October 3, 2011
Hot-Dry/Mixed-Dry

Abstract: 

This report summarizes hygrothermal analysis of specific attics constructed in California. The analysis was done using historical experience, published work in journals and trade publications, current building code requirements and WUFI hygrothermal simulations to assess benefits and risks associated with insulating the roof decks in both vented and unvented configurations. The focus of this report is on modifying conventional, ventilated attics, constructed with impermeable roof shingles (with fiberglass batt insulation on the ceiling plane) by adding fiberglass batt (or netted fiberglass or netted cellulose or spray applied fiberglass) insulation to the underside of the roof deck (i.e. on the slope) while leaving the attic air space ventilated to outdoors.

This report summarizes hygrothermal analysis of specific attics constructed in California. The analysis was done using historical experience, published work in journals and trade publications, current building code requirements and WUFI hygrothermal simulations to assess benefits and risks associated with insulating the roof decks in both vented and unvented configurations.

The majority of the configurations evaluated are well understood and have been addressed in previous published work or in the model building codes. However, the focus of this report is on modifying conventional, ventilated attics, constructed with impermeable roof shingles (with fiberglass batt insulation on the ceiling plane) by adding fiberglass batt (or netted fiberglass or netted cellulose or spray applied fiberglass) insulation to the underside of the roof deck (i.e. on the slope) while leaving the attic air space ventilated to outdoors.

The hygrothermal simulations suggest that adding air permeable insulation to the underside of vented attic roof decks constructed with impermeable roof shingles or impermeable roof membranes is a viable option in many climate zones in California. In climates with cooler wintertime temperatures and/or more night sky radiation, there is a potential for some increase in the moisture content at the inside face of the OSB roof sheathing. The risk of moisture problems is less when the attic ventilation rate is higher than the ceiling leakage rate. Attic moisture problems are strongly connected to indoor humidity levels and it is recommended that indoor humidity be actively controlled.

The key technical issue addressed is the effect on the moisture content of the roof deck sheathing when air permeable thermal insulation is added to the underside of the roof deck.

Specifically, Figure 1, Modified Vented Attic Asphalt Shingles – can be used in all CEC Zones except 16. More broadly speaking, Figure 1 can be used only in International Energy Conservation Code (IECC) Climate Zones 2B and 3B.

The roofing paper in Figure 1 can be substituted with an impermeable fully adhered impermeable membrane with no apparent moisture risk.

A ventilation ratio of 1:300 is recommended wherever attics are vented irrespective of climate zone and is a basic code requirement in the International Residential Code (IRC) for vented attic assemblies.

It is further recommended that field exposure experimental work be conducted to verify these results.

1. Introduction

This report summarizes hygrothermal analysis of specific attics constructed in California. The analysis was done using historical experience, published work in journals and trade publications, current building code requirements and WUFI hygrothermal simulations to assess benefits and risks associated with insulating the roof decks in both vented and unvented configurations.

The majority of the configurations evaluated are well understood and have been addressed in previous published work or in the model building codes. However, the focus of this report is on modifying conventional, ventilated attics (with fiberglass batt insulation on the ceiling plane) by adding fiberglass batt (or netted fiberglass or netted cellulose or spray applied fiberglass) insulation to the underside of the roof deck (i.e. on the slope) while leaving the attic air space ventilated to outdoors. The key technical issue addressed is the effect on the moisture content of the roof deck sheathing when air permeable thermal insulation is added to the underside of the roof deck.

2. Standard Configurations

Attic roof assemblies can be divided into two categories: vented and unvented. The basic principles of both approaches can be found in: BSD-102: Undertanding Attic Ventilation and A Crash Course in Roof Venting. A more detailed review evaluation can be found in: BA-1006: Building America Special Research Projects—High-R Roofs Case Study Analysis. The standard vented and unvented attic configurations are presented in Appendix A. The recommended limits of their applicability are specified and are based on current language in the Model Building Codes. The language in the Model Building Codes is based on previous published work (BA-1001: Moisture Safe Unvented Wood Roof Systems).

