系数的定位热误差和直线度热误差模型;以五轴龙门机床QLM27100-5X的移动进给系统为研究对象,验证了两种模型的精度。(4)提出基于最优阈值的温度变量分组方法,以分组得到的典型温度变量作为建模自变量,建立主轴热误差分片逆回归模型:以模糊聚类和相关性分析为基础,提出一种基于最优阈值的温度变量分组方法,解决热误差建模中温度变量分组过于依赖经验的问题;基于该方法获取最佳建模温度变量,提出分片逆回归热误差建模方法,解决了当前常见的热误差回归模型函数形式固定、做长期预测时精度不高的问题,提高了热误差模型的泛化性和外推能力;采用所提的方法,对卧式加工中心MCH63进行温度变量分组,找出最佳温度变量用于主轴热误差分片逆回归建模,分析显示模型精度良好。(5)以五轴联动机床QLM27100-5X和加工中心MCH63为实验平台,利用课题组自主开发的温度-误差检测系统及误差补偿系统,实时检测机床温度和热误差数据,并对移动轴单项定位几何误差、移动轴空间热误差及主轴轴向热误差三类模型进行了仿真验证及实验验证。
[Abstract]:With the deepening development of the transformation and upgrading of China's manufacturing industry, the demand for precision machinery products in all walks of life is increasing. As the basic equipment of manufacturing industry, the machining accuracy of CNC machine tools has a direct impact on the quality of the workpiece. In order to solve the above problems, this paper deeply studies the quasi-static spatial error modeling and error detection and identification methods of multi-axis CNC machine tools, and combines them with other methods. The main research work and achievements of this paper are as follows: (1) A space error modeling method based on incremental matrix is proposed. The actual coordinate transformation between adjacent bodies of machine tools is discussed in incremental form, and the total space of machine tools is solved by superposing the errors caused by the position and pose between adjacent bodies. Based on the proposed modeling method, the error sensitivity analysis model of NC machine tools is established by using the differential method, which provides the theoretical basis for accuracy allocation and local error compensation of NC machine tools. The error sensitivity analysis of two kinds of CNC machine tools is carried out. (2) The error detection experiment of moving and rotating axes of multi-axis CNC machine tools is designed, and the error identification model is established. The displacement error and rotating angle error of rotating axes are measured and identified separately by using the Ball-Bar instrument, and the error of RTTTR five-axis linkage machine tool is identified. A and C axes are taken as the research object, and the displacement error is measured and identified by axial and radial measurement methods. According to the specific characteristics of A and C axes, two axes rotation error detection experiments are designed and the identification model of rotation error is established. Improvement measures are proposed and a new identification model of moving axis error is established. The moving axis error detection and identification of five-axis gantry machine QLM27100-5X is carried out. The spatial geometric error and thermal error distribution caused by three-axis movement are discussed. Sensitivity analysis of spatial geometric error is carried out to verify the feasibility of the proposed identification method. Feasibility. (3) The geometric error support vector machine (SVM) model and the displacement thermal error model based on thermal deformation correction coefficient are established for the feed shaft of multi-axis CNC machine tools. The problem of neglecting the inconsistency between axial and radial thermal deformation coefficients of the moving shaft and the incomplete understanding of the thermal error characteristics is studied. Based on the analysis of the thermal deformation mechanism of the moving shaft, and considering the correlation between the linear expansion coefficient and the volume expansion coefficient of the crystal material, the axial and radial thermal deformation formulas of the lead screw are modified theoretically, and the temperature field of the moving feed system is non-uniform. Based on the uniformity characteristics, the location thermal error and straightness thermal error models based on the thermal deformation correction coefficient are established, and the accuracy of the two models is validated by the moving feed system of the five-axis gantry machine QLM27100-5X. (4) A grouping method of temperature variables based on the optimal threshold is proposed, and the typical temperature variables obtained by grouping are taken as the modeling independent. A piecewise inverse regression model of spindle thermal error is established. Based on fuzzy clustering and correlation analysis, a grouping method of temperature variables based on optimal threshold is proposed to solve the problem that the grouping of temperature variables depends too much on experience in thermal error modeling. The modeling method solves the problem that the function form of the common regression model of thermal error is fixed and the precision is not high when making long-term prediction, and improves the generalization and extrapolation ability of the thermal error model. Regression modeling and analysis show that the accuracy of the model is good. (5) Taking QLM27100-5X and MCH63 as the experimental platform, using the temperature-error detection system and error compensation system developed by our research group, the temperature and thermal error data of the machine tool are measured in real time, and the single positioning geometric error of the moving shaft, the spatial thermal error of the moving shaft and the error of the moving shaft are measured. Three kinds of models of axial thermal error of spindle are simulated and verified by experiments.
【学位授予单位】：南京航空航天大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TG659

【相似文献】