Seismic case for Integral Bridges. 1

[Roycronos]

Hello:

I'm looking for a method for the Seismic Design in Integral Bridges, principally I need to calculate the Earth Pressures (Passive and Active I think).

The problem that I have is about a iteration, because this case is not completly Passiva either Active, so I don't know how get the real coefficient and earth pressure value in Abutments.

Thank You for the Help.

Roy.

PD: Sorry my English, this is not my main language (Spanish is)

 

[bridgebuster]

I attached a paper with some info related to seismic. I have another collection of papers that discusses seismic. I'm trying to upload but it's giving me trouble probably because of the size.

I'l try again. Go to the CalTrans website you should find something there.

 

[Qshake]

Nice reference Bridgebuster!

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[bridgebuster]

Thanks Qshake

 

[Roycronos]

Thanks bridgebuster...
Really it's a great book for Integral Bridges, but i'm still with my problem, because I don't know how to apply the Earth Pressure in the Seismic Case on Abutments. For the Static Case, I used Design Curves Recommended by NCHRP 343 (1991), (Wall movement/Wall Height vs K). This give Coefficients for Earth Pressures in Active or Passive Movement, but it's valid only for Static Case (Using Iterations).

In the Seismic Case, I don't know if there is some similar method valid.

Also, I have another problem with Earth Pressure: I was working in Sap2000 Software, and I did Earth Pressure with "Springs", but now I need to do it with Area Loads (Equivalent), it's because I'm not applying the correct coefficients of AASHTO LRFD for Combinations in the Design if I use Springs, so I need Area Loads for Apply Coefficients for Combinations.

I hope a answer and thanks really!

Roy.

 

[Roycronos]

Another Question: Do You this kind of Bridges's Design with Iterations too??? (For the Earth Pressure Case (Active/Passive))

 

[Qshake]

If your bridge is short, say less than 200' in length, you can use linear springs at the endbent to represent the comression or passive pressure case. Naturally with such a high compressive spring you'll be able to move load out of the structure and into the fill provided the fill is not overloaded (yielded).

You will also have the stiffness of the pile (estimated at 40 kips per inch) and can be used for tension and compression.

The regular procedure is to do one of two things:
1. Apply on one full value for the compressive spring and pile at one endbent and the other is free to move. This represents the case where that endbent is fully engaging the passive pressure. It is also only one iteration. If the bridge is symmetric then you can be done.
2. Apply one half the full value of the compressive spring with full pile stiffness at both endbents. When you check your forces you have to multiply by 2 to make up for the half spring.

In both cases 1 and 2 above you need to check the spring to make sure you're not overloading. if so you need to artificially reduce the spring and re-run the analysis. What you use as the limit of your spring may be the backwall strength, or soil strength or some other weak link in the system.

This is for integral abutments only as for open seat abutments you have the matter of the expansion gap to close before you get any real benefit of the passive pressure.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[Roycronos]

My bridge have less than 200' in lenght (it's almost 100')

I have two questions to your last answer Qshake:

1) If I do these case, 1 or 2, so how I get the value of Coefficients of Springs for the Seismic Case??, because for the Static case, I apply the Tº(+ or -), Dead load and Live load, so 1 Abutment get Passive earth pressure, but the another abutment get passive earth pressure or active earth pressure???, and the value for this case (Static), are obteined from curves of Report NCHRP 343; But Seismic case, I don't know.

2) After to apply these loads and coefficients of Springs, I must to verify if all springs are in compression (for passive case), and how I'm applying the Coefficients of Combinations of AASHTO LRFD in the Earth Pressure???, because if I'm doing it with Springs, the elements like beams need to be majority for the design, and I'm not doing it in Springs with Earth Pressure.

 

[Roycronos]

PD: I have done the complete Bridge on Sap2000, so I have both Abutments and the Superstructure there...
It's a bridge of 2 lanes, so I put 2 Trucks...
But if I apply Tº and/or Seismic loads, so 1 Abutment must to get Passive Earth Pressure, but the another ?? (Active Earth Pressure i think), but both don't get the complete Pressure, it's only a lot.

For this, I use the Iterations, for get a value more real, with Curves from Report NCHRP 343 (for Static Case only this curve), I don't have something for Seismic Case.

If I do Complete Passive Earth Pressure in the Bridge, It's Oversized, so I can't do it.

 

[Roycronos]

Here is the Curve of the Report NCHRP 343.

 

[Roycronos]

First I apply only Dead Load, Live Load (truck) and Temperature, so I get a Movement of the Wall...
With that, I can get from the Curve a Coefficient for the Earth Pressure (I choose Medium Dense Sand for it), and like Temperature +, so I get Passive Earth Pressure Case.

