Based on simulation results, the high-mass
solar heating system being demonstrated
in Virginia will provide 70 to 80 percent
of a home’s heating load.
by DAVE STETS with JIM MCLESKEy
and MARSHALL SWEET
sOLaR TODAY®
JANUARY/FEBRUARY 2011
VOL. 25, NO. 1
Copyright © 2011 by the American Solar Energy Society Inc. All rights reserved.
used to model our single-story ranch style home
built on a crawl space. The crawl space was mod-
eled using Type 701 (Basement). We used Type
76 (Theoretical Flat Plate Collector) to model
solar thermal collectors that provide thermal
energy to the heatstore bed (Type 342).
Our team modeled six different-sized houses:
800 square feet (74 square meters), 1,000 square
feet (93 square meters), 1,400 square feet (130
square meters), 1,600 square feet (149 square
meters), 2,000 square feet (186 square meters)
and 2,400 square feet (223 square meters). Each
house was modeled with and without a heatstore
bed. The bed volume was varied for each house.
All simulations were run using weather data
for Richmond, Va. On models with a heatstore
system, 80 percent of the south-facing roof was
covered with solar thermal collectors.
Heat demand for each home was modeled
with an idealized heating system to maintain a
minimum temperature of 68°F ( 20°C) inside the
house. The same system was used for auxiliary
heating in models with heatstore. For homes
with heatstore, the auxiliary heating system was
designed only to keep the house from dropping
below 19°C (66ºF). In the model, if the tem-
perature of the heatstore bed was less than the
temperature of the home, then the radiant floor
would not activate. Similarly, we modeled the
system so that it’s only possible to add heat to the
system when the output temperature of the solar
collectors exceeds the temperature of the storage
medium. This avoids cooling the bed.
Results showed that the optimal size for a heat-
store bed in Richmond, Va., is between 530 and
565 cubic feet ( 15 and 16 cubic meters) and the
shape is square. This dimension varies for other
geographic locations. The simulation showed that
a bed of this size can provide 70 to 80 percent
of the heating load for the modeled homes. We
measured this effect as the amount of reduction
in needed auxiliary heat (figure 4). Smaller homes
having smaller heat demands saw the greatest
reductions. However, larger homes have a higher
we found the optimum collector area to be 75
percent of the total roof area, as long as all panels
faced south at a 37-degree angle. These simula-
tions involved many other variables, as well.
Initial calculations estimate a payback
of approximately 25 years for the optimized
system. The solar panels account for the bulk
of the cost.
Figure 3. TRNSYS Simulation System for a House
with Seasonal Solar Heat Storage
internal volume-to-wall area ratio, which allowed
them to achieve higher efficiencies. The higher
efficiencies were evidenced by a decrease in heat
loss from the home per square foot that was lower
in larger houses.
We also optimized the flow rates through the
size and number of solar thermal collectors used
to charge the heatstore bed. We found that high-
er flow rates increased heating efficiency, with
an optimal speed of 3 gallons per minute. Not
surprisingly, results showed that we reduced
needed auxiliary heating as we increased solar
collector area. Even though using more collec-
tors leads to greater reductions in auxiliary heat,
david Stets founded richmond bysolar ( bysolar.net)
in 2007 and installed the first solar thermal system at
Virginia Commonwealth university (VCu). He funded
the original solar heatstore research at VCu. Contact
him at dave.stets@bysolar.net.
James t. McLeskey Jr. is an associate professor of
mechanical engineering at VCu. mcLeskey holds a
Ph.d. in mechanical engineering from the university of
Virginia. He teaches courses in thermodynamics, fluid
mechanics and energy conversion systems.
Marshall L. Sweet is completing a master of sci-
ence degree in mechanical and nuclear engineer-
ing at VCu.
