The Fourth Amendment and Smart Grids

If we are not careful, smart grids are in direct collision with the bill of rights. Some smart grid activities define or enable business practices for balancing energy supply and demand. There is a direct link between commonly accepted business practices and some definitions of our constitutional rights. With the best of intentions, we may be casually removing significant barriers to some of our most cherished freedoms.

The Fourth Amendment to the United States Constitution is the part of the Bill of Rights which guards against unreasonable searches and seizures. During the American Revolution, British forces made extensive use of writs of assistance, a sort of general search warrant that could be extended and used without ongoing review. In response, the Fourth Amendment created a standard whereby government searches must be issues only on a discovery of probable cause, and specifically limited in location and as to the matters being searched for, based on specific information supplied to a court.

The Fourth Amendment is the most explicit source of any support for privacy that I can find in the Constitution.

Dr Orin Kerr is one of the most respected legal voices on Fourth Amendment issues. Dr Kerr blogged this week on the relationship between technology, common practices, and developing standards for reasonable search (see reference below). Specifically, Dr Kerr was exploring the ten year old Supreme Court ruling in Kyllo vs. United States that defines the limits of police use of high technology in warrantless searches.

In cartoon form (IANAL), police scanned houses with some sort of IR scanning system and noted a hot spot in the attic. From the hot spot, they deduced that the defendant was growing marijuana under grow lights in his attic. Kyllo asserted that this was a prohibited search under the 4^th amendment. The question was, in effect, is a non-intrusive search using high tech an unreasonable search. Clearly, if Kyllo had been growing the marijuana in his front yard, there would have been no dispute when police noticed this when on routine patrol. Previous rulings had stated that police fly-overs are legal searches because non-police could fly over the property and spot the plants; the property owner has no reasonable expectation of privacy applied to aerial views of his property.

In this case, the search was ruled unconstitutional; Kyllo won. The Supreme Court adopted a test designed to let the result change with social practice: “when . . . the Government uses a device that is not in general public use, to explore details of the home that would previously have been unknowable without physical intrusion, the surveillance is a “search” and is presumptively unreasonable without a warrant.” Because infrared temperature sensing was not in “general public use,” the thermal imaging was a “search” that required a warrant.

Dr. Kerr was blogging on whether under this standard, the search in Kyllo was still prohibited. Remote infrared temperature-sensing has become quite common in a wide range of applications. I heard an ad on the radio yesterday for a remote home thermometer enabling mom to take a sleeping child’s temperature from the door without waking the child. Thermal images of houses to reveal gaps in insulation have become common; many utilities will pay for them as part of energy efficiency efforts. The question was, then, is this high tech device now considered to be in in “general public use,” and if so, can the police use it without a warrant without violating the Fourth amendment.

And so, at last, I loop back to smart grids.

Some business practices we are defining, particularly in what we are calling Managed Energy, can routinely monitor the activity of every device in a home. If we establish these practices as general practice, have we eliminated any Fourth Amendment shield against the use of the same techniques by police?

Analysis of electrical power consumption reveals more than you might guess. Research a decade ago explored what engineers could learn from these signals. One anomaly occurred almost every day in a home somewhere between a half hour and two hours after the owners left each day. Further research determined that the family dog waited each day until it was sure that its owners were really gone for the day—and then climbed onto the warm waterbed. They were detecting the change in the pattern of water heater use. Further research demonstrated an ability to distinguish how much activity was on that waterbed…

When we define business practices for the smart grid, we are doing more than solving a a difficult engineering problem. We may be creating practices that re-define our precious constitutional rights. Privacy is more than a best business practice for smart grids.

Energy Collisions and Autonomous Appliances

Appliance manufacturers are moving beyond energy pain points to energy collisions. Utility-based energy standards are stuck on energy pain. Energy collisions can offer much more benefits to smart grids than can pain points; they can offer still more to the off-grid or near grid building. Collisions are part of a wide variety of autonomous energy behaviors we will see in the near future—if only the energy suppliers will stop blocking them.

