基于Kalman Filter-GM理论的数控机床热误差建模研究

发布时间：2018-08-02 10:09

【摘要】：随着“中国制造2025”计划的提出,我国开始实施制造强国三步走战略。而数控机床作为工业母机,在制造业中占有重要的地位,因此提高数控机床精度是非常有意义的。提高数控机床精度方法之一就是对数控机床误差进行建模分析和补偿,这篇论文主要围绕着数控机床的误差测量和建模分析来写。在数控机床各种误差源中,热误差是数控机床等精密加工机械的最大误差源之一。根据国内外研究者的研究内容,本文搭建了数控机床主轴热误差研究平台,主要从热误差测量、温度敏感点选择、建模和模型对比验证分析四个方面展开叙述,具体如下:(1)首先搭建数控机床热误差测量平台,然后通过模糊聚类分析和灰色关联分析对数控机床热误差测温点进行优化选择,确保数据的有效性和准确性,为接下来的建模收集和分析数据。(2)基于卡尔曼滤波-灰色模型理论(Kalman Filter-GM)建模,能够有效地减少噪声对观测值的影响,这样就可以得到观测值比较好的估计,从而确保基于Kalman Filter-GM的数控机床热误差的有效建模;在建立模型过程中,相比传统的最小二乘法估计参数,这种方法更加稳健,因此这种模型相对于传统的建模来说较好;(3)文中的残差修正模型是通过模拟值与建模数据建立的,这种方法可以让预测效果更好;基于卡尔曼滤波方法通过迭代变形数据分别构建了几个灰色模型,然后通过数据融合的方法把这几个灰色模型的数据进行融合,最终得到了预测数据的最佳估计值。(4)基于Kalman Filter-GM所建立的数控机床热误差模型,与传统的建模方法最小二乘法和最小二乘法支持向量机的数控机床热误差建模方法进行数据对比分析,最终得出结论,此种建模方法更好。
[Abstract]:With the proposal of "made in China 2025", China began to implement the three-step strategy of manufacturing power. As an industrial master machine, numerical control machine occupies an important position in the manufacturing industry, so it is very meaningful to improve the precision of numerical control machine tool. One of the methods to improve the accuracy of NC machine tool is to analyze and compensate the error of NC machine tool. This paper is mainly written around the error measurement and modeling analysis of NC machine tool. Among all kinds of error sources, thermal error is one of the biggest error sources of CNC machine tools and other precision machining machines. According to the research content of researchers at home and abroad, this paper builds the research platform of the spindle thermal error of NC machine tool, mainly from four aspects: the measurement of thermal error, the selection of temperature sensitive points, the modeling and the analysis of model comparison and verification. The details are as follows: (1) first, the thermal error measurement platform of NC machine tools is built, and then the temperature measurement points of thermal error are optimized by fuzzy clustering analysis and grey correlation analysis to ensure the validity and accuracy of the data. Collecting and analyzing data for the following modeling. (2) based on Kalman filter-grey model theory (Kalman Filter-GM) modeling, it can effectively reduce the effect of noise on the observed value, so we can get a better estimate of the observed value. In order to ensure the effective modeling of numerical control machine tool thermal error based on Kalman Filter-GM, this method is more robust than the traditional least square method in the process of modeling. Therefore, this model is better than the traditional modeling. (3) the residual correction model in this paper is based on the simulation value and the modeling data, this method can make the prediction effect better; Based on Kalman filtering method, several grey models are constructed by iterative deformation data, and then the data of these grey models are fused by data fusion method. Finally, the best estimate of the predicted data is obtained. (4) the thermal error model of NC machine tool based on Kalman Filter-GM. Compared with the traditional modeling methods, the least square method and the least square support vector machine, the numerical control machine tool thermal error modeling method is compared and analyzed. Finally, it is concluded that this modeling method is better.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG659

【相似文献】