Research On Thermal Deformation Monitoring Of Heavy CNC Column Based On FBG Sensors

Heavy CNC machine tools play a very important role in the manufacturing industry, so improving its machining precision has always been an important research topic attracting more and more attention. Research shows that thermal error accounted for about 45% of the total errors of the machine, while the higher accuracy, the higher the proportion. In order to reduce or later counteract the thermal error, current research on thermal error of machine tool mainly focus on analysis of temperature field, temperature-thermal error modeling, and so on. The machine tool is always a complex physical structure, and affected by unpredictable interference from outside environment, so it's difficult to reconstruct its temperature field. The accuracy and robustness of the built model is also a major problem, and its operation efficiency also affects the effect of thermal error compensation device.The thermal error of machine tool is the common effect of deformation of all the machine parts under the influence of heat, and the thermal deformation modeling is the basis of thermal error compensation techniques, by the model deformation can be predicted and compensated with the real-time temperature changes to achieve reduction in thermal error. From the point of view of physical structure, shape changes of column, cause beam bending forward, then lead the spindle thermal drift. In this paper, with the large structure – column for study, and its relationship with the thermal deformation, the research contents are as follows:(1) The distribution of the heat source of machine were analyzed, the measurement of the temperature field of the column by the optical fiber grating temperature sensor, and the synchronous acquisition of machine tool spindle thermal drift(thermal deformation in Y direction) by CCD laser displacement sensor, is also achieved. The correlation between the temperature field and thermal column error is analyzed too.(2) Not all the points are related to thermal errors, so the temperature measuring points are grouped by fuzzy clustering method, and then from different sides, partial correlation analysis, the maximum sensitivity and grey relational analysis method are used to select key temperature measuring points from the groups. And based on the cross among screening results, this paper presents a more balanced comprehensive strategy to optimization measuring points.(3) With the key temperature measuring points are obtained, multiple linear regression model, BP neural network model and GA-BP network model is established to analyze the relationship between temperature and thermal error. This paper verified the different measuring point optimization strategy, and then put forward advices for optimization of temperature measuring points and improvement of model-building methods.(4) In order to conveniently study on the mechanism of the column deformation affecting thermal errors of machine tool, the relationship between deformation and strain of cylindrical structure is derived. This paper also designs a real-time monitoring model of two-dimensional deformation of cylindrical structure, based on Fiber Bragg Grating strain sensors. With the using of monitoring model, bending characteristics of column structure can be shown online, which provided convenience for further study on the thermal error compensation.