PONDPACK ICPM TW and HW increment

Question

Hello Guys, 
 I have a complex ICPM and I got some warnings like "mass balance for routing volumes vary by more than 0.5%" and when I made a small Changes on Tailwater or Headwater increment, the flow and WSE changed significantly.And some of the results don't make sense. does anybody knows how to findout the best TW and/or HW increment for multiple ICPM? I am trying to attach the file 
 thanks

MPachlhofer · Accepted Answer

Adane, 
 I was able to narrow down part of the problem in your model setup. The instability that you see for the flow in ponds PO-1, PO-18, PO-19, PO-6, PO-17, and PO-2 is coming from the fact that you have 3 no volume ponds in series with very large channels connecting them. This can prove challenging because for small changes in head there are large changes in flow that occur, which makes the model results very sensitive. When this happens for three consecutive no volume ponds and outlet structures the solution for model becomes difficult to converge on. The instability that is characterized by the sharp up and down shape we see if the hydrographs is caused when the solver is iterating over the flows of the three ponds trying achieve a flow balance. Sometimes making the ICPM time step very small can help the model converge, but in this case, even with a small time step of 1 second, there is still a rapid fluctuation in flows because of the combination of channel sizing and consecutive similar configurations. The good thing is that when the flows get to the final outfall O-5 they are stable and the hydrograph is smooth. 
 In this situation I&rsquo;ve tried a few different workaround that involved different ways of connecting all the flows upstream of O-13 into the PO-3 pond, but the different configurations that I attempted to use changed the peak flow on the O-5 hydrograph and had some variance in shape on the receding from the original hydrograph. Based on the aforementioned information these workarounds probably wouldn&rsquo;t be viable in this modeling situation. 
 In similar cases using CivilStorm with the dynamic wave solver might be a better option for modeling this system because it was designed to handle more complex networks. What you could do in CivilStorm is model the no volume ponds and culvert outlets as actual channels with cross sections that define each change in the channel and that should be more stable than PondPack. 
 
 Regards, 
 Mark

MPachlhofer · Answer

Adane, 
 
This wiki article should answer all your questions and help you address the mass balance user notification that you are getting. 
 
 communities.bentley.com/.../9299.pondpack-user-notification-mass-balance-for-routing-volumes-vary-by-more-than-0-5 
 
Please let us know how the solution works for you. 
 
Regards, 
Mark

MPachlhofer · Answer

Adane, 
 
In my previous post I forgot to mention a things that I noticed. It appears that you're using version 08.11.01.51 of PondPack. The latest version is 08.11.01.56 and has some updates to it that might be crucial for getting this model to compute correctly. Here are some instructions for how you can download the latest version of PondPack. Make sure you download and install the latest patch set with that too. 
 
 communities.bentley.com/.../8175.how-do-i-download-the-water-and-storm-sewer-hydraulics-and-hydrology-products 
 
I also wanted to point out that the "reverse flow conditions encountered for one or more headwater elevation" is purely an information message and it's not something that you need to be concerned about. It's just there to let you know that you have some reverse flow occurring in some of your outlet structures. 
 
I'm still looking to see if I can find anything else in your model that needs to be addressed. 
 
Regards, 
Mark

MPachlhofer · Answer

Adane, 
 
Based on what I am observing your continuity error is due to your irregular cross sections shape. It looks like they are very small cross sections as far as depth of the channel and it's also very narrow, which is going to make it difficult for the implicit engine to find solution with a good continuity error for your model, especially when the flows become larger as the storms increase in intensity. If you make a profile for your system that starts at O-2 and goes to PO-1 you'll notice the hydraulic grade line shoots up starting at CS-3 and continues escalating through CS-1 because of their narrow shape. I have found a few changes to the calculation options that will help. 1) Change your calculation time step to 0.008 hours or 0.5 minutes and do that same with the calculation time step. 2) Change your computation distance to 0.5 feet. That should help lower the continuity error a little. 
 
I've also tried to changing the cross section type to a trapezoidal shape with a bottom width of 10 ft. and height of 10 ft. and that seems to help with the continuity. I'm guessing that the way you placed the cross sections in your model is the way they are in the field, so that might not be an option for you, but it helps with the continuity. 
 
Regards, 
Mark

MPachlhofer · Answer

Adane, 
 
I would expect the continuity error to drop significantly while just using a free outfall because it's an easier computation for the engine to settle on a solution for. Adding a tidal outfall that has reverse flow conditions adds complication to the model because it introduces another level complexity the solver has to workout. Rather than answer you last question let me ask you a few questions to get some more information. 1) Are the cross sections in the field as you entered them in your model or are they different? 2) What's the goal of your project? Are you modeling an existing system to see how it functions or are you modeling the system in hopes of designing or re-engineering part of it? 
 
Regards, 
Mark