30-05-2012, 12:18 PM
Parameter Estimation of a Synchronous Generator Using a Sine Cardinal Perturbation and Mixed Stochastic–Deterministic Algorithms
Parameter Estimation of a Synchronous Generator.pdf (Size: 690.41 KB / Downloads: 30)
VIRTUAL INSTRUMENT FOR SINC PERTURBATION
In order to carry out the proposed experiment, it is necessary
to develop a platform that has the ability to make the following
tasks: signal generation, signal amplification, and datalogging.
The testing platform consists of a virtual instrument, power
amplifier, and sensors. The virtual instrument is developed
in the LabVIEW software, and it handles one analog output
and three analog inputs from a data-acquisition card (DAQ).
The front panel of the virtual instrument [Fig. 1(a)] shows
the different data input controls that allow choosing the signal
parameters, such as its amplitude, time duration, and sampling
frequency. It also includes two areas for plotting of signals; one
of them shows the generated signal and the other one shows the
response signal for either machine magnetic axis.
PARAMETER ESTIMATION
The modeling of a synchronous generator is based on the
two-axis theory, where the machine three-phase windings are
represented by orthogonal windings along the d- and q-axes.
Each set of d–q-axis windings can be modeled as equivalent
electrical circuits, and they are shown in Figs. 4 and 5. In this
paper, one damper winding for each axis is used; this leads to
the classical synchronous-machine model. The development of
the d–q equivalent circuits for the synchronous-generator model
is dependent on the kind of test to be applied.
EXPERIMENTAL RESULTS
With the machine at standstill, the sinc time-domain signal is
applied to the d- and q-axis stator windings. In the d-axis test,
the stator and field current responses, together with the field
voltage, are simultaneously acquired with a sampling rate of
10 kHz. In the q-axis test, only the stator voltage and current
are logged. The magnitude of the excitation signal is adjusted
by the user through the virtual instrument to get a current of
about 1 A, which represents 5.4% of the nominal line current;
it was found that at this level, the current signals are less noisy
than at smaller levels. It is worth mentioning that one reason for
using the standstill tests for machine parameter determination
is because a test can be carried out at low excitation levels;
this is particularly useful in large generators due to limitations
of having high current amplifiers.
CONCLUSION
A novel excitation signal for the parameter estimation of
a synchronous generator has been successfully applied. The
sine cardinal perturbation was used to excite experimentally the
direct and quadrature axes of the generator to acquire the dynamic
information of the d, q, and field-winding currents which,
together with the excitation voltage signals, were employed to
estimate the parameters of the electrical equivalent circuits. The
fundamental parameters of the d–q-axis models were estimated
using GA and the Gauss–Newton method for a 7-kVA 220-V
60-Hz four-pole wye-connected synchronous generator.