Problem setting up a booster pump station with a storage tank control node.

Question

I have a booster pump station with 3 VSP installed. Its system pressure dependent, however, the pressure transducer is inside the storage tank. The storage tank has separate inlet and outlet pipes and when i try to run the model, it gives me an error "Cannot run VSP with a control node that has separate inlet and outlet pipes". I can run the model if I change those pumps to run as constant flow pumps. The booster pump station is installed directly next to the storage tank and the primary function is to ensure the pressure in the distribution system stays at its required level.

Jesse Dringoli · Answer

Hello Nick, 
 Thanks for providing the model and details. 
 My first observation is that you are using an old version (V8i SELECTseries 6) which is no longer supported per the Support Policy . We strongly recommend upgrading to the CONNECT Edition if possible. See more here: Downloading OpenFlows Software 
 My second observation is that the top-fill tank is not configured correctly. You have entered an inlet invert level of 345.5 ft, which means 345.5 ft above the bottom of the tank (262+345.5 = 607.5 ft.) This means that the tank is not able to fill since the HGL on the inlet side would need to be above 607.5 ft. If the 345.5 ft is the elevation of the inlet invert, the "Level (inlet invert)" needs to be set to 345.5 - 262 = 83.5 ft. See related article: Error or bad results when using top filling tank option 
 Beyond this, I am not entirely clear on how to reproduce the problem as the model you sent does not appear to align with your descriptions. Can you provide specific steps to reproduce, and confirm the scenario to look at? Here are a few replies to some of your statements: 
 Nick Sweitzer said: The pumps are VFD not VSP. They are set(existing system) to run in a constant pressure setting. 
 If the pump changes its speed to meet a target pressure, it would be modeled as a VSP in WaterCAD and WaterGEMS, by selecting "true" for "is variable speed pump" in the pump properties. In the model you sent, this is set to "false" (at least, in the scenario that was active, "MDD_2009".) 
 Nick Sweitzer said: These pumps are set to maintain a minimum pressure that is 82 ft above the floor of the reservoir (35.5 psi). 
 What is the "Reservoir"? The model you sent only contains one active tank "T-330(PUD)"and no active reservoirs. 
 Nick Sweitzer said: I can change the control node from the tank to a junction close by and the system runs just fine. 
 See my previous reply on how a VSP behaves when the control node is set to a tank. Indeed if you are trying to model a fixed pressure, you would need to select a junction element. However, if it nearby a tank, it may not be possible for the VSP to maintain that pressure so close to another hydraulic boundary. I am still a bit unclear on where the target pressure location is in relation to the pump and tank. In the model you sent, the pump is set to fixed speed with an initial status of off, so the tank drains out and causes the network to become unbalanced by about 6 hours. There are also no controls to turn the pump on or off based on model conditions. The scenario is using a control set that has all controls disabled. 
 Nick Sweitzer said: The booster pumps will turn on when the system demand requires them to be turned on, otherwise the 3 wells are supplying the distribution system. 
 Where are the "booster pumps" in relation to the VSP, tank and reservoir? In the model you sent, only one pump is active. 
 Nick Sweitzer said: The way the current system is set up, as target head, it shouldnt be cycling water through the overflow. 
 See above, I do not see the pump set up as a VSP with target head. Please identify the scenario or steps to reproduce. When you say "cycling water through the overflow", do you mean that flow is traveling from the pump into pipe P-15 and into the tank inlet? After fixing the inlet invert as mentioned above, the tank would only fill if the pump raises the head above 345.5 ft. 
 Nick Sweitzer said: Initially i wanted to set it up and run it in a "Static" condition, no pumps, no wells, just the tank and a fire flow. Im not even sure that is set up properly. 
 Is this the conditions I see in the model you provided? (MDD_2009 scenario) The pump initial status was set to Off, the control set used by its operational alternative has all controls disabled, and there is no source to re-fill the tank once the pump empties it.

