Research On Error Compensation Of Articulated Coordinate Measuring Machine

Now the higher request is placed on the precision machining technology along with the rapid development of modern manufacturing.It is considered that the accuracy of the measurement device is the factor that affects directly the quality of industrial products,so the current research is made to be focused on how to reduce the measurement error of the measurement instrument.In this paper,the articulated coordinate measuring machine,a non-Cartesian flexible coordinate measuring system,is regarded as the research object.And the measuring machine is influenced by multiple errors,and its serial link mechanism causes the angle error to be accumulated,transferred and enlarged step by step without the ideal accuracy of the measuring machine.First of all,the mechanical structure and working principle of the articulated coordinate measuring machine is to be discussed in the paper,and the various error source is to be analyzed that affects the accuracy of the articulated coordinate measuring machine.And the main error source that causes measurement error is to be determined at: structural parameter error,circular grating eccentricity error,probe system error and thermal deformation error.The error compensation technology of articulated coordinate measuring machine is to be studied on this basis.Then,the different types of structural parameter error and the eccentricity error of the circular grating angle sensor is to be analyzed,and correction formulas for structural parameter error and eccentricity error are to be derived respectively.Since it makes no influence on each other,a measurement error model is to be established to uniformly compensate for the two kinds of errors.Subsequently,the principle of self-calibration of parameters is to be introduced,and the structural parameter error and eccentricity parameter in the error model are to be calibrated by using simulated annealing algorithm.From the above analysis result,it is known that the average value of the length error of the measuring machine after parameter calibration is reduced by 0.1606 mm,and the standard deviation is reduced by 0.0761 mm.So the measurement error of the measuring machine is significantly reduced.Finally,on the basis of parameter calibration,the BP neural network with nonlinear mapping characteristics is used to compensate the coordinate values of the articulated coordinate measuring machine for the probe system error and thermal deformation error.The error source is quantitatively to be analyzed to determine the input parameter of the neural network,and the simulated annealing algorithm is used to optimize the weight and threshold of the BP neural network to enhance the accuracy of network output.It is indicated in the compensation comparison results that the average error of the three axes of the articulated coordinate measuring machine,after passing through two compensation,is respectively reduced to0.0237 mm,0.0177 mm,and 0.0186 mm.And it can be expected to be further improved on the accuracy of the measuring machine.