Mechanical resonant systems identification and compensation for the self tuning algorithms

Abstract

This paper is focused on the identification and the compensation of the two mass electro-mechanical system (electric motor and flexible load forming two mass system) with resonance. The methodology of this paper is based on the three-phase Permanent Magnet Synchronous Motor (PMSM) model with Field Oriented Control (FOC) and with the flexible load model. The resonant part is estimated with two approaches. The first one is the Recursive algorithm of the Least Squares Method (RLSM) and the second one is the Spectrum Analysis (SA), from their results the compensators are constructed in the discrete form for the implementation in a digital control system. In the case of the RLSM, the compensator is obtained as the inversion of the gained model of the resonant part in the discrete form. The compensator from SA is obtained as the result of iterative algorithm, where for each iteration there is gradual calculation of the compensator numerator and denominator The results of both experiments are contrasted with the resonant part of the model, to show the independency of the algorithms on the electrical part of the motor. The Pseudo Random Binary Sequence is used as the input signal for the identification experiments. Finally, the capability of the obtained estimates as the resonance compensators is evaluated in the final part of this paper Both presented methods are capable of the use for the self tuning algorithms.