Jesse Dringoli · Answer

Nick, setting the VSP target to a tank tells the program to try to maintain a constant water level in the tank (flow in=flow out). If you are actually trying to maintain a set pressure, set the target element to a junction instead, and set the desired target pressure or HGL. However, in order for this to work, the VSP needs to have direct control over the pressure at the node. If there is a tank between the pump and the junction, that is not possible as the tank would control the pressure. See more here: Using Variable Speed Pumps in WaterGEMS and WaterCAD 
 Perhaps you could include a sketch to illustrate where the pump is in relation to the tank, and elaborate on exactly how the system needs to operate?

Jesse Dringoli · Answer

Nick Sweitzer said: I uploaded another set of files. 
 I see the new files but I am not clear on the steps to reproduce. The scenario "TW Fire Base Physical" is set as steady state and is using a control set with no controls enabled. 
 Nick Sweitzer said: The booster pumps will turn on when the hydraulic grade in the tank is at 344.0 ft (82 ft above the floor of the tank). The well pumps will be called to operate when the hydraulic grade in the tank is 339 ft (77 ft above the floor of the tank) and they will shut off when the hydraulic grade is 342 ft above the floor of the tank 
 Can you confirm that "booster pump s " refers to the single pump called "B. Pumps"? It is only a single pump, so I assume it is using a pump curve representing the combination of multiple pumps. If you need to perform energy studies, you'll want to ensure to explicitly model all individual pumps. 
 Is the "Well pumps" the single pump called "T***** Pump"? 
 Do you mean that the tank will supply demands but when it gets low, the "booster pump" will turn on to fill it, but if the tank continues to drop, the "well pump" will turn on? The problem I see with the setup is that when the booster pump is on, it can fill the tank via the inlet, but the tank will not be able to supply water through the outlet at the same time, because the outlet is exposed to the downstream side of the pump. So, I think the pump would be supplying both the tank and the downstream demands. Similarly if the well pump was called to turn on, it may be "fighting" with the other pump and not be able to fill the tank because of the head already provided by the booster pumps. (or maybe you just want the well pump to supply demands in that case and not to fill the tank). I may have misunderstood. 
 Nick Sweitzer said: Im getting an error now that is calling out the booster pumps and telling me "Parallel VSPs are not allowed to have different pump curves". This is odd as there is only 1 b.pump being modeled. 
 I do see this message when setting both pumps to the same control node. This is because when you set two VSPs to the same control node, the program thinks you are trying to model VSPs in parallel with the logic mentioned in this article about lead/lag pumps turning on and off automatically. As seen in the article, the pump curve needs to be the same in this case. 
 If you were indeed trying to set both pumps as VSPs, this would not be possible because they are both discharging into the same pressure zone (see note in this article ). Even if you were to "trick" the software into working, they would essentially "fight" against each other and the hydraulics would likely not be able to solve. 
 Nick Sweitzer said: in the control sets i have Booster pumps and the 2 wells references a pressure inside the tank. if the controls reference the tank and the control node is not the tank is that a problem? 
 Having controls and a control node for a pump is not a problem and the elements can be different. The problem is that the validation does an initial check of whether the VSPs are in parallel and if the required criteria is met as explained further above. 
 Also note that your pressure-based controls would need to have a band. Currently you have them set to turn on/off based on the same setpoint of 36 psi. This could cause rapid oscillation as the pressure may immediately drop as soon as the pump turns off, causing the opposite control to immediately turn back on. See this and more here: Troubleshooting Controls in WaterCAD and WaterGEMS 
 Nick Sweitzer said: I added an overflow reservoir and put its elevation at the overflow invert elevation. 
 I see that the reservoir elevation is set equal to the tank top but that would mean that water would never flow from the tank to the reservoir. The elevation would need to be above the top in order for flow to travel to the lower elevation reservoir, but when the elevation reaches the top of a tank, the built-in altitude valve will close the pipe. To model the overflow you would need to set the reservoir elevation below the tank top. See more here: Modeling a Tank with an Overflow 
 Nick Sweitzer said: You mentioned you didnt understand where the regulated pressure needed to be, the pressure needs to remain above a specific value across the entire system, not just one particular node. 
 VSPs can be used to maintain a fixed pressure at a particular location, but you can only have one VSP or a station of VSPs in parallel pumping into a pressure zone, as they need to have direct control over that pressure (and not "fight" against another VSP for example). With the pressure maintained at a particular node, pressure elsewhere might still be below or above that value, due to differences in elevation and pipe headloss. If you simply have a constraint on your design that calls for a minimum pressure to be maintained, maybe you do not necessarily need to try to force the pressure via things like a VSP, but rather size pipes and use storage and control schemes to accomplish this? 
 Nick Sweitzer said: I also have a question about pump station capacity. The wells that i have added have a total allowable capacity dictated but water rights. Is there a way to make sure those pumps operate at a specific flow? 
 There are ways to force particular flow like using a negative demand or a FCV, but that is not advisable. It is always best to model the system as it actually operates. If the well has an "allowable capacity", you'll need to determine exactly how that is enforced and then model that. For example if there is an actual valve that throttles to keep the flow below a certain value, you could model it as a FCV but that is unlikely as seen in the blog post at the bottom of this article . If there isn't anything in place to keep the flow below a certain value, then that might mean that you need to design the rest of the system in such a way that the flow from the well does not exceed that value. 
 Nick Sweitzer said: The purpose of the booster pumps is to ensure pressure remains higher in the entire system therefore in the actual application (out in the field) the pumps are run on a "constant pressure" mode. 
 So, is only the booster pump working as a VSP? What about the well pump? See previous comments regarding the ability of a VSP to maintain pressure. Even if the well pump is not set as a VSP and the booster pump is, if the well pump is on at the same time as the VSP, then the VSP would no longer have direct influence on the target pressure and the hydraulics would likely not be able to solve.

