End of Life - Fatigue 1

[Fencer01]

Hello,

I am doing an end of life study for an offshore crane. The crane was initially designed for 25000 cycles with max load, but it hasnt gone through that in its life time. What i m wondering is if we were to do a thorough check of the welds and the rest of the structure and we find no cracks in it - can we reset the fatigue life to original 250000 cycles?

thanks
Gaurav

 

[metengr]

No. Fatigue is a time dependent failure mechanism based on loading. There is a reason for tracking use cycles on critical equipment. Having no cracks on x cycles of use means no current issues but you cannot reset the clock. The cycles continue to accumulate which means you must track your 25,000 cycle limit for sage operation.

 

[davefitz]

I would add that the fatigue life is a function of the number of cycles and the load per cycle, and can be considered independent of time, per se. If the actual , repetitive load were limited to about 25% of the original design load then ,perhaps theoretically, one could raise the fatigue life by a factor of 10, but I am not sure your insurance company would agree.

There were several major catastrophic failures of offshore drill rigs in the north atlantic in the 1960's - 1980's,involving many workers deaths . After analysis, some of the failures were attributed to fatigue failure at the welds, which led to the BS british standards methods of evaluating the fatigue life of welds . Some parts of this method are contained in the EU's PED, for example EN 12952-3 annex B tables B1,2,3. These tables to not account for corrosive effects that are expected in sea structures.

In any case, if you really need to revise the fatigue life, there is likely available a number of fatigue experts who specialize in welded sea structures, as the design of these platform rigs is now a developed art.

"In this bright future, you can't forget your past..." Bob Marley

 

[Fencer01]

Thanks guys,

I did think that we should not reset the clock, esp if the number of cycles have not been achieved.

However then my question would be "what happens after the 25000 cycle has been reached?" We do a thorough inspection and find there are no cracks or areas of concern. I guess we cannot reset anything, but continue to monitor.

cheers
Gaurav

 

[TGS4]

This is both a philosophical question and a very important engineering application question.

In general, it is considered that "failure" in the context of a fatigue analysis means that you have sufficient load/cycles to initiate a crack. However, that does not mean that the component or equipment has failed.

My recommendation for in-service equipment is to take a fracture-mechanics approach to cyclic life. Postulate a crack size, based on the minimum-detectable crack size for the NDE inspection technique that you will use. Then, using the cyclic stresses from the cyclic loading that you calculated from your fatigue analysis, perform crack growth calculations. Calculate the number of cycles it will take for the crack to reach a critical size (usually based on the crack behaviour changing from ductile to brittle, but there may be other criteria, depending on the equipment/component). Halve that number, and that now becomes your inspection interval.

Every inspection that you perform that doesn't find a crack resets the inspection interval (and hence the life). However, expect Murphy's law to bite you in the a$$, so always have a repair plan in-place for when you do find a crack!

The problem with fatigue in general, and fatigue curve specifically, is that it is quite random. A graph of test data showing stress vs cycles will look like a shotgun blast at the paper. Most published curves will either be a lower-bound curve or a mean curve with a design margin to bring it down to approximate a lower-bound curve. Nevertheless, there is inevitably test data that falls below the curve, so at best you can think of these fatigue curves as -2σ or -3σ curves, with the expectation that fatigue crack initiation is possible.

Now, for welded structures, fatigue failure is likely to occur in the welds, as welds behave differently than smooth bar steel. Unless your fatigue analysis used something like the Structural Stress Method (Verity Method), you should almost expect pre-existing microcracks in welds to exist and have grown. Depending on the NDE originally performed on the welds, it is possible that the life of the welds is an order of magnitude or more lower than what you would otherwise expect for non-welded construction.

 

[EdStainless]

All that I can add to TGS4's good comments is that we also take a statistical approach.
We do a baseline inspection before critical lifting equipment ever goes into service, and then we do a first inspection at 1/2 the estimated life. For some equipment that is 2 years, and with some it is 10 years, we never go beyond 10 years. And from there out we start reducing inspection intervals. Some critical (and old) equipment is on 6 mo inspection now.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[SJones]

@Fencer01

The subject about which you are enquiring is generally known as 'fitness for service.' On that note, you may care to study such documents as BS 7910, offering a prime example of assessment methodology.



Steve Jones
Corrosion Management Consultant

http://www.linkedin.com/pub/8/83b/b04

All answers are personal opinions only and are in no way connected with any employer.

 

[mrfailure]

Parts that ultimately fail under fatigue spend most of their lives accumulating fatigue cycles before cracking initiates. Crack growth per cycle also increases as the crack propagates. You run the risk of a crack initiating and growing to failure if you have reached design life cycles and your inspection interval is too long to ensure actually finding cracks before failure. Your design criteria assumes maximum load to ensure you do not ever risk reaching a point where cracking can begin and failure can occur.

