Low temperature assessment - Looking for advice - B 31.3 Stress ratio

[nick262b]

Hi everyone, thanks for taking the time to read this thread.

Background:

Production manifold produced Circa 1990 by hot isostatic press method and welding sections together.
Certified to B 31.3 as opposed to a pressure vessel code.
Material S31803.
Op temp -36deg and charpy testing would have been carried out at -36Deg. (the required pass may have been significantly higher than the code values at 70J avg but this is not certain atm)
My aim is to justify safe operation if one of the nozzles is chilled to -56deg.

Work done:

I have performed a pipe stress analysis on the piece using CAESER II in order to justify a reduction in allowable operating temperature. This will be done using the guidance given on the stress ratio in 323.2.2. I have calculated the code stresses in a sustained case with displacement, pressure and weight loadings as per the guidance (I also have thermal and hanger loads) and my model shows the code stress is well below the allowable stress for the material and would, according to the code, allow a large reduction in operating temp. I have worked to this code as this is what it was certified too.

Questions to the forum:

I have a few concerns with this method however and was wondering if anyone could shed any light as to the origins of this technique and whether it is really applicable for this case.

1. SIF's are not included, what allowances have been made for this? I would think SIF's and max stress intensities/local stresses are important in a brittle fracture scenario especially those at the base of the nozzle. Is this accounted for by using allowable stresses and safety factors in the code?

2. The stress ratio compares to a value of S from the tables in A1. If I calculate the value of Sa however allowing for fatigue cycles for the piece this is significantly less than S and would impact the amount of temperature reduction allowable if Sa was used. I could understand using a value of S if you were operating within the materials endurance limit but ignoring fatigue damage or cracks that may have grown over time seems risky?

3. The material in question is Duplex S31803 and although interpretation 23-01 states it is acceptable to be used with this approach I was wondering if I should have any concerns/anyone have any other opinions.

4. Would I be better of trying to justify this using a pressure vessel code such as ASMEVIII or PD5500? I was thinking of doing this in any case. I also would like to calculate the local stresses at the nozzle to compare to stresses required for crack tip opening etc. Does anyone know where I can find any methods e.g. Roarks.

Thanks for your time reading this and I would really appreciate any general guidance or tips people can offer.

Kind Regards,

Nick

 

[weldstan]

Do you have the actual yield strength of the material? Why would you expect that impact testing was done at -36 ( C or F) when the operating temp is -36? Generally the designer would specify a lower temp, especially when fabricated by HIP?

 

[nick262b]

Hi Weldstan thanks for the reply.

Sorry -36deg is the design temp thats a typo I'll change this additionally:

Unfortunatley we no longer have acess to the fabrication records or materials certs and so I cannot be sure of the exact material properties. I contacted the supplier of the manifold and they could only say that the manifold would of been tested to project specifications (which typically are also no longer avaliable!). Looking at other project specs that are avaliable these seem to specify testing at the design temp of the lines in most cases and not below.

e.g Some linepipe welding certified to 4515 (1984 revision) had a design temperature of -20degc and was tested at -20degc as per the code requirement. This I believe will be the case in all instances I believe. I am not sure if the BS 31.3 at the time specified additional extra degc on top of test temp dependent on thickness for example like the extra 10dec or 20degc in BS4515 but this could help us as the manifold is rather thick.

In short yes the design temp is -36degc, op temp is originally likely to have been much higher with the -36 as a nominal minimum as most of the wells were expected to flow hot. Now that we have some new cold gas wells the JT cooling across the choke valves is generating much colder flows than originally anticipated.

Cheers Nick

 

[weldstan]

I qualified or directed qualification of a number of WPSs on this dupelex SS at temps of -60 F and never had a problem with weld metal, HAZ and base metal as long as 34 years ago and none of it was HIP'd. Do your calcs and determine as % of SMYS. If the calcs show very low stress at design conditions, it will be safe at the new operating temp and it would appear that this temp is temporay, possibly due to Joules-Thompson cooling under specific conditions.

 

[nick262b]

Thanks for the advice Weldstan, its good to hear you have had some experience with this material at LT. You are correct in assuming this would be a temporary deviation from the norm although could go on for extended periods of time depending on how the well is flowing.