Jesse Dringoli · Answer

Nick Sweitzer said: When called to operate the well pumps will raise the distribution system pressure when the system demand is lower than the well pumps capacity, 540 gpm. 
 You can create controls based on pipe flow as a condition. Keep in mind that the corresponding action might immediately cause the flow in that pipe to change, so an opposite control may need to be based on a flow far enough away so as not to cause rapid oscillation. 
 Nick Sweitzer said: As the well pumps raise the system pressure the constant pressure booster pumps will slow down and will stop when the system pressure is higher than their constant pressure setting. 
 This is where you may have trouble. The VSP may not be able to properly calculate the required speed needed to achieve the desired pressure, when there is another pump discharging into the same pressure zone. If it does work, then this is indeed how it would operate - the booster pump VSP would turn itself off if it is not needed in order to maintain the desired pressure at the control node. 
 Nick Sweitzer said: Since the well pumps are VSPs but they are set to a specific flow, I may be able to set them to a constant flow setting. The booster pumps, I would imagine, need to be VSPs in this particular scenario. 
 I do not believe that this model will work with both pumps set as VSP - see my previous post and link about only one VSP station discharging into a pressure zone. Having both as VSPs will cause them to "fight". Are the well pumps truly working to produce a specific flow by adjusting their speed? 
 Nick Sweitzer said: as for upgrading the WaterCAD version, I don't own the license, its owned through my employer. 
 I will send you a private message with the names of your Site Administrators, who can help with the upgrade. It is important for you to use a Supported version, so that we can continue to provide you with technical support, so that you can realize the full value of the product (we have added many new enhancements), and so that you can benefit from any fixes and improvements that you might encounter when working with a complex model.

Jesse Dringoli · Answer

Nick Sweitzer said: The pumps are set to a specific flow with a SKADA system. So when those pumps turn on they turn on to a specific flow rate. 
 How does the pump achieve this specific flow rate? If it is not changing its speed to maintain this, perhaps you are actually looking at a calibration exercise (getting the pump flow result to match the measured flow rate). 
 Nick Sweitzer said: Is there a way to model a tank where the water level doesnt change? 
 This would be modeled with the reservoir element, which represents an infinite water source. It essentially acts as a boundary HGL. Or, you could use a tank and set the initial elevation to the desired value and run a steady state simulation.

