7 replies to this topic

[Luigi]

This is the case I am starting to look at and would like your always interesting comments and advice:

Steam at 150psi if sent to a pressure vessel with a design pressure of 70 psi, through a control valve. There is no safety valve on the vessel, so if the control valve fails, there is a potential overpressure. Trying to understand the protection criteria used in the design, I thought that it might have been considered the vessel as part of "system" under the concept of ASME that allows a relief valve to protect interconnected equipment.

There are two potential paths for relief, one doesn't have any block valve, but goes through a condenser and into a third vessel, with a safety valve set at 70 psi.

The other goes through a pump and reaches a column with a safety valve set at 70 psi. In this case, we have more than one block valve (normally open).

For the first path, even without a valve, I believe that there are other conditions to be met. For instance, we have to assure that under relieving conditions, the pressure on the vessel been analyzed won't be above (70 psi +10%).

For the second case, I think that relieving through a pump would be totally unpredictable, thus, unacceptable.

I would like your opinion over the applicability of the "system" concept for these two possible paths.

One additional question that derives from the first path is, assuming that the flow rate and pressures (and any other additional requirement) is complied and we accept it as an adequate relief path, would the presence of a normally open manual valve (not car-sealed) void it? That is what I would understand from the code but, wouldn't it be another case of double jeopardy?

Thank you for your input.

[latexman]

Environmentally, if a steam regulator fails wide open, do you want to blow just water to the atmosphere or water AND chemicals ?

[Luigi]

Latexman:

I am not sure I understand your point clearly, so I will add some more information I believe is related to your comment.

I am reviewing an existent design and I need to understand the criteria used for protecting the vessel from overpressure. The system has no hazardous products, but even if there was, the assumption (for the purpose of answering the qustions) should be that the existing safety valves are properly routed to a safe place.

Could you please elaborate a little more based on this?

[latexman]

In reality, the safety and environmental impact of the discharge of a PSV to atmosphere is tied so intricately with the pure technical solution that they cannot be separated. I know from past experience that a lot of engineers out there, act like they can, and want others to follow, especially their reviewer. In this experienced relief designer and reviewer’s opinion, the *right* solution cannot be reached unless they are considered together.

Therefore, I regretfully cannot elaborate further on my points if "the assumption (for the purpose of answering the qustions) should be that the existing safety valves are properly routed to a safe place".

[AA Mishra]

How come you don't understand the criteria to protect the vessel from overpressure

You check operating pressure, name plate capacity of safety valve and related alarms provided.

In case, control valve failure is bothering you so much, you change control valve from fail to open to fail to close.

Regards

[Dacs]

From the standpoint of a PSV protecting interconnecting vessels, it has been done with columns/reboilers/reflux drum system and absorber/stripper columns in series.

Having a common PSV to protect the system is one thing, having the PSV *sized* for the worst possible scenario for *all* equipment concerned is another.

I've seen a stripper system where a PSV is located in the column and another PSV is located in the reflux drum. While it goes against what I said above but knowing later that they've been sized for different scenarios, then it made sense.

Since you're reviewing what I would assume an existing system, you should take what I said above into consideration. BTW, the destination of the relief would be another issue and I also assume that this has been dealt with in the system appropriately.

BTW, relieving through a pump with more than one block valves is suicidal in my opinion.

[Qalander (Chem)]

From the standpoint of a PSV protecting interconnecting vessels, it has been done with columns/reboilers/reflux drum system and absorber/stripper columns in series.

Having a common PSV to protect the system is one thing, having the PSV *sized* for the worst possible scenario for *all* equipment concerned is another.

I've seen a stripper system where a PSV is located in the column and another PSV is located in the reflux drum. While it goes against what I said above but knowing later that they've been sized for different scenarios, then it made sense.

Since you're reviewing what I would assume an existing system, you should take what I said above into consideration. BTW, the destination of the relief would be another issue and I also assume that this has been dealt with in the system appropriately.

BTW, relieving through a pump with more than one block valves is suicidal in my opinion.

Dacs& Luigi(The OP) Hello/Good afternoon,

Although this seems to be a fairly old almost (DORMANT) thread being 2 years 11 months old;but interesting to comment.

Basically I am of the opinion that from legacy of existing system such scenarios are witnessed by many of us including me.

Undoubtedly have higher degrees of potential risk possibilities;wherever possible we must try our best to improve upon either through

• Thoroughly reviewing whole scenario in detail and possible haz-ops conducted having a team of operating personnel and design people;thus suggesting and implementing due safety in-place.
• Other option can be and should be to ensure that a fool proof car-seal or chain-lock open or shut condition is ensured for the most credible scenarios to mitigate(minimize) the risks in-place with slightest modification on exiting valves/fittings etc.

Hope this shows some way forward, although most presumably the issue might have already been attended. If OP can provide feedback!

[Dacs]

Oh God! I didn't notice that this was an *old* thread