March 31, 2017, 05:15 AM
InfoWater Fire Flow & Fire Flow Test results - Please Help!
Good morning All.
I am having a problem to correlate fire flow results simulated by InfoWater and the fire flow test results collected from the field on a 150 mm Cul-de-sac watermain. Here is the issue I am facing and will be grateful if anybody could help me to sort it out:
1) Field Fire Flow Test - It uses two hydrants, one is flow hydrant and another one is residual hydrant. During the field test it flows the flow hydrant and measures residual pressure at the residual hydrant which is 90 m away from the flow hydrant. Then they provides a plot for flow vs residual pressure that necessary does not show the available pressure at the point of discharge (but 90 m apart).
2) InfoWater Fire Flow Results - We typically assign fire node and assign fire at the selected node. InfoWater provides pressure available at the fire node for the specified fire flow. It also provides available flow at the minimum required pressure at the fire node, which is 20 psi in my case. I believe, InfoWater does not know or care where the residual hydrant is (that used in the field test). Which simply means InfoWater provides fire flow and residual pressure at the point of fire or fire node.
Now the problem is, InfoWater provides fire flow and available pressure at the fire node (flow hydrant where the water is drawn), whereas the fire flow test provides flow from the flow hydrant (location of fire) and residual pressure at the residual hydrant which is about 90 m away. As a result, on a Cul-de-sac road where water is coming from only one direction makes huge difference between the InfoWater results and the filed test results. In my scenario, InfoWater results indicate that @20 psi of pressure we can have only 1100 GPM, whereas fire flow test shows we have a 1840 GPM @20 psi. Please note that the location of measurement of 20 psi of pressure is not the same, but 90 m apart. If we deduct the head loss caused within 90 m of pipe at 1840 GPM is huge for 150 mm pipe, and we end up having negative pressure at the flow hydrant (fire node).
Now my question is how to consolidate these InfoWater fire flow results and fire flow test results in order to draw conclusions whether my water system is able to provide 1840 GPM of fire flow @20 psi pressure. Our ministry, MOECC requires minimum pressure of 20 psi at any point in the distribution system during Maxday demand plus fire.
I will greatly appreciate your feedback and comments on this.
City of Vaughan
March 31, 2017, 09:42 AM
If you are trying to replicate a field fire flow, there are several considerations you need to keep in mind that are also important of why there can be differences between field and model results:
1) You need to ensure that the field and model boundary conditions need to match. If there are differences in either the system demand or what pumps are running or tank levels, this generally will impact the model predictions especially when predicting results to 20 psi. Changes in what tank levels are when the fireflow is run or what pumps are running can have a significant impact in results. We recommend first checking your boundary conditions match as this is often a common cause for differences in results. You need to make sure that both the field and model operations are consistent to compare results. Verifying static (before the flow test pressures) pressures in the model are reasonable are a good initial indication if your boundary conditions are at least close, but not a guarantee of all boundary conditions when the flow test occurs. Make sure demands and boundary conditions match during the static and dynamic (flowing) comparisons.
2) In a field fireflow it is not realistic to use the flow hydrant to get the residual pressure and this is why a residual hydrant is used that is as close as possible to the flow hydrant. However as noted, this is not what the fireflow tool does because the field limitation of not practically being able to measure pressure at the flow hydrant is not a limitation in the model. But if you wish to replicate field result, it is often best to add a demand on the field flow hydrant equal to the flow observed and then compare pressures at the residual hydrant location by using a standard run as this would most closely replicate what was done in the field. The comparison would be done typically on the actual flow and residual pressure observed in the field rather than the field predicted flow to get to 20 psi as this is a direct actual comparison of field measured values rather than projected values. While there are ways in the latest version of InfoWater (12.3 Update 5 as of today) to see fireflow results and the impacts of pressures and flows for different results, it is often easier to simply run a manual fireflow in a standard analysis (or even use the multi-fireflow tool with the flow hydrant as the only junction selected) to do the comparison you are looking for.
3) Essentially what you are doing when you are doing this type of comparison is essentially a form of model calibration. Differences between the model and the field results can also be caused by connectivity differences in either the model or the field. If there is a closed or partially closed valve in the field not closed in the model you will generally see more headloss in the field (and thus less predicted flow at 20 psi) than the model. In your case it is the reverse which may indicate you have pipes that appear connected but are not which limit where water can flow and increase the headloss in the model that are not so limited in the field. The best way to find these in the model is to run the Network Review\Fix tools in the InfoWater-> Utilities menu. The find Nodes in Close Proximity and Pipe Split Candidate tools generally find the issues that can cause the most impact in fireflow results. You can also run a manual fireflow and look to see if any major pipeline that “should” be supplying water is not. Investigate these areas and you often will find either a Node in close proximity or a pipe split that “looks” connected but in reality is not. Fixing any model connectivity issues is essential to getting model and field fireflow results to match within a reasonable calibration tolerance of ~ 5 psi.
4) If all of your boundary conditions and demands match, you have eliminated connectivity concerns, and your model results are still under predicting the fire flows seen in the field, then there is a chance you may need to adjust your pipeline C-factors. However, make sure to eliminate all other factors before you adjust c-factors. Typically published new pipe C-factors are reasonable for pipe materials that do not internally corrode (such as plastic or any pipeline that is concrete or cement mortar lined.) Only unlined Cast Iron or Unlined steel pipe would be expected to corrode internally and be able to justify c-factors less than 120 and certainly less than 100. If you eliminate all other causes you may wish to review your pipeline C-factors used. Comparing the field to model pressure drop form non-flowing (static) to flowing (dynamic) is a great verification of the headloss observed and one of the best indicators of the quality of your pipeline c-factors. Example: Field Static pressure is 90 psi and residual is 40 psi so there was 90-40 or 50 psi of pressure drop observed in the field. How does that compare to the model pressure drop? The pressure drop differences between the model and field should ideally be within 5 psi or less. Contact use at firstname.lastname@example.org if you would care to discuss this further.
5) It is rare that you reach this stage (i.e. have resolved all issues 1-4) and still have significant differences in fireflow results. But, if you can eliminate all other causes you should then look at your pump curve definitions and key control valve setups to make sure these are accurate. With pump curves, be careful of design point curves. These are fine when you operate near the design point but because they use an assumed pump curve shape (shutoff head at 2X the design head and high head at zero at 2X the design flow) they may not accurately represent the actual pump performance away from the design point. Fireflow conditions would most definitely have the potential to push the pump curve away from the design point and could result in inaccuracies in model predictions. The solution would be to have each pump use a multi-point pump curve that had at least 7-10 points on the curve. This generally will best replicate the actual pump behavior along the entire flow range of the pump.
6) One other consideration is the accuracy of your elevation data. Model predicted pressures are calculated from the HGL of that point as Pressure (psi)= [Head (ft)– elevation (ft)]/2.31 (ft\psi). If your elevation data is inaccurate, this will impact the predicted pressure. Ideally, most models use elevation data that is accurate within a few feet. Be careful of using things like free USGS elevation data in areas where the terrain is hilly as this can result in large inaccuracies of the pressure data due to errors in the elevation data.
We hope this helps you with a few additional things you can do to identify what is causing the model vs. field differences in predicted fireflow. We would be glad to discuss this further if it would assist you in a WebEx session as well. If you would like to do this, please contact us at email@example.com and refer to this forum post.
Originally Posted by Vaughan-Mod
March 31, 2017, 09:48 AM
Thanks Patrick for your email great insight. very helpful!