Você está na página 1de 9

Location of Check Valve on Pump Discharge

alienitmeca (Mechanical)
(OP)
15 Jun 16 09:45
Hello,

When designing your piping layout, one would certainly have to use a fair amount of valves
of various types (check valves, gate valves, butterfly valves...) to choose one type over the
other demand some thinkig along with good engineering skills and operating conditions, My
question is included in this department:

when on uses a check valve along with a gate valve (at the outlet of a pump, inlet of storage
tank...) how to know when to put check valve then put gate after in it or is it the other way
around ? Could you help me how to decide in this matter.

Thnk you

LittleInch (Petroleum)
15 Jun 16 10:17
My way of thinking is that in the direction of flow you have the isolation valve and then the
check valve.

The reasoning is that the check valve is difficult to drain past and if you shut the isolation
valve you can lock in fluid between the two. Hence if you drain the pump, not all the fluid
leaks or can suddenly escape.

Do it the other way around and the fluid is able to flow from the isolation valve to NRV side
to somewhere downstream.

Inlet to a tank would be the same thing in direction of flow - isolation then NRV then tank.

Remember - More details = better answers


Also: If you get a response it's polite to respond to it.
4
pennpiper (Mechanical)
15 Jun 16 12:31
For check valves at Pump Discharge it should be: Pump, then Check Valve, then Isolation
valve
Check valves tend to fail more often the isolation valve. So if you have to remove the NRV at
a tank you will risk draining all of the fluid in the Tank if you place the isolation valve as Little
said.
Rule to remember is the isolation valve should always protect the greater inventory of
product

Sometimes its possible to do all the right things and still get bad results
bimr (Civil/Environmental)
15 Jun 16 13:43
Both a check and gate valve should be installed in the discharge piping with the check valve
placed between the pump and the stop valve to protect the pump from reverse flow and
excessive back pressure.

http://www.flowcontrolnetwork.com/pump-guy-mailbag...
LittleInch (Petroleum)
15 Jun 16 15:18
The tank I'll give you, but in reality how many times is an NRV ever removed from pump??

So in two posts you have two different ways of doing it - never simple eh.

Remember - More details = better answers


Also: If you get a response it's polite to respond to it.
europipe (Chemical)
16 Jun 16 06:10
I think you're alone now, little.
I always saw it like mr. Pennock sais.
Also, for discharge at a pump, in many cases a pressure relief drain is placed between check-
and isolation valve.
LittleInch (Petroleum)
16 Jun 16 09:08
"in many cases a pressure relief drain is placed between check- and isolation valve"

That's my point - if you don't put your pressure relief and drain between the two you can
run into problems. Put it the other side and you don't need to bother if the single PRV and
drain connection is on the other side.

From a purely process view, either location is fine - a lot depends on particular piping
layouts and orientations.

I've seen both and ultimately this is a view the designer needs to take, but there are no
"rules" to this, only custom and practice.

Remember - More details = better answers


Also: If you get a response it's polite to respond to it.
donf (Chemical)
16 Jun 16 15:22
I agree with Pennpiper (pump-check valve-isolation valve). In a parallel pump arrangement,
having an isolation valve between the check valve and THE OTHER pump allows the
defective check to be replaced, while the OTHER pump is pumping.
donf

LittleInch (Petroleum)
16 Jun 16 15:36
You guys have a lot of defective check valves.....

Remember - More details = better answers


Also: If you get a response it's polite to respond to it.
europipe (Chemical)
16 Jun 16 17:54
Check valves are never 100 percent reliable.
alienitmeca (Mechanical)
(OP)
20 Jun 16 10:55
Thank you everyone for your responses,

it's been quite an interesting discussion, very informative as well

BigInch (Petroleum)
20 Jun 16 17:10
LittleInch, you give up too easily.

The preferred arrangement should not be decided based on facilitating check valve, or any
other valve's maintenance. It is far more important to facilitate pump maintenance, as the
pump will statistically be the most frequent item requiring break out maintenance of one
sort or another. Pump, isolation valve and check is thus the preferred solution, as no
pressure will be trapped between isolation and check valves nor will any pressure be in the
maintenance's crews face when breaking out the pump. If you do it the wrong way, then
you must add the extra vent valve to release the resulting trapped pressure.

If there is any concern about draining adjacent piping or tanks, provide an appropriate
spectacle blind.
LittleInch (Petroleum)
20 Jun 16 21:51
BI welcome back, haven't seen you on here for a while....

Fully agree, makes sense to me.

Remember - More details = better answers


Also: If you get a response it's polite to respond to it.
BigInch (Petroleum)
21 Jun 16 09:43
smile Thanks.
I've had some other priorities to attend to lately.
Krausen (Mechanical)
24 Jun 16 14:52
It really does depend on the system layout. The pump + block valve + check (or meter +
block valve + check) sequence is the best choice IMO for reasons LI explained, but should
only be used in systems where the piping/equipment can be blocked in elsewhere without
suffering a major downtime. In my work we apply this layout about 80%-90% of the time.
We've had clients who are adamant on flipping the block & check, without appealing to
reason other than "we've always done it that way".

