The master's thesis deals with a new measurement method which allows the determination of
losses, efficiency as well as heating of a multiphase electrical machine for different operating
points using synthetic loading. The usual approach, originating from and established in classical
3-phase electrical machine technology, requires a mechanical coupling of the test machine with
a sufficiently large brake in back-to-back configuration. That complicates and makes the
measurement process more expensive and, especially for higher power machines, may even be
infeasible.
In Chapter 1, both the classical natural loading method (NLM) and its alternative the
synthetic loading method (SLM) are briefly introduced. SLM is limited to multiphase machines
with an even number of phase windings. It is able to estimate the losses of a permanent magnet
synchronous machine over a wide operating range. The SLM is particularly suitable for
multiphase machines whose stator winding consists of an even number of 3-phase winding sets.
SLM was therefore applied to a 6-phase permanent magnet synchronous machine (PMSM),
which has two 3-phase winding sets. During the measurement test, one of the 3-phase sets
operates in motor mode while the other set operates in generator mode. Most of the power flows
from the motor to the generator set through electromagnetic field, while only the power needed
to cover the drive losses flows from the DC circuit.
Due to the different operation mode of each set a substantial radial component of the
magnetic force can occur in SLM, causing an unbalanced magnetic pull (UMP). Moreover,
there is an additional radial load on the bearings, which can also shorten their service life. A
computer tool has been developed to evaluate the UMP, which is described in more detail in
Chapter 2. Starting with machine geometry and winding scheme a simulation model is created
using the MATLAB and FEMM. FEMM then performs a magnetostatic analysis based on the
desired current and rotor angle and provides information on the resulting UMP. The simulation
was performed for the aforementioned 6-phase PMSM, and based on the results, it was decided
that the UMP during the SLM is within acceptable limits (Chapter 3).
SLM as an alternative approach for machine loading test is presented in more detail in
Chapter 4. The physical background of SLM operation is described and its advantages as well
as disadvantages are listed. An equivalent circuit for 3-phase PMSM has been adapted for
v
6-phase PMSM, where iron loss has been included as well. This enables us to better explain
power circulation between the sets operating in motor and generator mode.
The measurement system, the practical implementation of the SLM, and the
measurement procedure are described in Chapter 5. The loss measurements were performed
using a 3-phase power analyzer in 2-wattmeter configuration. Thus, only two wattmeters were
needed to measure the input power of the 3-phase winding set, whereas the third wattmeter was
used to simultaneously measure the DC power covering the losses of the entire drive. Since
only one power analyzer was available, the measurements were made separately for each
winding set operating in generator and motor mode, respectively. It was necessary to ensure the
same operating conditions for both cases, what we achieved by setting the operating points
identically via the user interface. Additionally, we simultaneously measured the DC power as
well as the converter and machine temperature. This information served as an additional
confirmation that we were at the same operating points. To thermally stabilize the drive
component during the measurements, the machine and the converter were connected to a
common liquid cooling system.
The measurement results for SLM are presented in Chapter 6. Based on phase current
waveforms of the motor and generator set, we have confirmed the predictions that the
magnitude of the currents of the motor winding set is higher than that of the generator set.
Because of the way the synthetic load control loop is implemented, this is only true as long as
machine operates in constant torque region. Namely, in field-weakening region the current of
the motor winding set becomes lower than the current of the generator winding set. From the
power measurement of 3-phase winding sets and the DC circuit, the losses of the machine and
converter were determined. It can be seen from the results that the machine losses increase with
speed and increasing load. The losses of the inverter do not change with increasing speed up to
the field-weakening region and only change with increased load.
To confirm the performance of the SLM, the results would have to be directly compared
with those obtained with the NLM, which could not be realized within the scope of the master's
thesis. Nevertheless, based on the measurements, we can conclude that SLM is a good
alternative to NLM. The machine losses are mainly affected by the losses in the stator winding,
which depend on the stator current, the iron losses, which are due to a large magnetic field
density and the rotational speed, as well as the losses due to friction in the bearings. Because of
vi
the way the machine is driven in the SLM, we conclude that the magnetic field density in the
iron is lower than during NLM, hence the SLM somewhat underestimates iron losses. Due to
the increased radial components of the magnetic force resulting in additional bearing load, we
assume that the frictional losses in the SLM are higher than in the NLM.
|
