selecting proper piping design temperature for HP service per Code

[virgil1961]

I am working on design for high pressure and high temperature process. The problem is defining the design temperature for a piece of piping between a high pressure (4300 psig design pressure) and high temperature (design 1100F) reactor and downstream exchanger (design temp 750F with design pressure of 4300 psig). The normal operating temperature at the exit of the reactor is 300 to 500 F. However, should the quench water stream that enters into the bottom of the reactor fail, then the exit temperature of the reactor will be as high as 1000F. The piping spec. currently calls for an expensive high nickel alloy for this piece of piping (4300 psig @ 1000F design T and P).

However, there is limited corrosion data on this alloy material for this service, and that data available shows poor performance. Some people on the project dont' want to use extra thick or bigger piping for this service due to cost issues with the piping (length of piping is about 15 ft). Therefore there is a push on the project to change this material to C276, which has demonstrated more favorable corrosion resistance. But the problem here is that the temperature rating for C276 goes down to 750F which is lower than the worst case temperature excursion (to 1000F, some people want to ignore this case saying it's only for a minute or less and they want to measure the temperature with redundant thermocouples and replace the piping when that temperature is exceeded, other people are saying no, it can be as high as 5-10 minutes due to operator response time and one can not depend upon the TCs because they are controls and not a mechanical device). I could not find a mechanical device that could be put into the piping that would activate at 750F in event of thermal excursion to higher temperature (>750F). The C276 fails to meet code requirements, yield and stress fails, at these higher temperature excursions (>750F).Please advise on approach to setting piping design temperature, thanks!!

 

[Zapster]

What are the consequences if the line fails? Is this a life/safety issue or only a monetary inconvenience? If it is a life/safety issue, error on the conservative side.

 

[virgil1961]

What are the consequences if the line fails?
>>>>the equipment is located within a safety paneled area, we have had failures of high pressure components in the past and the safety panels are no gaurentee to hold the pressure. A vapor and liquid stream would result, which the human eye could not see. In the past the concrete wall 15 ft up in the air was indented with bullet like impressions about one quarter of inch thick. What would that do to a human? So I would say it is as much or more a safety issue, as it is a monetary issue.

 

[TD2K]

I'm not sure I agree with the statement 'you can't depend on the TCs as they are controls and not mechanical devices'.

I would agree if I had a temperature control loop that was maintaining the outlet temperature I would not want to use the same transmitter's signal to trip the system since the transmitter could be the device to fail. But I would consider having a second transmitter that would trip the system or at least, I wouldn't eliminate it out of hand. For example, temperature probes are used as trip devices in isomerization reactors to prevent runaway reactions where you need to stop the flow of hydrogen to the process.

Protective instrumentation systems are going to generate a lot of discussion how they have to be designed to ensure their reliability (do you use the plant DCS, do you install a redundant PLC like a Triconex, etc). How do you configure it, one out of one tripping the system, 2 out of 2, 2 out of 3, etc.

However, given that the length of piping in question is 15 feet and given the consequences, I'd go with the heavier wall pipe. Even at 4300 psig, you're going to spend more time in meetings than what the additional metal is going to cost. Another option, can you go with a C276 lining?

Given the consequences (1/4" indentations in a concrete wall) I can guarantee you the first time this fails it will be fixed properly. My advice is to do it right the first time. I wouldn't want to put in a system and find out later someone died as a result of a 'what-if' I ignored as the engineer because the bean counters didn't like the cost.

 

[virgil1961]

Thank you both for replies, I appreciate it. I was looking more for what the ASME code says what one has to do.

>>>I'm not sure I agree with the statement 'you can't depend on the TCs as they are controls and not mechanical devices'.

>>My feeble understanding of the ASME requirements for piping is that one can not count on control devices and have to depend on mechanical devices. The problem with TC's is that scaling can cover up their readings, and they can fail in field. I thought the code was specifically revised in the 1990s to take out temperature controls, operator error, control system failure, etc. ???

