The selection of the factor of safety to be used for various applications is one of the most important engineering tasks. On the one hand, if a factor of safety is chosen too small, the possibility of failure becomes unacceptably large on the other hand, if a factor of safety is chosen unnecessarily large, the result is an uneconomical or nonfunctional design. The choice of the factor of safety that is appropriate for a given design application requires engineering judgment based on many considerations, such as the following:
- Variations that may occur in the properties of the member under consideration. The composition, strength, and dimensions of the member are all subject to small variations during manufacture. In addition, material properties may be altered and residual stresses introduced through heating or deformation that may occur during manufacture, storage, transportation, or construction.
- The number of loadings that may be expected during the life of the structure or machine. For most materials the ultimate stress decreases as the number of load applications is increased. This phenomenon is known as fatigue and, if ignored, may result in sudden failure.
- The type of loadings that are planned for in the design, or that may occur in the future. Very few loadings are known with complete accuracy—most design loadings are engineering estimates. In addition, future alterations or changes in usage may introduce changes in the actual loading. Larger factors of safety are also required for dynamic, cyclic, or impulsive loadings.
- The type of failure that may occur. Brittle materials fail suddenly, usually with no prior indication that collapse is imminent. On the other hand, ductile materials, such as structural steel, normally undergo a substantial deformation called yielding before failing, thus providing a warning that overloading exists. However, most buckling or stability failures are sudden, whether the material is brittle or not. When the possibility of sudden failure exists, a larger factor of safety should be used than when failure is preceded by obvious warning signs.
- Uncertainty due to methods of analysis. All design methods are based on certain simplifying assumptions which result in calculated stresses being approximations of actual stresses.
- Deterioration that may occur in the future because of poor maintenance or because of unpreventable natural causes. A larger factor of safety is necessary in locations where conditions such as corrosion and decay are difficult to control or even to discover.
- The importance of a given member to the integrity of the whole structure. Bracing and secondary members may in many cases be designed with a factor of safety lower than that used for primary members.