Non-Intrusive Load Event Identification Algorithm Based on Color Coding

Non-invasive load monitoring (NILM) can help residents monitor the operation of household appliances and achieve the purpose of energy conservation and emission reduction. Load event identification is a key task of non-intrusive load monitoring. In order to enrich the characteristics of load events and improve the identification accuracy of load events, a load event identification algorithm based on color coding (CCA) is proposed. On the basis of retaining the basic waveform of active power, the three characteristics of active power (R), reactive power (G) and reactive power change trend (B) are fused to construct the color image of load events, and the image is trained and rec-ognized based on the AlexNet convolutional neural network with parameter adjustment. The experimental results show that this load event identification algorithm can stably and effectively distinguish the load events of different devices. Compared with three commonly used classifiers, the results show that this algorithm is superior to the traditional event classification algorithm based on power sequence. © 2016, J. Network Intell. All rights reserved.