System pressure drop in a closed Loop Heat Transfer System with branch connections

[Wazobia]

Hello guys

First of all I apologise if this has been covered somewhere else, I have done a bit of searching but haven't seen it.

As the topic suggests, what I have is a closed loop heat transfer system which heats up process streams around an oil & gas separation facility. The HTF (Therminol fluid) system consists of a pump(duty/spare), a burner unit, a storage tank for the HTF fluid, an expansion tank, pipeline which transports the fluid around the system, and a differential pressure control valve which modulates depending on the difference in pressure on either side of the system, i.e. the supply and return headers. A sketch is attached along with this post.


As the supply temperature (and therefore heat load) of the HTF is to be constant, the flows through each consumer, (X, Y, Z on the drawing) which are controlled by temperature control valves is expected to be the same (Consumer Z, a heater used for testing well fluids, will have an intermittent operation). Flow from the pumps is to be 200m3/h, at a supply pressure of 100psig. Pressure will be lost along the distribution piping, and within the consumers (X, Y, Z). The job is to work out the pressure at the return to the pump suction and check if it is enough to match the existing suction pressure as provided by the nitrogen blanketed storage tank.

My question surrounds the pressures at the different nodes, as marked on the sketch (A, B, C, D, E, F). Pressure at A is simple enough, it's the supply pressure (in this case from the burner unit) less the pressure drop up to A, which can be taken to be 5 psi. There is some pressure drop from point A to the discharge flange of skid X due to piping, equipment and the temperature control valve. Let's take that value to be 8 psi. If the pressure drop from A to B due to the flow of 180m3/h of fluid is 5psi, what is the pressure at point B?

I hope my query isn't too confusing. What I want to know is: is the pressure at B simply due to the pressure drop between A and B? Or does it have to take into account the losses within skid X too?

Thank you

 

[LittleInch]

A well written post with a diagram - we don't get that too often.

Is the presusre at B due to the pressure drop between A and B - Yes. The losses inside X are not relevant.

HOWEVER, the issue with a system like you have shown and described is the the actual flow through X can vary from 0 to whatever it will take when the control valve is fully open. Unit X has the highest differential presusre of the three units as it is closer to the pump discharge and pump inlet so has the highest pressure in and the lowest pressure out. Often what you find is that when the control valve controls on temperature only and not flow, your units X & Y can take all the flow leaving Z with very little, or if Y closes 50% then Z suddenly gets a lot more flow than it had before and not a constant flow.

There are a few ways around this - either use the control valve you have to maintain a fixed pressure downstream of the valve but upstream of the unit set so that your max flow is 20m3/hr (or whatever that unit wants) or insert some sort of flow meter (orifice plate is fairly simple and cheap) and add an override (low selector block) to your control valve logic to limit max flow to your max for that unit.

Your end valve also needs to be controlled such that it only opens when it needs to, i.e. presumably when all other valves are closed.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Wazobia]

Thanks. I've read enough posts here to realise how much a drawing can make a question easier to understand. Are you affiliated with "BigInch", by the way? He/she has been a regular poster here for a long time.

The temperatures and flows going into the process side of the skids X, Y and Z (these are heat exchangers) are not expected to vary much. The temperature control valves within the skids will act to maintain a constant outlet temperature of the process fluid and since the heating medium exiting the burner is designed to be at a fairly constant temperature (resulting in a constant heat load), the flowrate going into each skid should be fairly constant. Except if there is a significant process upset, of course.

The final control valve should, in theory, only open when one of the consumers is not in service. In reality, the total flowrate of heating fluid required at the temperatures specified (i.e. the total heat load required by the system) is less than what is being offered by the vendor, and hence that control valve will always be partially open to allow the excess fluid pass through. But it is set like you say, i.e. if the valves within the skids are closed, the pressure upstream of the differential PCV will increase and it will open more to allow the fluid pass through the system.

So the pressure profile through the system would be something like (pressure drop values are actual values obtained from calculations using detailed piping layout):


Pressure drop up to point A: 6 psi (at 200m3/h). Fluid at point A is at a pressure of 94 psig.

Pressure drop from point A to B: 7 psi (at flowrate of 180m3/h). Fluid at point B is at a pressure of 87 psig.

Pressure drop from point B to C: 1 psi (at flowrate of 40m3/h). Fluid at point C is at a pressure of 86 psig

Pressure drop from point C to Differential pressure control valve: 1 psi [approx] (at 20m3/h). Fluid at control valve inlet is at a pressure of 85 psig.

dP across differential pressure control valve: 4 psi.

And so on. From your earlier statement, I take it this the proper way to go about the calculations?

How does one account for the pressure drop within the heater skids? I'm finding it hard to grasp the concept of the pressure drop within the distribution piping being separate from that within the skids...this is my first time working on a closed loop pumping system.

And surely the pressure drop within the heater skids contribute to the total backpressure on the pump?

 

[Latexman]

Your system "seems" to be violating one of the design guidelines I use for acceptable "maldistribution" in a supply/return header design. That guideline is - the pressure drop in the supply and return header should be small compared to the pressure drop of each consumer. Your numbers indicate the pressure drop in the supply and return header is about the same order of magnitude as the pressure drop of each consumer. For example, [Δ]P[sub]AB[/sub] = 5 psi and [Δ]P[sub]AD[/sub] = 8 psi.

Search and read up on designing for acceptable "maldistribution" and designing manifolds.

Also, become familiar with the momentum affects in the header as flow is diverted to consumers and recombined from consumers.

Good luck,
Latexman

Technically, the glass is always full - 1/2 air and 1/2 water.

 

[Wazobia]

Thanks Latexman.

Due to the layout of the plant, some of the lines are quite long hence the (relatively) large pressure drops, compared to the consumers.

I'll look into the distribution like you said, always good to learn something new!

 

[ione]

Wazobia,
You have received valid advises from posts above. Further to Latexman's good point you can look for "primary/secondary loop piping". If you are not able to minimize pressure drop in the main header, compared to those of the secondary branches, you'd need a dedicated pump for each secondary branch (sized to match branch's flow requirement at branch's pressure drop).

 

[LittleInch]

For the Big Inch / Little Inch thing see this thread -

http://www.eng-tips.com/viewthread.cfm?qid=352450

For your questions and system -
You need to think of the headers as providing delivery pressure and flow to a point and allowing flow at a pressure back in to your system at a point, but yes what you laid out looks Ok to me providing that your end HX ends up with enough DP to be effective. Having different inlet and outlet pressures for each one can create issues over operation, especially at start-up.

I do't uderstand why your end valve is open during flow - just let the system run at a slightly lower flow rate.

Yes the pressure drop in the HX will contribute to the overall pressure losses that the pump sees.

In steady state this will run fairly smoothly and as you have more flow capacity than you need then you can cope with small upsets, but larger ones will start to cause problems if you don't try to limit max flow through one of your units.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way