New Daedalus

Notes on SmartGrid Domain Experts Workgroup, NIST, August 5, 2008
NIST (National Institute for Standards and Technology) started by making a strong claim for ownership in this area, citing Title XIII, 1305 of EISA 2007. NIST set out an aggressive agenda including a preliminary report at GridWeek on 9/24 and a NIST workshop on developing standards at Grid Interop in Atlanta November 11-13.

NIST wants to have in place tight working relationships with the target SDO’s (Standards Development Organizations) in place before 2009. NIST and the GridWise Architectural Council are working together to direct the standards direction toward e-commerce and interactions with building operations and with the building occupants. Some of these standards will be e-commerce focused, some will be looking to the Building Information Models, and to the energy models they support. I am excited that this might push these design approaches into continuing use during operations.

B2G Breakout (Building to Grid)

We quickly agreed that goal of the SmartGrid standardization efforts is to design the information exchange and informational interoperability to enable healthy markets to emerge around energy use in buildings. Success was defined as enabling buildings to trade their energy.

The group was in violent agreement that we needed to work on business to business interactions, and not on machine to machine interactions. Services inside the building would be coordinated by the business processes of the occupants. Grid messages would go to the business agent of the occupants. Interactions, including pricing and bidding, would be between the grid agents and the building agents.

But market development for what? The problems that need solving quickly include real time pricing and automated demand response. The solutions should encourage the development of distributed generation and local energy storage. The Pricing Models and Buying Models for Energy should also work inside microgrids such as the building or neighborhood.

We spent a considerable time defining the characteristics of live pricing. There was intense interest in moving beyond static prices to curves, i.e., if the prices move like this tomorrow, I will commit to energy consumption in a curve that looks like that. Automated contract execution and on-line exchange of tariffs are both desired.

One thing that everyone agreed is that automated metering infrastructure would never meet its potential unless full live real-time access to all meter information is made available from *both* sides of the meter should be the standard. Parties could then collect the data in accord with the schedule that made sense to them.

This has been about the business exchange – soon I will write about the attributes of these transactions and product differentiation on the grid.

My hydronic system failed this summer. Spare parts for the boiler, still more efficient than most on the market, are no longer available. It supported a hot water heater and two zones in my house. I am splitting out the water heater, moving to a tankless system. While the best boilers haven’t gotten any better, the price for an equivalent efficiency boiler has gone way down; the incentive to put everything on one boiler is gone.

I would like to add one or more zones to the house. The upstairs of this old house, with bedrooms and bathrooms, has never been conditioned. With the kids away, half of the bedrooms are only rarely occupied. I like keeping the rooms for a few more years, as long as kids show up for various holidays and vacations.

There are two frustrations with this process.

The first is the contractors. They arrive, already knowing which product they will install, based upon manufacturer incentives. The manufacturers seem to be responding to energy prices by offering dealers incentives on big systems, much as auto dealers are offering incentives on SUVs and Hummers. I tell them what I want over the phone, and warn them what the criteria will be; even so, they waste my time and their own. They decide the best fit based on incentives before they look at the space, they push the incentives, and they wonder why they leave without a contract. I cannot even imagine why they think I will accept fewer zones in twice as much space. Until the installers move beyond this attitude, home building system performance will remain abysmal. The worst part is that this must work, as these jokers stay in business.

The second issue is how little intelligence the control systems have. The home market appears to be dominated by systems that pretend to have taken the digital age into account. The thermostat is a nice flat screen. The time of day functions are easy to access. The actual control sequences are not as sophisticated as my last installation, which was constructed out of a complex nest of relays to squeeze extra energy out of each cycle. If this is what the American control companies are offering through their dealers, they deserve to lose to the Chinese.

The third issue is flexibility. I have been offered many controllers, but no flexibility in any of them. Single purpose systems are offered with different controllers than hybrid systems. Adding a third energy source is yet another decision. Each representative who comes by seems surprised when I ask for flexibility, and explain that it would be far too expensive. Based upon what their dealers, the home comfort system companies have not learned the essential lesson of the digital age, that simple product lines with large production runs are cheaper, and therefore multi-purpose re-programmable controllers will be cheaper. Every brand (and I have seen them all) is done a disservice by its local distribution.

So what do I think a standard, flexible controller would offer?