I beleive the method you have proposed is similar to that in the code.

I am not all that familiar with HIP'd materials but it sounds like you have some experience, just as general point have you found their properties to be generally superior to cast or forged products?

Cheers Nick

 

[DSB123]

nick262b,
You mention the design Code is BS 31.3. Are you sure it is not ASME B31.3?? Since the system has been in since circa 1990 (24 years) which is longer than the probable Design life of 20 years at the time I would suggest that you look at fatigue life usage from the operating history of the system. How many pressure and thermal cycles has the system seen and how many greater thermal cycles are you going to subject it to?. Also you mention that you have performed a Caesar analysis and determined the code stresses in a sustained case with displacement. This is not what B31.3 requires (if the Code is indeed B31.3) sustained stresses and displacement stresses are not treat together.
Are the nozzles composed of weldolets? I have seen one of these pull right out of the manifold when subjected to chilling.

SIF's should be included for the branches not forgetting the torsional - axial and pressure SIF's. Do not just use the In-plane and out of plane bending SIF's.

 

[nick262b]

Hi DSB123 thank you for your reply,

Yes another typo apologies its ASME B31.3 I’m just used to typing BS a lot of the time.

Your approach to fatigue is part of what I was asking about. In my assessment by displacement stress I mean that I have included wellhead movements at the end of the modelled flowline. Perhaps I have used the wrong analysis case I don’t think I have but may well be mistaken.

Section 302.3.5 which defines sustained loads and allowable’s states that sustained loads SL are ‘sum of longitudinal stresses due to sustained loads such as pressure and dead weight’ and that this should not exceed the allowable Sh from table A1. It then goes on to define displacement stress separately and states to use separate allowable’s SE which is based on a reduction in Sh depending on the number of cycles and loads etc.

The section I am using is section 323.2.2B which is a little different to the other sections and this states that you must calculate the ‘combined longitudinal stress due to pressure, dead weight and displacement strain (stress intensification factors are not included in this calculation)‘ and compare this to S.

I guess my question is should S be Sh directly from table A1 or SE?

Essentially in my model I am trying to calculate the maximum possible code stress at the node and compare this to a suitable allowable stress to use the figure 323.2.2B.

It is worrying to hear that you have seen nozzles ripped from lines but thankfully in this case the pieces are formed as one in the HIP technique and the only welds are between the shell segments. The 4” nozzles have a substantial thickness at the base, 26mm with the shell thickness been 30mm.

Cheers Nick

 

[weldstan]

Since the equipment has been operating for a long while, do check the branch to pipe, high stress locations with phased array UT to assure freedom from any cracking. Also check for corrosion pitting, wall thinning.

 

[DSB123]

The value os S should be the value of S from Table A-1 based on the minimum temperature of operation.

I still beleive you should be performing a fatigue assessment based on the operating history.

 

[nick262b]

Hi guys,

Weldstan yes I believe these sort of checks will be done as part of the planned inspection regime in anycase.

DSB123 yes I believe you are right and this is currently what I am doing using S. Using S must, I think, assume that there is no damage pre existing.

The verification is I believe essentially saying despite the temperature is lower and the material is more brittle it will be ok as the stresses are sufficiently low that cracking is not a worry (based on CTOD data or trials?). I also assume that the code must use sufficient safety factors in this ratio approach such that SIF's can be ignored.

I could also conduct a fatigue asessment based on the history of the piece as you mention and I am sure one will have been conducted at the design stage to check that the operating stress's give a suitable fatigue life. I do not think that in this case the temperature reduction has any effect on the already calculated fatigue case as it is based on displacemwent stress and in anycase the change in thermal stress's is minimal. The only cause I can see to redo the fatigue asessment would be if there is a concern that the material fatigues quicker at a lower temperature which I think I was trying to get at at the start with point 2. i.e. are there any changes to fatigue properties to consider when doing a LT asessment such as this.

All in all I think I am covered by using the method I have and will also do some local stress checks on the nozzle in anycase to see if there are any reasons to be concerned.

Thank you both for your time and advice.

Kind Regards,

Nick