Two factors that affect the performance of these assemblies are the temperature of the roof deck and the ability of moisture laden air to access the underside of the roof deck. In vented attic configurations it is expected that moisture laden air is able to access the underside of the roof deck. The function of the attic ventilation is to remove this moisture before it can damage the roof deck.

In unvented attic roof assemblies the primary approach to provide acceptable performance is to either warm the underside of the roof deck by insulating the top surface of the roof deck so that it remains sufficiently warm that excessive moisture accumulation does not occur or to prevent moisture laden air from accessing the underside of the roof deck by installing an air impermeable layer of insulation. The function of the air impermeable layer of insulation is three fold – to prevent air from reaching the underside of the roof deck, to provide an interior surface facing the attic space that is warm enough to control condensation and finally to provide sufficient vapor diffusion resistance to control moisture migrating through the air impermeable insulation such that moisture damage to the roof deck does not occur.

3. Modified Conventional Vented Attics

A conventional, ventilated attic (with fiberglass batt insulation on the ceiling plane) can be modified by adding fiberglass batt (or netted fiberglass or netted cellulose or spray applied fiberglass) insulation to the underside of the roof deck (i.e. on the slope) while leaving the attic air space ventilated to outdoors. Figure 1 shows the placement of deck insulation and the venting details necessary to ensure continued ventilation of the modified attic assembly; Figure 2 shows a range of deck insulation options. The modified conventional vented attic configuration is not well understood and is examined in detail in this study.

Figure 1: Venting Details for Modified Conventional Vented Attic

Figure 2: Insulation Options for Modified Conventional Vented Attic Assembly

Modified conventional vented attic assemblies are created through the addition of insulation at the underside of the existing roof deck. This insulation strategy aims to reduce heat transfer between the roof deck and the attic space. On hot, sunny days the roof deck insulation will have the benefit of reducing heat gain from the solar heated roof surfaces so that attic temperatures are lower, there is less heat gain to the living space and cooling loads are smaller. In this scenario, the roof deck temperature will be higher than it would be without insulation.

Conversely, on cool, clear nights, the roof deck insulation will reduce heat transfer from the attic space to the roof deck so roof deck temperatures are lower and the relative humidity at the deck and moisture content of the sheathing are both higher. Night sky radiation may cause the roof deck temperature to be much cooler than the air temperature and the potential for condensation will be greater.

A series of hygrothermal simulations were performed to assess the anticipated benefits (i.e. reduced attic temperatures and cooling loads) and risks (i.e. increased roof temperatures and elevated sheathing moisture content) associated with the proposed strategy and to determine the effect of various parameters:

  • Climate
  • Indoor Humidity
  • Roofing
  • Ceiling Insulation
  • Roof Deck Insulation
  • Ceiling Air Leakage Rate
  • Attic Ventilation Rate

This report summarizes the inputs for and results of the hygrothermal simulations.

The key parameter examined is the moisture content of the roof deck and the effect various insulation strategies have on the moisture content of the roof deck.

4. Hygrothermal Simulations

The WUFI Pro 5.1 computer model was used to simulate the effects of insulating the roof decks on the moisture and temperature conditions of the roof deck structure. WUFI is one of the most advanced commercially available hygrothermal moisture programs in use today. Its accuracy has been verified (by the Fraunhofer Institut Bauphysik in Holzkirchen, Germany – www.wufi.de) against numerous full-scale field studies of enclosure performance (roofs, walls, foundations, parking garage decks, etc.) over a number of years. Much of the field verification work supporting the model has been related to the hygrothermal performance of residential roof systems.

It is one of the few models in the public domain that can properly account for adsorption of water vapor, absorption/redistribution of liquid water, and night sky radiation. Given the appropriate inputs, WUFI calculates heat and moisture flow every hour under the influence of sun, rain, temperature and humidity (see also www.wufi.de). The analysis is, however, only as accurate as the assembly data, the material properties, and the interior and exterior conditions input.

