A residential attic model, contained in the finite element computer program FSEC 3.0, was empirically aligned with measured attic data from three roof research facilities in Florida and Illinois. This model was then used to simulate hourly space conditioning energy use, and roof and attic temperatures, for peak cooling days and annual weather, for Orlando, Florida and Las Vegas, Nevada. Results showed that, when compared to typically vented attics with the air distribution ducts present, sealed cathedralized attics (i.e. unvented attic with the air barrier and insulation at the sloped roof plane) can be constructed without an associated energy penalty in hot climates.
Historical Perspective
Detailed Simulations
A residential attic model, contained in the finite element computer program FSEC 3.0, was empirically aligned with measured attic data from three roof research facilities in Florida and Illinois. This model was then used to simulate hourly space conditioning energy use, and roof and attic temperatures, for peak cooling days and annual weather, for Orlando, Florida and Las Vegas, Nevada. Results showed that, when compared to typically vented attics with the air distribution ducts present, sealed cathedralized attics (i.e. unvented attic with the air barrier and insulation at the sloped roof plane) can be constructed without an associated energy penalty in hot climates.
Table 1: Summary of Annual Simulation Results for Las Vegas: Space Conditioning
Angel Park Subdivision
Following the favorable indications from simulation results, code variance was obtained and two test houses were constructed in Angel Park with sealed cathedralized (unvented) attics. These houses were tested and monitored to evaluate their cooling energy performance compared to a conventional vented attic house.
Short-term monitoring
Results from short-term testing in the late summer of 1996 showed that the unvented attic houses had cooling energy use savings over the conventional 1:150 vented attic house (an average of 19 percent). This was mostly due to sealing the attic and getting the air distribution ducts inside the air and thermal boundary of the building.
Tile top temperatures were hardly effected by the sealed attic. The 3 F maximum tile top temperature difference agreed well with the simulated prediction. The maximum measured plywood roof sheathing temperature increase of 17 F for the sealed attics is less than the temperature variation experienced by changing from tile to asphalt shingles of any available color. During the test period, the maximum measured roof sheathing temperature of 126 F for the sealed attics is well within an acceptable temperature performance range of wood-based roof sheathing (
These results set into motion plans to build an entire subdivision (Cypress Pointe) according to the Building Science Consortium (BSC) Building America Initiative (BAI) specifications.
Long-term monitoring
One of the two unvented attic houses (Angel Park Lot 24) and the conventional vented attic house (Angel Park Lot 22) were monitored between July 1997 and March 1998. Analysis of the cooling season data showed an average of 5% savings for the unvented attic house. This is illustrated in Figure 1. The data was somewhat challenging to analyze due to a number of factors including: 1) different occupant habits impacting cooling system efficiency and internal heat generation; and 2) the addition of a pool at the BAI house which necessitated the removal and reinstallation of the outside compressor-condenser unit. These challenges increased the level of uncertainty in the result. The expected result was at least 15% savings for the BAI house.
Figure 1: Regression of cooling energy consumption versus inside to outside temperature difference for the vented attic (LV22) and unvented attci houses
Figure 2: Average hourly temperatures for the month of August showing the large difference in environment where the air distribution system ducts reside
Figure 3: Hourly maximum bottom of roof sheathing temperatures for the vented (LV22) and unvented attic houses, peaking at about 20 F differential
Analysis of the heating season data showed heating energy consumption savings of over 50% for the BAI unvented attic house. This information is shown in Table 2 and Figures 4 and 5. In addition to the benefits of the sealed cathedralized attic, part of these savings were also due to the higher performing low-e windows, however, the exact effect of the windows is unknown since the use of window coverings varied with the occupants. We have found that most window coverings in the Las Vegas climate are kept closed during the summer, somewhat limiting the benefit of high performance windows, but use varies in the winter.
Table 2: Summary of measured heating data for unvented and vented attic houses in Las Vegas 7-28 February 1998
Figure 4: Regression of heating energy delivered, normalized for inside to outside temperature difference, versus outdoor temperature
Figure 5: Hourly average heating energy delivered, measured at the air handler unit, for the month of February
Figure 6: Hourly average temperatures for the vented and unvented attics for the month of February, showing the large difference between the environments where the air distribution ducts reside
Figure 7 is a high/low/mean plot of the temperature of the bottom of the roof sheathing of the unvented attic. At night, the sheathing temperature briefly approaches the indoor air dewpoint temperature of about 45 F, then sharply rises during the day. Moisture measurements were made of the roof sheathing during the winter and all readings were below 6% moisture content. . .
Download complete report here.