So later I apply all (with Earth Pressure), and I get a new Movement in the Abutment, so I get a new Coefficient and I calculate relative Error, if that error is less than 1%, so it's finished, else, i need more iterations.

It's some easy if I do it With Springs, But I'm not applying the Coefficients for Majority loads of AASHTO LRFD, so I need now do it with Area Loads (Pressure) in Sap2000, because there i can apply that coefficients, but it's some more difficult, because I need first to calculate a Equivalente Earth Pressure (Area Load).

 

[KDMengineer]

Qshake,
We do something similar in Tennessee. We assume an initial linear stiffness spring for the soil behind the endwall, first assuming that the piles are effective at both abutments, and the soil spring only being effective at one abutment (if there is no expansion joint at either abutment). We have come to the concensus that 40 kips per inch might be generous for the pile stiffness, however. We have started scaling that value back some, depending on pile type. Then we run Seisab and use the resultant abutment forces to calculate what the equivalent soil pressure behind the abutment endwall ends up as. If the resultant pressure exceeds our arbitrary limiting value, then we know we've assumed too large value for the soil stiffness spring to begin with. This is where the few consultants we farm out our bridge designs to screw it up (Tennessee is one of the few states that still designs most of our bridges "In House" at TDOT). Usually, the consultant thinks the way to address this is to add more piles to reduce the pressure on the soil behind the endwall. But what is really happening is that your stiffness model was not realistic to begin with. You see, it's not as if the soil behind the endwall will crumble and fail catastrophically, but rather once the soil approaches yielding it will throw more of the resultant seismic forces into the interior substructures. That's why we iterate until we come to a soil spring value that will result in a soil pressure behind the endwall that comes close to our limiting value. If you do this, you will notice with each iteration that as the abutment soil spring value (and the resultant pressure behind the endwall)decreases, the resultant forces on the interior substructure increases. That would be the most realistic result.
Hope this helps.

 

[crossframe]

That's pretty much the way a certain large West Coast DOT does it (or at least used to do it). When somebody asked how to reduce the soil stiffness behind the abutments, my cohorts wouldn't let me tell him to drill holes in the soil under the approach slab - the size and number determined by how much he needed to reduce the stiffness.

Those were the days...

 

[gwynn]

KDM

We use a similar approach as well. The main difference being that we run non-linear springs with a defined maximum reaction, which allows for fewer iterations.

 

[Roycronos]

KMD, after You finish an iteration with Springs, You use "Seisab" in order to find Soil pressure (equivalent) in the abutments right ?, but if I don't have "Seisab", do You have any other method for this?

gwynn, do You use "Sap2000" for run "non-linear springs" ?

 

[KDMengineer]

Roycronos,
We use Seisab because it's the most efficient (and available) way of generating a seismic output (multiple licences and we've been using it since it's been in a DOS format). We also have SAP2000 (now re-branded as CSIBRIDGE) but it is much more time consuming than Seisab (because the learning curve for CSIBridge is pretty intense). Couple this with only one floating license for CSIBridge amongst a floor of bridge designers and it's easy to see that for our purposes CSIBridge is not an attractive option for everyday seismic bridge analysis. However, if SAP2000 is your only tool, it can be done. By the way, we don't do non-linear springs at present, so I can't give any guidance on that score.
gwynn, I'd much appreciate some background knowledge on the use of non-linear springs.

 

[KDMengineer]

By the way, Crossframe:
You naughty, naughty bridge designer! I've fantisized about just doing something like that.

 

[Roycronos]

KDM,

what do you think about this method?

First I do an abutment with springs, and the another abutment, with Earth Pressure (like a load, with at-rest pressure value for the first iteration), after I need check the displacements in order to verify if it's Active, Pasive or At-rest condition and make the next iteration, if it's necessary, (with the correct earth pressure as appropiate), also I must to change the stiffness of the springs and earth pressure value in each iteration.

and, again i have the problem with the seismic case, because i'm not sure which is the condition and coefficient for the earth pressure...

 

[gwynn]

KDM,
The basic premise with non-linear springs is exactly what you outlined earlier in the thread. The soil doesn't fail catastrophically once a certain pressure is exceeded, it just gets pushed around a bit. Modifying the spring stiffness of linear springs is essentially an artificial way of accounting for the yield plateau. For things like this we use springs that are linear in the elastic range and perfectly plastic once yield is reached. The only real advantage in this instance is that it takes fewer runs.

Roy,
We do not use Sap2000, but if my memory serves me correct it was possible to either use non-linear springs or fake them through joint modelling with it.