Research On Error Compensation Technology For High Accuracy Circle Gratings Angel-measuring System

With the further development of social production, the accuracy and stability of the angle measuring system are put forward higher requirements. Based on high-precision angle measuring system Renishaw incremental circle grating, this thesis has studied multi reading heads structure and software compensation algorithm to improve the accuracy of angle measurement system. Then the principle of multi reading heads structure eliminating the errors and the specific scenarios are analysed based on the error modeling. And the effects of error compensation in high-precision circular grating angle measuring system and the application scenarios by numerical compensation algorithm, BP network error compensation and wavelet neural network(WNN) based on error backtracking and compensation algorithm from the perspective of theory and simulation are analyzed comparatively.By modeling the errors of circular gratings angle-measuring system, the model of components of system error, random error and installation error are derived. And the principle of multi reading heads structure eliminating the errors is derived. Then the effect of eliminating errors of multi reading heads under the conditions of installation inclining and asymmetrical installation is analyzed quantitatively.Piecewise linear, cubic spline interpolation, polynomial and harmonic compensation are analyzed and two iterative simulation experiments on four kinds of numerical compensation algorithms by discrete calibration errors of circular grating angle measurement are conducted. The effects and the application scenarios of the four algorithms, in which the best is harmonic compensation, are analyzed comparatively. And the principles for BP neural network algorithm are derived. Moreover, improving BP network by changing the training algorithm and adjusting network structure is done. Then two iterative simulation experiments of compensating errors are conducted. The compensation effects of BP network and numerical algorithms are analysed comparatively.In order to optimize the system design, the traceability error compensation method based on wavelet neural network is researched. And wavelet network methods are applied for measuring angle error circle grating traceable compensation. Moreover the compensation effect of its relatively narrow and broad wavelet network two algorithms are analysed comparatively. The validity of numerical compensation algorithm, BP network and WNN algorithms is verified by experiments, compensation effects and applicable situations of the three methods are compared fatherly. The compensation effect of narrow WNN is obviously better than the BP network with the both excellent features of neural network and wavelet transform. With fatherly experiments on generalized wavelet network and harmonic compensation algorithm, generalized wavelet networks can identify additional error component to achieve better compensation effect than harmonic compensation, it’s very suitable for dynamic error compensation.