Actual Fire Flow Much Lower than the Ones in the Model

Dear Community

I just had 16 hydrant tests throughout a City of about 30,000 to calibrate the water model.

When I compared the fire flows with the model, actual fire flows were lower by 700 gpm on average, 600 gpm on median. At one location, it was 2,500 gpm lower than the model result.

The hydrant testing company did not do a good job at the first round so we are requesting to re-do the tests (They measured the pressures at the flowing hydrants instead of residual hydrants).

Before the second hydrant testing, I would like to make sure to eliminate anything that can prevent a good model calibration.

What could be the reasons behind the lower fire flows in the field? There is no pressure zone in this City.

Recently, we discovered that the City has several critical locations with significantly lower fire flow than required.

Based on our model analysis, we are making some recommendations for the pipe network upgrades.

Because of that, I chose most locations that are suffering from low C values and low fire flow. All 16 hydrants branch out from 6" watermain.

Should I pick a couple with larger watermain?

Thank you in advance for your input!


  • Remember in your case the residual HGLs at the flowed hydrants are wrong. The real HGL in the main will be somewhat lower (maybe 8-9 psi) because of head loss in the hydrant lateral.

    It is also useful to conduct a heal loss (C-factor) test for old pipes so you have an idea of the lower limit of C-factors you're likely see.

  • Dear Tom and others in the Community

    I did not use the residual pressure data since they did not make any sense.

    To find the available fire flows, I ran a steady state with average water levels at the towers during the hydrant flow tests and the same demands as the one on the hydrant testing day.

    What can be the causes of having higher fire flows in the model?

    All the C-values of pipes have adjusted based on the pipe material and installation date. We put the low C-values for any pipe without such data.

    Tom, are you saying that the major reason for higher fire flow may be due to higher C-values than actual?