Identification of PM synchronous machines in the frequency domain by broadband excitation

Abstract

Modelling and identification of synchronous machines has attracted much attention during the past decades. Although the standstill frequency response (SSFR) method has become standardized (IEEE Std. 115) it has some disadvantages (e.g. it is a time consuming method). To overcome these, the use of broadband excitation has been proposed in the literature. In this paper the use of broadband excitation for the frequency domain identification of permanent magnet synchronous machines is investigated. The measurement set-up and method are discussed, with special attention to the selection of the applied excitation signals. Three broadband signals suitable for frequency domain parameter identification are discussed and compared in measurements. The obtained frequency domain data can be used in different applications. Examples are given in this paper. Furthermore the extraction of a parametric model from the nonparametric frequency domain data is briefly discussed. The identified parameters can be used to study the influence of saturation or can be applied in the simulation and control of permanent magnet synchronous machines. Measurement results demonstrate that broadband frequency domain identification can be applied to extract the parameters (inductances) of permanent magnet synchronous machines. Moreover, the reproducibility of the method is shown to be good as the variation in the parameters obtained in different measurement sessions is low. Future work on the subject will include a closer investigation of the variations in the obtained parameters, the use of higher order models and the application of more complex estimation algorithms

On the stator flux linkage estimation of an PMSM with Extended Kalman Filters

“…To study the behaviour of saturated PMSMs a broadband frequency-domain identification procedure was presented in [8], [12]. This method is an extension of the Stand Still Frequency Response (SSFR) as standardized in IEEE Std.115 and uses broadband voltage signals (mainly multisines) to excite the PMSM.…”

“…The machine impedance is obtained from the current response in the frequency domain and the machine inductance is extracted with a least-squares method. The approach has been tested both with a linear amplifier [12] and an inverter [8] to provide the excitation. In figure 3 the obtained inductances L d and L q for a Siemens PMSM are depicted.…”

Improved modelling of PMSMs taking into account conductive rotor parts

“…The machine terminals are supplied by small test signals where the frequency is varied stepwise and both the voltage and the current are measured [7]. In order to shorten the measurement time, the SSFR-test was modified by using broadband signals as described in [3]. Here, a particular type of broadband signals, the multisine test signals, are supplied to the machine terminals.…”

“…When modelling the PMSM in [3], the stator winding is considered only, which results in a first-order model consisting of two RL-networks as shown in Fig. 1.…”

“…The electrical machine parameters are estimated by using Adaptive Particle Swarm Optimization (APSO) assuring rapid and global convergeance. To follow up the first results discussed in [3], a second-order transfer function is used based on the enhanced model, improving the approximation of the model to the measurements.…”