Welcome to the new Home Performance Forum.

This site is read-only for archival purposes. Please visit the Home Energy Pros Forum to continue the conversations.

Individual device power monitors


  • David ButlerDavid Butler Posts: 3,889
    Noninvasive Load Monitoring isn't new. There's even a Wiki page. Through a process known as signature analysis, individual loads can be monitored (more or less) from a single-point connection to both legs of a home's power distribution network.

    I worked on a similar project in the late '80's based on EPRI research done 5 years earlier. Back then the required spectrum analyzer was a bulky board-level product. Now you can get one on a chip for a tiny fraction of the cost. So in that sense, it's time may finally have come. I recall Fraunhofer CSE in Cambridge, MA was doing NILM research several years ago (https://vimeo.com/18472128)

    The biggest challenge is software -- training the monitor to recognize the various loads that make up the aggregate signal, and to account for loads that vary significantly over time. As you might expect, the larger the load, the easier it is to identify and track. I'm optimistic that useful monitors will finally be available in the next 5 years.
  • David ButlerDavid Butler Posts: 3,889
    I should add that the NILM technique described in Paul's post relies on inductive coupling. That's new, but in my opinion, introduces an unnecessary compromise in terms of the signal-to-noise ratio.

    Even 30 years ago, we envisioned a chip-level solution that would be contained in a retrofit 240V receptacle as typically used for electric ranges and clothes dryers. Most homes have these receptacles, even those with gas connections.

    Still, if it can be made to work well, this approach avoids the cost/hassle of UL approval and eliminates a couple of components required to connect to the power line.