MDMT Questions 5

[jgibbs22]

Hey everyone, had a question regarding MDMT. If I am understanding the Code correctly, MDMT is a value the vessel can operate at and still have a high toughness so that the material is in a ductile region when it is failed. Is that right?

That being said if I operate below MDMT, theoretically is the vessel still able to operate and pass all the other Code requirements, but now if my vessel fails I will fail in a brittle manner (not leak before break)? Is this right?

Finally, when calculating leak before break on thicker vessels with an assumed flaw, does this calculation really care about MDMT? Can you have a leak before break condition with a vessel that is operating below it's MDMT?

 

[rneill]

This seems to be a variation of the old argument that since I'm not going to overpressure, impact load, or otherwise fail my vessel, it doesn't matter if I operate below the MDMT ...

Code does not permit "pressurization" at a temperature below the MDMT for a vessel - period. That would be a flagrant violation of code requirements on it's own so I'm not sure there is much point to talking about whether you are otherwise in compliance with the code. Note: "pressurization" for a Div 1 vessel would be 35% of the MAWP.

Whether you are above or below MDMT may not affect whether or not a flaw progresses to failure, although you are probably be more likely to propagate a crack to failure in a brittle condition. However, there are many flaws that could result in a catastrophic failure in a brittle condition but which might only have resulted in leak in a ductile condition. Take a look at the linked picture of a vessel that failed in a brittle manner.

Since code is concerned with the metal wall temperature, there is some latitude to consider what the metal wall temperature is relative to transient process temperatures or ambient temperatures. Regarding process temperatures, you need to consider normal and abnormal operation and auto-refrigeration effects. Regarding ambient, you need to consider ambient in view of normal process conditions and determine a reasonable design minimum temperature; this should also include consideration of what would happen in a shutdown in cold winter conditions and how you might start the vessel up again without "pressurizing" it while below the MDMT.

 

[unclesyd]

As the lawyers would say "prima facie evidence" that you should not operate a vessel in the brittle range of a material.

 

[SJones]

By the definition of ASME VIII Div 1, MDMT is the lowest temperature at which the item can experience MAWP and this is the key concept. Also, before we bang on about 'old arguments', let's not forget that UCS-160 is entitled 'Vessels Operating At Temperatures Colder Than The MDMT Stamped On the Nameplate'. A review of API 579-1 (2009 correction) that discusses the issue in terms of critical exposure temperature (CET) and minimum allowable temperature (MAT) could help unravel the issue a little more.

'Leak before break' is the demonstration that a flaw will only grow in such a manner that, by the time it traverses the wall thickness, it is of sufficient size to be noticeable and still of insufficient size to result in a catastrophic failure. It's not a simple process and is outlined in BS 7910 annex F.

Steve Jones
Materials & Corrosion Engineer

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

 

[rneill]

Okay, you're right there are a few exceptions that UCS-160 leads you to which include the 35% "pressurization" criteria I mentioned as well as a permissible reduction which is dependent upon the "stress ratio". If the wall thickness provided is in excess of that required then you can go below the MDMT as specified in Fig UCS-66.1. For example, if you are operating below the MAWP (but above 35%) then you would be allowed some margin below the MDMT.

I stand corrected but I'm not sure that was the scenario proposed.

Note: In Canada, CSA B51 provides some additional impact testing requirements that would apply over and above the requirements (and exemptions) that might be presented in ASME VIII Div I so it is always worth reviewing those requirements as well (in Canada).



If I'm not mistaken, UCS-160 provides certain conditions under which the vessel can be operated at less than the MDMT.

Okay, you're right to the extent that those rules do allow some reduction below the MDMT provided that the operating stress ratio is less than 1.0 (e.g., you are operating below the MAWP). Below the pressurization level I mentioned (35% of MAWP) then there is a blanket exception.

 

[jamesl]

It seems to me there is a difference between MDMT stamped on vessel nameplate and MDMT that a vessel can be rated for. Many times, the rated MDMT is lower than the MDMT stamped on nameplate. Therefore we have UCS-160.

 

[rneill]

Yes, the MDMT stamped on the nameplate will be the MDMT that corresponds to the MAWP stamped on the nameplate (as indicated by SJones, whom I'm going to give a star for catching me out).

You can have other design combinations of MDMT (and MAWP) that would typically involve a lower MDMT and MAWP that the code might also permit. In my opinion, when you know there are multiple such design cases, indicate more than one MDMT / MAWP combination on the nameplate so that this is clear to the end users that more than one possible operating scenario is okay. UG-20(b) discusses this option.

The other complicating factor is that some manufactures stamp the MDMT as the minimum temperature specified in the original PO or vessel data sheet. For example, this might be 0 C when based on the materials in use, this vessel might have a real MDMT of -29 C at the specified MAWP.

 

[vesselfab]

We design, fabricate, and stamp what they ask for.

the calculations say what it is really rated for without additional charpys.

