Water CAD EPS modelling 1

[Mike953]

We have set up a distribution model with 4 pressure zones and multiple zone boundary valves between zones. We cannot seem to get the zone boundary valves to operate as they do in reality. Currently zone boundary valves include two valves at each location one set to allow flow if the upstream pressure is met and one to allow flow if the downstream conditions are met. This works for steady state modeling but under EPS modeling we get all kinds of unbalanced errors. The program also looks at the previous run to help set up the next run. So the reservoirs in the upper zones are depleted as the water moves to the lower zones and wells in the lower zones are never called to start. I was wondering if anyone else experienced anything like this.

 

[77JQX]

I don't use watercad, but I do see this all the time, particularly with EPS modeling with a diurnal pattern. You get into a specific demand where the control valves fight each other, and they just cycle through one opening and the other closing. With the software I use, you can vary the solution options (max iterations, accuracy, damping, etc.), and that sometimes helps. But not always. Sometimes your model just sets up an infinite loop at the combination of inputs you've established. And that may in fact reflect reality - I've seen two parallel regulator valves do exactly this in real life. Make sure you have adequate dead bands between parallel valves so that one is clearly in lead, and the other in lag. Still, especially when the valves are far removed, what works with one flow (and friction) and pump combination and tank level set might not work with another.

You can also use steady state modeling to give you guidance into setting boundary valve conditions. You say you have one monitoring upstream conditions (sustain valve?) and another monitoring downstream conditions (regulator valve?). Presumably, only one of these needs to be active at once at your problem flow level. It might be possible to disable one.

Probably the best advice I can give is to consider the purpose of your model, and asking what you're trying to answer with this model run. Consider simplifying the model so that it only answers the questions you're looking at. Look through the individual model iterations where it hangs up, and see if you really need the offending controls active for your purposes. For instance, if you're looking at supply and demand mass balance, you can often do away with the pressure controls. Yes, the output pressures will be useless, but your tank levels will still be right.

Yes, it would be nice to have one model that would work in all conditions to answer all questions you may have. But in my experience this isn't always possible with complex models.

I hope this help. I feel your pain.