A home system should support hybrid systems, with enough abstraction so that multiple fuel sources are supported. A standard controller would balance the price and availability of each energy source installed to provide heating and cooling. It is common for systems to support an outside set-point to change from, for example, a heat pump to a gas pack. A proper system would tune itself, and be able to suggest what that outside set-point should be.

It should also be able to accept prices. For now, there is no live energy pricing in my area; I should be able to enter the price from my last electric bill, and the price from my last gas bill, and let it suggest another cutover point. If I add a thermal store, I should be able to include that in the same algorithm. It should not matter if the thermal store is driven by time of day prices and pre-heating (or cooling) or by a solar thermal unit. If I add photovoltaics, the system should be able to understand the availability and pricing of that as well. The system should be live pricing ready, ready to receive live price signals for any of the energy sources when they become available in my area. Clearly there should be a means to upgrade the system to support ADR (automated demand response signals) when they come to my area.

There is no reason for this to be more expensive. The controllers they are selling already have enough muscle power. The interface and system logic, while more extensive than today, would be less extensive than the multiple product lines I am being offered each evening.

Until the control vendors and home automation vendors offer products like this, than it is a sham they provide any sort of sustainability or energy control. If they are offering products like I want, than they should support hot lines to report the local dealers who besmirch their names with poor proposals. Like GM, sitting fat and happy on the no competitors assumption of generations ago, they will slow lose their customers and their companies.

And if you think the products are available, here in central Carolina, let me know. I will write that up later….

This was the post I started writing a couple days ago until the first paragraph just metastasized to fill up the page. Once we have more than a few electric cars in town, then those cars will be potentially the biggest stress on the grid.

The peak stress on the power grid starts during the afternoon, during heat-of-the-day air conditioning and work, but it continues through the early evening. Offices are still turned on. Programmed houses are kicking in with their air conditioning in preparation for their owner’s arrival. Families are cooking dinner. The power grid is still working nearly as hard as it can.

Now let’s posit the electric cars coming home, drained from a day of driving. Perhaps they were doubly drained, used to carry their office buildings during the afternoon brown-out. What will people want from their cars next….

• To sit in the garage overnight, slowly charging.
• To be ready to drive 15 miles in twenty minutes when I go get one last kid from athletic practice.
• To be at least half charged and ready for anything in two hours when the baby sitter arrives and mom and dad head out for an evening on the town.
• To quickly get to at least a 40 mile range in case I get an emergency call from the nursing home, and thereafter just be sure to be ready for the morning commute.
• To get a charge for 15 miles by 8:15 when I head to choir practice at church. Better make that 25 lest we stop for coffee afterward.
• It's two hundred miles to the beach and we plan to take full advantage of the expensive week-long rental by getting there tonight! Kids, grab your bags, we are leaving in 20 minutes. Oh, and the car needs a full quick-charge, no matter the expense.

Gasoline handles all these scenarios. Many of them involve discretionary electricity purchases during the early evening peak. We will never solve these problems at the level of machine-controls. We need time of day pricing, to allocate the scarce resource. Just as many restaurants offer Monday-night specials, we need day-by-day pricing, to encourage people to choose when to schedule their evening activities. Electric cars will require live power pricing, by the minute, and by the day.

Let’s consider driving the electrical car further into our lives, and further into our infrastructure. Sometimes I will want to charge my car when I am not at home. This will require that cars identify who they are at the plug.

• When parking downtown, I want to plug in my car. I may want to choose between a quick visit, for a cup of coffee, and an all-day back-to-school shopping event.
• The Green Garage™ offers locally generated wind power for re-charging at its own special rates that vary with the wind. Having been burned once, I want to check prices before I leave the car.
• When I go over to your house for dinner, I want to plug in. Being a polite guest, I of course want the charges to go onto my own bill.
• The whole family gathers in the next town for Thanksgiving dinner. All cars are drained, and need to recharge over the next five hours except for the college kid, who arrives at the last moment, and leaves as soon as he can. Grandpa decides to overrule all normal agreements and cover all the charges for cars plugged in at his house.

The technical feat of creating amazing batteries and lightweight materials, however astonishing and inspiring, will be undone without the capability easy interaction with the lives and aspirations of those who drive the cars. Electric cars will require powerful intuitive systems interfaces, able to learn their owner’s tastes and habits. These systems can only interact with the power grid through simple standard economic interfaces.

