Receiver Geometries for Solar Dish Concentrators
Abstract
This thesis studies receiver/reactor geometry for the ammonia
thermochemical storage system and dish concentrators, using the
case study of a masked 9 m2 dish. Experimental results proved
that altering the geometric arrangement – in this case the cone
angle – of reactor tubes within a cavity receiver significantly
affects the heat transfers within and heat losses from the
receiver. By consequence both the receiver efficiency and the
solar-to-chemical efficiency of the dish-receiver system can be
significantly increased. This is note-worthy as improvements in
efficiency due to reactor tube cone angle do not increase the
cost of manufacture of the receiver. Moreover, any increase in
receiver efficiency allows an equivalent decrease in mirror area,
and in a commercial plant, the solar collector field (mirror
area) can account for around 40% of total project costs.
Gains of up to 7.0% absolute in receiver efficiency and 6.4%
absolute in solar-to-chemical efficiency were demonstrated
experimentally in this thesis at a given solar altitude.
Experimental data was examined for three receiver configurations
with reactor tube half cone angles of 17.5º, 7.5º and 3.7º,
with the 3.7º configuration producing the highest receiver
efficiencies, the 7.5º receiver a close second, and the 17.5º
configuration producing the lowest efficiencies.
Solar-to-chemical efficiencies were however slightly higher for
the 7.5º configuration than the 3.7º, indicating that
system-level operation, especially operation of the heat
exchanger, was slightly more optimal for the 7.5º case.
Preliminary results integrating the receiver losses over the
course of a year indicate that the 3.7º receiver geometry would
reduce annual receiver losses by 25% compared to the 17.5º
receiver geometry.
The experimental data comprised 178 hours of “on sun”
operation of the 9 m2 dish concentrator by the author,
providing 311 steady state intervals of 10-15 minutes. The
experimental results were compared with simulations using an
existing computational receiver/reactor model, but gross
disagreement was observed. Simulated reaction extents were 2-4
times those achieved experimentally. Therefore a new energy
balance was developed for each of the three receiver geometries.
This included 3D modelling of radiation losses, along with
calculation of convection and conduction losses, and used the
knowledge gained experimentally of reactor tube and cavity wall
temperature profiles. These new energy balances were verified
against a substantial body of experimental data – out of 86
steady-state data points at optimal flow, calculated receiver
losses fell within the experimental uncertainties for all but
seven points. The energy balances allowed curves of predicted
receiver losses to be generated for each of the three receiver
geometries over a wider range of temperatures and solar
altitudes. These curves suggest that the 3.7º half cone receiver
would still have the lowest receiver losses under all operating
conditions, closely followed by the 7.5º half cone receiver. At
mid and low solar altitudes, natural convection losses dominated
over conduction and radiation losses for all three receiver
geometries.
Finally, conclusions on favourable flux and temperature profiles
for the reactor tubes studied in this project were drawn. A flux
of at least 2 W/cm2 within 100 mm of the reactor tube inlet
resulted in higher conversion efficiencies, due to higher
temperatures and reaction rates immediately at the tube inlet.