The mission of this group is to bring together utility professionals in the power industry who are in the thick of the digital utility transformation. 


Advanced Metering Infrastructure (AMI) -- Part 1, Roots

image credit:
John Benson's picture
Senior Consultant Microgrid Labs

PROFESSIONAL EXPERIENCE: Microgrid Labs, Inc. Advisor: 2014 to Present Developed product plans, conceptual and preliminary designs for projects, performed industry surveys and developed...

  • Member since 2013
  • 883 items added with 596,050 views
  • Apr 25, 2018


As I start this series, I have just finished a series of six papers on supervisory control and data acquisition (SCADA) and related systems. In the last paper in this series (link below), I identified advanced metering infrastructure (AMI) systems as one of several that could be reasonably linked to SCADA systems. The reason I am now starting this series has to do with my personal history.

I spent over 25 years of my career working with SCADA and similar systems, but after almost 20 years, I took a detour, and then I got lucky, thrice. I worked for a rather large company (Landis & Gyr) that started a merger with a much larger company (Siemens). At that point I had been in the same job for over 15 years, and thought it was time for something new. I enquired and found about an opportunity in Siemens, applied for this opportunity, and ended up working for an "advanced metering startup" within Siemens. There was just one problem, the startup did not have any products that were suitable for the U.S. electric metering market.

So, we reached out to our contacts in the metering industry and identified a small company that appeared to have an exciting new product. To make a very long story short, we bootstrapped the new company, helped them develop and market their new product, and created a successful product and business. The small company ended up being SmartSynch. This I was lucky the first time.

This was in the late 1990s, and five to ten years before viable AMI systems (as we know them now) emerged. Thus, this advanced metering system was for polyphase commercial and industrial (C&I) meters, not mass-market single-phase meters. However these meters are used for electric utilities' largest and most important customers. We developed features for this system that were extremely advanced for C&I meters of that time, and this served as a model for features that were (and are still being) developed for AMI systems a few years later.

I believe this series will cover the following subjects:

  • The precursor system (roots) of AMI, this paper which covers C&I metering systems
  • The governmental processes that boosted advanced C&I metering and then AMI into major markets
  • AMI systems developed to date (including C&I Metering)
  • Future of AMI and the Internet of Things (IoT) in general.

In the sections below we will explore the functions of the meter data management (MDM) process as it once existed, and then the major advancements in C&I-meter technology that were made shortly after the year 2000. Although I may mention some details of the products that I was involved in, my intent is to focus on the applications, not the details of any specific implementation.

2.Meter Data Management

Intrinsic to meter data management is a utility function called profile-metering. We will delve into this in more detail later, but the basic function looks at a customer's electric energy consumption in a series of short time periods (intervals) of 15- to 60-minutes. The series is called a demand profile and allows a utility to determine the peak demand (highest consumption of electric energy in any interval) and use this as a factor in billing that customer.

Before we dive into MDM, we also need to touch on two major advances in meters that were made between 1972 and 2000.

2.1.The First Smart Meters

As we mentioned earlier, the class of utilities' customers we are focused on are the largest and most important. These customers have a major impact on utilities through their peak demand. I'm not sure when utilities started measuring demand, but it was before I started my career in 1975. At that time, they were using demand recorders that recorded the demand profile on magnetic tape.

In the early 1980s, solid-state demand recorders replaced tape-recorders. In the late 1980s, these moved "under-glass" (became a part of polyphase meters).

A few years earlier (1970s) Bell 103A compatible modems were widely used, but these were large and slow (300 b/s). In 1981 the Hayes Smart-modem was introduced. This included the Hayes command set, which allowed complete control of a dial-up phone circuit, but it was still large and slow. This was the time that personal computers (PCs) were rapidly expanding their market. During this period the modem was pushed in several directions by the PC market: (1) higher speeds (1,200 and then 2,400 b/s) required for file uploads and downloads, (2) design incorporation into more highly integrated chipsets, and finally (3) modems resident on expansion cards for PCs.[1]

This final development (I believe around 1983) meant that these modems could also be put under-glass in polyphase meters, allowing the meters to be interrogated for demand-profiles (and other data) remotely.

