Valid paths not routing upstream trace correctly through features in multiple subnetworks (Sewer Tee junction)

[J_iaxin]

Hi all,

We are currently facing the 001890 bug on invalid subnetwork connectivity. Following up on Tracing flow into multiple subnetwork controllers (similarly in Downstream Trace - Sewer Utility Network) and the guidance in that thread, we implemented the recommended valid paths approach on our directional manhole terminals, but the results do not match our expected results.

Context

We are implementing a gravity sewerage utility network. In several locations, a single manhole acts as a subnetwork controller where multiple subnetworks (with different subnetwork names) are topologically connected through the manhole, which reflects real flow/connectivity in the field.

An example in the network:

Prior guidance in the linked thread included:

1. Assign the same subnetwork name to controllers that should trace together — not applicable here; these are genuinely distinct subnetworks that must remain connected.
2. Break traversability between controllers — not acceptable for our use case; operations require the current connected topology.
3. Use a directional manhole with valid paths to control internal flow between terminals (upstream/downstream/crossover) 

We tested Point #3 with our Crossover valid path (instead of Normal [All]), which is similar to what was suggested to us:

Esri’s documented scenario (from the thread / product guidance) assumes a directional manhole with multiple upstream and downstream terminals (e.g. two upstream + two downstream). However, in our case, we only have 3 ports, where the flows separate downstream.

Observed results

After switching valid paths to Crossover (confirmed applied successfully on the terminal configuration):

1. Upstream trace from the manhole/controller area only consistently recognises one subnetwork side (Kranji in our test case, seen in the screenshot at the start).
2. Upstream trace for the other subnetwork (Changi) stops prematurely — it does not continue to the upstream sewer segment that is topologically part of both subnetworks through the manhole.
3. We attempted an alternative terminal assignment on the Changi subnet downstream sewer (Outflow 1 → Inflow 1). That change causes subnetwork validation failures because the subnetworks are no longer self-contained.

In our case where traversable controllers with different subnetwork names (001890 scenario) that must remain connected, is there a supported way to obtain correct upstream and downstream traces for both subnetworks through a shared manhole/terminal?

[RobertKrisher]

 I suspect there are multiple things happening. Let's start with something simple. Connect the lines on that manhole as shown below, configure an upstream trace with the "use digitized direction" option, and run the trace. Then post a screenshot of the resulting trace (it should look correct).

[J_iaxin]

Ah, unfortunately we're still on UN version 5.

I believe the "use digitized direction" option is available from UN v7 onwards? Is there another test I could try instead?

[RobertKrisher]

It sounds like there is a loop upstream of this manhole (in actuality there might be many loops).  You can use the shortest path trace to help you identify it.

I recommend you copy your data out to a mobile geodatabase, upgrade it to v7, and test out the tracing with digitized direction. If you can get the digitized direction trace working, it will require significantly less data cleanup than trying to get all the terminals aligned to make consistent subnetworks. You can then use something like the Batch Trace tool or community sample to track your sewersheds.