I am attempting to model a 1-hour fire flow via EPS analysis. I have created a pattern in which the multiplier (flow) at a specific hydrant is zero up to 9AM. At this time, the multiplier switches to 1. Then at 10AM, I switch the multiplier back to zero to simulate the 1-hr fire. In the graph view of the pattern the multiplier jumps up to 1x at 9 AM in a segmented matter (aka does not gradually slope up) as it should.
However, when looking at the demand of the hydrant in an EPS graph, from 8 AM the demand constantly increases from zero until the actual imported demand value is reached at 9 AM (1x multiplier). Then, it follows the same slope in the negative direction until it reaches zero again at 10 AM, thus using the same amount of water as the 1-hr fire but essentially over a 2 hour period.
I understand that in the pattern creation tool I can set the multiplier to 0 at 8:59 AM so the demand jumps up to 1x right at 9 AM for an EPS analysis, but what is the more realistic scenario of the two? Or realistically does it stem from a combination of the two (i.e. demand gradually increases for 15 minutes until it reaches the required fire flow)?
Realistically, a firefighter will open or close a fire hydrant in about a minute.
Bentley Technical Support
If you want to capture the behavior of a 500 gpm fire that lasts one hour, set up a stepwise pattern like this.
0.900 0.0001.000 500.0001.990 500.0002.000 0.0003.000 0.000
And set the hydraulic time step to something like 0.01 hr. You will get a beautiful square shape of demand vs. time.
If you only use a 1 hr time step, the model will only calculate at the hour and it will give you those sloping graphs that aren't a good representation unless all you care about is what happens exactly on the our and can interpolate between those times.
Answer Verified By: Kyle Blommer
This link has information on the steps that Tom mentions above. An EPS Fire Flow Analysis Tool was also included in the CONNECT Edition releases of WaterCAD and WaterGEMS. I would recommend reviewing this tool in case it is something that will work for your modeling case.
Thank you all for the advice.