电主轴温升智能预测与实验研究

发布时间：2018-01-31 23:58

本文关键词： 高速电主轴温度预测温度场仿真分析实验研究　出处：《沈阳建筑大学》2015年硕士论文　论文类型：学位论文

【摘要】：电主轴是高速数控机床的关键部件,其性能的优劣直接决定着数控机床整机的技术水平。高速电主轴在旋转过程中,受到各种工况的影响,电主轴热特性不稳定将导致故障发生并减小其使用寿命,而电主轴的损坏基本是因电主轴发热引起的。并且电主轴的热变形还会使工作精度下降从而影响电主轴的运行性能,因此为了确保电主轴工作时的精度和寿命,需对电主轴的温升进行预测和控制。高速电主轴在运转的过程中电主轴温度会因受到环境温度、冷却液温度和运行时间等因素的影响而变化。而电主轴的温度场变化规律具有高度的复杂性和非线性,并且对电主轴内部温度的测量具有一定困难,为解决对电主轴表面温度变化规律和对电主轴内部温度场的预测,本文以型号为150MD24Y20电主轴为研究对象分别对电主轴的外部表面温度和电主轴内部温度预测进行研究。主要工作如下:通过对电主轴的表面温度预测方法的深入研究,在获得电主轴表面温度实时检测基础数据后,建立基于数据的电主轴表面温度预测模型,并分别采用人工神经网络中的BP算法、RBF算法建立电主轴温度预测模型对金属电主轴进行表面温度预测,两种预测模型有效地预测电主轴的表面温度的变化规律,预测结果表明:两种方法有效地解决了电主轴表面温度的高度非线性问题,但两种预测模型存在各自的缺陷导致预测精度有所下降。为了解决单一神经网络预测模型的缺陷,本文利用遗传算法优化的BP神经网络建立电主轴温度预测模型对高速电主轴表面温度进行预测,利用Matlab软件进行仿真研究,并将预测结果与实验结果进行对比验证预测结果的正确性。结果表明利用遗传算法优化的BP神经网络建立电主轴表面温度预测模型预测精度更高。本文基于损耗实验的电主轴内部温度场预测分析,建立了电主轴的有限元模型,分析了电主轴的主要生热机理和传热机制,利用现有电主轴设备测量出的电主轴损耗数据来计算转子、定子的生热率以及轴承生热率,根据相关经验公式计算出电主轴不同部位的散热系数,然后把计算得到生热率与散热系数加载到所建立的电主轴有限元模型上。最后,利用ANSYS软件仿真分析得到高速电主轴温度分布情况和温度变化规律。并将由有限元模型得到的温度场分布情况与实验结果相对比来验证ANSYS预测模型的正确性。通过对电主轴内部温度预测研究,分析了电主轴的温度场分布规律,为电主轴总体温度、瞬态温度预测提供理论基础。本文研究表明:利用遗传算法优化的BP神经网络建立的电主轴表面温度预测模型相比于单一的神经网络电主轴表面温度预测模型的预测精度更高,利用ANSYS建立电主轴内部温度场分布模型可用于电主轴内部温度变化趋势的预测,为电主轴热特性分析提供理论基础。
[Abstract]:Motorized spindle is the key component of high speed NC machine tool. Its performance directly determines the technical level of CNC machine tool. The high speed motorized spindle is affected by various working conditions in the process of rotation. Unstable thermal characteristics of motorized spindle will lead to failure and reduce its service life. And the damage of the motorized spindle is caused by the heating of the motorized spindle, and the thermal deformation of the motorized spindle will also reduce the working precision and affect the operation performance of the motorized spindle, so in order to ensure the accuracy and service life of the motorized spindle. It is necessary to predict and control the temperature rise of the motorized spindle. The temperature of the motorized spindle will be subjected to the ambient temperature during the operation of the high-speed motorized spindle. The temperature field of the motorized spindle is highly complex and nonlinear, and it is difficult to measure the internal temperature of the motorized spindle. In order to solve the change rule of surface temperature of motorized spindle and the prediction of internal temperature field of motorized spindle. In this paper, the outer surface temperature and the internal temperature of the motorized spindle are studied with the model 150MD24Y20 as the research object. The main work is as follows:. The surface temperature prediction method of motorized spindle is studied deeply. After obtaining the basic data of the real time measurement of the surface temperature of the motorized spindle, the prediction model of the surface temperature of the motorized spindle is established based on the data, and the BP algorithm in the artificial neural network is adopted respectively. RBF algorithm is used to predict the surface temperature of the metal spindle. The two models can effectively predict the variation of the surface temperature of the motorized spindle. The prediction results show that the two methods can effectively solve the highly nonlinear problem of surface temperature of motorized spindle. However, the two prediction models have their own defects, which lead to the decline of prediction accuracy. In order to solve the shortcomings of the single neural network prediction model. In this paper, BP neural network optimized by genetic algorithm is used to predict the surface temperature of high speed motorized spindle, and Matlab software is used to simulate the surface temperature of motorized spindle. The prediction results are compared with the experimental results to verify the correctness of the prediction results. The results show that the BP neural network optimized by genetic algorithm is more accurate in predicting the surface temperature of motorized spindle. Prediction and analysis of internal temperature field of motorized spindle based on energy consumption experiment. The finite element model of the motorized spindle is established, the main heat generation mechanism and heat transfer mechanism of the motorized spindle are analyzed, and the rotor is calculated by using the loss data of the motorized spindle measured by the existing motorized spindle equipment. The heat generation rate of the stator and the heat generation rate of the bearing are calculated according to the relevant empirical formula to calculate the heat dissipation coefficient of different parts of the motorized spindle. Then the calculated heat generation rate and heat dissipation coefficient are loaded into the established finite element model of the motorized spindle. Finally. The temperature distribution and temperature variation of high speed motorized spindle are obtained by using ANSYS software, and the temperature field distribution obtained from finite element model is compared with the experimental results to verify the prediction of ANSYS. The correctness of the model. The internal temperature prediction of the motorized spindle is studied. The distribution law of temperature field of motorized spindle is analyzed, which is the total temperature of motorized spindle. Transient temperature prediction provides a theoretical basis. Compared with the single neural network model, the prediction precision of the surface temperature prediction model based on BP neural network optimized by genetic algorithm is higher than that of the single neural network model. The model of temperature field distribution in motorized spindle by ANSYS can be used to predict the variation trend of internal temperature of motorized spindle, which provides a theoretical basis for the analysis of thermal characteristics of motorized spindle.
【学位授予单位】：沈阳建筑大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TG659

【相似文献】