Fire Flow Nodes Failing Despite Constraints Seemingly Met

Hi Kristy,

Are you using the default solver version, max trials and Accuracy in the calculation options?

At first glance it almost seems as if it is somehow seeing a 20 psi zone pressure constraint even though you have it set to zero, and stopping/failing because of the pressure at J-55. Did you perhaps set a local constraint at some point and switch back? Does it help if you globally set all nodes to use local constraints, then global edit them all to the same desired settings?

Note that we have a new version with some improvements specifically around fireflow, which you can read about here: What's new in WaterGEMS and WaterCAD CONNECT Edition Update 3.3 (10.03.03.72)

If this does not help, please provide a copy of the model for review: Sharing Hydraulic Model Files on the OpenFlows Forum

Hi Jesse,

I had not changed the local constraints, but I tried changing them as you suggested and changing them back, but it didn't help. I have attached my model.

Interconnection.wtg.zip

Thank you

Thanks for sending the model, Kristy. This was a nice fun one for a Friday afternoon.

It turns out that the model is experiencing an ill-conditioned matrix between numerical trials due to air valve AV-4 which is internally treated as a PSV because you have "treat air valve as junction?" set to "false". Even though the pressure is well above zero and the air valve is not "opening", there is ill-conditioning occurring as described and illustrated in the following article under the section "Single PRV or PSV with downstream demands" - Ill-conditioning User Notification for valves in a hydraulic model

When this happens, because the PSV is an internal element used to model the actual air valve, the ill-conditioning notification and run failure is not reported back to the user interface, but this is what is stopping the fireflow nodes early. You can see this if you morph the air valve into a PSV with the same setting and attempt to compute a standard steady state.

The solution to this is to set "Treat air valve as junction?" to "True" for AV-4. The pressure is high enough that it would not be opening anyways (at least note with the pump on). If you must keep this air valve open (perhaps for other scenarios?) then you can try changing the Accuracy calculation option to 0.0001, or use a different option from the "Engine Compatibility" calculation option. Regardless, make note of the situation described in the above article.

I have added a note about this situation to the following troubleshooting article: Understanding Automated Fire Flow Results

Lastly, I will report this model to our developers to see if the program can be amended in the future to handle this situation better or make the user more aware that this is the cause of the failure.

Thanks Jesse, this fixed it.

Thanks Jesse, this fixed it.