A simple duct system was installed in an attic test module for a large-scale climate simulator at a U.S. national laboratory. The goal of the tests and subsequent modeling was to develop an accurate method of assessing duct system performance in the laboratory, enabling limiting conditions to be imposed at will and results to be applied to residential attics with attic duct systems.

Steady-state tests were done at a severe summer condition and a mild winter condition. In all tests the roof surface was heated above ambient air temperatures by infrared lights. The attic test module first included then did not include the duct system. Attic ventilation from eave vents to a ridge vent was varied from none to values achievable by a high level of power ventilation. A radiant barrier was attached to the underside of the roof deck, both with and without the duct system in place. Tests were also done without the radiant barrier, both with and without the duct system. When installed, the insulated ducts ran along the floor of the attic, just above the attic insulation and along the edge of the attic near the eaves and one gable.

Air temperatures were measured from the ridge to the insulation surface along the center of the test module at all ventilation rates. For all tests, air temperatures inside the ducts, as well as attic air, attic insulation, and gable and deck temperatures, were measured and compared to the predictions of the model. Only average attic air temperatures were compared since the model did not include stratification. The ducts were placed along the eaves in the test module. This is thought to exacerbate stratification in these tests more than the placement of ducts in real attics would. The ducts along the eaves partially blocked the path for ventilation air to mix with attic air near the insulation between the ducts.

Despite adequate duct insulation, the duct system kept attic conditions cooler in summer and warmer in winter. Sincethe infrared lights were heating the roof above ventilation air temperatures at all conditions, increasing ventilation caused attic air and insulation surface temperatures to decrease. At the mild winter condition, compared to measurements with no radiant barrier attached to the underside of the deck but with the ducts installed, there was an average 37% increase in heat loss into the attic with the radiant barrier and ducts in place. This heating penalty varied randomly with ventilation rate in these tests. At the severe summer condition simulated in the tests, the radiant barrier decreased the heat gain through the ceiling. The average cooling benefit was 34% with ducts in the attic and 29% without them. Variation with ventilation rate was again random, but there was less variation than at the mild winter condition.

These tests in a climate simulator achieved careful control and reproducibility of conditions. This elucidated dependencies that would otherwise be hidden by variations in uncontrolled variables. Based on the comparisons with the results of the tests at the mild winter condition and the severe summer peak condition, model predictions for attic air and insulation temperatures should be accurate within ±10°F (±6°C). This is judged adequate for design purposes and could be better when exploring the effect of changes in attic and duct parameters at fixed climatic conditions.

Units: Dual