Are bypasses around a PCVs still okay? 6

[SrChemE]

We were having a discussion here at work regarding the need of a bypass around a PCV.
I have a 3" steam line with a PCV in between but a 6" bypass around it. Picture is attached. This is an old system, installed sometime in the 70s'. The relief valve downstream of this system, was sized based on the failure of the PCV steam passing through the 3 inch line. The bypass is 6 in, so I will have to lock it closed in the field. The relief valve is undersized if the steam would to pass through the bypass line.

Are we going away from installing bypasses around pressure control valves? Are there any rules regarding this scenario?

Links to resources and pdfs as well as answers are much appreciated.
Thanks for your help!

 

[Latexman]

I think the answer depends. Is an unchained bypass needed to minimize the operational shackles? Or, is it not needed? Answer that, then go forward. With process controls and automation over the years, and the associated lose of the number of operators used to partially justify all those upgrades, we don't have the manpower to use many/any by-passes these days (for very long). Btw, this is an excellent conversation to have!

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[pennpiper]

SrChemE,
Issue #1 - It may be just my eyes or a bad photo but that does not look like a 6" bypass.
You may have measured the outside diameter of the Insulation by mistake.
If you compare the size of the three valve handwheels, the bypass valve hand wheel looks smaller than the upstream and downstream valves.

Issue #2 - I don't think people have gotten away from using a Bypass Valve in this type of situation. If you do want to think about it you should first consider the potential cost threat to production/Operations while you shut-down a critical Steam system because you do-not have that bypass and it will take fifteen days to get a replacement PCV. Think about it!

Sometimes its possible to do all the right things and still get bad results

 

[georgeverghese]

Bypasses around control valves are useful when pressures upstream and downstream and flow demand are constant, else they arent of much use. Designers still throw them in most of the time even when flow demands are expected to fluctuate. Add an RO next to the bypass valve to restrict flow to the relief capacity desired.

 

[DSB123]

Agree with Pennpiper,
Think of the maintenance situation - do you want to shutdown the whole system or just implement the bypass when the PCV needs repair/replacemenet?? Far better to have the bypass.

 

[DGrayPPD]

I also agree with pennpiper and DSB123. Any time I see a control valve without a bypass I question it. Maybe there are certain situations where one isn't needed but my thinking is shutting down the whole system for an extended period of time is less desirable than the added initial cost of the bypass.

Also agree that it does not appear to be a 6" bypass.

And maybe I'm missing something but I'm very curious why the bypass was designed so high up and far away from the control valve station. Maybe it is just the limited field of view in the picture but I assume access to the bypass valve is difficult. Seems strange to me. But like you said, it was designed in the 70s so I don't really expect to find out why.

 

[LittleInch]

one of my favourite topics in design reviews / HAZOPS.

Not sure why this is in the backwater forum of safety relief valves and on in the chemical pross or pipelines and piping forums but hey.

I have the opposite view to d gray. Every time I see a bypass on a P&ID or in the field I either remove it, or try and remove it.

why?

My contention has always been that if you need a control valve to control something then you need that control, whether it is pressure, flow or any other variable. Adding a bypass line, especially an on/off bypass basically removes that control from the system. The system is therefore effectively uncontrolled and flow etc only limited by the size of the line and / or any single sized orifice you place in it's path, so in that case why have the control valve in the first place. If the system concerned is that vital to the plant, then install a second control valve in series or remove the control valve. As soon as you put a valve there then at some point someone can open it and all you carefully thought out control scenarios go out the window.

At the very least, you can limit the flow to the max for a control valve wide open, but that cannot limit pressure in the same way. Even a manual globe valve type arrangement for me doesn't work as you then need someone siting there watching gauges and reacting to changes in demand etc.

So if you know nothing about this valve or how it is flow limited, chain it or remove the handwheel and hang notices on it saying DO NOT OPEN.

In just about every discussion in a HAZOP, I've mainly managed to persuade enough people of the logic of this argument.

So the proper design (for me) is either to not have one or install a spare Cv if it's that vital. To not have critical valve parts stored on site implies some breakdown of the spares philosophy.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Latexman]

In my batch system world, we use the following general philosophy. On the systems that keep the entire plant running, like cooling water, steam, bulk monomer headers, process water (hot and cold), we have some parallel CVs as spare. But on individual reactor lines, it depends on many things, like reactor size and product line. A small pilot reactor usually gets no spare CVs or by-passes. On production reactors, product lines that can stop while processing with no bad effects, we have no CV by-passes. On product lines that can stop while processing and maybe have quality issues, we have CV by-passes. On product lines that cannot stop while processing or they may set up the reactor, we have some parallel CVs as spare. Then, as usual, there are the exceptions, like the monomer feed system that is either seldom used or is "time-sensitive". We do not use spare CVs or by-passes on these, because they will be plugged when you want to use them. LittleInch, this sounds less black and white than you sounded, is it?

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[LittleInch]

Each system will have it's own peculiarities, but I would tend to agree with your general philosophy. It's when people put bypasses in as a "standard" simply because that's what the last project did that I get annoyed.

I tend to deal with fluids which are not time sensitive so there are always exceptions.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[don1980]

I think engineers install far more bypass valves than needed, and when they are installed these bypass valves are almost always far larger than needed. A small bypass valve is needed in some services to warm-up the system during start-up, but in my experience it's extremely rare to see a process control loop being operated manually. Yet, almost every control valve, regardless of it's criticality or service, is installed with a manual bypass. It seems to me that that is a wasteful outdated practice which almost never gets challenged during process design.

 

[DGrayPPD]

Very interesting POVs.

 

[mk3223]

A different topics... it didn't make sense to have the bypass valve located at the unreasonable height overhead which was difficult to use if needed.

 

[Latexman]

At first glance, true, maybe the OP can find/tell some history; maybe not. For instance, what IF that valve used to have a chain operator? Seldom used or a policy shift could have removed it. I've always enjoyed talking to the experienced operators and engineers on why some things are the way they are.

Good luck,
Latexman

To a ChE, the glass is always full - 1/2 air and 1/2 water.

 

[e43u8]

In the continuous process, considering economical and operational points, following criteria could be applied in by pass consideration around CVs:

1-CVs size less than or equal to 6" are equipped with by pass line

2-CVs size greater than 6" will be provided with hand wheel and no by pass line is being considered

3-CVs in main utility lines are normally equipped with by pass line comprises a spare control valve

 

[Ehzin]

Well my only feedback I suppose is regarding the height of the bypass. I've never seen the calculations for it but I've heard of the following issues regarding thermal expansion.

The possible scenario that could justify this location is as follows:

Assuming the bypass was closer to steam trap lines and this is in fact a steam system; manual valves closed for maintenance of control valve. Steam traps are used to maintain condensate removal while bypass is open for bypass flow. Before control valve removal, line between isolation valves begin to cool while bypass stays hot. Eventually the difference of thermal expansion between the horizontal sections of the bypass and main line build high stress in piping that can't be handled by the limited flexibility in the vertical sections between the horizontal bypass section and main valve section resulting in overstress.

This scenario could be exacerbated by height difference between the two horizontal sections of pipe, the width of the station itself, and degree of insulation on the main line. The opposite scenario could also occur during normal operation if the bypass wasn't insulated and it was smallbore and fairly thin walled. Like I said, I've never seen calcs on this scenario, just heard about it. The mechanics exist under the right circumstances but I don't think having a bypass 15 foot in the air is how I would go about solving it.

Feel free to offer comments but I suppose this may be outside the reason of this conversation a bit.

Thanks,
Ehzin