Isolation valves allowed upstream of relief valves?

[packdad]

ASME B&BVC, Section VIII, Appendix M (I think) allows full-area stop valves to be placed upstream of pressure vessel relief valves, provided that certain provisions are met, such as keeping the stop valve locked open and keeping a person at the valve at all times whenever it is shut. While these provisions make sense, I have to admit that the idea putting a stop valve in front of a relief valve still makes me uneasy. Are there any "hidden catches" to this allowance that I'm not aware of? I'm no expert on Section VIII.

Note: Our engineering group has been asked to look at possible solutions for fixing and maintaining a relief valve with a long history of repair problems. That is, the valve will occasionally lift and then fail to completely isolate when it reseats. Adding an isolation valve would allow the relief valve to be worked on without taking the associated heat exchanger out of service. Obviously, there are other issues at work here, such as why the relief valve occasionally lifts and why it does not properly reseat. However, I'm exploring all options, and adding an isolation valve is one that has been put forward.

Thanks.

 

[davefitz]

I thought there was a code case that allows the use of a 3 way valve with a duplicate relief valve, but I was not aware of allowing a stop valve per se.

 

[jproj]

Given the overall need for safety, I would highly doubt if the code allowed for an isolation valve upstream of a relief valve.... even if it did, I would be very uneasy signing off on that decision! I think the arrangement proposed by davefitz would be acceptable, however. You may want to contact the relief valve supplier. They may be able to give you some preventative maintanence tips or even help to eliminate the problem all together.

Good luck!

jproj

 

[Scipio]

Yep, you can have an isolation valve upstream of a PSV, and it's allowed by code. Actually just went through an installation about 6 months ago with an AI during a plant turnaround with PSV's that had isolation valves. Basically you just have to make sure those valves are car-sealed open (CSO'd). As long as the car seal's in place, and the valve is a full-port design, the PSV never even notices the difference. One important thing to note, those CSO's must be shown on plant P&ID's, as well as referred to in operating procedures. They don't do any good if the operators don't know why they're there.

 

[davefitz]

It is my understanding that the ASME VIII non-mandatory appendix guidelines for use of a stop valve upstream of the relief valve is only aplicable if the local inspector approves: it is counter to the mandatory rules.

 

[Scipio]

Actually UG-135(d)(1) permits stop valves provided they are "so constructed or positively controlled that the closing of the maximum number of block valves possible at one time will not reduce the pressure relieving capacity provided by the unaffected relieving devices below the required relieving capacity", it's my understanding car-sealing with a proper lockout procedure falls under the 'positive control' umbrella statement. I suppose it is entirely possible, however, that exactly what constitutes positive control could be open to debate by various authorities.

 

[packdad]

davefitz -- You're right that Appendix M is a nonmandatory appendix. I had not noticed that. I knew we had to have approval from our inspector, but wasn't sure the basis. (Like I said, I haven't spent a lot of time reading Section VIII until now!)

At our plant, we have already *sort of* done this before. That is, we have placed stop valves in series with thermal relief valves on other heat exchangers. However, in those previous cases, we took credit, procedurally, for other relief paths whenever the thermal isolation valve was shut. In this most recent instance, we don't necessarily have that option.

I'm going to talk to our inpector this afternoon. I'll find out then what his interpretations and recommendations are.

What, exactly, is a "car seal"?

 

[Scipio]

A car seal is just a means of locking a valve open or closed. Usually they're loops of wire with a lead seal (originally used on the doors of railcars, which is where I'm told the name came from) which are threaded through tabs on valves preventing the actuators from being operated without breaking the seal. They can vary from valve to valve, a lot of ball valves, for instance, can be ordered with tabs for a padlock to make them even more secure.

I have to admit, though, I re-read Appendix M after my last post, and Davefitz may have a point. I always considered car seals positive control, but appendix M refers to car seals specifically, so the non-mandatory aspect of the code may take precedence here.

 

[jte]

packdad-

I don't think we have any PSV's at the plant I work in which do not have a block valve (boilers are an exception to this). They are CSO'd and shown as such on the P&ID's. We frequently have spare PSV capacity which allows for servicing them on the run by taking them out of service one at a time. For example, we have one large vessel which needs four PSV's and has five installed.

jt

 

[pmover]

after 3.5 years at an epc (engineer-procurement-construction) firm, i can confidently state that isolation valves upstream of psv's are acceptable, provided they meet criteria previously written Scipio and are clearly marked on P&IDs as CSO. i've seen many P&IDs with iso valves upstream of psv's and designated cso. i am not disagreeing with davefitz statements either, i do believe that the operator/owner needs to clarify with insurance representative regarding such matter.
-pmover

 

[packdad]

Thanks, all, for the input. FYI, I spoke with our inspector, and his opinion is that PSV isolation valves are, in fact, acceptable as per the Code requirements. This particular person's preference is for a lock & key on the isolation valve vs. a seal. He also said a three-way valve with a duplicate relief valve was fine with or without a Code case, but the three-way valve would also need to be locked in position to prevent it from being left in mid-position. (I don't think we're going the route of a 3-way valve, but I asked anyway.)

 

[davefitz]

scipio:
as your UG 135 statement reads, closing the maximum amount of stop valves POSSIBLE could mean closing of all car sealed valves, since that is a possible configuration, but in the case of a 3 way valve, it is not possible to reduce the relief capacity by any possible valve movement.

I think that the non mandatory appendix can be used with inspector's approval. In the design state, the EPC contractor would need a letter from the insurance carrier that they approve of the arrangement prior to final dwg approval by the site prof eng. ( I also worked for an EPC for 3+ yrs, and the code compliance was not uniformly followed by all lead engineers- spotty at best. )

 

[robsalv]

Scipio pretty much outlines the situation we have in our Petrochem plant in Australia. Virtually every safety valve here has an inlet and outlet block valve, with strict CSO procedures. The CSO must be shown on P&ID's and is maintained by chain and padlock in most cases [key interlocks in others]. Removal of the CSO chain requires a "safety critical equipment defeat" approval. The removal of the SV [if required] also requires a SCED to ensure that the equipment item is still being protected.

We have a lot of overhead/procedure/auditing invested in CSO-ing but on the plus side it allows SV's to be overhauled when required - rather than in bulk at shutdowns.

By the way, the isolation valves have a host of requirements - including full porting, and in the case of gate valves, ideally installed upside down... just incase the gate drops.