2015/10/26 by Nipun Batra, Batra, Nipun, Amarjeet Singh +3
Engineering · #Building Energy and Comfort Optimization #Energy Harvesting in Wireless Networks #FOS: Computer and information sciences #Green IT and Sustainability #Machine Learning (cs.LG) #Smart Grid Energy Management #Smart Parking Systems Research
paper · pdf · doi:10.48550/arxiv.1511.02900
openalex publication_date 2015/10/26 · openalex created_date 2022/08/27 · openalex updated_date 2026/07/28
In this paper, we investigate whether "big-data" is more valuable than\n"precise" data for the problem of energy disaggregation: the process of\nbreaking down aggregate energy usage on a per-appliance basis. Existing\ntechniques for disaggregation rely on energy metering at a resolution of 1\nminute or higher, but most power meters today only provide a reading once per\nmonth, and at most once every 15 minutes. In this paper, we propose a new\ntechnique called Neighbourhood NILM that leverages data from 'neighbouring'\nhomes to disaggregate energy given only a single energy reading per month. The\nkey intuition behind our approach is that 'similar' homes have 'similar' energy\nconsumption on a per-appliance basis. Neighbourhood NILM matches every home\nwith a set of 'neighbours' that have direct submetering infrastructure, i.e.\npower meters on individual circuits or loads. Many such homes already exist.\nThen, it estimates the appliance-level energy consumption of the target home to\nbe the average of its K neighbours. We evaluate this approach using 25 homes\nand results show that our approach gives comparable or better disaggregation in\ncomparison to state-of-the-art accuracy reported in the literature that depend\non manual model training, high frequency power metering, or both. Results show\nthat Neighbourhood NILM can achieve 83% and 79% accuracy disaggregating fridge\nand heating/cooling loads, compared to 74% and 73% for a technique called FHMM.\nFurthermore, it achieves up to 64% accuracy on washing machine, dryer,\ndishwasher, and lighting loads, which is higher than previously reported\nresults. Many existing techniques are not able to disaggregate these loads at\nall. These results indicate a potentially substantial advantage to installing\nsubmetering infrastructure in a select few homes rather than installing new\nhigh-frequency smart metering infrastructure in all homes.\n