Research On Dynamic Measurement And Modeling For Thermal Error Of Horizontal Machining Center Spindle

The thermal error is the one of the main error sources of the precision CNC machine tool, and it accounts for 40%-70% of the manufacturing errors. By means of thermal error compensation, product manufacturing error could be eliminated greatly, improving the machining accuracy of machine tools. Thermal error measurement, temperature measuring points optimization and establishment of accurate and reliable thermal error model are essential for the thermal error compensation.This paper analyzes thermal properties for spindle system for horizontal machining center, by modeling the motorized spindle and headstock, calculating the thermal loads and the thermal boundary condition of spindle system, in order to obtain the temperature filed distribution of headstock and the thermal deformation of motorized spindle tool chuck in running condition, which provides the basis of the dynamic measurement for spindle thermal error.The dynamic measurement system for thermal error has been designed, through building hardware system for displacement dynamic measurement and multichannel temperature acquisitions, compiling thermal error dynamic measurement software based on Lab VIEW. The system could synchronously collect the displacement and temperature signals, connecting the temperature hardware system with LabVIEW based on OPC technique. Using five-point method to measure spindle thermal error in workspace, and experiments are carried out under different conditions. All that depicted above lays a foundation for optimizing temperature measuring points and establishing thermal error model.This paper comprehensively analyzes the multiple correlations among the measuring points and the relation between the temperature and the thermal error. Improved fuzzy C-means (IFCM) clustering algorithm is applied to classify the temperature measuring points, which reduces the correlations for different classes and avoids the shortcoming of the fuzzy C-means (FCM) algorithm that is too sensitive for the initial clustering center to get global convergence. Sorting the temperature points as the grey synthetic degree of association of the grey relational analysis (GRA), comprehensively reflecting the relation between the temperature and the thermal at the perspective of the value of change and the rate of change. Using IFCM-GRA to optimize the temperature measuring points, which keeps robustness and accuracy of the thermal error model, the number of the temperature measuring points decreases greatly.This paper puts forward a multivariable correlative and combined thermal error model, which considers the speed of motorized spindle, the temperature of machine tool, coolant and environment. Through analysis of the process of numerical control machining and the principle of the material thermal deformation, this model sets the differential temperature, relatively with initial temperature, as the temperature variables. Partial least squares (PLS) method is applied to extract the principle components as inputs of the LS-SVM, forming the PLS-LSSVM thermal error combined model. This model is tested with thermal error dynamic measurement experiments, and is verified to have good accuracy and robustness.