Abstract:
|
HTS machines are a potentially capable option in order to address technical challenges
that the wind industry is going to face in the coming years. HTS winding
excitation allows to highly increase the power density of the machine, achieving
large increments in output power of a single wind turbine. Already several HTS
machine prototypes have been constructed and experimentally demonstrated, and
they are about to cross the frontier to become marketable product. However, one of
the main factors that limits this progress is the high cost of HTS conductors. Therefore,
reducing HTS usage plays a crucial role to make HTS machines a cost-effective
solution.
This research presents a topology optimization of an HTS field winding distribution
for a given HTS synchronous machine design. The optimization problem is
formulated in order to minimize HTS usage, so that the HTS field winding performance
can be maximized. Two widely used topology optimization techniques
are applied and compared, both using very different optimization methodologies
(gradient-based and discrete optimization). Results show that discrete optimization
adjusts better with the limitations imposed by the critical current condition
of HTS tapes. However, gradient-based presents faster convergence.
Inspired by the results presented above, a third algorithm is developed combining
the discrete optimization with a local optimization based on an on/off sensitivity
analysis. A short test is performed where the superiority of this approach is
confirmed. This later algorithm is then applied into two different case scenarios.
The first scenario simulates a 2,5 T HTS wind turbine generator. It evaluates
optimal field winding distribution for an HTS excitation employing SP4050-2G tape
from Super Power Inc. As a second stage, an additional power supply is connected
reevaluating the optimal field winding distribution. This implementation is based
on a well known method for reducing conductor usage by allowing multiple power
supplies to excite the field winding. Maximum HTS savings of 9,1% are obtained.
The second scenario comprises the SUPERWIND HTS machine geometry. Optimal
HTS designs are acquired as a function of 1G and 2G available HTS tapes. Whereby
it is obtained that 2G winding design employs substantially less HTS material
compared to 1G winding design, achieving 22,7% of savings. For completeness, a
final execution of the algorithm is performed applying the same conditions used in
experimental investigations. The results achieved validate the findings obtained in
this work. |