Has anyone else noticed data quality issues with Nebraska Public Power District's AMI system? I'm seeing gaps in 15-minute interval data and some readings that seem way off baseline. One client shows zero usage for 6 consecutive intervals on January 8th, then a massive spike that's 10x their normal demand. NPPD claims the meter is functioning properly, but this doesn't pass the smell test. How do you challenge utility meter accuracy when they insist everything is working correctly?
NPPD interval data accuracy concerns
Greg, I've dealt with similar data quality issues with PPL here in Pennsylvania. The key is documenting everything methodically. Request the complete interval data export, not just summary reports. Look for patterns in the missing or erroneous data - does it always happen at the same time of day or after weather events? Also request the meter testing reports and calibration certificates. If you can show systematic errors, utilities are more likely to investigate properly.
We had Duke Energy replace a faulty AMI meter in Cincinnati after showing them data that was clearly impossible - negative demand readings and consumption spikes during confirmed outages. The trick was correlating the bad data with external factors like weather and known operational schedules. Can you get your client's maintenance logs to verify what their actual usage should have been during those zero-reading periods?
Idaho Power had similar AMI issues about two years ago. What helped was requesting data from multiple meters in the same area to see if the problems were localized or systemic. If other customers are experiencing similar data gaps at the same times, it points to communication or system issues rather than individual meter problems. Also check if the missing intervals are being estimated by the utility and how that affects billing.
In Oregon, we've learned to always compare interval data exports with the actual bill calculations. Sometimes the raw data looks wrong but the utility is using estimation algorithms to fill gaps before billing. Other times, the data is being processed incorrectly. Eugene Water & Electric had issues where communication failures created false zero readings, but their billing system was substituting estimated values. The customer never saw the real problem until we dug into the detailed interval files.
Great suggestions everyone. I requested meter data for the entire transformer bank and found three other customers with identical data gaps on the same dates. That was enough evidence for NPPD to investigate their communication system. Turns out there was a faulty relay that was causing intermittent data collection failures. They're replacing the equipment and will restate bills based on estimated usage for the affected periods.
Greg, that's a great outcome. We see similar communication issues with Dominion Energy here in Virginia, especially during ice storms or high wind events. It's worth keeping a log of weather conditions when you notice data anomalies. Pattern recognition has been key to getting utilities to acknowledge systemic problems rather than isolated meter issues.
Sandra makes a good point about weather correlation. Madison Gas & Electric actually proactively flags interval data collected during severe weather events because they know communication can be spotty. It would be nice if all utilities were that transparent about data quality issues. Most of the time we have to play detective to figure out what's really happening.
This thread has been really helpful. I'm dealing with similar Xcel Energy data gaps in Minneapolis and will use these strategies. The weather correlation angle is particularly interesting since we've had several ice storms this winter that coincide with missing interval data. Sometimes the solution is simpler than we think - just need to connect the dots properly.
Ten years later and we're still fighting the same AMI data quality battles! National Grid here in Rhode Island has gotten better about acknowledging communication issues, but it still takes solid documentation to get them to investigate. This thread should be required reading for anyone doing interval data analysis.
Great historical perspective Anthony. Here in Idaho, we've learned to always request multiple months of data to establish baseline patterns before challenging any anomalies. Idaho Power is generally cooperative, but they want to see comprehensive analysis, not just isolated data points. The documentation approach discussed here has been our most successful strategy.