How to model pipe material ("event") changes between structures?

COSPNWGuy

I am really interested in hearing from the community how you are doing this. We generally split our water, wastewater, and stormwater mains at structures (i.e. fittings, tees, manholes, valves, catch basins etc.) which seems to be the main way the industry does it. We have always struggled with how to handle pipe material (even diameter) changes that occur in-between these features. This is happening more and more as construction costs increase and replacing whole pipe runs is becoming cost prohibited. Our pipes are turning into a patchwork quilt off different materials. It is very common to see a pipe segment composed of both concrete and PVC in-between two manholes. We have tried splitting mains at these transitions and track it in attribution, however people don't like all these "ghost" points that don't represent a structure\feature. I'd love to know if the utility network has any tools or ways of handling this (we have not implemented yet - we are still using geometric networks). The way I would like to solve it is by some kind of linear referencing system per pipe run, so that "events" whether it be material changes, diameter changes, and even different project "footprints" could be referenced without having to "slice and dice" our linear features. Thoughts\advice anyone? What are you doing to account for this in your GIS?

MattKennedy

Greetings, PNWGuy. The Esri Water Team looked at your question and we have these thoughts:

You don’t need to create ghost junctions between the sections of pipe as long as they are the same asset group/asset type. Connections between lines with the same asset group/asset type don’t require a junction in the utility network. To maintain the changing material types along the lines you will still need to split them into segments and attribute as needed. Changes in diameter typically have some kind of real world junction feature, that could be captured in the utility network but does not have to be as long as you have the same asset group/asset type between the line segments.

There is no industry standard for when you should or should not split your pipe segments. If you choose to spit the lines less often you will have fewer features in your data set which can be beneficial in that you limit the overall size of the dataset. However, if you choose to split the lines more often, you will have more granularity in your data and have more detailed (accurate) results when running analysis such as traces or reports on pipe defects based on material, diameter, pressure rating, etc...

Regarding establishing a linear referencing system to model pipe events: This would be extremely complicated to manage in a water distribution network. If it were a couple large transmission mains it might not be that bad, but otherwise we would not recommend this.