The benefits of Advanced Metering Infrastructure (AMI) can be wide ranging, but they stem from a complex and long-term value chain and are easily misunderstood. And easily underestimated.
An interesting study regarding peak demand reductions after implementation of AMI and Time of Use (TOU) rates at a utility in Ontario was recently completed by the University of Waterloo.
I will explain here why I think it is misleading that the study led Science Daily to publish a summary headlined “Study shows big smart meter investment yielded 'very small' electricity savings.”
The study, which dealt with residential customers only, found a 2.6% reduction in on-peak demand (and 2.4% reduction in mid-peak demand) after implementation of TOU rates. The study factored out impacts of variations in weather on electricity demand across the time intervals before and after onset of TOU rates, utilizing advanced statistical algorithms.
It is fascinating to see the hype generated by the study (including U.S. utility trade press outlets that repeated the Science Daily headline, which make the study’s conclusions sound universally applicable), in comparison to the relatively humble scope of the study’s conclusions.
After reviewing the full study and corresponding with its authors, who kindly gave me clarity about details not initially clear to me, it is fair to say that the conclusions of the study are fully justified only for the 20,000-meter utility in Ontario for which the study was conducted--and only if that utility considers a 2.6% to 2.4% reduction in demand to be “very small.” This remains an open question because the utility involved is anonymous.
The study’s authors acknowledged in their conclusion that “no single residential electricity demand model is universally applicable,” and added that “policy makers should insist on clear, reproducible methodologies…The results of an analysis should only be considered reliable if adequate supporting metrics are provided.”
While the conclusions of the study may also be helpful elsewhere in Ontario where AMI and TOU and Demand Response programs are fully in place, similar applicability for utilities in the U.S. is doubtful, given how the level of use of AMI is lower, let alone the level of residential TOU and DR.
I asked Entergy’s Raiford Smith, VP of Energy Technology and Analytics, about the applicability of this type of study of one utility’s TOU and DR results, for other utilities, He said: “There are so many factors that differentiate one utility’s situation from another, it is hard to generalize about the value of peak demand reduction. For example, the amount of load reduction could vary based on the nature of the customer’s load, weather, the type of demand response program, the ratio of peak to off-peak TOU pricing, variations in consumer awareness, and the size of residential electric bills relative to overall household income. Any of these factors could significantly impact the achieved demand reduction.”
When asked whether he would consider a 2.6% peak demand reduction “very small,” Raiford Smith said, “Any time resources are constrained, a reduction in peak demand has value. After all, at $1,000 per MWh, any reduction will provides some level of economic relief to the system. If the grid is under a lot of physical stress, most system operators will tell you they’ll take all the load reduction they can get. As such, one could easily be happy with a 2.6% reduction, depending on the location of the load reduction and the capability of the grid to meet customer demand.”
AMI vs. Demand Response
Some AMI-related misunderstandings and benefit estimation shortfalls of recent note seem to stem from mismatches between AMI and Demand Response, as well as TOU.
As has been highlighted in a T&D World article our Smart Utility column titled “How Well are Utilities Leveraging the Potential Value of AMI Data?,” the penetration of peak-reducing customer engagement programs, including demand response and TOU, not only involves a smaller set of utilities than the set involved in AMI, it also, for each utility where both AMI and customer engagement programs are in place, involves a lower penetration of customer engagement programs than of AMI—to the tune of single vs. double digit percentages.
When the meter-reading function is replaced by AMI, the employees previously involved in meter-reading typically remain in the employment of their utility, so there is no immediate reduction in payroll expense associated with automation of the meter-reading function.
We should also consider the situation regarding remote shut-off of meters once AMI is in place—in some states, such as Massachusetts, some utilities installed AMI along with a new capability to perform remote turn-on and turn-off of meters, but were still required due to regulatory concerns, to dispatch crews to knock on doors and let customers have one last chance to pay their delinquent electric bills. (Admittedly such concerns are reasonable, e.g. considering elderly customers and those relying on electricity for heat in regions with cold winters)
The fact is, the path upon which AMI and related initiatives have set our industry is the right path, but it is a long path. Some of the supposed shortfalls will be viewed in the future as equivalent to someone criticizing the supposed efficiency improvements to be expected from Personal Computers, within months of PC’s having been invented.