Research On Quasi Static Volumetric Error Modeling And Error Identificationin Of Multi-axis NC Machine Tool

With the further development of the transformation and upgrade in Chinese manufacturing industry,the demand for precision machinery product continuously rises in various industries.As the basic equipment of manufacturing,the working accuracy of NC machine tool directly influences the workpiece quality.Therefore,working accuracy has been a hot point for research.However,accuracy and error are some kind of paradox.Revealing the mechanism of error production and making study on error modeling and error controlling are key scientific issues that must be solved.To solve the problems,this dissertation intensively studies quasi static volumetric error modeling and error identification of multi-axis NC machine tool,and verifies the correctness and feasibility of proposed methods.The main research work and achievements are:(1)Based upon incremental matrices,a new volumetric error model for multi-axis machine tool has been advanced.The actual coordinate transformation between adjacent structures is discussed in the form of increment.Through the superposition of errors caused by position and orientation between adjacent structures,the volumetric error can be calculated,which provides theoretical basis in revealing main error sources.According to the model,an error sensitivity analysis model is established with the theory of differentials,which is helpful for accuracy assignment and local error compensation.With the two modeling methods,volumetric error models of three-axis CNC machine tool and five-axis CNC machine tool with the structure configuration as RTTTR have been built,and error sensitivity analysis has been made,which proves the feasibility of the proposed methods.(2)Experiments are designed for error measurement in moving axes and rotating axes,and error identification models are established.Displacement errors and angular errors are identified respectively with ballbar.To study the errors of the rotating axes as A and C in five-axis machine tool of RTTTR,displacement errors are obtained when measuring in axial and radial directions.According to the characters of he rotating axes as A and C,two types of multi-axis motion have been designed,and angular errors are identified.In view of the disadvantages in nine-line identification method as strict requirements in selecting measuring points,accumulated error existing in error calculation and perpendicularity influence contained in the measured value of straightness,improvement approaches are proposed and a new error identification model is established for moving axes.On the platform of machine tool QLM27100-5X,researches of error measurement and identification in motion axes has been carried out,then the distribution of geometric error and thermal error in working space is discussed,and error sensitivity analysis is made to testify the feasibility of the error identification method.(3)A support vector machine(SVM)model of geometric error and a thermal deformation correction model of displacement error are advanced for multi-axis NC machine tool feed axes.On account of the less samples and strong nonlinearity in geometric error,theory of SVM is introduced in geometric error modeling for moving axis.In present research,the inconsistency in axial and radial thermal expansion is ignored,thus the dynamic characters of thermal error can not be fully grasped.To solve the problem,based upon the mechanism of thermal expansion,the relation between coefficient of linear expansion and cubical expansion is discussed.Then the thermal deformation formula of ballscrew has been modified.Finally,based upon the characters of ballscrew nonuniform temperature field,the models of thermal positioning error and straightness error are built.The mobile feed system of machine tool QLM27100-5X is studied to verify the model accuracy.(4)A temperature variables grouping method is proposed based on an optimal threshold.With the method,the representative temperature variables are selected for a sliced inverse regression(SIR)thermal error model.Firstly,with the theories of fuzzy clustering and coorlation analysis,a temperature variables grouping method is advanced based on an optimal threshold,which can resolve the problem that temperature variables grouping depends to much on experience.Then,a SIR model is established for thermal error prediction,which solves the problems of conventional regression model as fixed function form and low accuracy in long term prediction.The SIR model has good ability of generalization and extrapolation.Finally,with the proposed methods,optimal temperature variables are found out after temperature variables grouping on the platform of the machining center MCH63,and a SIR thermal error model is established.Simulation analysis shows the high fitting precision of the model.(5)A temperature and error detecting system and an error compensation system have been developed.With the platform of QLM27100-5X and MCH63,three models of geometric positioning error in moving axis,volumetric thermal errors caused by three moving axes,and spindle thermal error are verified through simulation and experiment.