Controlability of system with a flat pump and system curve

[lewismg]

I am sizing a control valve for a reflux system that has been poorly designed. The valve controls temperature in the column (cascaded to flow) by throttling flow downstream of a centrifugal pump. Unfortunately, an inappropriate pump has been bought for this application and it will be operating inefficiently at the left side of it's curve. Also, the dynamic losses in the system are low so the system curve is very flat. Results in the following Cv cases:

Min Cv Case: Q = 5.31 GPM, dP = 48.2 psi
Normal Cv Case: Q = 10.6 GPM, dP = 45.8 psi
Max Cv Case: Q = 13.8 GPM, dP = 45.1 psi

My contention is that because the difference in pressure drop is so small, the pump will be sensitive to small changes in control valve pressure drop and will float along it's curve. Can a system like this be controlled?

The 30% CV pressure drop / dynamic loss (or 15 psi) rule of thumb ensures that the pressure drop across the CV is not too low. Is there a similiar rule for a case in which CV pressure loss >> system dynamic losses?

Any help is appreciated.

Mike

 

[TD2K]

I don't see a problem. The control valve % opening will essentially set the flow if you think about it. The flat head pump curve just means a more or less constant inlet pressure to the control valve.

I had a cutter pump system (heavy fuel oil blending) which had a pretty flat pump curve which was controlled similar to yours, it worked fine.

 

[d23]

lewismg

Operating on the left side of the pump curve offers two more questions:

If you replace the pump will the power (efficiency) improve enough to justify replacing?

Will operating on the left side of the curve cause premature failures?

Good luck!

 

[lewismg]

TD2K,

Yes I agree that a system can be controlled if you have a flat pump curve. But what about if both the system and pump curves are flat?

Note: when i refer to system curve, this neglects the pressure drop across the control valve.

Regards

Mike

 

[d23]

lewismg

If you have a very flat pump curve then you have a very very limited amount of system control. This flat curve should let you know that an extremely minimal amount of wear on the pump will require an immediate pump repair in order to meet your system curve. Before pump failure(s) become an issue you may want to redesign your pump or consider possible adjustments to your system (curve.) By designing a pump to operate at it’s BEP you may save enough on system efficiency to help justify the replacement.

This is just my personal opinion.

 

[25362]

A reflux pump is generally also delivering flow out of the system on LLC of the reflux drum. The system characteristic of the parallel stream may not be flat at all.

However, when the system head curve is flat sometimes by-pass control is used, especially on those high specific speed (axial flow) propeller pumps were the power consumption drops or doesn't markedly change with increasing flow rates.

This is not to be confused with minimum flow by-pass (kick back) protection systems used on most (other) types of (medium and low specific speed) centrifugal pumps.

 

[TD2K]

If both curves are flat, then essentially the control valve sees a constant dP. If you need more flow, the control valve will open to get the additional Cv. The pump's output is constrained in effect by the control valve, it's not analagous in this case to it hunting all over it's operating curve.

Note, if you were trying to control the pump by discharge pressure with a flat curve like this, you would have problems as small changes in set-point would produce large changes in flow as the pump established a new operating point.

Sundyne pumps commonly have a very flat head curve (till it drops off the cliff at the far right), flow control works fine on these.

As mentiond, depending where you are operating on the curve, you can have increased maintenance problems. Are you above the continuous minimum flow?

 

[fdomin]

What is the overhead product?
If it is a pure component and overhead pressure is also controlled, using a reflux set-point controlled by overhead temperature will lead you to control problems in the column as condensation temperature will be ruled by overhead pressure.

 

[lewismg]

Thanks for all your responses. They were very informative and have given me an idea of the best way of controlling a reflux system.

I agree that the system (albeit an inefficient one) can be controlled even though both curves are flat - the control valve will react to flow (not discharge pressure) and open/close accordingly. To answer your questions:

TDK2: We are operating the pump (Total rated capacity = 26 GPM) under continuous minimum recycle (12.2 GPM). This was a vendor decision and I am assuming that it was to move the operating point to the right side of the curve (We only require 13.8GPM).

fdomin: This is reflux for a Reformer stabilizer column - largely propane, ethane and butane. You mentioned that control would only be a problem with pure components.