Tank in Series with Known Demands Issue

Question

Hi, 
 I'm trying to setup a scenario where a supply (Reservoir 1) fills Tank 17 at a specified flowrate. A FCV is used on the line between the reservoir and Tank 17 to set the Flow, 168-GPM. Tank 17 fills from the bottom and has a separate pipe for the outflow. The Tank 17 outflow pipe tees splitting flow into two separate directions. The first is to Tank 12 (16,000-ft away). Tank 12 has a separate inflow and outflow pipe. The Tank 12 outflow pipe has a known demand (86-GPM). The second branch for water exiting Tank 17 tees to a junction with a known demand, called Tank 22 Pumped Demand, 71-GPM. 
 The summary graphs for Tank 17 and Tank 12 show that they appear to fill in a logical time period. Tank 17 is the first tank online and fills in 15-hrs. The graph shows it oscillates between filling and emptying until Tank 12 is full, 29-hrs. Everything beyond that generates user notifications which create concern that flow is not actually making it out of Tank 17. Numerous system disconnected errors and pressures below physically possible pressure in system as well as notification that FCV can't deliver flow. The notifications show Tank 17 and 12 both fill but the pipe segments between show no flow in the properties dialog boxes. 
 A scenario was created to have two demands setup on the system as mentioned above. The results are not realistic. The hydraulic grade for the system appears to be completely off and it also shows Tank 12 actually flowing back toward the other junction with a known demand. 
 Using the network navigator to trace downstream, starting from reservoir 1 the highlighted route stops at Tank 17. Are we missing something to have flow coming out of each Tank? 
 Should a scenario like this run as a steady state or EPS. 
 Is there a better way to model Tank 17 filling at a specified rate. Other forums discussed the PSV and the FCV which has been modeled. 
 Any sage advice would be greatly appreciated. We also tried creating a smaller model but run into similar issues. Below is an image of the scenario, but slightly modified in the actual model. 
 Here is a link to Onedrive with the actual waterCAD file if anyone has time to review. 
 https://eesportal-my.sharepoint.com/:f:/g/personal/brett_schlanger_ees_us_com/Eqj2gbBnATRMmyuUQ0k39kcBeBH9o_YpjKi29n4I3uXq4A?e=UfUjwe 
 Thank you, 
 Brett

Scott Kampa · Accepted Answer

Hello Brett, 
 The controls on the tank seem to be hindering that. Instead of using the IF/THEN/ELSE statement, I created two separate controls for each of the two pipes. For instance, for P-10, I have a control that opens the pipe if tank T-17 is more than 5% full and another that closes it if it is less than five percent full. (I did the same for the controls on P-11, though these never really came into play, as the tank never got above 5% full.) 
 With this setup, there will be some flow into T-12. However, as soon as there was any flow out of T-17, that dropped the percent full value for the tank below 5%, thus closing the pipe again. In the mean time, T-12 will become empty again, thus making it so that the demands can't be satisfied. This happened less frequently, but it still occurred. 
 Assuming the controls are accurate, one thing to consider having a higher initial elevation in both tanks. T-12 particularly should not become empty, or start empty, because of the demand downstream. 
 The other thing to review is the flow going into the reservoir compared to what is going toward the demand. You have an inflow of 100000 gpd at node SOURCE. Three-quarters of the flow goes to the reservoir, while only about 25000 gpd of flow moves in the direction of the demand, which is 80000 gpd. This will mean that most of the demand will need to come from the water stored in the tanks. Even if you start the tanks at a higher elevation, you may still see the tanks become empty and see the disconnected demand node message. 
 Regards, 
 Scott

Scott Kampa · Answer

Hello Brett, 
 The tank T-12, which feeds the demand at junction J-5 DEMAND is empty at the start of the simulation. With no water in the tank, there is nothing available to satisfy the demand at the junction. You will want the initial elevation for the tank to be higher than the minimum elevation of the tank. 
 In addition, I would include controls on the tank so that it doesn't become empty again during the run. If that happens, and no other source of flow, the " disconnected demand node " message will occur at other time steps. 
 Regards, 
 Scott

Tank in Series with Known Demands Issue

Top Replies