'UK-DALE': A dataset recording UK Domestic Appliance-Level Electricity demand and whole-house demand

Many countries are rolling out smart electricity meters. These measure a home’s total power demand. However, research into consumer behaviour suggests that consumers are best able to improve their energy efficiency when provided with itemised, appliance-by-appliance consumption information. Energy disaggregation is a computational technique for estimating appliance-by-appliance energy consumption from a whole-house meter signal. To conduct research on disaggregation algorithms, researchers require data describing not just the aggregate demand per building but also the ‘ground truth’ demand of individual appliances. We present ‘UK-DALE’: an open-access dataset from the UK recording Domestic Appliance-Level Electricity at a sample rate of 16 kHz for the whole-house and at 1/6 Hz for individual appliances. This is the first open access UK dataset at this temporal resolution. We recorded from four homes, one of which was recorded for 499 days, the longest duration we are aware of for similar datasets. We also describe the low-cost, open-source, wireless system we built for collecting our dataset. I. BACKGROUND & SUMMARY Energy disaggregation researchers require access to large datasets recorded in the field in order to develop and test disaggregation algorithms but it is not practical for every researcher to record their own dataset. Hence the creation of open access datasets is key to promote a vibrant research community. Researchers at MIT led the way by releasing an openaccess smart meter dataset in 2011 [4] and more datasets have subsequently been released by researchers around the world. At the time of writing, the only open-access disaggregated dataset recorded in the UK is the DECC/DEFRA Household Electricity Study [2] which has a sample period of 2 minutes. We present the first open access UK dataset with a high temporal resolution. We recorded from four homes. Every six seconds we recorded the active power consumed by individual appliances and the whole-house apparent power consumption. Additionally, in two homes, we sampled the whole-house voltage and current at 44.1 kHz (down-sampled to 16 kHz for storage) and also calculated the real power, reactive power and RMS voltage at 1 Hz. In home one, we recorded for 499 days and individually recorded from almost every single appliance in the home resulting in a recording of 54 separate channels (although less channels were recorded towards the start of the Fig. 1. The system diagram and the three major components of the system: (top left) the data logging PC; (top right) the sound card power meter (which uses the sound card on the data logging PC to record the output from an AC-AC adaptor and a current transformer (CT)) and (bottom) the ‘RFM EDF ecomanager‘ which uses a Nanode to communicate over the air with a set of individual appliance monitors (IAMs) and current transformer (CT) sensors. dataset). We will continue to record from this home for the foreseeable future. We recorded from three other homes for several months; each of these homes recorded between 5 to 20 channels. Figure 1 provides an overview of the system design and Table I provides summary statistics about the dataset. This dataset may also be of use to researchers working on modelling the electricity grid; exploring the potential for automated demand response or researching appliance usage behaviour. A longer version of this paper is available [6].