This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

Skeletonization

Based on the law of continuity if there is a series of pipes with 10 nodes for example and there is no demand at any of the nodes on the pipe the flow that passes through the 1st node in that pipe series would be the same as that passing through the last node in that series.

But then say if node 5 in that same series of pipes has a demand. Then the flow that passes through the last node would then be the flow at the beginning of the pipe series minus the demand at node 5. 

Or in other words when you assign a demand to a node that node then becomes a point where water is viewed by the model as being consumed. 

Question?

Now say in reality the water is not consumed at node 5 but there is a branch downstream of node 5 with multiple customers which produces this demand at node 5. If it's the case that I do not need to know the pressures/pressure drops within the branch downstream of node 5 I could simply assign the demand associated with that branch to node 5 and eliminate that branch from my model right? 

Or in other words what you need to consider when deciding whether or not to remove a branch is whether  or not you will need to extract information (pressures) from that branch. If the answer is no you could pretty much remove that branch from the system as long *** you assign the branches demand to the upstream node right ? 

  • how much demand can you distribute among nodes before it significantly affects the accuracy of your model?

    It sounds like you may be essentially asking about model accuracy vs. level of skeletonization. I do not think there is a single answer for this as it depends on the system, what you need to accomplish with the hydraulic model, and the degree to which you simplify the network.

    Using your example (I cannot read the text in your screenshot) if you add together the demands from the two junctions in the upper left and upper right corner and place the sum as a node at the crossing point, then the flow in the upstream pipe would be the same as it would have been if you explicitly modeled the two demand nodes, so the pressure at the node at the crossing point should be the same. But, you would not be able to check the pressure at the two demand junctions that are no longer in the model. If the headloss is relatively low and the pipe is relatively short and/or large and the junction elevations not significantly higher, then this may not be significant. You would ultimately need to use your engineering judgement and use the functionality of the software to try different levels of simplification (or demand distribution) to get a good understanding of the impact on the results.

    Here is a related article: Is there a recommended level of Skelebration for different modeling conditions?


    Regards,

    Jesse Dringoli
    Technical Support Manager, OpenFlows
    Bentley Communities Site Administrator
    Bentley Systems, Inc.