 

[WKTaylor]

I asked this same question of Dr Lincoln, USAF aircraft Fatigue expert ~1985.

IF All defects remain below the detectable level that was considered the 'threshold' for analysis, then yes, in-theory, the clock can be reset. HOWEVER... the Statistical probability still exists for undetected damage.

The problem is that manufacturing is imperfect, operations and service environments can be unpredictable and cracks are all rogues... not to mention that corrosion and SCC can be wild cards waiting to pop-up. Miss one rogue crack, or corrosion-spot, in thousands of holes/fillets/edges/surfaces/etc and the wing still comes off the jet... most likely killing the crew. This is compounded by the numbers of aircraft built and the numbers of fatigue/strength critical locations.

Worse, yet, a study of NDI specialists found that the very best of them could consistently find cracks ~0.2--0.3" long... and the worst could consistently miss cracks exceeding 1.0". I confirmed this dismal fact in-the-field many times... I was usually better at detecting cracks [and defects like corrosion] than 95% of NDI techs... because I knew what to look for and where-to-look and how to look and had 'fresh eyes'.

NOTE.
A friend lost his nephew in a power-generation plant accident ~10 years ago. A section of steam pipe at an 'S' curve was eroded/pitted internally... undetected by external inspection and an occasional/haphazard US NDI. WHEN the side of the elbow failed, his nephew/crew-mates were walking-by and happened to be in the fragment path. His nephew took a salad-plate size elbow-fragment in the side of his chest that knocked him against a wall. He was dead in a split second. Many of his crew suffered serious hits from smaller fragments and steam injuries.

Regards, Wil Taylor

o Trust - But Verify!
o We believe to be true what we prefer to be true. [Unknown]
o For those who believe, no proof is required; for those who cannot believe, no proof is possible. [variation,Stuart Chase]
o Unfortunately, in science what You 'believe' is irrelevant. ["Orion", Homebuiltairplanes.com forum]

 

[FennLane]

I don't agree that the clock can be reset of defects are below a specified length.

This assumes that you have studied the defect tolerance of the structure and understand the fatigue crack propagation.

There is some confusion in the original post an that is whether or not the crane has been designed for 25000 or 250000 as both number are quoted.

The issue must be how has the fatigue limit been arrived at? What is the anticipated loading spectrum and what is the accumulated damage of it has only been loaded to 90% of its load.

As stated fatigue is stochastic and without much more data I think to rate the crane for Maximum Load Cycles is unhelpful.

The only way to fully understand the condition of the crane is to conduct a study examining loads and cycles. Typically, the average lifted load of a typical offshore crane does not stray significantly from 3.2Te with a variance of 0.75Te.

Although the weight of the load does play a part in the analysis, it is more pertinent to study the stress range. Calculating the difference between the existing residual strain on the structure, the added stress during an actual lift and the spare lifting capacity provides an accurate picture of the fatigue experienced.

Because of the relatively low average load, the number of lift cycles has a greater effect on the fatigue life of structure than the lifted load itself. This also means there are many large capacity cranes operating at a fraction of their design limits. Lower stress ranges provide longer fatigue lives and occasional lifts at maximum capacity will not significantly affect their fatigue life.

Cranes that predominately operate towards the outer regions of their radius range will result in greater stress ranges and therefore reduce their fatigue life compared to those operating at closer ranges.

All of this can be plotted on an S/N curve with fatigue strength and the number of cycles along each axis with the Miner’s Rule then used to calculate the damage through variable stress ranges.

A critical part of analysing fatigue is determining an accurate life cycle. In the majority of cases gaining reliable information to determine the crane’s life cycle can be a challenge. Many of today’s cranes have a recording load indicator system with a full history of each lift, detailing the weight and the position the crane was in when this took place. This is ideal for fatigue analysis but without this information it can be difficult.

It is still possible to establish realistic life cycles for any particular offshore platform crane even without a recording load indicator system. Knowing the type of work, when drilling programmes have taken place and even the deck layout makes it possible to establish the likely lifting frequency, average weights involved and where loads were being moved to and from.

It can take some effort and some skill but it may be possible to carry out some analysis and decide when replacement is required.

 

[tbuelna]

The crane was initially designed for 25000 cycles with max load, but it hasnt gone through that in its life time. What i m wondering is if we were to do a thorough check of the welds and the rest of the structure and we find no cracks in it - can we reset the fatigue life to original 250000 cycles?

Is this a typo? For ten times the number of load cycles (250,000 vs 25,000) the applied load would need to be significantly reduced.