Research On Temperature Measuring Point Optimization And Thermal Error Modeling Of Large-scale Gantry CNC Machine Tool

The thermal error compensation is an economical and effective method to improve the machinetool precision. A large gantry CNC machine tool QLM2365is studied. Focused on the characteristicsof temperature field and thermal error of the large machine tool, the temperature measuring pointoptimization and the thermal error modeling of the existing approuches is improved. The main studycontents are as follows:The structure feature and heat resource of CNC machine tool QLM2365is emphatically analyzed.The location of temperature sensors on the machine is studied, and the temperature andthermal error were collected.The key temperature measuring points are selected by fuzzy clustering and partial correlationanalysis. Aimed at the characteristics of temperature field and thermal error of the large gantrymachine, fuzzy clustering is improved by gray system theory. And grouping optimization oftemperature measuring points is analyzed by gray correlation coefficient, instead of correlationcoefficient. Different optimization results are applied to RBF network modeling. Compared to theperformance of different network modeling, it is showed that fuzzy clustering based on graycorrelation coefficient is more suitable for large machine tool.Aimed at defects of RBF neural network trained by typical clustering algorithm, RBF neuralnetwork is optimized by genetic algorithm(GA), and GA-RBF network modeling is obtained. RBFnetwork modeling and GA-RBF network modeling are applied to thermal modeling of the large-scalegantry machine. Comparison with the fitting and forecasting results of thermal modeling is showedthat GA-RBF network modeling is better for predicting thermal error of the large-scale gantrymachine.Machine tool measuring points optimization and thermal error modeling sysytem based onMATLAB is developed. The system integrates the different methods of measuring point optimizationand modeling, to rapidly analyze CNC machine tool test data and select thermal error modeling.