Aircraft Bridge Design Guides 1

[tjclose]

I am looking for the latest information regarding designing bridges to resist loadings produced by aircraft. I will be looking at a two-span, "tunnel" bridge which spans across a two lane highway and is continuous with the abutment walls and the center pier wall. Each cell of the tunnel will contain one direction of roadway traffic.

The bridge is located in a runway landing zone therefore there will be large vertical gear loads due to impact and large longitudinal gear loads due to instantaneous braking. I am looking for the appropriate resisting width distribution factors for such a model.

I was looking at general purpose software which allows for user defined moving loads, but they all seem to allow the moving loads to be placed on frame elements only. This will not allow for the capture of an effective resisting width which could be determined if the structure was modeled with meshed plate elements. I could use meshed plates and manually move the gears through load cases but I was trying to avoid that method.

If anyone has a list of aircraft bridge design guides or has successfully utilized a particular software package for this type of problem, I would appreciate your input.

Thanks

 

[ludvik]

I don't know about aircraft bridges in particular, but I might be able to help with your moving load analysis.

As you point out, moving loads in most structural FE packages can only be applied to beam elements and not shells. A method often used to dealing with this is to use some beam elements in your model if you have them (representing girders or something), or to put in some extra beam elements purely for the purpose of applying your moving load.

You can put in a string of rigid elements between nodes in your shell mesh. Each element should have moment and axial force releases configured so that they do not contribute to the stiffness of the structure, and only transfer vertical loads to the shells. This is not difficult to do.

BTW: I do not understand why structural software firms don't develop their moving load analysis more....there are a bunch of things moving load analysis won't do which it should. Another big one is multiple lanes. LARSA is promising that it will have 2-D influence surfaces in some future release which will help a little, but still, every single bridge ever designed needs a moving load analysis - there should be a market for a sophisticated tool.

Hope this helps.

 

[zliazmei]

Yes,There is a problem with FE packages.

But what I would firs is to check a 2D frame model with stiffnesses of meter lenght of structure. All packages will do this.
2 spans are no so much staticaly indeterminate. There are a few places in the structure where maximumm effects will ocure. So you can manually move the gears through and it will not take a long time.

If you can tell us something about geometry of the structure and loads , and type of package you have it will be good.

Zlia Zmei

 

[Dinosaur]

Ludvik,

I don't agree that all bridges need a moving load analysis. This has been done for over a hundred years with influence lines. All you need is some reasonably conservative estimate of the 'distribution width' and you are off. If you have ever tried to 'code up' a moving load analysis you would understand why there are not many out there and that they are often limited.

Getting back to the problem . . .

Tjclose,

I have been told that MDX uses a grid analysis module to determine moments and shears in plate girders composite with a concrete slab. The girders may be straight or curved and the geometry of the problem is allowed to be quite complex, but on these problems the input is complex as well. You don't get something for nothing. MDX also allows the user to input a user defined vehicle and define lanes for the vehicle as well. It seems to me that proper use of these tools could solve many of your problems.

I am no expert on aircraft loading, but I would be most concerned for the 'touch down loads' when you would have maximum impact both vertically and longitudinally. Since the superstructure is continuous with the abutments, the longitudinal loads should be handled through the slab and resisted in a passive soil pressure at the abutment. The downward impact load is the big concern. One problem is that if the loads are large enough, you will have to take precautions against uplift at the abutments.

Sorry I don't have more answers, but I think this may help.
Good Luck

 

[boo1]

The Aircraft Engineering forum might be a good place to inquire about the design impact loading of aircraft.

Cheers

 

[Qshake]

The placement of most bridges in a airport layout exclude the actual runway and thus are not subject to the high impact loads a suggested. It is far more common to have bridges on taxiways, where the plane is limited in speed which in turn will minimize impact.

 

[tjclose]

Dinosaur:

I agree that the "touchdown" loads are of a high concern. I have already done some preliminary work on the longitudinal load resisting mechanism and determined that it is really a soil-structure interaction problem between the frame action of the tunnel/bridge and the percentage of passive pressure that will be mobilized based on the frame deflections.

I plan to use soil springs whose stiffnesses vary with depth applied to the abutment walls to capture the soil-structure interaction. Again, resisting frame width becomes a large issue.

Qshake:

This case is somewhat unique in that this bridge will be located within the runway landing zone.

 

[wiktor]

tjclose,
Some comments on the above discussion;
1. For the typical loads refer to the following documents:

