I'm reviewing Xcel Energy interval data for a Minneapolis client and seeing some unusual demand spikes that don't match their operational schedule. The 15-minute data shows 340 kW spikes occurring at random times - 2:15 AM, 6:30 AM, 3:45 PM - with no pattern I can identify. These spikes are costing them about $8,000 annually in demand charges. How do you typically investigate what's causing unexpected demand peaks when the client swears their equipment isn't running at those times?
Question about interpreting demand spikes in 15-min interval data
Christine, first thing I'd check is if there are any large motors or HVAC systems with automatic controls. Sometimes programmable thermostats or building automation systems can trigger equipment at odd hours for maintenance cycles or pre-cooling. Also look for any backup generators that might be doing weekly self-tests. We had a similar case in Springfield where emergency lighting systems were cycling on for testing and creating unexpected demand spikes.
Another possibility is power factor correction equipment cycling on and off. Sometimes capacitor banks can create apparent demand spikes if the metering isn't compensating properly. Also check if the facility has any tenants or submetered areas that might have their own schedules. I've seen cases where a small tenant's equipment was creating large building-wide demand peaks because of how the electrical system was configured.
Thanks Elmer and Albert. Turned out to be exactly what Elmer suggested - the building's fire pump was doing weekly automatic tests every Tuesday at 2:15 AM. The client had no idea because it's managed by an outside fire protection company. We're now working with them to reschedule the tests to occur during on-peak hours when the demand charge is already high. Simple fix that will save thousands per year.