I just listened to most of a lecture by a CarnegieMellon professor sponsored by @FIATECH that was supposed to be about Instrumenting Civil Infrastructure to Increase its Resilience, Adaptiveness, and Self-monitoring Capabilities. It actually had very little to do with civil infrastructure, and considerably more to do with invading private property.
It actually was about figuring out how to instrument a house for the purpose of monitoring (and controlling) with virtually no recourse to already existing systems, network protocols other than 802.11x, and deliberately doing so without any interaction with the resident.
I have philosophical, technical and even moral qualms with this line of research.
First, although I do not want to put words in the professor's mouth, I had the very strong feeling that the reason that he wanted to do this energy monitoring (and certainly later controlling) covertly is because he recognizes that the typical citizen will not be willing to stand for such an invasion of his property without significant compensation.
The purpose of smart grid is NOT to save the consumer money, but to make the power companies more efficient (read more profitable). Nobody is saying that smart grid technology will bring the cost of power to the consumer down...in fact, I am being charged a special fee from the two power companies I pay electric bills from for their smart grid initiative. As a homeowner, I am certainly not going to allow covert monitoring of my power consumption (the professor showed he could correlate from electric data when the bathroom was in use) unless there is a HUGE payback to me personally for being so invaded.
Technically, like many academics and beltway bandits, the professor is working with home made sensors, except for a couple of low cost temperature sensors from HOBO. He showed a two board EMF sensor that he said could be produced for about a dollar. Now, I've spent about half my career as a product manager and/or manufacturing engineer in sensors. My estimate for such a combination board would be on the close order of $15 going to $5 in 1 million quantity. It might be possible to get the COST down to a buck, but the price will never be. The professor showed a great capability for theoretical statistics and correlation based on consumption in both the time and frequency domains. He obviously didn't do any work with any of the existing sensor companies on how to design his needed sensors.
This is a serious fault. Several years ago at the PCSF forum, I met the then-head of the CERT program (which is run by CarnegieMellon) and asked him if he'd ever been inside a process plant. He said NO. I then proposed that I could spend a day, gratis, taking him to five or six different types of process plant within a 30 mile radius of CarnegieMellon. He declined.
If we are truly going to have a robust bidirectional smart grid, we need to have researchers who know the science but who also know the business, and the body of research that R&D departments like Emerson, Endress+Hauser, ABB, Siemens et al have amassed.
My moral qualm is the fact that the professor clearly was intending that this monitoring be done covertly...and I recommend to him the Milgram Protocols. If you don't know what they are, I suggest you google Stanley Milgram and find out. Scary. Real scary.
