WaterGEMS Model Set Up For Tank Problem

Question

Hi, 
 I am working on a WaterGEMS model and do not have things set up correctly to model the scenario that I would like to see. 
 
 The model looks like the above image and the setup that I need to run should be fairly simple in my opinion, but I do not have it set up correctly currently. I will also attach the WaterGEMS model if anyone is able to make changes or edits and send back to me with your suggestions. 
 1) The tanks, pipes, and junctions are all located where they should be. Pipe and Junction information is accurate. I don't think I have my tanks set up correctly. 
 2) I need to determine what water surface elevation in the tanks will produce a flow of 9,600gpm at J-5 (each tank has a peak flow of 2,400gpm so the total at J-5 is 9,600gpm). 
 3) J-5 can essentially discharge into the atmosphere (or another tank if the model needs it). I just need to be able to generate 9,600 gpm using only the head available in the tanks. The tanks must all have the same water surface elevation and they are 26'x13'. The maximum water surface elevation in each tank is 5.22'. I currently have everything based off an elevation of 100' so max tank elevation is 105.22'. 
 I'm just not sure where to go from here. Do I need to somehow apply an inflow to each tank of 2,400gpm and run that through the system to see how much the tank fills up? Also not sure if I should be using the Steady State or EPS? The piping for this model is all flat and after J-5 it can be assumed that we essentially discharge into the atmosphere (or a very large tank that will not affect design). Any help on how to get a model like this set up would be helpful. 
 Thanks, 
 WGEMS.zip

Jesse Dringoli · Answer

Joshua, 
 Can you provide more background on why you need to do this? What are you ultimately trying to achieve by modeling this, or what decision do you need the model to help you make? What can you assume about the tank inflow? 
 If the flow is driven by gravity, then it sounds like you need to define the downstream boundary hydraulic grade so that the model can solve the flow generated by the hydraulic grade differential between the upstream and downstream boundary conditions. If J-5 represents a discharge to the atmosphere, you could model the boundary condition with the D2A element . The orifice equation may need to be used to determine the "typical flow" parameter associated with the respective "typical head" (the pair is used to define the flow vs. headloss relationship at the D2A location). The model will then solve the flow in the pipes based on the D2A configuration (orifice size) and the tank hydraulic grade. So, you could then adjust the tank initial elevation and run steady state simulations until the flow is 9600 gpm. 
 Joshua Hesting said: Do I need to somehow apply an inflow to each tank of 2,400gpm and run that through the system to see how much the tank fills up? 
 Tank outflow is based on downstream hydraulics, so you would not simply add inflow to the tank (negative demand) equal to the outflow you want. The outflow in your case would be controlled by downstream fixed demands and/or downstream boundary condition (reservoir, tank, D2A, etc). 
 Joshua Hesting said: Also not sure if I should be using the Steady State or EPS? 
 If you are trying to determine what the tank levels will "settle" on if the outflow is assumed to be 9600 gpm and the inflow (negative demands on the tanks) is a certain value, you could set the initial level somewhere in the middle and run an EPS with a small timestep and a total of 9600 gpm as inflow into the tanks (a good start might be 2400 each, as negative demands). Then, run the model and graph the tank level and downstream pipe flow. Observe what the tank level is when the downstream flow reaches 9600 gpm. 
 There are a few possible problems with this approach: 
 
 Tanks in close hydraulic proximity can be troublesome as described here . If there is a slight differenced in hydraulic grade between them, the numerical solver will generate a certain flow between them in order to balance energy, which can then cause the tank to fill up quickly in a single timestep. This is why a very small timestep is needed if you will use an EPS (instead of a manual steady state approach with trial-and-error) otherwise you'll see rapid oscillation of flow and tank level between the tanks. 
 The tanks cannot be all exactly the same hydraulic grade, because there will be some headloss (however small) in the pipes between them: P-1 through P-7 in the screenshot shown. Tank T-1 will need to be slightly higher than T-2 due to that headloss, which will need to be slightly higher than T-3, and so on. It will likely be close, but not exact, and it would require you to adjust the tank inflow slightly in order to achieve exactly the same "settled" level. Another approach is to approximate the four tanks as one equivalent tank, in which case you can simply add the 9600 gpm as a negative demand on that tank and neglect the headloss in the piping between them (or lump it into the first pipe downstream of the single equivalent tank).

Yashodhan Joshi · Answer

Hello Joshua, 
 Adding a few points for consideration here; 
 1. In steady state analysis, the "Elevation (Initial)" is the hydraulic grade at the tank. You won't see any tank level fluctuations in steady state simulation. For that you would need to run an EPS study. This will help you understand how the tank is emptying / filling. See the article here for more details: Understanding Tank Operating Range 
 2. WaterGEMS is a demand driven software. If you have a demand of 9600 gpm at the junction J-5, it will be satisfied. What can change is the flow through each of the tanks as there are bound to be some differences due to headloss in the connecting pipes as Jesse suggested. 
 If you are back calculating the tank levels or ascertaining tank levels for observed flows you might have to consider a few additional parameters such as; 
 - Is the outflow from each tank i.e., 2400 gpm always constant through time? So, if you ran a 24 EPS scenario will this outflow be constant? 
 - What is the source of those tanks? Any upstream pump/s filling these tanks? What are the inflow rates into these tanks? 
 With this inflow and outflow information you can build mass curves for the tanks in some spreadsheet application like Excel which can help you understand how your tanks are fluctuating through time for the given volume (26' x 13'). Corresponding to this you can determine the tank level at different times. 
 However, this approach is based on the assumptions that there is demand downstream of 9600 gpm which would force 9600 gpm through the common header. For the individual tanks you can have FCV's placed downstream of each tank to force 2400 gpm downstream.