If you are flowing into a system that is not easily isolatable (e.g. injection into a mainline
that runs 24/7), the pump + check + block valve is the better option due to having to shut
down an entire mainline in order to service a defective check valve.
BigInch (Petroleum)
25 Jun 16 09:11
Then you have the same problem when your block valve needs to be repaired.
georgeverghese (Chemical)
27 Jun 16 03:51
In many cases, a check valve has a process safety function, and this may in some cases
where the consequence of a check valve failure poses some safety and loss of production
risks, demand some means of isolation on BOTH sides of the check valve to run a leak
integrity test and have it removed for repair where necessary. Else you have a large volume
of gas or liquid to dispose off before this test and removal can be done.

The periodic scheduled verification of a check valve's integrity in such services has increased
in importance in the last 10-15years or so. On the other hand, manual isolation valves are
not process safeguarding devices.
flukeynub (Mechanical)
18 Jul 16 16:08
This is the exact question I've been asking some of my firm's senior engineers to shed some
light on in the last two weeks without much luck.

I debated this question with myself at my last job (was an operations engineer) and decided
at the time I preferred pump/check/iso mainly because we saw many more problems with
check valve failures than pumps/iso valves. I had to redesign an acid scrubber spray system
to allow for check valve replacement as process system up time was significantly more
important than ease of pump maintenance. Not to mention, for some reason, none of the
pump/check/iso configurations at that site had a vent which is why I even thought about
this in the first place.

Thanks for the opposite thought process perspective!


lilliput1 (Mechanical)
18 Jul 16 17:12
Pump, check valve then isolation valve so check valve can be serviced without draining the
line. The isolation valve is normally left open all the time. It is closed only when pump or
check valve servicing is required. The check valve insures no back flow when pump is
stopped and there is a parallel pump operating. Typically for chilled water service, the chiller
automatically starts its associated pump and the pump isolation valve is open.
bimr (Civil/Environmental)
18 Jul 16 21:44
Check valves are not considered to be reliable so process safety engineers do not use check
valves for process safety functions.
2
JJPellin (Mechanical)
19 Jul 16 13:02
I work in a large US oil refinery. We have about 1600 centrifugal pumps. All of them, except
for two have the check valve between the pump discharge flange and the block valve. This is
a requirement of our standards. Perhaps some references could help.

API Recommended Practice 686; Second Edition

“A check valve shall be installed in the discharge line of all pumps, compressors or blowers,
whether centrifugal or rotary, unless there is no possibility of a reverse flow or pressure
surge (such as water hammer) under any conditions. The check valve shall be located
between the machine discharge flange and the discharge block valve.”

Hydraulic Institute Standards for Centrifugal, Rotary & Reciprocating Pumps, fourteenth
edition

“A check valve and a stop valve should be installed in the discharge line. The check valve,
placed between the pump and the stop valve, is to protect the pump from reverse flow and
excessive back pressure.”

Piping Design for Process Plants; Rase and Holmes; 1963

“Place check valve between discharge nozzle and gate valve to prevent liquid backup when
pump stops running.”

I have worked with many thousands of centrifugal pumps over the past 27 years. I don’t
recall ever seeing the check valve installed outside of the first block valve.

Johnny Pellin
JJPellin (Mechanical)
19 Jul 16 13:41
I pulled another couple of books off my shelf.

Centrifugal Pumps; Karassik and Carter, 1960

“Generally, both a check valve and a gate valve are installed in the discharge line. The check
valve is placed between the pump and the gate valve and protects the pump against reverse
flow in the event of unexpected driver failure.”

Pump User’s Handbook; Bloch and Budris, 2004

“Gate valve should not be between check valve and pump.”


Does anyone have a published reference or standard that shows the check valve outside of
the block valve?

Johnny Pellin
bimr (Civil/Environmental)
19 Jul 16 14:58
Agree with JJPellin. Have never witnessed any other installation.

If there is a requirement to eliminated the pressure trapped between isolation and a check
valve, a small drain line w/valve is included.
lilliput1 (Mechanical)
19 Jul 16 15:22
Pressure between check valve and closed block valve would only be a concern is the trapped
fluid is colder than ambient and it expands as it heats up.
BigInch (Petroleum)
19 Jul 16 16:07
All those references ... and there is no accompanying logic to justify the requirements???

The hydraulic facts are that a check valve installed downstream of the pump and block valve
is equally adept at stopping backflow to the pump as a check valve upstream of the block
valve ... unless the block valve is already closed, but that would mean no backflow is
possible at all
smilesmilesmilesmilesmilesmilesmilesmilesmilesmilesmilesmilesmilesmilesmilesmilesmiles
mile
europipe (Chemical)
19 Jul 16 16:18
BI, I think like most of us a checkvalve is no reliable option to close a system, a blockvalve is.
So exchange a checkvalve is easier to do, like mr. Pellin sais.
JJPellin (Mechanical)
19 Jul 16 16:18
We fail check valves quite often. Some of our check valves are designated as critical checks
for mitigation of process risk associated with back-flow. Those valves often have to be
pulled and inspected on some defined frequency.

