Subdivision epanet model help. 1

[notsocivil]

Hi all,

I am a new engineer and I am trying to model a distribution network for a subdivision. The layout is basically a teardrop with spurs coming off the teardrop. The spurs end with fire hydrants. I am using epanet2 to model the pipe line and I am absolutely clueless. The required fire flow demand per county is 2500 gpm. Based on the hydrant flow test next to my connection the available flow at 20 psi is 7053gpm. The results from the flow test at the residual hydrant are static = 55psi , residual = 50psi and a total flow of 2466gpm. I have converted the flow test to a pump curve and then attached a reservoir (hgl = 2.31*55 =127ft) and pump to my network. Is it correct to assign a demand of 2500 gpm to each hydrant throughout the subdivision and check the computed pressure? The total max demand for all the houses is approx. 120 gpm. Also, The pump is connected to a junction that represents a cross I am confused as to what demand to assign to that junction. Thanks in advance for any responses and help.

 

[francesca]

You should just have to assign fire flow to the highest hydrant/at the end of the longest deadend line (i.e. the worst case) and not all simultaneously. However, you do need to maintain regular demands along with fire flow.

As a new engineer, you should be working under a PE who should be able to answer these questions for you.

 

[eddieng27]

notsocivil,
I might not understnad your model set up, but generally you would not use the fire flow results as the flow input to your distribution system at that node. You need to know where the system is supplied and how. In suggesting a "tear-drop" shape, are you saying that you have only one supply connection? You could model the supply conditions with a tank at the supply connection and a hydraulic grade line that you manipulate in a way that your model will produce the observed flow conditions (2466 gpm [or 2500 gpm ... remember your significant figures] @ 50 psi). Also, as "francesca" suggested, when evaluating the system, check that the lowest pressure at any node is at or above 20 psi (that's how you find the "worst case"). If the subdivision dwellings are relatively uniformly spread throughout the water distribution area, you could reasonable distribute the maximum demands evenly at each node (likely won't make much difference for a fire flow in your situation). As an aside, you should question the fire flow test data somewhat as NFPA requires at least a 10 psi drop at the residual fire hydrant; your data indicates only a 5 psi drop. In this situation you'll likley not have any problems using this data because it seems as though this distribution system is quite strong in that the determined flow was achieved with only a 5 psi drop (the flow testers would have likely needed to open more hydrants to get added pressure drop). You might also review the situation of the other components in the water system (supply pumps running, tank levels, PRVs feeding into the system). When modeling fire flows it is usually wise to "stress" the system some by lowering tank levels 10 feet (or possibly half full ... we want fire flows at the end of a fire as well as at the start) and without other supply into the modeled area. Remember that this is a conservative approach so evaulate your results accordingly.