View Full Version : Leveraging Hourly Meter Data

March 11, 2015, 01:53 PM
I work for a small utility with an advanced metering infrastructure (AMI) system in place. I have access to hourly meter data of all 5,200 of our meters. I am trying to leverage this data to more accurately create demand data. Is there any way to create diurnal curves for each meter from my AMI data rather than applying a global demand pattern to the entire system?

Anyone with experience leveraging frequent meter data for demand allocation?

Thanks in advance!

Shawn Huang
March 26, 2015, 12:03 PM
I think it is a bit too much to let each demand junction to have its own pattern. The typical way people are doing is to group all meters according to consumption characteristics such as residential, commercial, etc; and deride patterns from different groups of meters.

Stewart Sargent NZ
April 13, 2015, 06:43 PM
Shawn has a good point... and I agree. Also... if you can rationize the meter data into themeatic groups it is always then much easier to:
a. spot anomalies lik;e bad data, leaks, etc.
b. use some of this group patterns for design projections and different loading scenarios.

Basically I alway find it is worth doing some analysis (and then groupings, synthesis etc) on the data using XLS and ArcGIS BEFORE I start using the data for modelling. In other words see what the data itself will tell you.

Hope this helps.

Patrick Moore
June 12, 2015, 05:12 PM

It is possible to make diurnals for each meter, but as noted by others, completing the work to do it may out way any benefit you may gain. Assigning a diurnal for a pressure zone based on a mass balance of the inflow and outflows and water tank changes in a zone is what most modelers do for diurnal patterns. The key is that all of the flows in and out of a zone must be known.

If you do wish to try to leverage the hourly meter data here are a few things to keep in mind:

1) Will multiple meters be assigned to a given node? If so a method to calculate the combined pattern for all meters or a way to easily assign a separate pattern for up to 10 demand sets must be developed. This unfortunately is not easy to automate. If a separate node for each meter is used, it is much easier to automate, but this can create a very large number of new nodes and pipes to the model. Both options have significant challenges
2) Once patterns are created, the biggest challenge is often in getting them in the format that InfoWater wants them. For instance for a 24 hour pattern for just 225 meters, the Pattern detail table will balloon to over 5400 line entries. it also take a little bit of excel wizardry to take a table of values for diurnals and quickly convert those into the format needed in the Pattern Detail and pattern header tables where the patterns are stored in InfoWater.

Incidentally, I have modeled in the last 4 months a recycle water system that had hourly meter data for roughly 225 individual meters. This was the first model I had ever calculated individual diurnals for. Calculating the diurnals was easy, but getting them into a format that could be copied directly into the Pattern table in InfoWater was the greater challenge. InfoWater stores Pattern data in two separate tables. One is the Pattern Header table and one is the Pattern Detail table. It took a fairly extensive use of Excel wizardry using several vlookup features to accomplish this, and it was not without challenges even for 225 meters! We decided to make an individual node for each meter in this model which greatly simplified the creation of patterns and assigning of patterns. But I don't know if this is very practical for a potable water system that may have thousands of users.

One thing your data is very useful for though is to get a sense of how classes of users typically use water for planning purposes. hourly meter data like that could be very useful in planning studies for your system.

Anyway, good luck! It is possible, and you can do it, but it may take a lot of effort and thus may have a high cost without improving model results enough to justify the effort. I suspect it will happen someday, but it has the potential to dramatically increase the data stored in any model that uses it.

As this is an emerging area, there are model tools being developed right now like Innovyze Demand Analyst (product is currently available) that will make this much easier to do in the near future. However, be aware that these type of tools will typically require an additional expense of software to use. Like any expense or effort, a utility (especially a small one like yours) would have to decide if that expense will produce enough benefit to justify the cost. However, tools like this will eliminate the challenges noted above associated with doing this manually.

Best of luck to you. Please let us know if we can be of further assistance at Innovyze Support.

Patrick Moore
Innovyze Support

Eland Afuang
June 14, 2015, 05:48 PM
Hi Ngaue,

You may want to look at InfoWater's DemandAnalyst (http://www.innovyze.com/products/demandanalyst/). DemandAnalyst runs on top of ArcGIS as an add-on to InfoWater using smart metering data to automate the calculation and allocation of base demands and associated diurnal patterns (per usage type).

Kind regards