Too many energy suppliers are stuck on models of direct control. When they accept using prices, they want to use them to create direct control. (There is a name for this in Economics—drop me a line or comment if you know what it is…). To use a cartoon version of this approach, if I knew that when the yellow light comes on, I will be charged $1000 / minute to run the air conditioning, that yellow light is as good as an on / off switch. This mechanistic approach, seeking only for the right price that will achieve direct control, will not get any better results than the direct control of the 1980’s

The appliance manufacturers have a more engaging vision, in which they can compete as to how well they engage the consumer in better energy decisions. Most appliances can run in high and in low energy modes. The low energy mode may use half the energy but take three times as long. If time is money, this approach asks the question, “But how much?” Do you want that shirt clean and ready in 30 minutes [high energy mode]? Is it OK if it takes 90 minutes [low energy mode]? Is it OK to wait for 10 minutes until the energy price drops? How about 45 minutes? How about seven hours to get the overnight energy prices—or the wind-sourced energy?

The appliance manufacturers know how to do this already. They are starved for information. They want not only information about the price now, but predictions about price in the future. They want to compete on how well they can communicate energy decisions to the consumer.

This model of autonomous response can reach past the relatively low energy appliances. The intelligent thermostat may want to cool more now to in anticipation of higher prices later. The Plug-In Electric Vehicle (PEV) must support the household schedule while deciding when to charge.

There is renewed focus within the autonomous appliance community on energy profiles to support this model. Energy profiles as defined by Open Smart Grid efforts or by ZigBee have a simple model of energy use, low energy mode, turn off mode and ramp time. Building systems and appliances have a more complex mode. That washing machine may use no energy while filling, and then plenty while agitating the clothes. If an appliance understands its own energy profile, it may start filling its tank five minutes before the price drops—and time its final spin to complete before energy prices step up.

And then they began talking about systems working together to avoid energy spikes…

One of the foundational approaches in networking is collision sensing and detection (CSMA/CD) on a shared bus. Nodes on a network can transmit message whenever they want. Each is responsible for detecting when another node is transmitting at the same time, called a collision. When a collision is detected, each node waits a random period of time and then re-transmits. You may recognize this pattern as what humans do in conversation.

Today’s appliance manufacturers are talking about comparing energy profiles avoid the spin cycle and the refrigerator’s compressor cycle from running at the same time. With almost no degradation of performance, these autonomous systems can begin to shape the overall load profile of a building—or of the green neighborhood.

This approach can provide a smoother, more predictable load to the utilities. When combined with price responsiveness. It could produce a very predictable market. It really becomes interesting, however, when it applies to off grid buildings, and something that I call near-grid buildings.

If a building is running on site-generated energy, it has very distinct energy budget. That budget is not merely aggregate energy over 15 minutes, but hard, clear limits on maximum energy use at any moment. That upper limit may be continuously varying as, say, the speed of the wind changes. Continuous autonomous load shaping, based upon detailed energy profiles, may be critical to a distributed energy future.

To get there, we must get beyond price as a proxy for direct control. To paraphrase General Honore, don’t get stuck on pain.

General Relativity and Control Systems Standards

I suspect most of my readers can just about remember light speed, the 100 foot barn, and the 110 foot log from learning about relativity. The barn had doors at each end, and one set would close the instant the other doors opened. The challenge was to transport the log through the barn. The answer had to do with light speed and collapsing space, so that as one got close enough to light speed, the log shortened, and it could fit through the barn. It was a simple enough calculation as to how fast one could go to make the log shrink how much. When each of us had completed the math, the professor sprang the surprise on us: "OK, what is happening from the perspective of a cockroach on the log?"

I haven’t been writing much recently, because I have been writing all of the time. The national smart grid roadmap is a project being completed in double time. The EPRI team is diverse and whip smart. The workshop participants are opinionated and have hundreds of millions on the line. I would be surprised of the process was not contentious.

