The 15 Minute Myth: Why More is Not Always Better
Well, it has happened again. I’ve just had an interesting meeting with a client that is more than enthused about the prospect of his company implementing an advanced metering infrastructure (AMI).
Well, it has happened again. I’ve just had an interesting meeting with a client that is more than enthused about the prospect of his company implementing an advanced metering infrastructure (AMI). He was actually salivating about the prospect of what he could do with that 15 minute interval metering data from all of his customers.
It seems to happen more times than not. My clients have become enamored with the marketing of the capabilities of an AMI system. Much like the auto industry, AMI system vendors and meter manufacturers are proudly promoting the capabilities of their advanced metering systems. While they have refrained from making a big to-do over the machismo aspects of owning a REALLY fast Corvette, Porsche or Maseratti convertible, they are very actively promoting their capabilities for Time Of Use, Interval Data, kWh, kW demand, tamper, voltage, and outage reporting.
It really takes me back to my younger days, days when, as a young man, I was pouring over all of the statistics on high performance vehicles in Motor Trend and Car and Driver. Zero to 60 in 3.4 seconds with a top speed of 144 miles per hour. Man, if I could just get my hands on something like that, I could be zipping down the road with lightning fast acceleration and cruising speeds approaching Mach 1. Being blessed in life, I finally achieved my dream, and as with so many dreams, reality set in.
Zero to 60 in 3.4 seconds sure sounds nice, but it’s really tough to attain when that sweet little blue haired lady in the lane in front of you with her left turn signal permanently flashing is accelerating at the rate of two blocks per hour. That laser fast 144 mile per hour top end was kind of difficult to achieve sitting in the longest parking lot in America, the Washington D.C. beltway. On the rare occasion when I could cut loose, there always seemed to be a guy in a uniform with a Smokey the Bear hat and blue bubble gum machine lights on top of his car, giving me very expensive advice on how I could increase fuel efficiency by driving at a slower speed. Insurance was a magnitude higher than a four-door sedan, vehicle maintenance was often and outrageous, and one of the Z-rated tires that were needed for those sustained speeds exceeding 140 MPH cost more than an entire set on an SUV. The cost of ownership was more stunning than the sleek appearance of the car.
The myth of 15 minute interval data is that using it for all of your meters makes sense. All customers are not created equal and with a good number of customers there is no business value in 15 minute data. One argument I often hear to collect 15 minute data on every meter is, “Hey, the AMI system is capable of collecting it and it’s only going to cost me some additional SAN storage to keep the data for future reference”. That is much akin to saying that my car can go 144 MPH and it’s only going to cost me a little more in gas. AMI systems are much akin to that sexy, go fast, car. The dream of 15 minute data for every meter is alluring, but there are several other costs associated with consistently going that fast or collecting that amount of detail on a broad basis.
Let’s take a look at a hypothetical electric utility (HEU). HEU has 500,000 meters. As with most electric utilities, 10 percent of its customer base is Industrial and Large Commercial accounting for 40 percent of its revenue. Both of these customer classes are no-brainers for 15 minute interval data. As a matter of fact if you are going to maximize demand response 15 minute interval data is mandatory for Real Time Pricing (RTP) and Critical Peak Pricing (CPP) rates.
On the other end of the meter class distribution curve we have 10 percent of the meters on a General Service (GS) commercial rate. These are those accounts that you really wonder if it even pays to serve, including railroad crossing lights, school crossings, stop lights, warning flashers, street lights, CATV boosters, etc. - very low load customers that have load shapes that are more flat than cylindrical. Do you really see the need for 15 minute interval data on all of these meters? Daily reads should be more than adequate.
In the middle of the distribution curve we have 70 percent of our meters that account for around 55 percent of the revenue: residential and small commercials. What is the need for 15 minute interval data in this meter segment? A case can be made for 15 minute interval data in these classes of customers. You can also make a much better case for simply collecting hourly data or collecting daily data with a statistical sample collecting 15 minute interval data. Let’s think about it for a moment. A high use residential meter may register 1200 kWh over an average month. That comes to 40 kWh per day. Do you really see value in dividing that 40 kWh into 96 separate readings? Twenty-four hourly readings almost seems like overkill.
Generally speaking, when managing Meter Data Management Systems you will need one staff position for each 3-5,000 meters with 15 minute interval data and one staff position for every 50,000 meters collecting interval data. Given these staffing requirements the following table indicates the required staffing levels.
As you can see, the level of staffing required to operate the system and perform Validation, Editing, and Estimation (VEE) is approximately seven times higher to obtain 15 minute interval data on all meters as opposed to more efficient data collection routines. Based on staffing requirements the most efficient routine would be 15 minute intervals on Industrial and Large Commercial, hourly for Small Commercial and Residential and daily for General Service low load meters. This would provide staffing levels one-sixth of 15 minute intervals for all and require only one-third of the data storage and requirements.
Another issue is throughput on the system. On a daily basis how long will it take to perform VEE and have billing quality data once the data is loaded into the MDM? Throughput is a direct function of the quantity of data, so it would take one-third as long to process the daily meter readings using the 15 minute/hourly/daily scheme than it would take to process all 15 minute data.
Overall, the myth is that 15 minute interval data use for all of your meters makes sense. Not all customers are created equal and the reality of the situation is, you can spend a boatload of money to collect and store a ton of information that will most likely never be used. While there is business value in measuring time of use blocks, taking it down to the 15 minute level doesn’t always make sense unless it’s being used for research purposes. Like that super sleek, fast sports car, 15 minute interval data can be very appealing and tempting. But is it truly going to work for you on a daily, more practical basis? Probably not.
About the Author: Dave Turner is Senior Vice President of Energy, Gestalt, and John Burke is Senior Consultant, Gestalt.