Efficient Detection of Appliance Consumption Pattern by Using Level-CrossingSampling

Abstract

In modern countries, the concept of using smart meters grows quickly. The smart grid stakeholders need to be provided with a comprehensive metering data collection and analysis. Time invariant is the conventional data sampling method. As a consequence, a significant amount of excessive data is collected, stored and processed. This deficiency is overcome by the use of level-crossing sampling technique. It allows data compression in real time. Subsequently, modern adaptive rate techniques are used for data segmentation and extraction functions. The related characteristics for the usage habits of appliances are then used to classify them.It is realized by the use of the mature classification technique of Artificial Neural Network. The findings achievesa 4.5-fold increase in compression and the processing capacity of the proposed system while preserving 95.9 percent accuracy of identification.