Jesse Dringoli · Answer

Hi Nick, I am not sure about a VFD that achieves a flow based on a "frequency" and will ask around. 
 You mentioned that the pumps have a "maximum allowable" flow though, which may suggest that the pumps operate at a constant speed with a flow control valve that throttles to keep the flow below the target value...? You you tried the approach of using a constant speed pump with a FCV? FCVs are not used often and in this situation would seem to waste energy though so I recommend reading the following: 
 Using a Flow Control Valve (FCV) 
 (Blog) Death To Flow Control Valves

Jesse Dringoli · Answer

Nick Sweitzer said: I made some adjustments to the model over the last week. I have a question about "tricking" the model. I assume the booster pumps will raise the distribution pressure to a certain hydraulic grade. If I put a tank, that has a very large diameter (so when water flows out the water level doesnt change ever) can i ignore putting the booster pumps in? I'm not running an assessment on if the booster pumps will work or not, they have worked for 10+ years. I want to know fire flow conditions and if the system will meet those parameters. If I put an "Infinite tank" in the system to mimic the constant hydraulic grade, those booster pumps are set to maintain, can i safely assume this is a way to make the system a little less complicated? I'm still having trouble getting everything to work properly and if I can simplify the system in any way and have it still accurately model what is going on, that would be beneficial. 
 Using a reservoir would be the same as using a tank with a huge diameter - it would model a constant hydraulic grade where flow can both enter and leave. 
 If you have pumps that you know will produce a constant downstream hydraulic grade regardless of system conditions, then you could try using a reservoir set to that elevation, replacing all the pumps and upstream piping. This would produce that constant hydraulic grade. You would need to use a check valve on the exiting pipe if you need to model the pump check valve that would prevent reverse flow through the pump in cases where the downstream hydraulic grade is higher. If there are cases where the pump station needs to be turned off, you could have controls to close the pipe instead. 
 Nick Sweitzer said: What are the potential drawbacks of modeling my system this way? 
 The main drawback of using a reservoir is that it assumes the pumps really would always produce that exact hydraulic grade, which may not necessarily be the case during all conditions. In other words, you risk potentially missing out on problems that would prevent the pumps from producing the constant hydraulic grade. 
 Also you would not be able to assess pump energy cost and other details of the performance.

Scott Kampa · Answer

Hello Nick, 
 Can you confirm what scenario you are looking at for this setup. The active scenario in the model provided on Tuesday appears to be a fire flow run. This will help us make sure we are looking at the correct setup for your questions above. 
 Regards, 
 Scott

Jesse Dringoli · Answer

Hi Nick, 
 Some general comments while we wait for clarification about the model: 
 Nick Sweitzer said: How should I model the well pumps? They are set to turn on when the water level in the tank drops down to 77 feet above the floor and to turn off when level raises above 80 feet. However, I cant set them to have a control node as a tank with separate inlet and outlet. As well as I cant have them control a singular node because the 3 pumps are different sizes and have different pump curves (cant run them in parallel). 
 The limitation with the top fill tank and different pump curves is in regards to a variable speed pump which, which set to control a tank, will attempt to keep it at a constant water level. If you want to set controls to turn a pump on and off based on high and low levels, you would instead set the pump as a constant speed (is variable speed = false) and use logical controls to accomplish this. You can use the control "wizard" to create the same on/off controls based on tank level. See: Using Controls, Conditions, Actions and Control Sets in WaterGEMS and WaterCAD 
 You can also see an example of this setup in the Example1.wtg model included in the Samples folder within the installation folder:

Nick Sweitzer said: Now, moving onto the booster pumps...I dont believe a constant reservoir will work for this situation as its more complicated than I originally figured. Im not necessarily concerned with energy cost and performance, but I want those booster pumps to maintain a system pressure that is ~36 psi. Those booster pumps work independently of the well pumps. The booster pumps turn on when a system pressure drops and turn off when the system pressure rises again. The well pumps turn on when the water level in the tank drops. However, the booster pumps have the pressure transducer that is at the bottom of the tank (I believe) and the well pumps are all controlled by 1 set of floats (I believe). The well pumps WILL raise the distribution system pressure because they are adding water to the system. 
 The reservoir approach would produce a constant pressure for the entire simulation, but it sounds like these pumps actually turn on and off, and when they are on, they produce the 36 psi. Is that correct? If so, you could still model it with controls to open and close the pipe leaving the reservoir (to essentially turn the "pump" on and off). I may not be correctly understanding how the system operates though so further clarification on the logic may help.

Problem setting up a booster pump station with a storage tank control node.

Top Replies