LRFD and aircraft structure design 1

[amorrison]

Does the "new" civil engineering bridge and structure design method called load and resistance factor design (LRFD) have anything to say to aircraft structure design.
Where "standard" design uses a single safety factor - LRFD seems to be a method were the load factor (handles quality of "load" knowledge) and resistance (quality of structure itself - connections etc.) factors are treated seperately to give more optimized structure.

 

[der8110]

Civil engineering codes do not apply to aircraft structure, though there are many areas of common application.

Multiple factors are nothing new in aviation, such as fitting, bearing, and casting factors, in addition to the 1.5 limit/ultimate factor that always applies.

"Quality of load knowledge" is typically handled by the usage of the data, the chart indicating Level I for first estimates, then Level II being the minimum for initial certification, finally Level III being usable for sustaining efforts. In addition, statistics are generally accounted for in s-N curves for fatigue and da/dN data for crack growth.

Stress concentration factors and stress intensity solutions for fracture mechanics are other areas where multiple factors are compounded.

 

[amorrison]

Some background - copied from

http://www.tfhrc.gov/pubrds/mayjun99/geoscan.htm

A Primer on ASD and LRFD Design Methods

The following provides a basic explanation of the differences between allowable stress design (ASD) and load and resistance factor design (LRFD) methods. The common concept behind both design methods is to compare applied forces with available resistances to ensure that a certain level of reserve capacity is available to account for the uncertainty in both the loads and resistances. This reserve capacity provides confidence to the engineer that his/her design is safe against poor performance — or worse, catastrophic failure. The method of defining and quantifying these uncertainties is the fundamental difference between these two methods of design.

ASD, structural elements such as structural foundations, bridge beams and girders, or earth-retaining walls are designed to support, or resist, anticipated service loads, including vehicular live loads, superstructure dead loads, or lateral soil loads. To account for the possibilities that structural elements are overloaded during their service life and that the materials providing resistance to the load are not as strong as expected, engineers apply a global safety factor on the resistance side of the design equation to ensure that the structural elements are large enough to account for all uncertainties in design. In this way, global factors of safety account for the uncertainty in both loads and resistances. The general forms of the equation appear as follows:

General Design Equation: Resistance provided (R) > Loads applied (G L)


ASD: R / F.S. > G L, where the Factor of Safety (F.S.) = 1.5 to 3.5
Although the ASD approach ensures that the supporting design element is sufficient to carry potential overloads, the approach does not supply the designer with two vital pieces of information. The total capacity of the supporting element cannot be ascertained with ASD, and therefore, the mode of failure cannot be predicted with certainty. Often, this means that the global factors of safety are set at overly conservative levels.

In some cases, global factors of safety are not conservative. This may be difficult to imagine since structural elements do not frequently fail. However, rather than attributing this to the quality of the analytical method, this can, in large part, be attributed to the fact that engineers employ judgment and experience in the design process. The ASD method does not provide a rational means to define the level of safety of the design element.

In LRFD, uncertainties in both applied loads and structural and material resistances can be better discerned when they are separated and studied individually. Likewise, if safety factors can be applied in the design equation, both on the load and resistance sides, the designer can better use analytical tools to establish the total capacity of design elements. The designer can more accurately predict dead loads such as the weight of concrete and steel in the superstructure; however, they may apply a more conservative load factor to transient or vehicular live loads.

The general form of the LRFD equation takes on the following simplified appearance:


LRFD: N R > G m L
In this equation, resistance factors (N) are values less than one to account for the uncertainty that the materials providing resistance may not be as strong as anticipated. Load factors (m) are values greater than one to account for the possibility that overloads will be applied to the element during its service life. With the LRFD approach, the designer can better assign margins of safety to each portion of the design equation as suited to the level of confidence with which each load and resistance can be predicted. Therefore, designs can be based on risk and reliability concepts. By calibrating the load and resistance factors to an overall margin of safety, designers can ensure that all designs have prescribed margins of safety against failure.

 

[rb1957]

this sounds a little like probablistic design, as opposed to deterministic design (which uses factors to allow for variability).

simply, deterministic design says if we factor our maximum stress by 1.5 (and other special factors) and the result is less than ftu (or buckling allowables, etc) then we're ok.

probablistic design says there is a probability distribution for applied load, and for material strength, and for anything else that affects the strength of the part/structure. muiltply these distributions together and ensure that the chance of failure (applied load exceeding part's strength) is less than 1e-7 (or -9).

personally, i'd want to see alot of the underlying substantiation of these probability distributions before i believed it. it also sounds like a lot more work for probably very little gain.

 

[swearingen]

Obviously the factors have to be researched intensively. Fortunately for the CE structures field, there are lots of data to support them.

"it also sounds like a lot mork work for probably very little gain" - rb1957

It depends. In bridges, where the dead load to live load ratio can be extremely high, LRFD can return a considerable savings. In fact, AASHTO (American Association of State Highway and Transportation Officials) codes specify the use of this method for roadway bridges for precisely this reason.