|Posted on Dec 16, 2008 03:24:41 PM | Wayne Hale | 12 Comments ||
Old joke: “You see the glass as half empty, I see the glass as half full, but an engineer sees the same glass and says ‘it is overdesigned for the amount of fluid it holds.’”
When an engineer starts out to build something, one of the first questions to be answered is how much load must it carry in normal service? The next question is similar: hom much load must it carry at maximum? An engineer can study those questions deeply or very superficially, but having a credible answer is a vital step in at the start of a design process.
Here is an example. If you design and build a step ladder which just barely holds your weight without breaking, what will happens after the holidays when your weight may be somewhat more than it was before you eat Aunt Martha's Christmas dinner? You really don’t want to throw out your stepladder in January and build a new one do you? Obviously you would should build a stepladder that can hold just a little bit more. Don't forget what might happen if you loan your stepladder to your coach-potato neighbor who weighs a lot more than you do? Can you say lawsuit?
So how do you determine what your stepladder should hold? Do you find out who is the heaviest person in the world and make sure it will hold that person? Probably not. Better, pick a reasonable number that covers, say, 95% of all folks, design the ladder to that limit and put a safety sticker on the side listing the weight limit. Yep, that is how most things are constructed.
But that is not all. Once you determine normal or even the maximum load it is a wise and good practice to include a “factor of safety”. That means that you build your stepladder stronger than it needs to be. This helps with the idiots that don’t read the safety sticker; it also helps protect for some wear and tear, and it also can protect if the actual construction of your stepladder falls somewhat short of what you intended. So you might build your stepladder with a FS of 2. That would cover 95% of all folks with plenty of margin for foolish people that try to accompany their friend climbing the ladder; or when your ladder has been in service for 25 years (like mine), or when your carpenter buddy builds the stepladder with 1/4” screws rather than ½” screws like you told him to.
Factors of safety are not pre-ordained. They have been developed over the years through experience and unfortunately through failures. Some factors of safety are codified in law, some are determined by professional societies and their publications, and some are simply by guess and by golly. Engineering is not always as precise as laypeople think.
“ . . . the choice of an appropriate factor of safety is one of the most important decisions the designer must make. Since the penalty for choosing too small a factor of safety is obvious, the tendency is to make sure that the design is safe by using an arbitrarily large value and overdesigning the part. (Using an extra-large factor of safety to avoid more exacting calculations or developmental testing might well be considered a case of “underdesigning” rather than “overdesigning.”) In many instances, where only one or very few parts are to be made, overdesigning may well prove to be the most economical as well as the safest solution. For large-scale production, however, the increased material and manufacturing costs associated with overdesigned parts result in a favorable competitive position for the manufacturer who can design and build machines that are sufficiently strong but not too strong.
As will be evident, the cost involved in the design, research, and development necessary to give the lightest possible machine will be too great in most situations to justify the selection of a low factor of safety. An exception is in the aerospace industry, where the necessity for the lightest possible construction justifies the extra expense.”
"Some general considerations in choosing a factor of safety are . . . the extent to which human life and property may be endangered by the failure of the machine . . . the reliability required of the machine . . . the price class of the machine.”
Standards for factors of safety are all over the place. Most famously, the standard factor of safety for the cables in elevators is 11. So you could, if space allowed, pack eleven times as many people into an elevator as the placard says and possibly survive the ride. For many applications, 4 is considered to be a good number. In the shuttle program the standard factor of safety for all the ground equipment and tools is 4.
When I was the Program Manager for the Space Shuttle, there were a number of times when a new engineering study would show that some tool either could be exposed to a higher maximum load than was previously thought, or that the original calculations were off by a small factor, or for some reason the tool could not meet the FS of 4. In those circumstances, the program manager – with the concurrence of the safety officers – could allow the use of the tool temporarily – with special restrictions – until a new tool could be designed and built. These “waivers” were always considered to be temporary and associated with special safety precautions so that work could go forward until the standard could once again be met with a new tool.
In the aircraft industry, a factor of safety standard is 1.5. Think about that when you get on a commercial airliner some time. The slim factor of safety represents the importance of weight in aviation. It also means that much more time, engineering analysis, and testing has gone into the determination of maximum load and the properties of the parts on the plane.
For some reason, lost in time, the standard FS for human space flight is 1.4, just slightly less than that for aviation. That extra 0.1 on the FS costs a huge amount of engineering work, but pays dividends in weight savings. This FS is codified in the NASA Human Ratings Requirements for Space Systems, NPR 8705.2. Well, actually, that requirements document only references the detailed engineering design requirements where the 1.4 FS lives.
Expendable launch vehicles are generally built to even lower factors of safety: 1.25 being commonplace and 1.1 also used at times. These lower factors of safety are a recognition of the additional risk that is allowed for cargo but not humans and the extreme importance of light weight.
It is common for people to talk about human rating expendable launch vehicles with a poor understanding of what that means. Among other things, it means that the structure carrying the vast loads which rockets endure would have to be significantly redesigned to be stronger than it currently is. In many cases, this is tantamount to starting over in the design of the vehicle.
So to the hoary old punch line: Would you want to put your life on the top of two million parts, each designed and manufactured by the lowest bidder?
Tags : engineering, risk, safety