Fusebox dialects: Central appliance recognition for a smarter grid

More intuitive and direct information can help engage end-consumers in the smart grid. Ongoing research in the field of appliance recognition could prove to be a solution by recognizing and reporting energy usage at the appliance level in a cost efficient manner. In order to embed such a system on a smart grid scale, a thorough understanding of how the appliance profiles data will "look" when measured from a central point, like the fusebox, is needed in order to build robust algorithms that can be recycled between buildings. In this paper seven appliances with different characteristics are evaluated in four different buildings. The results show that appliances with a more reactive load are more affected by already present appliances on that phase. In extreme cases the centrally measured fluctuations of smaller appliances are perceived as consecutive on and off switches. Care has to be taken to these effects when designing a practical appliance recognition system to be applied on a wide smart grid scale.