HigherEd 2.0

The factor of safety is a somewhat mysterious but always necessary recognition that as engineering designers we do not always know everything we would like to know. Former Defense Secretary Donald Rumsfeld provides some unexpected wisdom here. He might say: “You design using the information you have right, not the information you might want or wish to have some day.” Our goal is to design now, based upon the best information we have. And we recognize that there is uncertainty in our information, that is, in the parlance of Rumsfeld, “known unknowns”.

The point is that as engineering designs, we never (never) have complete information. We have to do the best we can with the information we have, and then build some wiggle room into the design to account for any uncertainties. This wiggle room is often implemented as a factor of safety.

Typical uncertainties involve:

1. loading parameters: what’s the maximum expected load? how many cycles of loading can we expect? what’s the mean and standard deviation of load parameters? what’s the physical mechanism (torsion, bending, etc.), and how severe is each?
2. structural geometry: how accurately (and at what cost) can we manufacture these components? will the geometry evolve over time due to wear (or lack of proper lubrication)? what is the impact of an evolving geometry? are there any installation or assembly uncertainties?
3. environment: how fast will this component oxidize or corrode? will the environment have excessive vibrations (this relates to load parameters too)? are we in an earthquake zone (are we designing in California or Missouri)?
4. safety: will humans use this structure (i.e., is this an air conditioning duct, or an airplane)? what is the relative cost of a “safe” design vs. the cost of litigation related to loss of life in a “failed” design?
5. cost: how much money can I throw at all these problems (how much can I justify to my boss)?

So there are lots of sources of uncertainty, and despite our best efforts we only have so much control over most of these.

The factor of safety, then, establishes a guideline for avoiding failure. But what is engineering failure? Structural failures could be:

1. fracture: a structure literally breaks into pieces
2. plastic deformation: a structure permanently deforms into a configuration not suitable for operation
3. excessive elastic deformation: parts deform too much and create interference or otherwise inefficient operation
4. wear: a component changes geometry due to wear, to the point at which is no longer functions properly (perhaps it vibrates due to excessive clearance)
5. corrosion: a component is weakened due to corrosion or oxidation or other chemical action

and many others as well. So the definition of “failure” is really important for engineering design.

So how do we set a factor of safety? Consider the consequences of failure, and the cost of avoiding failure. Two examples:

1. An air conditioning duct. If the duct experiences a structural failure (say, gets a hole in it for some reason), no loss of life is likely. Replacement cost is probably not too significant. Essentially, there is no significant penalty for failure, so we can design the structure right on the margin. We can use exceedingly thin, light structures for the duct–this will also be the cheapest, of course–with very little fear of serious consequences. (Note: I’ve thrown out the heat transfer aspect of this problem and focused only on the structures part. There will certainly be some interplay between material thickness of the heat transfer properties).
2. An airplane. This is serious business, and so in the aerospace industry the factors of safety tend to be very high. There is certainly some uncertainty in the load parameters, the models used for design, and in fact the basic physics of flight (for instance, unsteady aerodynamics experiments and models). But the driver for high factors of safety is obviously the consequences of failure, which involve massive loss of life. This is also why the industry is so closely regulated by the FAA, in both OEM and maintenance practices. In addition, the aerospace industry is one of the leading champions of design paradigms like six sigma.

This is a huge topic, and you don’t have to work very hard to find lots of good book out there.