Topic for open discussion:
|
Send
AWE notes and topic replies to editor@upperwindpower.com
|
?next? |
NREL Exaflop Multi-Physics Multi-Solver Validation for Kite Networks Dec. 5, 2020, note by Dave Santos of kPower High unit-count AWES Kite Networks easily pencil out to ~100GW scale unit plants. Is this realistic? Single-unit
kites have a simple network topology. Many-unit many-connected network
topologies are long established in classic kiting and AWE theory, as an
obvious scaling path. Based on many historic giant kite cases,
unit-kite scaling limits are fairly well understood as operational,
ranging to ~1000m2 and ~10MW.
Simply
building an extreme-scale AWE plant to test is not realistic, nor is
slow empirical scale-up, given ecological urgency for clean energy. The
obvious extreme-scale engineering practice is to simulate extreme-scale
plants both numerically, and by subscale prototypes. Heuristic "napkin"
calculations suggest dramatic predictions a high fidelity sim should
validate. Would aerodynamics operate at extreme scale as expected by
mountain-wind interaction observations? Would extreme-scale polymer
lattices stay within mass-scaling exponent limits? Would weather
modification effects be manageable; benign or even favorable? What
would dynamics be across all wind and load conditions? Could the
network loiter efficiently during calm by reverse-pumping and gradual
sinking? And so on.
NREL,
in partnership with other National Labs, intends to continue scaling
HAWTS by depending on multi-physics multi-solver simulations preformed
on exaflop parallel-processing supercomputer due in 2022. Under NREL's
DOE Wind R&D Roadmap mandate to develop AWE, and its general
mission to stay abreast of wind engineering-science progress, the
simulation of extreme-scale AWES networks should be developed in
conjunction with simulation of larger HAWTs. This will also allow
extreme-scale AWE to be compared with extreme-scale HAWT farms, so that
planners can decide how the choice between these radically different
wind technologies compare.
The
basic requirements for NREL to integrate AWE into the exaflop computing
project are to extend the HAWT wind field model to about 2000m high,
and to support the extreme-scale kite networks spring-mass "rag and
string" lattice polymer dynamics. These reasonably minor extensions of
the intended computation scope, if done now. There would also be a
parallel design project to develop AWES virtual models along with the
HAWT models. In fact, both the HAWT and AWES sim camps are already
working on current generation modeling, and can begin to integrate
their work. In AWE there is a particular need to rough-vet a zoo of
promising and doomed concepts.
Outdoor
subscale model testing of extreme-scale AWES networks will be fast and
cheap at ~1/50 scale. Not only would the respective aerodynamics be
usefully similar, but many operational and detail-engineering issues
could be worked out. A lot could be learned even at wind-tunnel scale.
Exploiting the large return-flow plenum of some the world's largest
wind tunnels (UMD is a special opportunity). Just as tall buildings are
aerodynamically validated in normal wind-tunnel practice, so will new
HAWT and AWES designs be early tested. by 2022, the subscale modeling
might be at ~1/10 scale, and working well.
Close-correlation
of computed sims with subscale modeling is the best possible validation
outcome by NREL research, so that extreme-scale AWES platforms can be
built, and perhaps be the key game-changer in the historic world
transition to 100% clean energy.
52 images: Request of editor@upperwindpower.com for an email with with the 52 images for this post. |