The real problem, though, is no one thinks of the cockroach. Each player on the multi-disciplinary team sees the problem set up the way that they want things to work. Power grid engineers see homes and offices as just one more set of slow devices to turn on and off. Homes and offices see the grid as a secretive and not very reliable partner they have to work with. Green and sustainable energy folks seem to see the laws of thermodynamics as as much a social construct as are the tariffs and business procedures of the grid. Utilities executives see distributed generation as an inefficient way for middle class hobbyists to get their obsessions paid for by those less well off.

The cockroach was moving every bit as fast as the log he was sitting on. While an observer saw space, and the length of the log, contracting, the cockroach was sitting on the log and saw it remaining at 110 feet. The cockroach actually saw the barn getting shorter still, and not likely to let the log pass. However, the cockroach also saw was time dilation instead of space dilation. To the cockroach, the two doors no longer open and close simultaneously, giving the log just enough time to slip through.

And that is the problem with the smart grid. The grid operators do not see the problems of the buildings. The building owners do not see the problems of the grid, because they are hidden by the rules and market design. Venture capitalists do not see a path to profitability in funding projects with years of indecision by the utilities built into the sale cycle. “If only those others would learn about how hard my problems are…” None of them will embrace the perspective of the others; they happen to have other jobs.

Today, I have been wrestling with “Architecturally Significant Interfaces”. Grid architects tend to see the world as late 60’s open plan houses, with no proper rooms to divide the houses activities. Open up the kitchen to the dining room and living room. (I wonder how much great rooms are responsible for the tendency to eat take-out in front of the TV.) Open up the master bedroom to the great room as a loft; it is open and honest, and who cares if it scares the kids. Heck, pry the doors of the bathrooms, so everybody can interact, no matter what they are doing.

A good architecture divides the house into rooms, and thereby defines how people live there. It does not determine the furniture or the wall paint. The conceptual model of the smart grid (read it yourself, chapter 3) describes the functions of the grid and the buildings and people who participate in it. The Architecturally Significant Interfaces could define how information is handed between them; if selected correctly they will free up those in reach room to innovate, without concern for those in other rooms. If we end up with an open floor plan, we will have a mess, wherein in the name of openness we will need a family meeting to before we can decide to change anything.

Relativity—it relies on acknowledging different perspectives. Without acknowledging a few architecturally significant interfaces, the smart grid will assume a perspective held by no one. And that will be a prescription for failure.

Transactive Energy and Little White Lies

As I head off to the second smart grid interim roadmap workshop (whew – that’s a lot of pairs) I think back to one of the participants in the Business and Policy track that I led with Lynne Kiesling. Several members, bunched together in the participants, were from the Edison Electric Institute, the association of share holder owned utilities. They peppered us with detailed questions and countered transactive smart grid scenarios with valid objections. It was on the second day, however, that I recognized the thought behind many of their concerns. They feel that they are asked to subsidize pretend transactions. When I say “buy power from your neighbor’s solar cell”, they hear “so we’ll put in $1,000 of equipment so you can buy $100 of power; the cost of which will be subsidized by all the other customers.”

Well, they’re right.

One reason that there is so much inefficiency in electricity is because utilities are asked to be providers of all sorts of social services. No cut-off of electricity in winter in the north. Subsidize rates to the poor. Smooth rates throughout the year. None of these ideas are bad; it is bad that we have no idea what they cost.

Now we add in feel-good electricity generation: Subsidize new energy. Provide reliability to back up the intermittent energy sources. We do need to plan for tomorrow, but do these investments make sense? How large are they? Are we increasing the base rate for electricity which we will then subsidize down for the poor to pay for the hobby power affected by the well-to-do?

These are legitimate questions. Murky accounting that hides the costs—costs to utilities, costs to customers, and costs to society—is always a bad policy. Transparency is good. Occulting is bad.

Transactive energy on the smart grid actually supports the concerns of the Institute. They wish to know what things actually cost. Transactive energy will reveal those costs. As a society, we may decide that we want to subsidize particular energy sources. We may wish to prime the pump for new energy. We may wish to impose carbon taxes on other energy.

Whatever we decide, we should do it in the light of day. In a free country, we should not hide public policy behind a cloak of murky accounting. Let’s make our decisions in the light of day. Transactive energy lets the sun shine in.