[an error occurred while processing this directive]

Automated Demand / Response causes new difficulties and problems. When the utility enables systems that have been suppressed, they all turn on at once. The resulting spike of power use can be more destructive than the original Demand event. Utilities have tried to offset this with zone-by-zone re-activation, which lessons the big system spike but may be as destructive to the local zone or substation. They also use random setbacks, wherein some systems wait are set to turn on right away when released, and others to wait some number of minutes. In practice, this precise control just leads to a series of usage spikes.

As we seek to extend Demand / Response to more systems in the house, we run into what we might call Knowledge Problems, problems of diversity and understanding. (Economists use the term knowledge problem to refer to the fundamental problems of central planning in an economy; many of their discussions apply just as well to the problems of distributed control of systems.) Washing machines cannot cease operations until they insure that cleaning products will not destroy the current load. A freezer can turn off for hours with no problems unless it has just completed a defrost cycle; in which case it must ignore the load shedding request. Even the simplest systems rely on internal domain knowledge not available to grid operators.

Direct Demand / Response require customers to cede autonomy and privacy to the utility. Loss of autonomy is a barrier to long-term participation acceptance of Demand / Response. The building occupant may have a critical customer event on-site. The homeowner may have weekend guests and wish optimum amenity. The children in the house may have chicken pox and spend the day in warm oatmeal baths. For whatever reason, opt-in and opt-out of Demand / Response must be under local control.

Direct control may also create a privacy liability. When utilities perform direct control of building systems, they may enumerate these systems and their operation. The resulting database will reveal more information about the customers then the utility wants to possess. It is only a matter of time until records documenting, say, hot tub use become part of a divorce proceeding. Today, insurers are mandating that IT departments purge credit and identity information from commercial systems as a condition of continued coverage. Utilities may soon find that Demand / Response databases are a liability as well.

Economic interactions are clearly becoming the preferred solution to these problems. Standards are simplifying the interaction with the building to one of price. Every brownout is merely a pricing problem.

Life cycle does more than lifestyle to determine energy usage. Homes with small children have different energy profiles than empty nesters. Life-cycle trumps life style in energy use except in the most extreme cases. Extreme energy savings are not ever going to be a mass phenomenon. People would rather get to the beach an hour earlier, and get the complaining kids out of the car and in bed on time than they would drive for greater mileage on the trip. These facts are not likely to change. The occupants of homes and buildings will be as bad at managing building systems as is the power company.

[an error occurred while processing this directive] The alternative is that the devices must manage themselves. As new buildings, with internal systems for energy generation, storage, and conversion join the ever more digital house, the knowledge problem of centralized control of even the house becomes greater. Sooner or later, we will come to economic negotiations between systems in the house, as the way to manage the internal energy budget, and to respond to the external needs of the grid.

We can start simply. The hybrid heat pump / gas furnace is today set to switch between energy sources at a given outside temperature (mine is 38 degrees). It easily has enough processing power to re-compute this based upon spot prices in electricity. I could easily enter the monthly prices for natural gas from my monthly bill. Suddenly, my heating system is an energy day trader, working for me.

I could ask my dishwasher to run itself, and manage its own budget for the month. I could also set service standards that the dishes always be clean before dinner the next day. This leads to a relatively simple and consistent user interface.

I could tell my solar panel to sell to the grid whenever the price is above a certain amount, and to store any excess energy. The grid might consistently outbid the dishwasher-and that's OK. If so, the dishwasher would still run only at night.

I could tell my whole-house storage system to buy power at any price until it has four hours on hand. Thereafter it might buy whenever energy is below a target price. I could let it take bids from the household systems and devices, or from the neighbor. This system would need, of course, to charge an appropriate mark-up based upon its inefficiency of storage.

If we develop the right sort of abstract business interface between the power grid and our buildings, we can also use it between buildings, or within buildings. Most throw-away cell phones have more computing power than it took to go to the moon. Surely, our embedded systems can do a little day trading...