http://www.boeing.com/assocproducts/aircompat/777rsec7.pdf

http://www.boeing.com/assocproducts/aircompat/7474sec7.pdf

You will be also able to get data for different aircrafts from the same web site.
2. Touch down theorem - just drop it! The landing aircraft is much lighter than the one taking off (for the same model). The details are in the publications above. One should also consider the lift, which is the only reason why the aircraft is flying. At the "touch down" it's just the touch, and newer the full weight of the airplane applied.
3. The links above present the loadings for the design of the runway and the current equipment. The LA Airport is already using heavier loads in their designs foreseeing the trend in the industry.
4. It is advisable to consider crash loading to avoid potential liability in case of major accident. It could be a typical aircraft landing at regular approach speed, say 140 – 150 knots, at 15-degree angle, no landing gear. The bridge should be capable of taking this impact with no collapse requirement, to avoid secondary casualties.
5. It is advisable to avoid “hard spots” in the runway. One will be created when the tunnel bridge will be located just under the pavement, with no runway bed (sub base) on the top. As one-foot difference will not provide for major difference in the construction cost, I will strongly advise going slightly deeper with the bridge.
6. The load distribution is the same as for any other moving loads. Influence lines for plane frame will take care of the longitudinal moments and forces and you will be able to locate the positioning of the plane for the maximum loading.
7. For the transverse distribution, create the FE model and load it with the design plane at the position to produce max longitudinal moments (see above). It will give you max transverse to design your slab. It could be used (with the soil springs) for the longitudinal loads.
8. The horizontal loading will need some additional analyses. The runway pavement will take most of the loads. There will be some transfer through the sub base material, but not to the tune to be of the major concern.

Good luck wit h the design!

 

[tjclose]

Wilktor:

Thanks for all of the input. It was really helpful

I have a question regarding your proposed method of using the frame analysis to locate the position for maximum longitudinal stresses and then use a FE model with the plane located at that location to determine the transverse distribution of the longitudinal stresses.

I would think that the width of slab resisting the longitudinal forces would change throughout the length of the slab since the slab stiffness is changing between supports. It seems that it would be possible that the resisting width at a maximum longitudinal force location could be larger than the resistng width at some location where the force is less than that of the maximum. The result could be that the force per foot width is larger at a location where the longitudinal force is less than that of the maximum.

You advise using a crash landing load case with no landing gear. ACI 343 recommends designing for the failure load of the landing gear. Have you some experience as to which is the most realistic scenario?

ACI 343 recommends a 100% impact factor for bridges located in the touch down areas of runways. This is different comment #2.

It seems that you have some great experience with this type of problem. Thanks for your help. Any further comments would be appreciated.

 

[wiktor]

tjclose:
I was unaware of the requirements of ACI 343, but these are very realistic and likely represents "hard landing". 100% impact will likely get you close or at par to 15% descent angle and no gear (i.e. crumpling body), as the contact areas for overloaded gear are are much smaller. I would still check the crash scenario, just to avoid potential liability.

The width of the slab resisting horizontal forces could be analysed as a shield resisting concentratd loading in plane. The load could be distributed at 45 degree, but you could have more accurate distribution from your FE model.

My experience is not directly related to the aircraft loading. I've designed a tunnel bridge under the derailing track, where the crash loading was a major factor, as the assumption was that the railway cars loaded with ore could pile up. Also, for some other tunnels I've used external explosion overpressure as a major loading, with no collapse as a requirement. Not surprisingly, the whole desin was controlled by this load.

 

[Dinosaur]

Wiktor,

I am surprised to hear you consider the 'Hard Landing' load combination significant. After all, you clearly stated that the 'touchdown theorum' should be abandoned. If you didn't think I was refering to the increased effects of impact loading, what did you think I was refering to? When designing highway bridges we include up to 30% impact for trucks moving horizontally. Certainly an aircraft descending would generate a greater impact load.

Dinosaur

 

[wiktor]

Dinosaur,

The aircraft descending should not produce big loads, assuming normal landing. The ACI provision is targeted at so called hard landing with 100% impact, wich means almost 1 g downward acceleration. Also, this load would be concentrated just on the landing gear, where crah loading should be rather spread on the footprint of the body of the aircraft, or at list subtancial part of it. I was simply unaware of the existance of the ACI requirements and I misunderstood your meaning of the "Touch Down" load, which is and was correct.

 

[tjclose]

Wiktor, Dinosaur:

It seems that a "typical" load case would be that of a hard landing which would include a 100% impact factor on the gear loading. The "extreme" load case would be that of a crash landing which would be the failure load of the landing gear.

Wiktor:

Regarding my comment regarding the width of slab resisting the longitudinal forces. I was referring to the resultant longitudinal slab forces (shears, moments)due to the vertical gear loads - not the longitudinal applied loads.

I still think that the width of slab resisting the gear loads would change throughout the length of the slab as the gears travel over the length of the slab since the slab stiffness is changing between supports.

Anybody:

What software are people using for a problem such as this. We are having LARSA coming in for a demo next week. Any comments?

 

[ludvik]

LARSA is alright. I have used it fairly extensively for seismic analysis of bridges and have found it to be a reasonable performer.

The single most useful feature of LARSA is its spreadsheet interface. All of the input and output has both graphical and spreadsheet interfaces which makes for much easier presentation/post processing of results.

The latest version, LARSA2000, has a bunch of cool features. The standout feature for my money is grouping - it allows you to group your components into a whole directory tree. This makes organizing a large model much easier. It is fully object oriented too which is nice.

On the negative side is are a couple of things. LARSA2000 has a number of features that are only 99% implemented, which means that these functions either don't work or only half work.

The interface is rife with bugs, and it crashes a fair amount. They are working very hard at fixing them, but it is still a pain in the neck. Bugs are probably my single biggest complaint. LARSA98 still has a few bugs which haven't been worked out. All of the bugs I have come across have been in the interface and just caused the program to crash or a file to be corrupted. I have not found any bugs in the analysis engine.

My only other comment would be that their help desk is, as the name suggests, helpful. If you have a problem they are good at sorting it out for you.

HTH