Parts of the network are disconnected due to demand shortfall

Question

1777.watercad.zip 
 Hello guys.I have a complete water distribution system.The main challenge I am facing is how to meet the demands at specified nodes after running a criticality analysis tool.Can you guys please provide insights and assist me so that the model above can run successfully.I have set up measures and controls so that the tank will never get empty.After doing a criticality analysis on the model the hydraulic engine outputs unreasonable results.I think my question is, after running the criticality analysis(by unchecking the check valve option) and understanding which segments are facing outages then what next?Is there a way that I can incorporate during the design in order to avoid disconnection of certain segments due to demand shortfalls? I have worked on setting up new calculation options for the engine but still, I get unexpected results when I look at my flex table. 
 Please note that I don't want to use Pressure Dependent Demand or emitter coefficients and instead I only want to work in Steady state mode with fixed demands as they are.None of the elements is disconnected from the main model as seen in the network navigator toolbar.

Mal Sharkey · Accepted Answer

Hi Allan, 
 I don't think pressure dependent demands will help you much, because the pressure at your demand nodes is quite high. I would just suggest these changes: 
 1. Your tank is empty at the start of the simulation, and this causes the "disconnected demand node" errors. Set the tank "Elevation (Initial)" value to a number between the base and maximum tank elevations - e.g. 1,490 m 
 2. Your pump controls should leave a small buffer between the tank-full and tank-empty levels. For example, you could set something like this: 
 IF T-1 Percent Full <=25% then PMP-3 Pump Status = On 
 IF T-1 P ercent Full >=95% then PMP-3 Pump Status = Off 
 3. You should check your pump capacity against the demands. Currently the pump can't deliver enough water from the reservoir to the tank, so, if I run an extended period simulation, the tank empties after a couple of hours. 
 Regards, 
 Mal

Scott Kampa · Answer

Hello Allan, 
 I would recommend review the WaterGEMS Help topic called "Criticality Results," which prove a good resource for the questions you have. This talks about how to review the results from criticality from a number of different analysis types, including cases where PDD are not used. You can find this Help topic by going to File > Help > WaterGEMS Help (in CONNECT Edition) or Help > WaterGEMS Help (in V8i). Choose the Search tab and search for "Criticality Results." 
 Regards, 
 Scott

Jesse Dringoli · Answer

To add to Tom and Scott's reply - I see that you have posted a file, but it is only the .WTG file. The .SQLITE file would be needed in order to open the file. If you are using CONNECT Edition Update 1, you can go to File > Save to Package. 
 If you are encountering a specific error or issue with Criticality, please elaborate on what you are seeing. 
 If you are asking for general advice on what to do after a criticality analysis to improve things, please elaborate on a specific challenge you are having. For example are you seeing a large system shortfall when a certain segment is out of service, but do not know what to do to prevent such a large shortfall? This may largely a matter of engineering judgment. For example there may be a trade-off analysis, comparing the cost to loop the system in such a way to avoid the shortfall when the segment is out of service, compared to the benefit and chance that the segment will be out of service. If the segment in question in the one just downstream of your main pump station for example, it may be very costly to mitigate this, as it may involve installing a backup pump station. In other case, the "solution" to high demand shortfall may be to add a loop; piping that provides another path for flow, to go around the segment that is out of service. 
 You may find further useful engineering advice in the book Advanced Water Distribution Modeling and Management

Scott Kampa · Answer

Hello Allan, 
 There doesn't seem to be a criticality study in the model. Do you have a version of the model where a criticality study is set up? 
 I did look into this a little, creating a criticality study. Since there are no valves in the model, I needed to set this up with the "Consider Valves" option unchecked. Basically, this set each pipe as its own outage segment. When I computed criticality for this, there were indeed cases where the demands could not be met. However, this would be as expected in many cases, based on the layout of the system. 
 For instance, consider a case where pipe P187 is closed. If this pipe is closed, the demand at junction (Point)-75 will not be satisfied. There is no other path for the flow to get to the node. So in a case like this, I would expect there by be times when the demands cannot be satisfied if there is an outage at a given pipe or segment. 
 In the future, if there are valves in the model, you may want to include these, if for no other reason than to minimize the number of outage segments. However, given the linear setup of the system, I would expect some shortfalls. 
 All that being said, I noticed that if you compute the model on its own (not including a criticality study) that there are user notifications related to disconnected demand nodes. If this is ultimately what the issue is, the reason for the disconnected demand nodes is because the tank T-1 is empty. Because this tank is empty, this results in pipe P52 being closed. This link has some details on this. To resolve this, I would change the initial elevation of tank to a value between the minimum and maximum elevation. When you do this and compute the model, the disconnected demand nodes no longer occur. The following link has more information on this issue: Disconnected Demand Nodes user notification when computing model . This link mentions an empty tank being a possible cause of the issue. 
 If the issue is with running criticality, I would still set the tank to a higher water elevation. Criticality uses the scenario as a basis for the analysis, so having a viable solution is needed. Setting the tank level to a higher elevation will still result in outages, since there are some demand nodes with only one path in which to satisfy the demand, but you will at least be computing criticality with a better starting point. 
 Regards, 
 Scott

Parts of the network are disconnected due to demand shortfall

Top Replies