Working on a manufacturing client here in Indianapolis and their IPL interval data is showing these weird 15-minute demand spikes that don't match their actual production logs. The spikes are hitting 847 kW for exactly one interval then dropping back to normal 420-450 kW range. This is happening 3-4 times per week on their Schedule LGS rate. Plant manager swears nothing is cycling on/off during these times. Anyone run into phantom demand spikes like this before? Could be a metering issue but IPL is saying the data is accurate.
Weird 15-min spikes showing on IPL demand profile - anyone seen this?
Greg - I've seen similar issues with older GE meters. Sometimes the CT connections get loose and you get these false readings. The fact that it's exactly one 15-minute interval suggests it might be a communication glitch between the meter and IPL's head-end system. Have you requested the raw meter logs directly from the site? That would tell you if it's happening at the meter level or in the data transmission.
This screams voltage sag compensation to me. When voltage drops for whatever reason, current spikes up to maintain the same power load. If there's a voltage regulator or capacitor bank switching nearby, you could see these kinds of anomalies. ComEd had a similar issue last year where a faulty capacitor bank was causing false demand readings across multiple customers. Check if the spikes correlate with any voltage irregularities in the power quality data.
Terry and Yuri - great points. I'm going to request the voltage data from IPL and see if we can get the raw meter logs. The client is getting hammered on their demand charges because these spikes are setting their monthly billing demand. We're talking about an extra $3,200/month in demand charges if this keeps up. Will post an update once I get more data from IPL.
Greg, make sure you're looking at the time stamps carefully. I had a case in Youngstown where AEP's interval data was showing demand spikes that were actually happening during maintenance windows when the plant was supposed to be shut down. Turned out the meter clock was drifting and the intervals were getting shifted by 15-30 minutes. The "phantom" spikes were real equipment starts but logged at the wrong time.
Adding to Jim's point about time drift - Duke Energy down here has been having issues with their AMI system where daylight saving time transitions cause interval data to get duplicated or shifted. If these spikes started around March or November, that could be your culprit. Also worth checking if the meter firmware has been updated recently. Sometimes new firmware introduces bugs in the interval recording.
Update: Got the raw meter data from IPL and you guys were right - it's a communication issue. The actual meter readings are normal but something in the head-end system is duplicating certain intervals and assigning phantom timestamps. IPL is working on a fix and they've agreed to re-bill the past 3 months without the erroneous demand spikes. Thanks for all the troubleshooting help!
Great outcome Greg! This is exactly why we need to dig into the raw data instead of just accepting what the utility gives us. I've got a similar case brewing with Xcel Energy up here in Minneapolis where interval data doesn't match the customer's actual load profile. Going to use your approach and request the meter-level logs first before assuming the data is correct.