Most of the time, the process downstream of the block valve can only be taken out of
service at a major turnaround (every 4 to 6 years). I have never heard of any problem with
stored pressure between the check valve and the block valve. If the check valve is outside of
the block valve and it fails, I have to shut down a process unit (potentially costing millions of
$US). If the check valve is inside of the block valve, I might have to use a short spool with a
bleed valve to eliminate the possibility of stored pressure trapped between the valves. This
is not a problem for me. Putting the check outside of the block valve seems to have a huge
potential downside (unit shutdown) with no real upside. The fact that every standard I can
find goes to the trouble to specify this arrangement suggests that a lot of people have
reached that same conclusion.

Johnny Pellin
bimr (Civil/Environmental)
20 Jul 16 00:05
Regarding: Pressure between check valve and closed block valve would only be a concern is
the trapped fluid is colder than ambient and it expands as it heats up.

How do you relieve pressure when you line break the pipeline for maintenance?

Link

lilliput1 (Mechanical)
20 Jul 16 01:00
One could slowly unbolt the flanges so the fluid leaks out to relive the pressure. Or maybe
the pump needs to be serviced because fluid has already leaked out and relived the
pressure!
bimr (Civil/Environmental)
20 Jul 16 01:25
Last time I checked, it is 2016 and employees working for responsible corporations have line
breaking/opening procedures for safety of the employees working on piping.

https://www.google.com/?gws_rd=ssl#q=Line+Breaking...
lilliput1 (Mechanical)
20 Jul 16 01:51
bimr,
You are right. Steam pipes and hot water pipes need to be able to relieve pressure so as not
to cause injury when disconnecting piping or equipment. Provide drain valves and blow off
valves at strainer. Locate between equipment and block valve.
LittleInch (Petroleum)
20 Jul 16 07:57
I can certainly see the point here and maybe what it highlights is the difference between
process plant type pumps and pipeline type units. Use of check valves for a process function
is different to using it for a pressure break / safety function. You can take credit for a check
valve to reduce back flow, but not to seal.

Having said that sometimes they do work very well and the key point is to look at your
particular system and piping layout and make sure you can vent and drain between a check
valve and its downstream block valve. If your particular use of check valves is such that they
are seen as critical to the process, used in anger a lot and hence more likely to fail, then
suitable isolation needs to be in place and for many this will be between the pump and the
pump isolation valve. For others it might not or the check valve is there for an extreme
event and not used in anger very often so would be expected to last for a long time and can
be upstream or downstream.

So the key answer (for me) to the original question is look at your systems and understand
the issues over space, potential for removal or servicing, isolation requirements and impact
on the process and the ability to drain sections of piping when you install check valves.

Remember - More details = better answers


Also: If you get a response it's polite to respond to it.
georgeverghese (Chemical)
21 Jul 16 05:43
Whilst on this topic of check valves, though not the topic of discussion, some OpCos have
stopped the use of wafer style check valves for at least 10years now when in flammables
service.
europipe (Chemical)
21 Jul 16 07:12
George, can You explain why, and why 10 years?
MickMc (Mechanical)
21 Jul 16 08:13
It is common for wafer style valves not to permitted in hydrocarbon service as in fire case
the stud bolts are unprotected and liable to fail first.
europipe (Chemical)
21 Jul 16 08:38
Ah, the fire situation, what is OpCos?
BigInch (Petroleum)
21 Jul 16 16:29
Perhaps it means operating companies.
europipe (Chemical)
21 Jul 16 16:47
That's a wide understanding
georgeverghese (Chemical)
22 Jul 16 04:12
Yes, long bolts on wafer style checks tend to loosen more during a fire...

At least in Shell, wafer style checks were removed from their piping classes since the year
2000 or so.

OpCos = Operating Companies.


LittleInch (Petroleum)
22 Jul 16 06:37
I've seen pictures and videos of long bolt wafer checks and wafer valves during a fire
compared to other bolts and it's alarming. Unless there is no possibility of a pool fire
underneath it or you fire protect it, this would / should come up in a HAZOP or similar safety
review of the plant design or just be banned for new installations.

See this - slide 10 is the key one comparing a "normal" flange and a long bolt one.
http://slideplayer.com/slide/4731877/

This is a one page version of the same thing https://www.google.co.uk/url?sa=i&rct=j&q=...

Remember - More details = better answers


Also: If you get a response it's polite to respond to it.
europipe (Chemical)
22 Jul 16 07:48
What about spect. blinds then?
They have also longer bolt lenghts.
LittleInch (Petroleum)
22 Jul 16 08:28
Not as long as those being noted in the reports.

Mostly they refer to flangeless valves, but clearly the longer the bolt the worse it gets in
terms of expansion in the event of fire.

My guess would be if you more than double the length of the bolt then you start to run into
that issue, but haven't done any calcs on that.

Remember - More details = better answers


Also: If you get a response it's polite to respond to it.

Você também pode gostar