Using a District Meter Area (zone), calculating water loss, and displaying results

TravisAndersonPE

I'm am trying to set up a map in AGOL with DMA (zones) so I can calculate the water loss in these areas and display it. Our basic setup is we have each of our meters GPS'd and that info is updated monthly to AGOL to reflect the new consumption in the attribute table. Our meters are symbolized with different colors to show which zone that particular meter resides in. Master/Zone meters are also color-coded to easily distinguish them from customer meters. I have also created a layer so I could draw polygons that represent each of the zones where we have our meters. I then use "Analysis" to run the aggregate tool, which totals the "consumption" field in the attribute table for each meter for each zone. It works great if you are only looking at zones that are independent of one another. But there are cases where we have one zone down stream from another. Therefore you have two master/zone meters to contend with in the summation of the "Consumption" field. 

The following pic is kind of tough to see, but illustrates my issue. I have Zone 1 in magenta and Zone 2 in green. The master/zone meters are the meters with the yellow crosses where I denoted Meter 1 and Meter 2. If I wanted to calculate the water loss in Zone 2, that's fairly simple visually. Total the consumption of all the meters in Zone 2, subtract that from the Master/Zone Meter 2 and I have my loss. To use the aggregate tool I assigned the master/zone meter a negative value. That worked great if the zones are not connected in anyway. But now I'm trying to determine the water loss in Zone 1 (magenta), and I have the issue of two master/zone meters. The negative in the aggregate calculation will no longer work. I somehow have to subtract out Meter 2's consumption from Meter 1. I initially included Meter 2 in Meter 1's zone, but again, the negative now hurts me instead of helps. Does anyone have a suggestion or idea on how I can accomplish that?

Appreciate in help/advice.

Travis

 

MartinPflanz

Travis,

You are embarking on a subtly complex and fun numerical puzzle to solve.  It may help to take into consideration the details on the variables at play in the water network when solving for an unknown, such water loss.  

• Customer Meter data - Inspect the duration over which the metered flow is averaged.  Volume through the meter divided by number of days / hours / minutes total.  i.e. 200 gallons per day or 6000 gal/month does not translate to a reality of 8 gallons used at a constant rate every hour every day, but rather some peaking (see diurnal curve) has occurred during each unique day.  To improve accuracy here, one would need instantaneous meter read at all meters to have a good handle on what was actually "metered" at a point in time.
• Master Meter data - same approach as above, instantaneous vs averaged, then knowing, since water is incompressible, do the measurement read clock times for the master meter line up with the customer meters? Also, during the master meter sample data period, were any unknowns of water demand occurring?  Was the water department flushing any hydrants?  Was the fire department fighting a fire?  Was a contractor drawing water off to fill a tanker truck through a construction meter? Was there any city/town unmetered irrigation systems that were on at the time? (these could contribute to the overall water loss inadvertently)
• Deciphering Types of Water Loss sources - water loss or unmetered unknown water can be a result of several factors:  meter inaccuracy (+/-5% depending on meters, each meter may be at a different age or accuracy, traditional leaks through pipe cracks, illegal taps, leak past a valve that is not fully seated, very low continuous flow through a meter (aka small toilet leak).  AWWA M36 WATER AUDITS AND LOSS CONTROL PROGRAMS, FOURTH EDITION is a great reference on water loss.
• Leak Behavior - a leak through a pipe crack, is at its basic behavior, just water exiting an orifice which thankfully has a defined equation (search fluid exit orifice loss equation).  To solve for the flow leaving an orifice, one needs pressure (which varies based on where and when in the network) as well as size and type of the orifice itself.  So you can see there may need to be some assumptions made based on pressure at each point in the network and size of leak to measure what the water loss is (assuming other inaccuracies such as meter accuracy and interconnects with the system have been taken out of the picture).  Sum this up by the thought that as pressure increases, water loss flowrate increases to a threshold where the orifice becomes a bottleneck.  
• Leak Location & Quantify - if the instantaneous delta between master meter and all customer meters was known, and making an assumption that the difference is only related to water loss through one or more cracks in a pipe (not other reasons), the number of leaks and location of leaks would remain unknown without further investigation (sonic devices etc fun toys see in action!)
• Interconnections - unmetered interconnections in the network such tees or open valves between metered areas create a path where unmetered water can flow in or out of that network, rendering all of the effort to focus on improving accuracy of any of the above variables for not

With all of these variables at play it, it can end up being a complex task to calculate by hand.  I have seen many water utilities gain understanding of their water loss by investing time in a well calibrated hydraulic model.  Ideally arriving a a model that, run over an extended period of time, closely matches real word conditions.  At that point, some of today's hydraulic modeling software's, given enough known variables can solve for plausible and potential leak regions.  Keep in mind even hydraulic model results can have many unique plausible scenarios that could mathematically match observed pressure and flow rate (ie. is it 2 large leaks or 6 smaller leaks near by).  

Typically I have seen water utilities most effective when they tackle water loss through a combination of approaches: 

• Proactive valve operation programs (to improve resiliency when leaks occur and minimize downtime/water loss)
• Proactive water loss education programs with the public (spot a leak -> report a leak)
• Meter replacement programs (to improve accuracy on the customer delivery points themselves)
• Using a well calibrated hydraulic model to reliably predict system behavior and identify areas to investigate

Esri has many well defined solution to support field operations, so I encourage you to check out the GIS for Water Utilities page and scroll down to the Business Solutions section and check it out.

Talk again soon!

Martin