TABLE OF CONTENTS For performance-based designs, an understanding of the uncertainties can be important. This appendix provides an overview of the methods that can be used to evaluate and combine uncertainties. Usually, the uncertainty will be dominated by a few key parameters. Significant contributors should be evaluated. A sensitivity analysis as discussed in Section 10.5.6.2 might identify these parameters that most significantly contribute to the uncertainty.
In order to determine the best way to treat the uncertainty, the following steps are taken:
The most significant parameters should be identified.
The types of uncertainty (e.g., theory or model) that exist should be identified.
It should be determined if it is crucial to treat the uncertainty quantitatively. Only quantities about which the uncertainty has scientific significance (i.e., is capable of reversing the acceptability of the final solution) need to be treated quantitatively.
The appropriate methodology or tool should be selected for the job.
Uncertainty is encoded on crucial variables and propagated throughout the analysis.
Treating every uncertainty quantitatively in a complex performance-based design might be impossible. The time required and computational complexity is prohibitive. However, it is unnecessary to do this. When dealing with uncertainty, one of the most important challenges is to identify and focus on those uncertainties that matter in understanding the project scope, and in making decisions about the fire safety tools, methods, and design options.
Scientific Uncertainty versus Statistical Uncertainty. Statistically significant refers to a mathematical calculation...
