UK energy disaggregation meet-up

The field of energy disaggregation has expanded so much since I started my PhD in 2010, with new companies and research groups joining the field nearly every week. As a result, it's becoming increasingly difficult to keep track of who is working in this space, even when those people are working in your own country. The NILM workshops are starting to address this by bringing the international community together in Austin, Texas, although I'm sure the inherent cost of travel will prohibit some people attending the workshop.

For this reason, I've been talking with Peter Davies of Green Running, and we are keen to organise an event to provide the opportunity to meet other people working in the field of energy disaggregation in the UK. It will likely be 1 day event held this July in London, free for anyone to attend, and feature a range of presentations and demos, as well as many opportunities for networking. We have a location in central London already to host the event which should hopefully encourage attendees.

Please let us know via email (osp@ecs.soton.ac.uk or p.davies@greenrunning.com) if you would be interested in attending this event, and also feel free to pass on this information on to anyone else who might be interested.

NILM 2014 workshop schedule released

The schedule for the NILM 2014 workshop in Austin has just been released. The workshop will include 3 sessions of paper presentations, a set of lightning talks and poster presentations, an invited keynote from Shankar Sasty, and a demo NILMTK. I'm really excited about the workshop, and look forward to seeing you there!

Training disaggregation algorithms without sub-metered data

I'm keen to include an unsupervised disaggregation algorithm (one that doesn't require appliance data for training) in NILMTK. At the moment, the toolkit only include two supervised benchmark disaggregation algorithms, which I think really limits its usefulness. This post is intended to be a first step towards a simple, intuitive and robust approach to learn models of household appliances using only household aggregate data. I would be really interested in any feedback from the community regarding any improvements or extensions.

Extracting step changes and clustering via a Gaussian mixture model

The approach can be summarised as follows:

1. Extract a set of step changes by taking the differences between sequential aggregate power readings
2. Take the absolute value of these differences such that both positive and negative step changes are identical
3. Discard small step changes (e.g. < 200 W) since there is too much noise at this range to extract any meaningful structure
4. Discard large step changes (e.g. > 3000 W) since these are most likely generated by multiple appliances changing state simultaneously.
5. Cluster the remaining set of step changes using a Gaussian mixture model

Evaluation using data from real households

I applied this approach to data set of aggregate data collected from real households. Unfortunately, the data set does not contain any sub-metered data from such households, so no quantitative results can be provided regarding its accuracy. However, a visual inspection of the extracted step changes and identified clusters shows some encouraging results.

This approach worked very well on some houses, such as the one shown below. The plot shows a black and white histogram of the extracted step changes, in which peaks corresponding to appliances are clearly visible at roughly 1000 W, 1800 W, and 2300 W. The plot also shows coloured probability density functions (PDFs) corresponding to the clusters found. Interestingly, the clustering algorithm successfully finds the three appliances, as shown by the cyan and yellow curves. However, it's also worth noting that many other clusters were found which do not correspond to appliances.

Click to enlarge

There were also households in which no structure was present in the extracted step changes, and as a result none of the clusters correspond to individual appliances, such as the plot shown below. This is likely due to a large amount of measurement noise in the aggregate data, or a number of appliances with highly variable step changes.

Click to enlarge

Conclusions and future work

These experiments have shown that even a very simple model can successfully learn models for appliances using only aggregate data. However, it has also shown that the performance is likely to vary widely between different houses. An important challenge that has not been tackled here is that of labelling identified clusters, e.g. cyan cluster = lighting, red cluster = noise.

Thesis code release

Today I'm releasing the code I wrote for the experiments in my thesis. The code includes an implementation of a Bayesian hidden Markov model, and its application to various appliance modelling tasks. The implementation is in C# and uses the Infer.NET framework for Bayesian inference. The input data I used came from the Tracebase and REDD data sets, and therefore I have not included it in the code release. As a result, I think the BayesianHMM class is probably the most useful to the community.

The code can be found via a link on my publications page.