HIPPS vs PSV

[maddocks]

Can anyone provide reliability data for standard spring operated PSV's? The SIS or HIPPS system is meant to provide a near identical level of reliability to a PSV, however, published data on probability of failure on demand for PSV's seems to be scarce. Our client refuses to provide an acceptable level of risk (understandably). So, I'm leaning to a SIL-3 but the methods seem a little vague.

 

[Latexman]

No reliability data, but we use SIL-2 for a single RD or PSV and SIL-1 for a RD/PSV combination.

Good luck,
Latexman

 

[CHD01]

See the following article and Table 2B at the following website. The author gives his own SIL recommendation and the table 2B gives failure data on sping and pilot valves. I did not notice any data on rupture discs - but I did scan it rather quickly. I intend to read it more carfully later.

http://people.clarkson.edu/~wilcox/Design/hipps.pdf

I don't agree with everything said in the article and there may be some minor errors in it but the article is interesting and complete. To me HIPPS is somewhat like putting the wagon before the horse. I mean you go through relief sizing methods to identify relief concerns before the money is spent to develope instrumentation or other means to mitagate the required rates initially calculated.

Also Case 2211's primary use that I have seen is to avoid a relief valve for installation on an ASME vessel; this is why the fabricator is involved.

Of the hundreds of relief valves I have done there are only maybe 3 where I had to deal with a system that could not be relieved normally; and only about 10 or so where we applied Case 2211. We still first try to relieve a system without reliance upon instrumentation, etc. to reduce rates. Then the economics fo a release, etc. dictate what else is needed. If the industry preferes to identify HIPPS as a stand-a-lone method thats OK I guess. What is important I agree is to improve upon the probability analysis of processing systems being protected and our methods of protection. I just see this work (HIPPS) as filling a void but not replacing pre 1996 relief sizing methods.


The more you learn, the less you are certain of.

 

[maddocks]

I have the same data presented in another sistech paper, but am looking for independent verification of the values.

I would agree that normally we try and protect the system with PSV's properly, however, in this case, we are bringing in a high pressure new gas source (1440 Psig) into an existing plant with a lower design pressure of 1045 Psig. The pressure control valve fail-open scenario comes up with a huge relieving volume, so much so that the existing PSV's are inadequate and the existing plant flare system would be unable to handle this high flow. This is what drove me into investigating a HIPPS or a SIL-3 system to ensure vessel protection.

 

[CHD01]

Understand. Some of the control valves have failed open rates that are absurd - basing flow sometimes on the internals being gone. Could you do something as simple as installing a restrictive orifice for the control valve? I have often done this; of course then the orifice becomes subject to operations control procedures.

The more you learn, the less you are certain of.

 

[maddocks]

I considered that, but under normal flowing conditions, I can't tolerate a large pressure drop in the valve. It will only be used during start-up, pipeline drawdown, and restricted flowing situations where we can't move the full volume of gas. This situation is really starting to bug me...

 

[CHD01]

Pilot valves have a lot more relief capacity than a conventional valve with the same inlet flange connection - you might want to investigate whether that would help; plus you get around inlet pressure drop that way also. If the nature of the release is manageable a rupture disc by itself may also work - but I presume you don't want that.

You know what, maybe you could use 2 valves, one for normal operation and another for the gas addition control valve. You could perhaps use a 3-way valve to isolate normal and start-up operations.

The more you learn, the less you are certain of.

 

[maddocks]

I thought about multiple control valves in parallel,but if the DCS goes bonkers, then the possibility of a multiple failure still has to be considered. I'd like to try and convince the pipeline group to set their MOP down to the same as vessel pressure. Looks like I have several choices:

1.) Drop pipeline pressure so the fail-open flow doesn't exceed PSV capacity.

2.) Install a SIL-2 or SIL-3 system with associated costs.

What about installing two(2) PCV's in series? Is that acceptable?

 

[CHD01]

I was thinking of a manual 3-way valve (AG full bore ball valve) with mechanical link and/or key-lock system to guarantee positions. This is code allowed. This consists of one valve with one inlet port and two outlet ports with only one outlet port open at a time.

If you take control valves options can you not get around DCS failure with one valve FAIL OPEN and other FAIL CLOSED?

Yes, 2 control valves in series are allowed, done alot to step down high pressure steam to lower pressures.

On pipe, rememeber that B31.1 allows relief set pressures above MAWP of pipe based on number and magnitude of overpressure events in a year. I don't know if this applies to you or not. If it does you may have solved your problem; because then you have code approved reason to lower your vessel relief valve set point to MAWP of vessel and pipe code with pipe relief valve set above MAWP of pipe code.



The more you learn, the less you are certain of.

 

[CHD01]

Sorry, but I sent that message before I was ready.

You could perhaps use two 3-way valves using AG full bore ball valves. The upstream valve would have one inlet port with two outlet ports and the downstream 3-way valve would have two inlet ports and one outlet port. Between one of the upstream 3-way valve outlet ports and one of the downstream 3-way valve inlet ports you would introduce the high pressure gas through its own 2 port control valve. The the mechanical links would connect both 3-way valves to control which ports are open and closed.

By the way why are two engineers working on their own time on a Saturday, must be an indication we are socially deprived. My excuse is I'm in between jobs looking for work on line after my present employer completed the contract he had that I was working on the last 3 years.

The more you learn, the less you are certain of.

 

[maddocks]

No kidding, in this day and age, I can't always turn it off like everyone says I should. Socially deprived? For sure. %-)

 

[JoeWong88]

Dear all,
Good discussion.

Latexman,
If the system demand for SIL-3, single PSV only provide SIL-2. Do you provide additional PSVs online or other means to increase SIL level ?

Should pilot operated PSV having lower SIL compare to spring loaded PSV as additional pilot involve ?

Thanks in advance.



JoeWong
Chemical & Process Technology