Although I would argue differently (in a later paper in this series), many point to these first smart-meters as the senior members of the Internet of Things (IoT).

2.2.MDM Process

The MDM process was developed before the meter product developments described above had fully been implemented. When profile metering via tape-based demand recorders was initially used, metering engineers noted that the data was not perfect. Occasionally one or more demand intervals was missing or corrupted. Fortunately, periodic meter readings were also available. If only one interval was bad, in a profile, and the meter reading before and after the period were available, estimating the missing interval was easy. However if multiple intervals were missing in the same period, how did the engineers fix it?

Eventually the industry defined a formal process called meter data management. This consisted of three sub-processes:

  • Validation: In its simplest form this might be verifying that the sum of the demand intervals matches the difference between the meter reading before and after the period in question. But what if it doesn't? Then we go to the next step.
  • Estimation: This is probably the most complex process. In a utility there are many processes for estimating what the load profile should be for a given customer over a given period (used for load forecasting). If several intervals are missing or suspect, this process was used for estimating what they should be, and this (again) was verified by the meter readings and adjusted if required. But what if the meter readings are also suspect or don't exist due to a meter anomaly? I'm not going there right now, but there are similar methods for handling this.
  • Editing: once we have used standard processes to estimate all load profiles and readings, the incorrect data needs to be edited. By this time the original data is probably in an "invalidated" table in a database in the metering master station. Although editing this should be pretty simple, it would be good if the edited data were automatically validated (again) before it is passed to the "validated table" in order to catch erroneous data entry or transfer errors.

The above "VEE" process is greatly simplified. For a really good and thorough article on this process by a highly qualified author (not me), go through the link below.

3.Advanced Communication & Applications

When we left C&I metering in the previous section it was using dial-up modems to download meter profiles and other data. Even at 2,400 b/s, this was a slow process. When I entered the metering field, large utilities were performing weekly reads. For large utilities with a few thousand C&I meters, this required large modem banks near the metering master. The process for installing, replacing or troubleshooting a polyphase meter was also rather cumbersome. Primary meters are usually mounted at the intertie near or in the primary substation for a large facility. In order to physically access the meter, the utility meter technician must make an appointment, and be accompanied by a facility electrician. If the phone line to the meter needs to be installed, tested or diagnosed, a technician from the incumbent local exchange carrier (ILEC, a.k.a. the local phone company) must also be present.

Diagnosing a communication issue can also be very problematic. In addition to the above issues, the ILEC will frequently insist that the utility test the meter on the bench (requiring a meter replacement) before they diagnose the phone line. When they do diagnose it, they "ring it out" with a single audio frequency in the middle of a full-duplex dial-up circuit's 300–3300 Hz frequency range. However a typical 2,400 b/s modem from this period used 1200 Hz and 2400 Hz frequencies, and a common issue with these circuits is high-frequency fading. So even though the phone technician's "ring-out" works fine, the circuit may not work well at the higher frequencies used by the modem.

3.1.Wireless Communication

Given the problems with phone lines, the meter industry was really ready for wireless communication. When the "Advanced Mobile Phone System" (AMPS or analog cellular) began to be deployed in mid to late 1980s, one of the variants was a "bag phone". Because this provided a dial tone and ringtone (and matched other characteristics of a phone circuit) it would work with the modems in meters and the modem banks near the metering master. Thus, utilities started using these in lieu of landlines.

Although the bag phones were not perfect, they were better than dealing with all of the ILEC issues, and the utility could mostly control installation and trouble-shooting on their own. So, AMPS became the best solution for a while. Meanwhile wireless carriers started converting to digital cellular in the 1990s, and in 2002 the FCC announced that the wireless carriers were not required to support AMPS after 2008. I will not describe the scrambles utilities went through to replace their bag phones before the AMPS blackout (although I was heavily involved in it).

Then there were two-way paging systems. This is what SmartSynch originally used (Motorola ReFlex Technology). Although we were able to build a successful business around this, the technology always had coverage issues. The carrier (Skytel) worked closely with us on major projects, and these were somewhat successful. Smart-phones largely displaced 2-way paging, and it was never able to reach a critical mass. Eventually SmartSynch started offering digital cellular as an alternative.