Climate Data

Outdoor Climate

While WUFI 5.1 Pro includes climate data files for dozens of cities around the world; however, the only Californian city included is San Francisco.  Custom climate files for other cities were created from .csw files provided by the CEC:

  • Sacramento, CTZ 12
  • Red Bluff, CTZ 11
  • Fresno, CTZ 13
  • Riverside, CTZ 10
  • Los Angeles, CTZ 9
  • Palm Springs, CTZ 15
  • Palmdale, CTZ 14
  • Blue Canyon, CTZ 16

It should be noted that the .csw climate files do not include rain data.  This is not a problem for the roof simulations, as the roofing materials do not absorb any rainwater; however, the generated climate files should not be used for other applications  where rainwater absorption may influence predicted hygrothermal performance.

For the purposes of this study, it is important to consider the effects of night sky radiation. WUFI Pro 5.1 uses counter radiation (emission from the roof cladding to the exterior) to predict these effects. Counter radiation was estimated using the sky temperature data from the .csw files. The generated climate files include this estimate of counter radiation.

Indoor Climate

Indoor climate conditions were predicted using EN 15026, one of WUFI Pro 5.1’s built-in indoor climate models. This model uses simple functions to relate indoor temperature and relative humidity to the outdoor temperature.  Table 1 shows the indoor temperature and relative humidity assumed for various outdoor temperatures.  In BSC’s experience, this indoor climate model makes reasonable predictions of indoor conditions with a minimum of assumptions and inputs.

Simulations were performed using assumptions for normal and high indoor RH (Table 1) conditions to assess the impact that indoor RH has on the hygrothermal performance of the proposed retrofit. As noted earlier the impact on roof sheathing moisture content of and the RH at the inside face of the roof sheathing were of particular interest.

Table 1: Assumed Indoor Temperature & RH vs. Outdoor Termperature

Outdoor TemperatureIndoor TemperatureNormal RHHigh RH
deg Cdeg Fdeg Cdeg F%%
-30-2220683040
-20-420683040
-101420683040
03220684050
105020685060
206825776070
308625776070
4010425776070
5012225776070

Material Properties

It is often not convenient (or even possible) to determine the many material properties necessary for hygrothermal simulations so WUFI Pro 5.1 includes a database of several hundred common materials. Hygrothermal computer models are only as reliable as their input data, and it is advisable to measure and use key material properties whenever this can easily be done. However, this project considers the potential performance of a 'typical' CA house hence no actual materials are available to be measured. For the purposes of the simulation of the proposed roof retrofits, we have used the generic material properties from WUFI Pro 5.1's 'Generic North American
Materials' database.

Air Movement

The WUFI Pro 5.1 computer model has the capacity to account for air movement through an assembly; the algorithms predict the amount of heat and moisture that are introduced or removed as a result of this air movement. For the purposes of this study it is essential that several air movement mechanisms be considered:

  • Air leakage through the ceiling (i.e. from the living space into the attic)
  • Ventilation of the roof (i.e. attic air changes facilitated by soffit and roof vents)
  • Air movement through the space between the underlayment and the roof tiles when present

Ceiling Air Leakage

For this study it was assumed that ceiling air leakage accounts for 50% of the house air leakage. Three different house air leakage rates were considered: 3, 5 and 7 ACH50. These were divided by 20 (to convert to ACH4 values), then multiplied by 0.5 (to get the leakage through the attic relative to the house volume) and finally adjusted for the ratio of the house volume to the attic volume (assumed 4:1). Table 2 summarizes the assumed whole house and ceiling leakage rates.

Table 2: Assumed Natural Ceiling Leakage Rates (Relative to House Volume & Attic Volume)

House Air LeakageCeiling Air Leakage
ACH 50ACH14ACH4 (Vhouse)ACH4 (Vattic)
30.150.0750.30
50.250.1250.50
70.35 0.70