4 replies to this topic

[naghizad]

Hi all,

I would appreciate some input regarding best practice for setting MDMT in gas systems.

In a dry high-pressure nitrogen storage system (around 60 barg), gas is let down through a control valve into a downstream header lower pressure. One scenario considered is a control valve fail-open case (not a formal emergency blowdown).

Should a control valve fail-open case normally be treated as a governing basis for MDMT selection of downstream piping and vessels?

Or is it more common practice to:

• Treat low temperatures as local to the restriction/PSV tie-in region

Assume:

• Dry gas (no liquid flashing)

• No cryogenic service

• Downstream overpressure protected by PSV

• Normal restart controls in place.

I’m interested in typical industry practice rather than specific simulation results.

Thanks in advance.

[Pilesar]

My judgment for this has not always matched my fellow engineers. I generally defer to them if they have more primary responsibility for the answer after giving some argument. With that warning, my personal opinions:

1) MDMT is based on metal temperature, not process temperature. How cold would the metal pipe be using average temperature from inside of the wall to the outside for the section in question? Ambient air would warm the metal. The metal pipe has mass which will want to maintain the same temperature. Gas heat transfer to metal is not very high rate. How long would extra-cold process fluid flow in this section before operator made necessary adjustment?

2) With failed control valve, would the temperature of the gas after the valve decrease or increase? Would not the downstream gas be warmer with less pressure drop? Is the piping rated for the full pressure? Overpressure would be a greater concern than low temperature here.

3) If there were boiling liquid at the bottom of the pipe cross-section, then the hazard would be greater considering uneven cross-section temperatures.

4) MDMT is for metal temperature at a specific pressure. Lower pressure technically allows for colder metal temperature than MDMT. See the ASME Code for Pressure Piping, B31.3. Note particularly Fig. 323.2.2B "Reduction in Minimum Design Metal Temperature Without Impact Testing". 

5) I think perhaps the concern is "the metal is normally cold, then the pressure increases which causes stress the piping is not designed for." If so, then assess the hazard. Reduce the maximum pressure if needed. Change piping material if needed.

6) The major suppliers of commercial nitrogen have standard designs for their storage and distribution systems. At some point, there will likely be a transition of material from stainless to carbon steel. When I did hazard assessment for plant distribution nitrogen systems using sch 40 carbon steel pipe, we evaluated each piping section spec for suitability to expected conditions. I do not remember specifically considering control valve failure, but that would have been one of the questions we answered. 

[naghizad]

Thanks. This is very helpful. Agreed that MDMT should be based on credible metal temperature at coincident pressure not min temperature alone.

In our case very low temperature are predicted locally at the valve outlet during early transient. Further downstream (long large pipeline toward distribution network) temperature recover due to heat pickup and lower velocity( demand driven) based on simulations. Overpressure protection is with PSVs; downstream piping ((150#) is not fully rated.

We are therefore treating MDMT issues local to restriction and immediate downstream. This means rating control valves (needs cryogenic service where required) and downstream piping to stainless steel including PSV tie-in. On the other hand we intend to manage vessels shell/restart risk via nozzle protection and procedurally (no re-presurisation while cold) to avoid upgrading full vessels to stainless steel.
 

Appreciate the reminder about B31.3 MDMT versus pressure basis.

[Pilesar]

We are therefore treating MDMT issues local to restriction and immediate downstream. This means rating control valves (needs cryogenic service where required) and downstream piping to stainless steel including PSV tie-in. On the other hand we intend to manage vessels shell/restart risk via nozzle protection and procedurally (no re-presurisation while cold) to avoid upgrading full vessels to stainless steel.

In flare systems, I have seen stainless used for the first portion (20 to 100 ft maybe) of piping after PSVs based on expected temperature recovery.
Cryogenic valves, as I understand them, are long stemmed to remove the mechanism from super-cold conditions to avoid valve freezing. These are usually not required unless you are dealing with truly cryogenic liquid conditions.
The temperature versus pressure curve for piping has a similar curve for pressure vessels for specific steel compositions. Tracing the path for operating temperature and pressure alongside the metal design curve will show where there may be a problem. Sometimes there is no reliable way to avoid the hazardous zones and the vessel should be replaced. I agree with procedures for no cold repressurization if that keeps the vessel from passing through hazardous zones. Procedures must be enforced strictly where it is applicable. Operators naturally want the plant up and running again as soon as possible. 
There has been much increase in knowledge about steel behavior in cold service in the past 50 years. If you have an older plant, the original design should be checked where cold could be an issue. The rule of thumb "carbon steel is good to -20F" is not always true. I was surprised to see some steel pressure vessels in service made from material that would not even qualify today for temperatures near the freezing point of water. Grandfathering pressure vessels due to proven service history is a risk that needs thorough deliberation and should not be assumed the only choice.

[naghizad]

Appreciate your inputs. Noted as well on cryogenic valves  since this is dry gas JT cooling, we will review requirement for long stem/bonnet with the vendor Thanks.

We are therefore treating MDMT issues local to restriction and immediate downstream. This means rating control valves (needs cryogenic service where required) and downstream piping to stainless steel including PSV tie-in. On the other hand we intend to manage vessels shell/restart risk via nozzle protection and procedurally (no re-presurisation while cold) to avoid upgrading full vessels to stainless steel.

In flare systems, I have seen stainless used for the first portion (20 to 100 ft maybe) of piping after PSVs based on expected temperature recovery.
Cryogenic valves, as I understand them, are long stemmed to remove the mechanism from super-cold conditions to avoid valve freezing. These are usually not required unless you are dealing with truly cryogenic liquid conditions.
The temperature versus pressure curve for piping has a similar curve for pressure vessels for specific steel compositions. Tracing the path for operating temperature and pressure alongside the metal design curve will show where there may be a problem. Sometimes there is no reliable way to avoid the hazardous zones and the vessel should be replaced. I agree with procedures for no cold repressurization if that keeps the vessel from passing through hazardous zones. Procedures must be enforced strictly where it is applicable. Operators naturally want the plant up and running again as soon as possible. 
There has been much increase in knowledge about steel behavior in cold service in the past 50 years. If you have an older plant, the original design should be checked where cold could be an issue. The rule of thumb "carbon steel is good to -20F" is not always true. I was surprised to see some steel pressure vessels in service made from material that would not even qualify today for temperatures near the freezing point of water. Grandfathering pressure vessels due to proven service history is a risk that needs thorough deliberation and should not be assumed the only choice.