The current generation of C&I polyphase meters offers both direct digital cellular and AMI network communication under glass.


The following applications for C&I meters were implemented shortly after 2000.

3.2.1.Profile Metering

The primary reason C&I meters exist is to provide detailed metering information, including:

  • Load profiles with 15-minute (or shorter) intervals for power, reactive power (volt-amps reactive or VARs) and other power parameters
  • Registers that store persistent readings of energy (kWh) and reactive energy (kVARh) consumed
  • For utility customers with self-generation, load profiles and registers that measure energy flow into and out of the facility
  • The ability to reset registers via communication, including as part of a register-read transaction (read and reset)
  • Spontaneous reads of profiles or registers reasonably quickly (in seconds to a minute)
  • Spontaneous reports of events, including tampering or outages (see next subsection)

Duration of scheduled read intervals will depend on a specific utility's practices.

One of the most important advantages of profile metering is that it allows tariffs (rate schedules) to change easily. Measuring reactive energy (kVARhr) allows large customers to be billed for excessive uncompensated reactive consumption. Measuring power flow into and out of a facility allows special tariffs and rules to be implemented for facilities that self-generate.

3.2.2.Power Quality

Power quality monitoring should be supported in two modes. The first mode is the spontaneous reporting of serious power quality events, where the meter reports these events without being interrogated.

The second mode allows interrogating each meter for an archived file of power quality events. The maximum number of events that can be stored in this file depends on memory size, and how this memory is allocated. Each entry in the file should be time-tagged and may contain other data (like voltages).

Recommended power quality functions are described below.

Outage: An outage is defined as a voltage drop below 50% on any phase for a user defined time (from 1 to 10 minutes). Outages should be reported immediately.

High/Low Voltage: A high/low voltage event is the voltage deviating from the normal voltage on any phase by a user defined percentage (from +/- 5 to 20%) for a user-defined time (from 1 to 30 minutes). Events and related voltages should be recorded.

Voltage Unbalance: A voltage unbalance event is a deviation of any of the phase-to-phase voltages from the average voltage by a user-defined amount (from 2 to 6%) for a user-defined time (from 15 to 30 min.). Events and related voltages should be recorded.

Momentary Interruption/Voltage Sag: A momentary interruption/voltage sag (MIVS) is defined as the voltage on any phase deviating from the normal voltage on any phase by a user defined percentage (below 80%) for more than 3 cycles (50 ms). A MIVS event will have occurred when a user programmable number of MIVS events (1 to 10) have occurred within a user defined time window (from 1 to 60 minutes). All MEVS should be recorded and a MIVS event should (optionally) be reported.

3.3.Digital Cellular Communication

Digital cellular communication, if properly implemented is extremely reliable in the short term and long term (excellent backward compatibility). Thus, if a meter does not have an anomaly, multiple communication sessions can be used to retrieve all demand intervals buffered by the meter at any time.

Modern digital cellular (3G, third generation or later) is based on code division multiple access (CDMA) technology. This technology is intrinsically secure, and carriers add additional layers of security via encryption. CDMA spreads the output of several transmitters over a single radio-frequency band using a pseudo-random noise (PN) code that is created using a closely-held algorithm. Thus over-the air interception and decoding is virtually impossible. Additional layers of encryption are used when data is sent over the Internet.

Author's Note: This is a message from 2022. I am editing this post since I'm using it in an upcoming post. In the above paragraph, note that 3g was just retired (in 2021) and was replaced by 4g. I wasn't involved in the AMI upgrades required by this technology sunset, but I'm sure it was fun. 


No discussions yet. Start a discussion below.

John Benson's picture
Thank John for the Post!
Energy Central contributors share their experience and insights for the benefit of other Members (like you). Please show them your appreciation by leaving a comment, 'liking' this post, or following this Member.
More posts from this member

Get Published - Build a Following

The Energy Central Power Industry Network is based on one core idea - power industry professionals helping each other and advancing the industry by sharing and learning from each other.

If you have an experience or insight to share or have learned something from a conference or seminar, your peers and colleagues on Energy Central want to hear about it. It's also easy to share a link to an article you've liked or an industry resource that you think would be helpful.

                 Learn more about posting on Energy Central »