Multiple Intersection of Simulated System Curve and Pump Curve 1

[IanVG]

Been scratching my head far too long on this issue. I have a model of my campus chilled water district system in Pipe-Flo. Everything is currently modeled at the same elevation. When I generate the system curve vs. pump curve I get the following graph:

I've been reading in open system with static head inputs, it's possible to have a negative head with a positive flow at a point in the system (i.e. I assume suction force acting on a body of water). But this system is closed. I modeled each building with a fixed 10 dP device (which is not correct) flow control device (I believe this allows the variation of the Cv value to achieve the flow setpoint) and the MHRL (most hydraulically remote loop) I modeled by a fixed 0 dP (i.e. completely open control valve). Changed the 0 dP control valve to a positive value (e.g. 5 dP) did not change results.

Any thoughts on maybe why this is happening?

As a side-note, I noticed that the pipe friction factor (some dimensionless unit) on some pipes shows this semi-odd behavior:

Without having dug into it further, the pipe friction factor seems to shows an odd spike at low flow values. Not a discontinuity by any means, but maybe this behavior is contributing to the odd pipe network analysis results.

 

[GBTorpenhow]

Not familiar with the software referenced but a system curve should be smoothly quadratic - unless you have control valves that are modeled as changing positions with flow in some specified manner programmed in somehow. System curve for a closed system should start at (0,0). Also negative head/head loss does not make physical sense.

On the friction factor, that's the transition from laminar to turbulent flow that you are seeing.

 

[IanVG]

GBTorpenhow: Thanks for the input. Right, the laminar to turbulent flow transition for the pipe friction factor makes sense. Thank you.

Right, something weird is happening to produce a system curve like the one I am showing. This system behavior occurs even when I change the valves to fixed Cv devices. So I don't think it's something to do with the control valves. I modified the model to remove all fixed dP devices and replaced them with curve dP devices. This fix produced a smoothly quadratic system curve.

 

[1503-44]

Valves have a Cv vs position curve. Control valves adjust their position to yield a flow and pressure according to their feedback signal.

Your actual system curve when operating using control valves may be the first one.
System curves only have a parabolic shape when control valves are not functioning. A backpressure control valve at the end of a pipeline will hold outlet backpressure constant for all flow rates, thus the entire "system curve" will be modified to make that condition true and your pumps will vary their discharge head (within their curve capacity) to try to make it work. Where such system curves do not intersect the "system curve" (your first graph), the pumps are not able to cope at the corresponding flow rates unless the control valves are adjusting the pump's discharge head to the indicated values.

in your other post I think I told you that you would have to initially deactivate all control valves.

--Einstein gave the same test to students every year. When asked why he would do something like that, "Because the answers had changed."