- Author
- Puchovsky, M. T.
- Title
- Improved Computer Algorithm for Growing Fires.
- Coporate
- Worcester Polytechnic Inst., MA
- Report
- Thesis, February 1991, 171 p.
- Keywords
- computers | algorithms | flame spread | combustible solids | fire codes | physics | evaluation
- Identifiers
- WPI/Fire Code; PU 7004; GROW
- Abstract
- The focus of this work is the development of a computer algorithm which will eventually replace the existing Growing Fire algorithm in the WPI/Fire Code. In order to do this a simple algorithm was developed so as to minimize computer time. The Growing Fire algorithm is a flame spread model which simulates an objects burning rate as the flame spreads from the objects center. The flame spread rate predicted by the Growing Fire is a function of the flame itself and the surrounding environment. The utility of the Growing Fire is limited since it was developed from specific burn tests of the polyurethane foam, PU 7004. Thus, results are only accurate for PU 7004. The objective of this thesis is to develop a growing fire algorithm which is applicable to a wider range of fuels. The new algorithm entitled GROW predicts horizontal flame spread over non-charring fuels. It was developed from existing flame spread theory which divides the fuel surface into a number of segments and then considers the transient radiative heat transfer impinging at each segment. GROW includes approximations for flame size, heat transfer to the unburned fuel and the rise in surface temperature of the unburned fuel as a function of time. The flame is modeled as a cone. Forward flame radiation to the unburned fuel is the only form of heat transfer considered. The concept of an ignition temperature for semi-infinite solids is considered since the flame front advances to the segments which have reached their ignition temperature. Results produced by GROW were used to analyze an existing flame spread and mass loss rate theory. The evaluation was conducted using existing open burn experimental results for three non-charring polyurethane fuels. Experimental results corresponding to steady state radiant panel tests could not be used. Experimental results consist of mass loss rates and flame position as a function of time. The flame shape, mass loss and flame spread results produced by GROW were compared against the experimental data. The effect of the user defined physical parameters such as number of segments and length of time step were investigated. An appropriate initial flame size for the fuel samples considered was determined. A previously developed alternative to the Growing Fire was also evaluated. Results from GROW show that existing theories using radiation as the only form of heat transfer do not accurately model the flame spread process. It appears that additional physical phenomenon such as lateral conduction through the fuel needs to be considered.