AASHTO LFD Design for Bridges 2

[bucknell06]

I am involved with a small, municipality-owned bridge project. The structural engineer has informed us that we will be using AASHTO Load Factor Design design (per AASHTO 2002 Standard Specifications)and that review of the design will be performed by the state transporation agency.

Our geotechnical explorations indicate the need for pile supported abutments. The piles will be end-bearing on relatively shallow bedrock. Our firm has traditionally used Allowable Stress Design for similar projects. The structural engineer has provided us with loads and load combinations to be used in the LFD design.

My understanding is that by October 2007 LRFD design of all bridges with federal funding will be required. From a geotechnical perspective (and in terms of analysis, reports, and budgets) what is the difference between ASD, LFD, and LRFD? Additionally, what are your experiences with structural engineers and adjusting to this transition in design methodology?

 

[jdonville]

bucknell06,

I recently attended training provided by the Ohio DOT addressing the AASHTO LRFD vs. ASD question. FWIW, Ohio DOT has exceptions to AASHTO that they want their consultants to follow. One of these is the design of pile foundations. Effectively, the Ohio approach will be to use DRIVEN to evaluate the unfactored driving resistance provided by the soil (i.e., no partial factors for skin friction vs. end-bearing).

This suits me just fine, but is at some odds with the AASHTO approach. No guarantees that it won't change, however.

As we haven't actually done any significant number of LRFD-based projects in Ohio (apart from some few isolated pilot projects), the structurals are also trying to feel their way forward.

Ask again in early 2008, and you'll get a better answer!

Jeff

 

[VoyageofDiscovery]

Hi bucknell06,

My understanding of LRFD, is the philosophy of designing to the capacity of the soil or founding material while applying a strength factor (less than one) due to the uncertainty of that founding material. I would recommend purchasing a copy of AASHTO and understanding the requirements in the foundations section.

HTH

VOD

 

[PEinc]

ASD uses service (unfactored) loads and safety factors for the materials.

LFD used factored (up) loads (load factors > 1) and ultimate stresses, I believe, like ultimate stress design.

LRFD used factored (up) loads with reduced strength (reduction factors < 1) for materials in order to account for variability of the materials.

It seems to me that if an engineer does not know the differences between ASD, LFD, and LRFD, then the engineer is not ready to design a bridge.

 

[bucknell06]

Thanks for your responses. Jdonville, I used DRIVEN while in school and it seemed to be a very simple, reasonable, and rational program. Seems like Ohio is taking a good approach (i.e. allowing geotechs to use a familiar program). I have searched this forum for LRFD topics and have not encountered a whole lot in the way of experiences from geotechnical engineers. It made me wonder just how familiar geotechnical engineers as a profession are with the concept.

Voyageofdiscovery, I have already taken steps to purchase the AASHTO manual. Does anyone still use the AASHTO 2002 Standard Specifications Manual (LFD approach)? According to AASHTO, this manual will no longer be updated and the 2006 version is based on LRFD. Thus, the project specific approach is based on an outdated manual...


PEinc, I'm not sure if your last comment was directed at me or not, but I will admit I am just a lowly E.I.T and am by no means designing the aforementioned bridge. I covered the drill rig during explorations and have been slightly involved in the report preparation. I do not have much experience with pile design, but I anticipate that I may be required to perform analyses in the future. I merely hoped to spark a discussion on the topic so that I could learn from the experiences of others. I have read up on, and understand the concepts of the three design methods, but am more interested in how they modify the 'deliverables' required of a geotechnical consultant. My assumption is that with more complicated and/or less familiar design approaches, more interaction with the structural engineer is needed = more analysis time, more meetings, higher budgets. Is this the case?

 

[jdonville]

bucknell06,

It would certainly seem that the initial learning curve will introduce additional costs to the design process.

Moreover, since LRFD applies geotechnical resistance factors based on 1) the method of analysis and 2) the method of determining the strength and material properties of the soils, it will probably take a little time for geotechs (in jurisdictions not currently using LRFD) to learn how to fine-tune the investigation plan to collect the appropriate type and number of samples and/or perform the appropriate type and number of tests (insitu and lab) in order to ensure a reasonable economy of the overall design.

I have heard rumors that Ohio DOT expects the initial LRFD projects to cost more than traditional ASD-based projects as the structural and geotechnical designers revamp their design processes and muddle their way through the various wrinkles. If nothing else, LRFD would seem to require that the structural and geotechnical designers spend more time talking with each other - which is a good thing - in order to iteratively refine the overall design.

A complication in Ohio is that the DOT just this week released new subsurface investigation guidance (the first major revamp in 20 years), requiring geotechs to check and double-check THAT new spec, as well.

An updated ODOT Bridge Design Manual (for LRFD) is expected shortly.

And AASHTO will be releasing the 4th edition LRFD specs any day now.

Finally, FHWA is promoting a new "best practices" training program (Subsurface Investigation Qualification) which is great, but is yet another (albeit small) new factor to be thrown into the new (to Ohio) LRFD mix.

Like I said, check back in a year.

Jeff

 

[VoyageofDiscovery]

bucknell06,

I am a Canadian engineer so I do not know the details of LFD, we use Limit State Design which is another similar method of design.

