Title :
Application of the Welch-Method for the Identification of Two- and Three-Mass-Systems
Author :
Villwock, Sebastian ; Pacas, Mario
Author_Institution :
Baumueller, Nuremberg
Abstract :
This paper deals with the measurement of the frequency response of the mechanical part of a drive for the parameter identification of a plant. The system is stimulated by pseudorandom binary signals. The measurement of the frequency response is part of a system identification procedure being carried out during an automatic commissioning of the drive. For the calculation of the frequency response of the mechanics, the Welch-method is applied for spectral analysis. The Welch-method is known from the fields of communications and measurement engineering. This paper addresses the application of this powerful method for the identification of electrical drives. Investigations have pointed out that the pure utilization of conventional identification strategies does not yield satisfying experimental results. Experimental results presented in this paper point out clearly the efficiency and flexibility of the proposed Welch-method. This paper contains many practical aspects and realization details that are important for their implementation on industrial systems. Although in principle, commercial software tools can be utilized for identifying the parameters of the plant, this paper addresses the implementation of the necessary identification algorithms on the embedded control electronics of the drives. The utilization of the Levenberg-Marquardt-algorithm yields excellent results for the identified parameters on the basis of the measured frequency response data.
Keywords :
electric drives; frequency response; identification; machine control; spectral analysis; Levenberg-Marquardt-algorithm; Welch-method; automatic commissioning; commercial software tools; electrical drives; embedded control electronics; frequency response; identification algorithms; parameter identification; pseudorandom binary signals; spectral analysis; three-mass-system; two-mass system; Electrical equipment industry; Frequency measurement; Frequency response; Mechanical variables measurement; Parameter estimation; Power engineering and energy; Software algorithms; Software tools; Spectral analysis; System identification; Drives; Mechatronics; Modeling; Signal Processing; System Identification; mechatronics; modeling; signal processing; system identification;
Journal_Title :
Industrial Electronics, IEEE Transactions on
DOI :
10.1109/TIE.2007.909753