Introduction

Moisture Control Strategies Presently Employed

Proposed Moisture Control Strategy

Developing the Algorithms

Using the Algorithms

An Example

Comparison with Existing Methods

Conclusions/Future Work

References

Computer simulations are quite effective in predicting the ability of a roofing system to prevent problems with moisture accumulation. However, it is necessary to set up, run and analyze a computer simulation in order to determine the results. Algorithms were therefore developed in order to predict the moisture control performance of a roofing system without having to perform and analyze the results of a computer simulation. These algorithms can then be included in a fast, user-friendly program, or performed using a hand calculator, which makes the information available to a much wider group. This will enable the roofing professional in the US to quickly and accurately determine if a roof constructed with a given type of membrane, insulation material and deck will be moisture-tolerant in a given location on a building controlled to a specific indoor relative humidity, without the need to set up and run a computer simulation.

The algorithms were developed using a database of 600 simulations. Five different climates were analyzed: Bismarck, Chicago, Knoxville, Miami, and Seattle. These were selected to represent the range of heating degree days (HDD) seen in the continental US. Indoor relative humidities of 40%, 50%, and 60% with an indoor temperature of 68°F were used in the study.

The range of roofing configurations evaluated included 1-inch and 3-inch thick wood fiberboard, 1-inch and 3-inch polyisocyanurate (PIR) insulation, and a 3-inch composite of the two. Four metal decks with permeances of 0.64, 1, 5, and 10 English perms were included. Two values for membrane absorptance of 0.7 for a white roof and 0.1 for a black roof were also used. All possible combinations of the above parameters were simulated using the finite-difference model. Table 1 shows the roof properties and environmental conditions analyzed [1].

This database was analyzed for each of the quantifiable moisture control requirements to develop the predictive algorithms. Multiple linear regression was done using combinations of first, second, third order and inverse terms of each of the variables to develop the necessary correlations. See Reference [9] for details regarding the production of these algorithms.