“Using the measured performance data from the power supply of a household, the algorithm serves to illustrate the use of individual appliances," explained Wilfried Elmenreich, who is developing the technology at the Institute for Networked and Embedded Systems in collaboration with Dominik Egarter.
The development saves costs: Firstly, because it is not necessary to use any additional measuring equipment to determine the frequency of appliance use, and secondly, because knowledge of the consumption increases the user’s energy awareness. The technology represents an integral part of the ‘smart grid’ concept.
Mathematical methods provide the background to the development: The algorithm presented is derived from a model based on Markow chains and a sequential Monte-Carlo method (particle filtering) for the state estimation of the appliances. The researchers were able to illustrate that the algorithm works at an accuracy level of 90 per cent in typical households.
The paper was published in the journal IEEE Transactions on Instrumentation and Measurement.
Reference: Egarter, D., Bhuvana, V. P. & Elmenreich W. (2014). PALDi: Online load disaggregation via particle filtering. IEEE Transactions on Instrumentation and Measurement, pages 467 – 477, 64(2).
Related articles and links: