Research On Resolver Decoding Technology

As an absolute position sensor with high reliability and high precision,resolver has been widely used in aerospace,robotics,and other industrial fields.The output of resolver is an analog signal that needs to be decoded by resolver-to-digital converter to get motor speed and angle.However,resolver output contains errors,which can not be suppressed by dedicated decoder chip.Based on a thorough study of the errors in the decoding system,a recursive least square-dual second order generalized integrator(RLS-DSOGI)decoding technology is proposed to improve the decoding accuracy of the resolver.The feasibility and effectiveness of this method are verified by simulation and experiment.Firstly,in this thesis,the working principle of resolver is introduced,and the influence of the imperfect signal on the angle measurement results is deeply analyzed.The interference formulas of four common errors,such as amplitude error,phase error,bias error and harmonic error,are derived.Secondly,the DSOGI decoding technology was proposed due to traditional Type II tracking PLL has no error suppression function.The DSOGI has the characteristic of whose output is a set of signals with equal amplitude and orthogonal phase.Therefore,it could suppress amplitude error and phase error,improve the decoding accuracy effectively.A RLS-DSOGI decoding technique is proposed to reduce the bias error in the double generalized two order integrator method.The recursive least square method is introduced to fit the signal path,then the signal is reconstructed by the identified ellipse parameters,the ability to suppress the bias error is improved.Aiming at the uneven distribution of data points caused by the variation of motor speed,a speed weighting factor is introduced to make the data points uniformly participate in the fitting calculation.In view of the fact that data points are densely overlapped and occupied computing resources during motor stops,the forgetting factor is optimized by the angle function,which reduces the computation pressure and optimizes the iterative process.Finally,a Matlab discrete simulation model is built to verify the theoretical analysis effect of the new decoding scheme.The decoding results of the DSOGI decoding technique and some existing methods such as synchronous reference frame-based phase locked loop and the double synchronous reference frame-based phase locked loop is compared,and the validity of the the DSOGI to suppress decoding error of resolver is verified.The error identification and signal correction performance of the recursive least squares ellipse fitting technique were checked,the dynamic response speed,overshoot and steady-state error of the RLS-DSOGI decoding technique are tested.The influence of three kinds of errors on decoding algorithms is measured.Meanwhile,an experienmental platform is set up for testing,the decoding results of the RLS-DSOGI are compared with the decoding results of a decoder chip AD2S1210,which verifies the effectiveness of the proposed decoding technology.