With regards to applicable code, ask the structural engineer what the State requires and then confirm it with the State Transportation Agency.

Regards

VOD

 

[Panars]

I have a number of comments.

1) To address bucknell06's original comment, LRFD is a big change for geotechnical engineers and it will take a while to get comfortable with it. While the definitions PEinc gives for ASD, LFD, and LRFD are accurate for structural design, the LFD and LRFD design process is very similar for foundations (at least in the AASHTO code). For foundations, LFD has performance factors that are multiplied by the nominal capacity of the foundation element. The same concept is used in LRFD but the terms are resistance factors and nominal resistance. Note however, that even though the process is the same, the values for the factors are different, as each has been calibrated to the relevant load factors for the code.

2) In my (albeit limited geographic) experience, LFD design of foundations is rarely done. Even if the superstructure of the bridge is designed using LFD, the foundation is usually designed by ASD with unfactored
service loads.

3) Since your situation involves piles driven to end bearing on rock, the structural capcity of the pile will probably control what load can be put on it. This calculation in LFD is different than in ASD. The capacity in LFD will appear to be much larger than in the ASD method, but it has to be compared to the larger factored loads.

4) I would disagree with jdonville that using Driven to calculate the nominal resistance of a pile is an exception to AASHTO LRFD. (Disclaimer: I work for Ohio DOT) Driven calculates nominal pile resistance (or ultimate capacity) using the same procedures that are in Section 10 of AASHTO LRFD. Driven doesn't apply the resistance factors, but the engineer can apply the resistance factor to the Driven results. The exception regarding piles that Ohio DOT is taking to AASHTO LRFD involves the resistance factor when dynamic testing is done on the piles.

5) Yes the AASHTO 17th edition (ASD and LFD) is no longer being updated. However, Ohio DOT will continue to use it for the design of rehabilitation and modifications to existing bridges. I believe many other DOTs are doing the same, as there are many difficulties in trying to combine the two approaches to the same structure.

 

[bucknell06]

Panars,

Thank you for your excellent post. As you mentioned, since the pile is end-bearing on rock, the structural capacity of the pile governs the design. The main difference that I see between ASD and LFD is that in calculating an allowable load for a given pile section (assuming steel H piles), rather than applying a safety factor determined by the engineer, a performance factor as indicated by AASHTO is applied.

So for this example, a performance factor of 0.5 is multiplied by the ultimate capactiy as indicated by AASHTO for piles end bearing on rock. This is opposed to dividing the ultimate capacity by 3 (or some other chosen S.F.) as in ASD. As Panars pointed out, the calculated allowable capacity determined via the LFD approach seems higher, but it must be compared to the up-factored load based on the load combinations provided by AASHTO. Assuming a load factor of 1.3, the net safety factor for the LFD design would be 1/0.5 * 1.3 = 2.6. This compares well with a safety factor in ASD of 2.5-3.

The disconnect that I see is between the structural engineer and the geotechnical engineer. I do not purport to be knowledgable in the area of structural engineering, but it seems as though the differences in the LFD (and possibly LRFD) method are largely on the structural engineering side. If I am reading Panars post correctly, for the geotechnical engineer everything up to determining an "allowable capacitiy" (old term) or "factored resistance" (new term) is the same. The basic difference between the two methods and ASD is that resistance factors are applied in lieu of a traditional factor of safety. The difference between LRFD and LFD is in the actual resistance factors (and the name given to them). Please feel free to correct me if my assumptions about the basics of these two methods are incorrect.

Sorry for the long post. I find this all to be very interesting. FYI, I graduated from college last May and we studied only traditional ASD design in my Geotech classes. I did have a course in LRFD steel design.

 

[jdonville]

Panars,

I only meant that DRIVEN could still be used under the ODOT LRFD procedures insofar as there are no different partial resistance factors applied to the calculated geotechnical skin and tip resistances under the ODOT method as I currently understand it. Rather, the entire geotechnical capacity as determined (for example) by DRIVEN is multiplied using a single reduction factor in order to compare with the factored load.

As I understand things, the 'strict' AASHTO LRFD spec requires the factoring of both the tip and side geotechnical resistance using reduction factors that are dependent on the method of analysis used to compute each component, which would limit the usefulness of DRIVEN for LRFD design.

Jeff

 

[Panars]

bucknell06-

You're summary of the geotechnical design process for LRFD is correct in my opinion. Not much has changed in the fundamentals until you get to the point of applying the safety factor. You say you graduated last May, would that be from Bucknell University? I am a native of the NE Penna myself (now in Ohio).

jdonville-

The AASHTO LRFD spec does not use partial resistance factors (i.e. separate resistance factors for the side and tip resistance.) If you look at Equation 10.7.3.8.6a-2 (Ugh, I hate the reference numbers in AASHTO) you will see that the same resistance factor is applied to both side and tip resistance. Also, if you look at Table 10.5.5.2.3-1 you will see that the resistance factors given are for both skin friction and end bearing. For a Driven analysis, you would apply a resistance factor of 0.35 to the ultimate capacity (nominal resistance) if the soils were clay or mixed clay and sand. You would use a resistance factor of 0.45 if the soil was all sand. But as you now know since you took the course, for Ohio DOT you would use a resistance factor of 0.7 since Ohio DOT dynamically tests piles at every project.