基于历史数据的进给轴运动响应模型研究及应用

发布时间：2018-01-09 14:28

本文关键词：基于历史数据的进给轴运动响应模型研究及应用　出处：《华中科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：为提高曲面加工的轮廓精度,需要降低数控加工中进给轴的跟随误差。为此需要对轴的运动响应特性建立较为准确的模型,常规Simulink环境下的框图模型和机电联合仿真模型,需要复杂的有限元计算,难以用于实时的运动控制。本文基于历史数据对轴运动响应进行建模,以线性迭代模型描述轴输出在相邻两周期内本周期实际速度与上一周期内速度跟随误差、上一周期实际速度的关系,速度跟随误差是同一周期内指令速度与实际速度之间差值,从历史数据中提取线性迭代模型的线性迭代系数。这种线性迭代模型可以在运动控制过程中对轴的响应进行同步预测,可以在实时的控制过程中改进运动控制的性能。对于一个具体的轴,其响应特性的可重复程度是研究基于历史数据的建模方法的前提。设计了重复运行试验,分析了轴在同一套控制参数下,同样的指令速度下,其实际速度曲线的重复性,采用了五点三次平滑法对实际速度进行平滑处理,去除采样数据中的噪声,得到实际速度曲线。重复实验结果表明,轴速度曲线的不重复范围是编程速度的0.1%,因此利用历史数据建立轴的响应模型是可行的。研究了影响轴实际速度的因素,及各因素对实际速度的影响程度,确定了以相邻两周期内上一周期速度跟随误差、上一周期实际速度为基本变元的线性迭代模型,采用最小二乘法从历史数据中提取线性迭代系数。进行了实际速度预测有效性验证实验,对预测速度和实际速度进行比较,预测速度与实际速度差值最大值在编程速度的5%以内,平均值在编程速度的0.5%以内,结果表明响应模型具有较高的速度预测精度。研究不同类型轴、不同类型机床对跟随误差的影响程度,进行了跟随误差预测有效性验证实验,其中线轨机床的预测稳态跟随误差与实际稳态跟随差值占实际稳态跟随误差的2%以下,结果表明响应模型具有较高的跟随误差预测精度。研究基于预测跟随误差的前馈方法,进行了圆加工实验,编程速度为1000mm/min圆加工中,加前馈后圆半径误差减少31.5um。该结果表明,基于预测跟随误差的前馈方法能有效提高圆加工精度。
[Abstract]:In order to improve the accuracy of contour machining, the need to reduce the error of feed shaft in NC machining. So we need to establish a more accurate model response characteristics of shaft, block diagram and the co simulation of the mechanical model of conventional Simulink environment, need complicated finite element calculation, which is difficult for the real-time motion control. The model of the in response to the motion based on historical data, using linear iterative model to describe the output shaft in two adjacent the actual speed and cycle period on a cycle speed following error, a relationship between the actual speed of the cycle, the speed of tracking error is the difference between instruction speed in the same period with the actual velocity, extracting linear iterative coefficient linear iterative model from the historical data. The linear iterative model can respond to the axial movement in the control process can be predicted in the control process of synchronization, real-time The performance improvement of motion control. For a specific axis, the response characteristics of the repetitive degree is the premise to research the modeling method based on historical data. The repetitive operation of experimental design, analysis of the shaft in the same set of control parameters under the same instruction rate, in fact, are repetitive speed curve. The five point three times smoothing method on the actual speed of smoothing, remove the noise in the data sampling, get the actual velocity curve. Experiments show that the range of axial velocity curve does not repeat 0.1% programming speed, therefore the use of historical data to establish the model of shaft response is feasible. To study the factors that affect the actual speed of the shaft the influence of various factors on the degree and the actual speed, to determine the two adjacent period on a cycle speed following error, a cycle of the actual speed of the basic linear iterative model variables, using the minimum Two linear iterative multiplication coefficient extracted from the historical data. The actual speed prediction validation experiment, the prediction speed and actual speed is compared with the actual prediction speed maximum speed difference value within the programming speed of 5%, the average value of 0.5% in the programming speed. The results show that the model has high response speed prediction precision. Different types of shaft, the influence degree of different types of machine tools on the tracking error, the tracking error prediction validation experiment, the prediction of steady state line rail machine follow error and real steady with the difference in the actual steady-state tracking error is below 2%, the results show that the prediction accuracy error following response model is higher. Feedforward prediction method based on tracking error, the circle processing experiment, the programming speed is 1000mm/min round processing, feedforward radius error is reduced by 31. after 5um. results show that the feedforward method based on the predictive following error can effectively improve the precision of circular machining.

【学位授予单位】：华中科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TG659

【相似文献】