Georgia Power's smart meter rollout in Atlanta has been an absolute mess. They're using Landis+Gyr meters that seem to have firmware bugs causing wild consumption spikes in the data. I've got commercial clients on Schedule TOU-8 seeing phantom demand peaks of 500+ kW during overnight hours when facilities are closed. Georgia Power keeps saying it's "data transmission errors" but these are showing up on bills. Anyone else dealing with similar AMI gremlins?
Georgia Power AMI implementation disaster - anyone else experiencing this?
PPL in Harrisburg had similar issues with their Sensus meters in 2014-2015. Random demand spikes, negative usage readings, meters going offline for days at a time. The billing system couldn't handle the bad data and just estimated everything. We had to file formal complaints with the PUC to get PPL to credit back overcharges. Document everything - screenshot the online usage portal, save monthly bills, track when meters go offline.
AEP Texas had a nightmare rollout in Corpus Christi. Smart meters were double-counting usage during certain weather conditions - something about heat affecting the wireless communication. One client got a $18,000 electric bill for a month that should have been $4,500. AEP eventually admitted the meter defect and issued credits but it took 6 months and threats of legal action. The old mechanical Schlumberger meters were far more reliable.
Indianapolis Power & Light switched to smart meters in 2013 and we're still finding billing errors. The AMI system creates 15-minute interval data but their billing software can't properly handle daylight saving time transitions. Every spring and fall we get clients with demand charges calculated incorrectly during the time change weekend. IPL's customer service has no clue how to fix it - they just keep saying "the computer calculates it automatically."
Alabama Power in Mobile finished their AMI deployment last year. The accuracy is better than the old mechanical meters but now we're seeing new types of errors. Time-of-use periods aren't being applied correctly for customers on Rate OSS. Summer peak hours (2-7 PM) are sometimes calculated as off-peak, saving customers money they shouldn't be saving. It's great for clients but creates audit liability if Alabama Power ever figures it out and wants back-billing.
Entergy Mississippi is just starting smart meter deployment in Jackson. After hearing all these horror stories I'm dreading what's coming. Their current mechanical meters are at least predictable - they might be slightly inaccurate but the errors are consistent. Smart meters seem to create unpredictable billing chaos during the transition period. How long does it typically take for utilities to work out the AMI billing bugs?
Xcel Energy in Sioux Falls had about 18 months of billing issues after AMI rollout. The meter data management system couldn't handle peak demand calculations properly, especially for customers with solar panels. We saw negative demand charges, usage readings that went backwards, bills with 47 different line items that made no sense. Xcel finally upgraded their billing software in 2016 and most issues resolved. Patience and documentation are key.
PG&E in San Jose had massive smart meter problems in 2010-2012 that led to a state investigation. Overbilling, phantom usage, meters that couldn't be read remotely despite being "smart." The CPUC forced PG&E to hire independent auditors to validate meter accuracy. Turned out many Sensus meters had calibration issues right out of the box. Georgia Power might be facing similar problems - suggest clients request independent meter testing if bills seem unreasonable.
Fresno sees similar issues with PG&E smart meters. The interval data looks impressive until you analyze it closely. We found meters recording 0.1 kW usage every 15 minutes for facilities that were completely shut down - no equipment running, breakers turned off. That phantom usage adds up to 35 kWh per month in "vampire load" that doesn't exist. PG&E claims it's within meter tolerance but 35 kWh monthly on an empty building is clearly a meter defect.
Huntsville Utilities finished their smart meter rollout in 2015 with minimal issues. They partnered with EPRI to do extensive pre-deployment testing of meter accuracy and billing system integration. The key was replacing meters AND upgrading billing software simultaneously. Most utilities roll out new meters but keep using 1990s billing systems that can't properly process AMI data. That's where the errors come from.
Idaho Power in Boise just announced they're starting AMI deployment next year. After reading this thread I'm going to recommend all my clients request baseline meter accuracy tests before the old mechanical meters get pulled. That way we have documentation of any pre-existing accuracy issues and can separate old meter problems from new smart meter problems. Prevention is better than trying to untangle billing errors after the fact.