Hi all.

I hope this makes sense and that I haven't got myself totally confused (quite possibly)...

A vibration survey has been carried out on a line, at a number of locations in the system there are noticed vibrations/displacements. Therefore, a fatigue assessment has been requested.

Obviously, the system vibration/displacements were measured at hot conditions. Therefore, how do you model this in CAESAR II? I know that as soon as a displacement is applied in the input it constrains the specific nodes in those degrees of freedom for all loadcases regardless of whether the displacment is called into the loadcase or not.

What I'm asking is if from ambient to hot position the pipe has a moved to say a position 10mm, 10mm, 10mm (dx, dy and dz respectively) and the vibration displacement is noticed to be 5mm and 6mm (dx and dy respectively) by entering the 5mm and 6mm in the displacement box, at the specific node location, results in restraining the pipe to that position whereas in reality would it not be the following?

dx = 10mm +/- 5mm giving 15mm or 5mm
dy = 10mm +/- 6mm giving 16mm or 4mm
dz = 10mm

I believe I should consider the fatigue at the hot position as that would give largest displacement but am now doubting myself as cant seem to do it.

Unless I have to undertake multiple runs as follows:

1 - normal run no applied disps to give system movements.

2 - apply system movements as D1 loadcase and add the vibration displacements as D2 loadcase and build loadcases to add the D1 +/-D2

W+T+P+D1 - hopefully will give similar results to run 1 results
W+T+P+D1+D2
W+T+P+D1-D2

Then do relevant loadcase subtraction to do the FATigue loadcase.

Any insight/help very much appreciated as scratching my head here, its a large model so the above approach could be problematic.

Many thanks.

The displacements you're seeing are the effects of the dynamic load. You have to simulate this load to obtain the same displacements in CAESAR II - only then can you perform the fatigue analysis.

Start with "what is causing the vibration", and try to model that loading event. Make adjustments to the model/loading such that the CAESAR II displacements match the field observations.

You must to trim your input so that the results to be equal with the mode
and the displacement measured at hot conditions.
Consider that the fatique result only one part the total fatique is.
D = n1/N1 + n2/N2.....
Maybe you have additional fatique parts n2/N2 from thermal stresses
Thermal shock etc..)

Richard.

Appologies for the delay in response, another job took priority and then with the Christmas holiday have only now got back on the task.

The vibration is deemed to be caused by slug flow, following process condition changes. Therefore i have adopted the time history dynamic approach to model the conditions...

The system is quite large so the slug will impact each elbow at a set point in time from the start of the slug flowing. I have created a set of graphs for each elbow and these data points (time and corresponding load) have been added into the DLF/Spectrum Generator which creates a shock table file. Each elbow is defined as its own force set.

When I run the analysis and watch the dynamic motion, the loads are applied at the elbows all at the same time, whereas i am expecting there to be a delay between each elbow being loaded, am i expecting CAESAR II to do something it cant or have I missunderstood the approach (quite possible).

A further problem is how the fatigue assessment is carried out on the run, the technical reference guide pg 5-22 onwards has been used, this states (pg 5-25) that the stress type can be changed to EXP, OCC, FAT etcetera but this doesnt appear an option, is it only an available option for response spectra rather than time history?

Regards.

From what you describe above it sounds like you have converted your "time pulses" to DLFs. If so, I think you're running a response spectra analysis instead of time history. You need your input reviewed, please open a Support Request on eCustomer and send in your job files.

You can change the "stress type" on the "Static - Dynamic Combinations" tab of the Dynamic Input.

Cheers for the response Richard.

Realised where I went wrong now, the manual could have made it a bit clearer as there is a two stage process to getting the time-load profile to store into the memory.

With regard to the fatigue issue, there definetely isnt a drop down tab type menu on the static-dynamic combinations screen, see attached PDF file.

So after all that I still cant carry out the fatigue analysis, its back to the drawing board for me!

Cheers.

Ah, it's there for "Response Spectrum" analysis. I still think you should send your job in via eCustomer so we can review it. We may have some suggestions.

Afternoon.

I sent the files in via the e-mail address on the CII home screen Help Tab (ppmcrm@intergraph.com - is that the same as the ecustomer approach, if not how do I do that?).

Sent the files in on Wed 09/01/2013 16:59 (UK time).

Got a delivery receipt for the e-mail so it arrived but not heard anything back from coade/interserve (as yet).

Maybe there will be a response waiting for me Monday morning.

FYI - With the North Sea Oil fields becoming depleted the issue of slug loading is something we are getting asked about a lot as many of the lines are multi phase flow. The company I'm currently at are wanting to get to grips with analysing the effects dynamically as well as statically for engineering reasons.

We're really interested in getting to grips with the CAESAR II approach and getting a sound knowledge base on the assessment methodology, there seems to be a few pitfalls with getting the time time history runs going, as always not enough time/budget to really get to grips with the problem.

Fridays = early finish for me so off home soon, have a good weekend.

Cheers.