A wide variety of mathematical models exist which attempt to assess the risk of mold growth in building envelopes. Model predictions often fail to adequately represent observations from field experiments. We examine ways in which one model, the mold resistance design (MRD) model, can be improved to better align with empirical data. A database of laboratory mold growth results was assembled from the literature, which included not only the European spruce and pine on which pioneering models were based, but also materials commonly used in North America, such as oriented strand board (OSB) and plywood. Next, this database was used to test modifications to the core temperature (T) and relative humidity (RH) dose factors in the MRD model. These dose factors together create a total dose based on average T and RH in a 12-hour period. The accumulation of these doses over time yields a prediction of mold growth and the original model uses a linear mapping to relate cumulative dose to mold rating. Improvement to the dose-response correlation was achieved by implementing a sigmoidal map of cumulative dose to mold rating to increase model predictive power. In addition, improvements were made for unfavorable conditions for mold growth to better align with known biological processes. Instead of a decline in the predicted mold rating, a recovery function was devised such that the predicted mold rating remains constant during unfavorable conditions and increases only upon resumption of favorable conditions contingent on a delaying period. Each modification is shown to improve predictions for the database materials, with promise for improvement in field studies. However, even these improvements still exhibit significant variation in predictive accuracy across our large database of mold growth measurements.
| File Size : | 1
file
, 1.5 MB |
| Note : | This product is unavailable in Russia, Belarus |
| Number of Pages : | 10 |
| Product Code(s) : | DBldgsXV-C078 |
| Published : | 2022 |
