数控机床热误差补偿模型稳健性理论分析及其应用技术研究

发布时间：2018-12-11 11:22

【摘要】：作为衡量国家制造业和综合国力水平的标志,高档数控机床在现代工业中的地位日益加重。在机床多种误差源中,热误差在机床总误差重的比重可达到40%~70%,因此,控制并削弱热误差对提高精密机床的精度非常重要。国内外研究表明,热误差补偿技术重点围绕着温度敏感点选择、热误差建模和补偿等三个方面给予研究。虽然目前已有比较完善的理论体系和补偿实施技术,但仍然存在一些影响其广泛应用、且亟需解决的关键性问题,如温度测点优化过程中温度变量之间的共线性影响以及模型预测稳健性不足等。针对上述热误差补偿技术的缺陷,本文从温度敏感点选择、热误差建模和补偿等方面提出了相关理论和方法,并进行了大量实验数据分析,最后以实际机床为例进行了热误差补偿实验验证。论文主要的研究内容如下:1)提出了一种数控机床稳健性热误差建模方法,将以灰色关联度为核心的温度敏感点优化方法和以主成分回归分析为核心的建模算法配合使用,共同对温度敏感点优化所需遵循的大权重、低耦合和少布点策略进行处理,既简化了测点优化的复杂度,又有效地提升了模型的预测精度和稳健性。2)采用传统的模糊聚类结合灰色关联度的温度敏感点优化算法,对全年分季度的机床热误差数据给予了研究,阐述了该方法选择出的温度敏感点存在变动性特征及其对模型精度影响。3)研究了多元线性回归算法、时间序列分布滞后和主成分回归算法的模型精度和稳健性特征,并总结得到三种模型在热误差建模补偿中的适用范围。另外,提出了基于主成分回归分析的分布滞后模型精度提升方法。4)介绍了热误差补偿技术的软件实施方法,包括数控平台、热误差测量集成系统、最佳温度敏感点的优化选择、热误差数学模型建立、热误差补偿与精度评定。评定结果表明补偿效果显著。
[Abstract]:As a symbol of national manufacturing industry and comprehensive national strength, the status of high-grade CNC machine tools in modern industry is getting more and more serious. Among the various error sources of machine tool, the proportion of thermal error in the total error weight of machine tool can reach 400.Therefore, it is very important to control and weaken the thermal error to improve the precision of precision machine tool. Studies at home and abroad show that the thermal error compensation technology focuses on three aspects: temperature sensitive point selection, thermal error modeling and compensation. Although there is a relatively sound theoretical system and compensation implementation techniques, there are still some key problems that affect their wide application and are in urgent need of solution, For example, the collinearity effect between temperature variables and the lack of robustness of the model in the optimization of temperature measurement points. In view of the defects of the above thermal error compensation technology, this paper puts forward some related theories and methods from the aspects of temperature sensitive point selection, thermal error modeling and compensation, and carries out a lot of experimental data analysis. Finally, the experimental verification of thermal error compensation is carried out with the actual machine tool as an example. The main contents of this paper are as follows: 1) A robust thermal error modeling method for NC machine tools is proposed. The temperature sensitive point optimization method with grey correlation degree as the core and the modeling algorithm with principal component regression analysis as the core are used together. In order to simplify the complexity of the optimization of temperature sensitive points, the strategy of large weight, low coupling and less point placement is used in the optimization of temperature sensitive points. Moreover, the prediction accuracy and robustness of the model are improved effectively. 2) the thermal error data of machine tools are studied by using the traditional temperature sensitive point optimization algorithm based on fuzzy clustering and grey correlation degree. The variability of temperature sensitive points selected by this method and its influence on model accuracy are described. 3) the model accuracy and robustness of multivariate linear regression algorithm, time series distribution lag and principal component regression algorithm are studied. The application range of three models in thermal error modeling compensation is summarized. In addition, a method of improving the accuracy of distributed lag model based on principal component regression analysis is proposed. 4) the software implementation method of thermal error compensation technology is introduced, including numerical control platform, integrated thermal error measurement system, Optimal selection of optimal temperature sensitive points, mathematical model of thermal error, thermal error compensation and accuracy evaluation. The evaluation results show that the compensation effect is remarkable.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TG659

【参考文献】