OpenFlows | Water Infrastructure Forum - Pump From Tank to Junction Downhill
BMPCreated with Sketch.BMPZIPCreated with Sketch.ZIPXLSCreated with Sketch.XLSTXTCreated with Sketch.TXTPPTCreated with Sketch.PPTPNGCreated with Sketch.PNGPDFCreated with Sketch.PDFJPGCreated with Sketch.JPGGIFCreated with Sketch.GIFDOCCreated with Sketch.DOC Error Created with Sketch.
Question

Pump From Tank to Junction Downhill

by
NKG
Expert

Good evening.  I know this is trivial but I am having a mental block and hope you can help me out.  There are three questions with one doing validation on a model, the second setting up the model optimally/suggestions on calculation and third hydropnuematic tank.  

First the scenario (water system):  Water is stored in ground storage tank.  Below the base of the tank, about 5', sits booster pumps that pump into a distribution system that goes downhill to a series of homes.  The homes sit on roughly the same elevation.  It obviously has a demand and minimum requirements.  So to simply I am pumping downhill immediately out of the tank.  So if we assume that the junction (which represents the homes is the datum elevation.  The tank base can be 105' from datum.  Pump can be 100' from datum.  Assume 45' water in storage tank and the pump at the middle of the curve is doing 100 gpm at 100' feet of head.  Assume 1000' of 4" pipe.

Question 1 -  (Here is my mental block)  I know how to typically write Bernoulli's equation but for some reason the concept of pumping downhill is throwing me.  I know what to expect of the numbers but without "forcing it" I am just having a mental jump rope of where did I go wrong.  So if you were to write the equation out; what would it look like.  When I do it my numbers come out negative for most of the values of Q.  Ultimate goal is to develop not only the pressures and flow at the junction but to also develop a system curve to verify the system curve in WaterGEMS.  

Question 2 - When generating a system curve with an assumed pump, does that pump apply no matter what the curve is of that pump or in order to truly get an accurate pump curve, without knowing the pump, take the pump element out and put two junctions with negative demand on the one and positive demand on the other and run the scenario that way?  As for the end junction (demand junction) does that need to be a PDD with a set minimum required psi/ft of head in order to generate a system curve?  

Question 3 - When using the HT element, I have it set to treat as a junction as false.  When doing that I can get a system curve but I also get a notification saying my pump is exceeding the max operating point and when I look at it, the pump is doing not only the demand of the junction but also the in an out of the HT element.  But if I treat it as true, the pump curve looks great and I do not exceed the pump anymore.  If Qin is equal to Qout of the tank in steady state analysis, shouldn't only the demand of the demand junction be demonstrated?   


Accepted Solution

[quote userid="21367" url="~/products/hydraulics___hydrology/f/haestad-hydraulics-and-hydrology-forum/196581/pump-from-tank-to-junction-downhill"]Question 3 - When using the HT element, I have it set to treat as a junction as false.  When doing that I can get a system curve but I also get a notification saying my pump is exceeding the max operating point and when I look at it, the pump is doing not only the demand of the junction but also the in an out of the HT element.  But if I treat it as true, the pump curve looks great and I do not exceed the pump anymore.  If Qin is equal to Qout of the tank in steady state analysis, shouldn't only the demand of the demand junction be demonstrated?   [/quote]

Hydropneumatic tanks are usually only modeled during a transient simulation, which is where the treat as junction = true option is helpful - it models a HT that "floats" on the system and operates at normal operating pressures, hence for simplicity is is treated as a junction for the initial steady state period.

If you're only interested in a steady state or EPS (not transient in HAMMER) and need to model a hydropneumatic tank (for example for a smaller scale system where these are used and you need to see the impact), then most likely you'll want to choose "false" for "treat as junction" so that they actually operate as a tank. In this situation the system head curve will behave just like if you had a tank element. Meaning, the system head curve would show the head needed to "lift' water to the tank's current elevation for a range of flow. The HT in this case separates the upstream and downstream system - the flow into the tank is based on the upstream hydraulics and the flow out of the tank to the downstream system is based on the downstream system hydraulics.

On the other hand when "treat as junction" is set to "true", it acts as a junction and the system head curve can indeed be entirely different. For example if there are only fixed demands downstream, then there is no storage "buffer" and therefore no system head curve since the pump flow can only be equal to the sum of the downstream demands (see "System Head Curves in Closed Systems" in this article)

If you are seeing a "better" system head curve with the option set to "true", then you might have another tank further downstream and therefore the system head curve can vary the flow. Check to see what the calculated hydraulic grade is in the "results" section of the HT properties and compare it to the initial HGL that you had set it to when using "false" for "treat as junction".

If you are running an EPS and looking at the system head curve after time zero, then the HT might be filling very quickly resulting in a very high HGL which would significantly impact the system head curve results since the pump would need to add a lot more head to "lift" the water. The following article has more explanation and tips: Modeling hydropneumatic tanks in an EPS in WaterGEMS and WaterCAD