Remaining Fatigue Life 2

[bridgebuster]

We're performing a fatigue analysis of an early 50's riveted, non-composite bridge. We're following AASHTO Standard specs and the 1989 Guide Spec for Fatigue Evaluation.

From our analysis, most of the girders for the existing condition have zero fatigue life. When we analyze the bridge assuming a new composite deck our fatigue problems for the most part disappear. girders; we're finding that most of the girders have an infinite life.

Someone has raised the point that the new condition cannot have an infinite life - we can get philosophical and say that since the world will end it can't be infinite life - but rather the remaining life is somewhere between zero and infinity.

However, the way I interpret AASHTO the clock is reset after the retrofit. In the early 80's I did fatigue analyses on some bridges in New England using the client's guidelines, which predated the AASHTO guide spec. Their guideline had a method for determining remaining fatigue life after performing a retrofit. However, this method is not in AASHTO.

Has anyone run into this situation? Any thoughts on the matter?
Thanks

 

[cloa]

You might try the Bridge Engineering forum

http://www.eng-tips.com/threadminder.cfm?pid=607

 

[Ron]

BB...fatigue life is obviously related to level of stress. If your analysis concludes that you have zero fatigue life remaining, then start looking for cracks. If a rigorous inspection shows no cracking, then perhaps your analysis was conservative and you can reset the clock, but that doesn't mean you can throw away the cyclic effects that have occurred...you just change the stress level to something more tolerable to the member, thus giving it a theoretical infinite fatigue life. Even with the retrofit, I would recommend additional periodic inspection of the members that had shown a zero fatigue life by analysis.

Keep in mind that cyclic loading is cumulative across a vibration spectrum. If you have vibration, you induce stress, even at low amplitude. If you have cyclic structural loading (which obviously you have)the amplitude is high but the frequency is low. Either case presents a material performance conundrum....in the end, you resort to engineering judgment based on your best information and document such in the event of a failure in the future.

I would also characterize and categorize each member for its failure criticality and establish the inspection protocol for such. Realizing that most states have established inspection protocol and frequency of inspections, you just have to impress upon them the need to increase the frequency and supplement their protocol for the site specific conditions.

 

[bridgebuster]

Thanks for the reply Ron. I agree with your comments. We don’t want to spend money if a retrofit is unnecessary. One of the recommendations we made is to increase the frequency of inspection. However, these are built up girders and if one or more interior plates were to crack, would anyone see it? It raises another issue (can of worms): The owner is on the hook for increased monitoring of several hundred riveted bridges.

What we’re debating internally is more philosophical. We know the fatigue life isn’t zero because the bridge is still standing but on the other hand AASHTO is vague (more like silent) about whether a retrofit negates all previous fatigue cycles. So claiming infinite life may not be accurate either, even though the retrofit reduces the stress range.

The AASHTO formula for fatigue life is somewhat comical in the sense that it can give a negative fatigue life, which would mean that the bridge failed before it was built.

 

[Forensic74]

Does your firm not typically handle older bridges? I find this surprising.

Just because it's standing doesn't mean there aren't a ton of flaws and cracks you cant detect. Just because it has zero fatigue life doesnt mean the bridge is coming down.

 

[jorton]

Perhaps the structure is experiencing smaller stress ranges than that used in your analysis. Have you considered using some strain gauges to verify your analysis? Using instrumentation data could significantly lower the cumulative damage to date.

Another option is to estimate the cumulative damage to date and the additional damage expected over the next X number of years with the retrofit using the Palmgren-Miner rule. If the expected total damage is small compared to the cumulative damage you can reasonably assert that the retrofit will greatly reduce the likelihood of future fatigue cracks.

I'm not familiar with the 1989 Guide Spec, but I've found "A Fatigue Primer for Structural Engineers" by Fisher, Kulak and Smith to be very helpful.

 

[bridgebuster]

MM - we typically do bridges from very small to very long. I agree zero doesn't mean it's coming down. But the question that's being raised is: What is the remaining life (AASHTO punts on the issue)? Some one has raised an objection to claiming infinity. I'm just wondering if others have faced this conundrum.

One possibility - although impractical and probably impossible - is to create a time history of various types of trucks that may have crossed over the bridge during the past 50 years. It's been done before but on a much shorter time scale.

 

[Forensic74]

Sounds like the owner wants more assurance than can be provided on a 60 year old bridge that is constructed out of god knows what type of steel with pre fatigue details.

I would not make any sort of assessment on the remaining life of this structure. Anything you say will be misleading. I did a case study on an old riveted bridge failure over a large river from the 30's. There are four other nearly identical bridges built within the same 2 years downstream that have not failed or fractured. It's really a roll of the dice and inspection is the only way to mitigate IMHO.

 

[rb1957]

engineering and philosophy don't mix ... "infinite fatigue life" means that the cyclic stress is beyond the endurance limit. if you want to be practical about this, and you're looking for a 60 year service life, then you can say that based on your fatigue calcs the life is >60 years.

what would a 1/1000 yr earth quake do to your bridge ? presumably it's designed for a 1/100 year quake/storm ?

 

[bridgebuster]

Thanks for your thoughts everyone.

If the notion of infinite life continues to cause agita perhaps, the best remedy is to suggest re-issuing the report with a statement that the fatigue life is within the life span of the new deck. In the 80's I did some fatigue analyses based on Miner's equation but AASHTO doesn't recognize this method. If I can change my analysis to the LRFD specifications my problems disappear because rivets are a Category C rather than D; the fatigue life would be infinite from day one.

Regarding inspection, the owner's policy is to perform 100% hands-on inspection of fracture critical and fatigue prone details during biennial inspection. As I mentioned earlier, to recommend increasing the frequency of the inspection could open a can of worms - if it's done here then it has to be done everywhere applicable (as a side note, the Federal government is looking to increase the time between inspections nationwide to save money.)

Fatigue is a strange bird. Several times in the 90's I inspected a bridge - 5 span continuous steel girders that with integral steel box beam pier caps, the roadway is curved but the girders were set on chords - that was retrofitted for fatigue damage. Cracks developed at the ends of the cross frame plates, which at the time of original construction weren't connected to the girder flanges, and cracks continued to develop after the retrofit, which involved connecting the plates to the flanges. The classic repair - didn't work; go figure. We weren't hired to do any analysis but I suspect the retrofit should have been - if it wasn't - designed taking into account the curvature of the structure.

rb - regarding the earthquake. Back in the 40's no one in this area would even think about an earthquake. Several years ago we did a seismic vulnerability assessment, which indicated that the bridge could be in trouble - primarily due to it being founded on Type E soil with moderate potential for liquefaction.

However, after the August east coast earthquake no damage; but that's not to say it isn't vulnerable.