Viscous Fluid Tailpipe Pressure Drop

[RJB32482]

I haven't done the calculation yet, but we are looking at redesigning the relief piping for a positive displacement pump where the fluid is silicone. The relief condition is just blocked discharge at a maximum rate of 9.8 GPH. Now since this is a very viscous fluid, it will have a high Renyolds number at this flow. So I have three general questions:

1. When determining the tailpipe pressure drop, is standard practice to use the worst case flow or the maximum flow that the valve can put through at relieving conditions (usually higher than worst case flow)

2. When this flow occurs, do you have to account for it since the flowrate is so low? Or can you consider the pressure drop low since the flow is so low (below 10% set P of 75 PSIG).

3. The constant backpressure is 0 PSIG. If I do get a superimposed backpressure > 10%, can I just derate the relieving capacity of the valve (lower Kb factor) and see at the tailpipie pressure drop if I still get adequate relieving capacity? Or do I need to increase tailpipe size to lower pressure drop?

Thanks

 

[Latexman]

If it is "very viscous", a low Reynolds number and laminar flow is more likely.

First thing you have to realize is, this is not a Code application. You are protecting the pump, which is not a Code piece of equipment. Pipe is usually not protected.

1. Use the pump flow.

2. I don't understand this question.

3. It's not a Code case, so you can do either one that will give the required capacity.

Good luck,
Latexman

 

[Latexman]

Oh yeah, ask for liquid trim for the PSV. And, if you pipe the discharge to suction or to the source tank and if this silicone has solids or will dry/adhere to the PSV spring, consider a bellows PSV.

Good luck,
Latexman

 

[RJB32482]

Looking at a Moody Chart, friction factor decreases as Re increases. Thanks for the other helpful information. I know our relief manual wants the less than 10% backpressure rule, but derating the valve if possible should solve it if its not a code issue.

 

[pleckner]

I would argue that you are not protecting the pump but are actually protecting the piping. PD pumps typically come with internal PSVs to perform the function of protecting the pump but these will not be size properly to protect the piping.

I disagree that you don't have to protect the piping. Althoug I am not up as much on ASME piping codes as I am on the BPV codes, from what I've read in the scope of ASME B31.3 and exclusions you do indeed fall under this code. This piping Code will direct you to PSVs as required by ASME Section VIII, Div 1. There is a lot of back-and-forth reading in B31.3 and there are many variances it allows for but if you are going to have to install a PSV, then you need to follow ASME Section VIII, Div 1. The PSV will be a liquid trim as suggested by Latexman but it should conform to those rules which govern installing a proper PSV for the application.

If you intend to use a conventional style PSV rather than a balanced bellows PSV, then you must be very cautious not to exceed the 10% variable back pressure rule. Yes you can derate the PSV but by what? The cpacity of these valves fall off super fast when the variable back pressure exceeds 10% of the set pressure. API RP520 used to publish a chart and the fall off is quick. They left the chart out in subsequent revisions to discourage the practice of trying to derate these valves. They would prefer you just put in the balanced bellows type.

By the way, this is a different issue than the superimposed back pressure. If you expect the superimposed back pressure to be "x" and you are using a conventional style PSV, then the spring will be designed so that it will open at your set pressure when the superimposed back pressure is "x". If the superimposed back pressure is "x+y", then your PSV won't open until the pressure in the system reaches "x+y". This can be permissable in ASME B31.3 to some degree, unlike ASME Section VIII, Div.1.

For a PD pump, you need to size the PSV based on the maximum flow this pump can put out. The discharge line loss would be based on this flow. The PSV "rated" flow really has no meaning here becauseyou can never flow anymore liquid out the PSV than what the pump can deliver.

 

[Latexman]

Phil,

Pardon my bluntness, but I know for a fact that some companies do not accept the cheap ass reliefs that come on some PD pumps. They cannot be removed and tested on the bench under controlled conditions without a special set-up that mimics the area they fit on the pump.

I have personally seen the holes in a cinderblock wall from the bolts on the head of a VERY popular gear pump that did NOT open at it's set pressure. In fact, those holes so thoroughly impressed the assistant plant manager that I did not get a raise that year, even though the pump was installed in the 1970's and PMI'd every year since then! I was unlucky enough to be the production engineer when the incident occurred. Not only did I add a high quality PSV, I also switched to a different pump vendor just to get that pump out of my unit.


Good luck,
Latexman

 

[pleckner]

Great story Latexman. I do question the vendors of PD pumps about the type of PSVs they use internal to their machines but again, that's why I always put a PSV on the line because those internal PSVs can't always be trusted.