>>>>>I would agree if I had a temperature control loop that was maintaining the outlet temperature I would not want to use the same transmitter's signal to trip the system since the transmitter could be the device to fail. But I would consider having a second transmitter that would trip the system or at least
>>the design has two separate transmitters feeding the same control system, hence error lies with single control system and/or possible operator error.

>>>>>>>, I wouldn't eliminate it out of hand. For example, temperature probes are used as trip devices in isomerization reactors

>>>I thought the code (again I don't have a handle on the ASME requirements for piping) handled reactors (vessels) differently than piping requirements. I don't have a problem with the ASME stamping on the reactor, or controlling and monitoring the reactor with the method per your statement. The piping, however, is another matter and the concern here is as stated above. I think that the reactor and piping are treated separately per ASME?

>>>>>>> to prevent runaway reactions where you need to stop the flow of hydrogen to the process.

Yes I agree , thank you with doing it right the first time. But unfortunately, I still need some back up from code as the design team are getting a lot of heat that the right way is go the cheap way. I will look into c276 lining, that is a good suggestion, I don't know if it's possible or if it's even more expensive.
Thanks, keep the replies coming!! Thank you.

 

[metengr]

Have you reviewed ASME B31.3 Process Piping? I would recommend you specifically review Paragraph 302.2.4 Allowances for Pressure and Temperature Variations. This will provide the guidance you need.

 

[davefitz]

First, there are less expensive materials than high nickel alloys that can be designed for 4300psig at 1100 F. The material choice would also have to include other process imperatives, such as corrosion, and you have not yet defined the corrosion environment.

Second , the downstream piping could be rationalized as being a lower grade alloy if you included the following:

a) the lower grade alloy would need to have a high enough UTS and YS at the elevated temp ( following failure of the spray attemporator) which would prevent short term failure if the pipe overheated.

b) there would need to be an automatic shutdown logic in the control system, using redundant thermocouples and failure mode sequences of valves etc that would lead to a fail safe shutdown of the system if the downstream piping overheated

c) a second, redundant , independent means of cooling the downstream piping can be provided- it is not normal to consider the case of simultaneous failures of independent systems.

d) the piping supports and clearances would need to be adequate for the overheated case

e) process vessels downstream of the piping that are recieveing the overheated fluid also need these considerations applied.

 

[virgil1961]

Thank for for your posts.

We have reviewed the ASME B31.1
Paragraph 302.2.4. Now people are fighting about what the design temperature of the piping should be. The first part of this addresses design pressure and design temperature. Some people are reading this as saying you can select a lower design temperature and using these allowables you can have the piping good to a higher temperature rating. My suppliers are saying this a bogus way to look at the code and that they rate it on the higher temperature, not the lower temperature.

Those advoating the lower design temperature of 750F, where they get this number is not clear (they say they can pick whatever design temperature they want, regardless of short term excursions basically. Others are saying no the process tells you what the temperature is)This number just happens to be the same number of downstream equipment temperature rating which is also 750F. These people are saying that at 750F the material is good to 1100F, others are counting why not make it 1100F as design temperature?
And that the code allows for excursions. They are also advocating replacing the piping every time the temperature of 750F is exceeded!! I say this is designing for failure.....(you can tell which side of the equation I am leaning on, I am trying to be impartial here)

The people advocating 1100F as the design temperature because that is the abnormal (excursion during loss of quench water) temperature that the piping can see even for very short periods of time. So the question now is what is the design temperature? 1100F, 750F, or something else? HELP!!? Thanks in advance again.

 

[davefitz]

B31.1 is pretty clear in this regard- you may never overheat a B31.1 pipe above the max temperature for which an allowable stress is supplied in the B31.1 stress tables. This piping is external piping in the workspace, and if it fails it has a high liklihood of leading to injury.

The only possible stretch of this code is if the overheat is for a time shorter than the time needed to raise the midwall metal temp above the max temp for which an allowable stress is provided in the stress tables.

I looks like you will need to prevent overheats, or shop around and use a different design code, or buy the higher priced spread.