I am a software engineer based out of Northern California working for a systems integrator.
My primary focus is app development and overall systems architecture.
I found Neo4J on a a project to try and identify potentially loss of liquid based off interconnected tank flows, and found it a great realtime solution for what I was looking for.
Glad to be here,
Welcome to the community, we're glad to have you join us!
Could you share more about that use case? I understand it may not be possible or permissible to do so, which is totally fine. But interconnected tank flows sound like a classic graph routing problem space, so I'm curious.
Either way, happy to have you here. Let us know if there's anything we can help with.
My client possessed a system that Identified connected tanks, one relationship at a time, and a system that keeps track of the tank flow at any given time based off changes in levels.
We used a graph to connect each of those related tanks into communities, and then used the flow information on each of them to do a mass balance of the entire system. That allows us to identify if any particular network is net positive or net negative in terms of liquid. Knowing that information can point them to a specific subset of tanks that may be losing liquid due to an open valve or something like that.
We group everything into Weakly Connected Communities and then pull all of the data back out grouped by those communities. We then look at the net flow in each community, against its tuning weights and biases to identify if we should trigger an alarm that an area is potentially losing volume.
It currently is just realtime, and I haven't implemented anything yet to add historical data, but that will probably come in the future.