A news story from last October by ABC affiliate KFSN in Fresno, Calif., reported on a public meeting called by State Senator Dean Florez to hear complaints about Pacific Gas & Electric (PG&E) electric bills based on “tens of thousands” of newly installed smart meters. One citizen asked, “How did my bill go from $300 to almost $1100? It didn’t make sense.”
When I asked companies that make smart-meter chips about the complaints, they expressed puzzlement. The meters have specs, they told me: 24-bit precision and accuracy across a wide temperature range. Besides, how hard is it to digitize 60-Hz voltages and currents?
THE STANDARDS IN QUESTION
Meter standards can be found in international IEC 61036 (see “A Measure Of Opportunity Awaits In Electric Meters”). The spec includes environmental requirements such as how much power the meter itself can dissipate and how much voltage it must tolerate, along with performance guidelines such as accuracy and electromagnetic compatibility (see Table 1 and Table 2). But IEC 61036 isn’t the end of the line for meter standards.
According to a white paper published in May by the Electric Power Research Group (EPRI), the American National Standards Institute (ANSI) developed new standards with more stringent accuracy requirements in the late 1990s.1 Under those ANSI C12.20 standards, typical residential solid-state electricity meters must be twice as accurate as the old electromechanical meters. In addition, they must continue to meter down to 0.1 A (24 W), versus 0.3 A (72 W). EPRI also said that 100% manufacturing test and verification is the common practice, and utilities vary in using 100% acceptance testing or sample testing.
METER MAKERS & HYPOTHESES
Landis+Gyr, a long established, global meter supplier in Switzerland, builds PG&E’s smart meters (Fig. 1). The wireless-mesh radio link, which comes from Silver Spring Networks of Redwood City, Calif., operates in the 900-MHz band. According to Landys+Gyr, the meters cost PG&E $100 each.
Assume the chip-based meters are as innocent as newborns. Could the old spinning-disc meters have been under-measuring? Their design dates back to the 1890s, after all. The meter movement is based on a reluctance, or eddy-current motor. The familiar horizontal metallic disc rotates in a field supplied by a permanent magnet. Induced fields from ac voltage and the supplied current create a torque that rotates the disk.
Eddy currents from the disc’s rotation through the permanent magnetic field retard that rotation. The ac line voltage and the power being drawn both affect the disc’s rotational velocity, which is therefore proportional to the power demanded by the circuits connected to the meter. The disc rotations increment a multi-dial clockwork mechanism that maintains a record of energy consumption. Commonly, one disc revolution represents 7.2 Watt-hours (Wh).
Even though the design may have met specs when it was first tested, over time, the lubrication in the clockworks could gum up the works and cause low readings. Countering that hypothesis, I have at least one contrary data point. When we sized our home solar system, we based it on our typical consumption, measured using an old fashioned meter that had hung on the house for at least 30 years. The goal in sizing was to account for a 4× differential in electricity rate between peak and off-peak, while keeping our net outlay for electricity at essentially zero dollars per year.
(Under present tariffs, the utility discounts but doesn’t pay us for electricity. So far, over three years, we’ve “donated” only a few tens of dollars worth of surplus electricity to the grid, implying that estimates based on the old meter have been validated by recent experience. We could only love our smart meter more if it dispensed twenty-dollar bills and made the tomatoes ripen faster.)
At one point, it seemed possible that the bad public-relations results of the meter installations could be attributed to bad communications. The KFSN news story said that in the area most affected, the Public Utility Commission had allowed the utility, which charges based on a five-level rate system, to significantly increase the rates at the top levels—with very little public awareness. Thus, while at level 1, Fresno customers pay only 11 cents per kWh used. By level 5, it rises to 44 cents. On hot days in the San Joaquin Valley, with the air conditioning on, usage can quickly escalate into the higher tiers.
SACRAMENTO, WE HAVE A PROBLEM
For the most part, consumers—at least the ones who’d seen their bills escalate dramatically—weren’t buying those explanations. On May 10, the San Jose Mercury News reported that the number of cutomers with smart meters who received innacurate charges “may be as high as 23,000,” even though PG&E had earlier insisted that only “a few” customers had received incorrect bills.2 While the company could not say “how many of those customers were overcharged, how many were undercharged or the total sum of the inaccuracies,” it was going to conduct “a major overhaul of its customer service efforts.”
That may signal the start of a needed public outreach and display of humility, but consider some numbers. PG&E has already installed 5.5 million smart meters (Fig. 2). It is adding around 10,000 a day. From that base, PG&E has said there were 43,376 cases where smart meters were involved in “some kind of problem.” The company also said 23,000 meters were installed improperly, but didn’t explain how it is possible to improperly install a standard form-factor meter that snaps into a standard service-drop socket. Later, PG&E said that all 23,000 of these were gas meters.
There are some jokes for late-night TV there, but let’s just attribute that to an overly hasty database query. Of the electric meters, 9000 couldn’t connect with the wireless network, and 11,376 didn’t retain consumer usage information. It could be a line-of-sight problem or a quality control problem on the flash-memory side. Neither of those possibilities is an obvious cause for a high bill, though.
So it’s still a mystery. In speaking to the reporters, PG&E did stress that 99% of its smart meters have had no problems. But, the Mercury noted, “One percent still means that 50,000 customers are being impacted, a figure \\[PG&E spokespersons\\] admit is far too high.”
Since this drama strikes close to home, even for utility customers outside of California, readers interested in tracking developments can access PG&E’s progress reports at www.pge.com/SmartMeterCPUCreports. Don’t be put off by the volume of pages (over 600), as it’s a collection of all monthly reports since August 2006. The “Issues” and “Risks” summary pages make interesting reading, but they do not provide any sudden insights into the roots of the problems.
1. “Accuracy of Digital Electricity Meters,” Brian Seal and Mark McGranaghan, EPRI, www.smartgridnews.com/artman/uploads/1/smart_meters_epri.pdf
2. “PG&E Acknowledges Thousands Of Inaccurate Utility Bills,” Dana Hull, San Jose Mercury News, www.mercurynews.com/breaking-news/ci_15057399?source=rss&nclick_check=1