 

[SJones]

It's not only about Charpy testing, it's about knowing whether you can have a 'cold restart' amongst other things. If people moved to the CET and MAT approach it would be much clearer, more helpful and possibly cheaper if they didn't jump straight for the -46 deg C MDMT without really thinking about what MDMT actually is. I guess it's just easier all round to get an absolute minimum temperature from blowdown calculations and, hey presto, it's immediately the MDMT even though there is no pressure after a blowdown.

Steve Jones
Materials & Corrosion Engineer

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

 

[SJones]

It also brings about issues like this one:

thread292-259316

Steve Jones
Materials & Corrosion Engineer

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

 

[rneill]

Regarding blowdown, my approach is to do the blowdown calculation from normal operating pressure to 35% of the MAWP and then to set the MDMT at the temperature thus calculated. Of course, you have to be able to warm up again prior to a restart.

Perhaps this is a bit more conservative than code since code does allow some temperature suppression with any pressure drop below MAWP, but I find this is relatively simple to explain and implement and due to the non-linear nature of the temperature decrease on blowdown, it greatly increases the minimum temperature you design to relative to designing for blowdown to atmospheric pressure.

On occasion, since blowdown is a transient condition, I have also allowed the EPC firm doing the work to complete detailed metal wall temperature analysis (dynamic over time) to predict the lowest metal wall temperature that would be expected rather than the process temperature predicted.

 

[SJones]

The 'proper way'

Steve Jones
Materials & Corrosion Engineer

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

 

[PVinspector]

Rneill,

Where did that 35% of MAWP for "pressurization" come from? Just curious, I have never heard of a value for that concept.

 

[SJones]

UCS-66.1. Technically, rneill's criterion is still not fully precise because it is operating solely on stresses arising from pressurisation. Other contributing mechanical stresses also need to be assessed if one is being accurate.

Steve Jones
Materials & Corrosion Engineer

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

 

[mikecorr]

Regarding autorefrigeration, both API 579 and 581 require consideration of the fluid's atmospheric boiling point when estimating the CET, if the vessel contains liquid. In addition to normal and upset operating conditions, does one have to also consider loss of containment - say a flange leak that would expose flange bolting and possibly the vessel external surfaces to lower temperatures?

 

[SJones]

thread378-260241

Steve Jones
Materials & Corrosion Engineer

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

 

[mikecorr]

Sorry, but I am not seeing the connection between my question and the referenced thread. Let's say you have a vessel containing liquid ethylene (BP -155F). If a flange leaks long enough, the flange bolting and possibly adjacent vessel surfaces could approach that temperature. The vessels normally have MDMT no lower than say -50F. Such leaks certaintly do occur at 1200+ psi. Is this a scenario to be considered when estimating the vessel CET for an API 579 Brittle Fracture assessment? Or are we suggesting that this a not credible?

 

[metengr]

mikecorr;

Is this a scenario to be considered when estimating the vessel CET for an API 579 Brittle Fracture assessment?

No. If you review API 579 there is a broad definition of what envelop for CET - startup, shutdown, upset and standby conditions. If you have a leak and auto refrigeration occurs this does not fall under the above categories. If a leak develops an assessment is made to either operate or shutdown. However, this is a separate event, similar to a hazard analysis.

The bottom line is this, the Design Engineer has the ultimate responsibility to evaluate process scenarios AND atmospheric conditions that would result in the lowest metal temperature under the conditions I mentioned. One also has to balance the fact that you can design for -320 deg F and see what clients are going to pay for this.

 

[mikecorr]

Thanks for your comments - that is a very reasonable approach. The only other reference that makes me hesitant is API 581 (Risk Based Inspection), which states, under brittle fracture (Section 21.6.3):

Determine the minimum temperature tmin, that the component may be subjected to during operation, use the lowest of the following:

1) The minimum design temperature
2) The minimum temperature as estimated by the process engineer, including upsets
3) If the vessel or pipe is filled with a pressurized liquid, the boiling point of the liquid at atmospheric
pressure. For example liquid ammonia has a boiling point of –33°C [–28°F] and propane has a boiling point of –40°C [–40°F].

This language is much broader. I do not personally believe that a loss of containment scenario needs to be addressed in a FFS or RBI, but I'm trying to come up with a reasonable argument to defend that opinion. The best I can come up with is to have bolt-up and pressure test procedures to minimze the chance of a flange leak and monitoring (leak detectors, video cameras and operator rounds) to minimze the leak duration, should one occur. I was wondering if anyone else had attempted to make a similar case.

 

[SJones]

If a flange leaks long enough, will it not trigger gas detection systems and a shut down? The connection to the piping thread was one of using leaks to set minimum design temperatures and then catering for what causes the leak, the size of leak, the duration of the leak and so on.



Steve Jones
Materials & Corrosion Engineer

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