Open operation (One Pump running and second pump starts 60 sec delay)

Question

Dear Team, 
 
 I have a scenario, in which one pump is running and seconds pumps start 60 sec later. I have not considered the pump tripping and i need know will there be any pressure variation during the starting of second pump. When i run the analysis for 300 secs, after 60 sec the pressure goes more than the pump head and continue for 300 secs. Theoretically when second pump starts after a delay, a sudden pressure peak should come at 60 sec & continue for few more seconds and then the pressure has to be less or equal to the pump head, but it is not happening. Please clarify why the pressure is (more than the pump head) continue till my analysis? 
 Note: Pump pattern is assign for starting of second pump.

Regards, 
 Abdul

Scott Kampa · Answer

Hello Addul, 
 This may be related to the specific speed used for the pumps. In the transient analysis, the program uses the pump results for the head and flow (or the nominal head and flow for a pump startup event) as the basis for the four quadrant curve defined by the specific speed. If the specific speed is not accurate, this may account for he difference in results. More information can be found here: Specific Speed and Pump Flow Calculations for a Transient Simulation. 
 You can find some additional information on why the pump flow and head may not be settling on the expected results by looking at the troubleshooting section of the pump start-up wiki . 
 If that doesn't help, we may need a copy of the model files to look into this further. 
 Regards, 
 Scott

Jesse Dringoli · Answer

Hello Abdul, 
 When you say "more than the pump head", what do you mean? Generally speaking when the second pump turns on, the pump that was already running will shift its operating point accordingly, and both pumps will tend to run at the same head (assuming the pumps and the parallel piping headloss is roughly the same). 
 If the second pump starts too quickly, I would indeed expect a pressure surge as the pump first starts up, and then eventually the pressure would settle down to the expected "nominal" operating point. Meaning, the flow and head should match the flow and head seen when running a steady state simulation with both pumps on. So, I would recommend that you first try doing this (if you have not done so already), to observe the pump operating points in a steady state condition. Then, be sure to use the observed head and flow for the pump in question, as the "head (nominal)" and "flow (nominal)" input fields for the pump (which are required when selecting the pump start transient pump type). 
 Next, consider the "control variable", which is used to control the "startup" by torque or by speed. See more about the differences here: Pump Startup occurs too quickly / initial upsurge too severe 
 Even with this configured correctly, there are a few other factors that can occur in the model to cause the final operating point of the pump (at the end of the transient simulation) to not match the expected operating point. See the eight numbered items in the following article under the section "Why is the Transient Simulation not Settling Exactly on the Nominal Head/Flow I entered?": Modeling a Pump Start-Up Transient Event in HAMMER 
 If this does not help, please provide a copy of the model, exact steps to replicate the problem, and a description of what you see: Sharing Hydraulic Model Files on the OpenFlows Forum

Scott Kampa · Answer

Hello, 
 This appears to largely be occurring because of results during the initial conditions. You currently have local high points in the model and downstream reservoirs that are at a lower elevation than these high points. As a result, the pump that is on is unable to add enough head to the system to raise the pressure above the high point. If you were to look at profile DM-PLANT-3 (either by going to View > Profiles or looking at the transient profile at the beginning of the simulation) you will see that the hydraulic grade is below the ground surface elevation. 
 The typical solution for this would be to include air valves to set the property field "Treat air valve as a junction?" to False. What will happen is the pump will see the air valve as a boundary and will often add more head to the system. I will recommend reviewing the following link for more information: Modeling Air Valves at High Points in WaterCAD and WaterGEMS . 
 I would recommend making sure that the initial conditions are set up correctly before computing a transient analysis. In your model, you have negative pressures at the high points. Is this actually occurring? Is the pump head and flow correct compared to field data? 
 If you decide to include the air valves in the initial conditions, there is a section in the link above with information on considerations when using HAMMER. 
 Regards, 
 Scott