Active power-based event detection algorithm for real-time load monitoring systems

Non-intrusive load monitoring (NILM) has recently become a promising and trending topic in energy management. Although several NILM-based solutions were proposed, a commercial-grade universal solution is yet to be developed due to inherent limitations in the event detection and appliance identification algorithms. This work proposes a novel real-time event detection algorithm specifically designed to be effective in NILM-based solutions developed for residential loads. Real-world experimental results indicate that the proposed algorithm can detect a change in the operating state of a device with a 98 percent accuracy within 1.35 milliseconds, on average.