Potential Research Topics
Although CFAST is not currently under active development, there still remain issues related to the existing model that could use further validation. Potential topics are listed below. A constant theme is the fact that the empirical correlations describing the plume, ceiling jet, vent flows, etc., were developed independently, and it has been left to the model developers to "sew" these various components together is a consistent way. If these correlations were implemented in their original forms, there would be discontinuities in temperature and potentially violations of the basic conservation laws. Hence, all of the basic correlations have been modified, and these modifications have never been rigorously evaluated in isolation.
The goal of any research related to these topics ought to be:
- Modification of the CFAST Technical Reference Guide, if necessary.
- A change in the CFAST source code, if necessary.
- Several verification cases added to the CFAST Software Development and Evaluation Guide.
- At least one comparison with full-scale experiments added to the Evaluation Guide.
We suggest that those interested in pursuing one of these issues contact us and propose a course of action. We would assist in whatever way we can.
The empirical model of flow through a ceiling vent was developed by Len Cooper at NIST in the late 1980s and early 1990s. It is similar to other models developed around that time, but there are few validation data sets (that we know of) that we can use to validate how it is used in CFAST. One particular issue related to flow through ceiling vents has to do with "plug-holing"; that is, the relative amount of gas from the upper and lower layers of the lower compartment that is entrained into the vent flow. A simple correlation based on the Froude number was developed using FDS to indicate how much of the lower layer air is entrained into the flow. This simple result has not been thoroughly evaluated experimentally. See the CFAST Technical Reference Guide, section "Horizontally-Oriented Vents (Ceiling Vents)".
CFAST uses Heskestad's correlations of plume temperature, velocity and entrainment. These correlations were developed for unobstructed plumes with no upper layer. However, in CFAST, we must compute the centerline temperature and velocity of the plume both below and above the interface between the lower and upper layers. The CFAST Technical Reference Guide contains a variation of Heskestad's plume temperature correlation for the region above the interface. The basic idea is that the temperature of the entrained gases are not ambient but rather that of the upper layer. This assumption has not been rigorously validated experimentally. See the CFAST Technical Reference Guide, section "Plume Temperature and Velocity."
Much of the experimental work on ceiling jets was performed under relatively large, open, unconfined ceilings. However, in real fires ceiling jets are usually bordered below by the upper gas layer. CFAST uses the empirical correlation of Alpert and Heskestad, with a modification to account for (1) the upper gas layer, and (2) the centerline plume temperature. These modifications have not been extensively validated in isolation. In other words, CFAST has been compared with full-scale room fire experiments, but it is difficult to assess the validity of the modifications of the ceiling jet correlation based on the overall comparison of hot gas layer temperature. See the CFAST Technical Reference Guide, section "Ceiling Jet".
There is a simple empirical suppression model implemented in CFAST based on the work of Madrzykowski and Evans (CFAST Tech Guide, section "Fire Suppression"). This model is based on a small number of experiments and its applicability to a wide variety of fuels is questionable. A good research project would be to either more rigorously validate this model, or suggest its removal. We have kept this model in CFAST in the hopes that it can be extended to a wider variety of fuels. If it cannot, it should be removed.