If a performing electric car were to arrive today, with adequate batteries at reasonable cost, it could well push today’s non-transactive energy infrastructure over the edge. Usually I write about intelligent building agents; when I write about the power grid, it is to discuss transacted energy purchases between those agents and an intelligent transaction grid. Today, I am going for those transactions on that grid, but leaving out the building. But first, a little on the building with cars.

There a lot of hopeful scenarios in which peak shaving is enabled by commuter cars plugged into office buildings. Peak shaving, initiated by what are called Demand-Response (DR) signals from the grid, is when buildings lessen their electrical demands to avoid peak periods of energy use. The story goes that we will go to work, and plug in our cars. When the DR event arrives, the building will run off the combined car batteries, reducing demand on the grid.

DR is very important for today’s grid, because the power supplied at the peak is the most expensive and usually the dirtiest to generate. I have seen numbers suggesting that as much as 17% of the grid’s capacity is used for less than 120 hours per year. If we manage peak electrical use, we have effectively grown the power grid for free.

Cars and their batteries, however, will never be an effective peak shaving tool for office buildings. Leave aside for the moment all HR-related issues associated with employers paying for commuting costs, and look at the people. Peak load occurs in the afternoon, and extends into the early dinner hour.

If I live some distance from my employer, will I be willing to end each day with a low charge on my car? Only until the first day I run out on the way home, perhaps because of an unanticipated need to attend a school event for my children, or to attend to a medical issue for my parents, or even to pick up some supplies for a social event. In any case, the first time it happens, I will resolve to park away from the building thereafter.

If I live close to work, I will arrive with my car already charged up. DR participation, always in the afternoon, will leave me always wondering whether I am subsidizing the company. The first time I am turned down for a raise, this thought will begin festering into a general resentment of my employer. Sub-vocal mutterings with phrases such as “blood-sucking leeches” come to mind.

Whether I live far away or whether I live close in, sooner or later I will leave early to head off for a summer (most DR events are during warm weather) weekend at the beach and find that despite my plans, my employer and its building have drained my car.

No, we cannot turn to electric cars to solve the DR needs of our office buildings. Not if actual people are involved. Perhaps if we make sure that our grid is intelligent and two-way transactional we can see a way past this.

I will try to write soon on what intelligence is needed, in grid and car, for more realistic use of more than a few electric cars.

At the University of Kassel in Germany, researchers are assembling a reliable power grid from a number of unreliable components. Kombikraftwerk (Combined Power Plant) is a grid assembled from 36 biogas, wind, solar and hydropower plants in a distributed network. The project was designed as a demonstration project to prove that it is possible for the German power grid to be reliable even if based entirely on non-traditional power sources.

This is a demonstration (again) of the old principle that you can gain additional reliability and availability from multiple technologies then you can from any single technology. While it is certainly possible that with additional research, development, any one (or two) of the technologies could be made ever so much more reliable, such efforts soon run into the age-old 90% problem. (This is usually expressed as “After you've done 90% of the project, you have the other 90% to do.”) Achieving each additional increment of reliability from an existing technology usually requires large amounts of additional effort.

A problem in scaling Kombikraftwerk will be the fallacy of large scale control. As the size of community to be orchestrated increases, the complexity of orchestration increases. Sooner or later, the Kombikraftwerk will fail due to the deep integration and direct control  of power production that appear to be embedded in its model.

The other apparent problem in Kombikraftwerk is the fine tuning of energy production to meet actual rather than anticipated needs. This requires additional spin reserve (Plants that are operating but not supplying the grid) to handle surprises. Some spin will always be required, but it is easy to imagine scenarios requiring much less than we require today.

These issues are just the sort that the abstract interoperability and intelligent end nodes envisioned by the GridWise Architectural Council (GWAC) will solve easily. GWAC is working toward abstract e-commerce style interfaces between each component of the grid, including generation, transmission, distribution, and end customer face. Both sides of each interface are assumed to be intelligent peers, able to defend their internal missions.

The GWAC smart grid is the simultaneous optimization of the diversity problem (which combines a number of unreliable technologies to produce a reliable cloud) and the complexity problem (it is difficult to control a mix of systems with different operating characteristics into a single large-scale control system) and the money problem (how do we fund this in such a way that each innovation can be rewarded).

Generating systems can signal their operating postures and capabilities using abstract messages. These interfaces hide the underlying diversity to prevent the overall grid management from becoming too complex. New technologies for storage and generation can come to market faster, and make money faster, because they need only interface to the simpler abstract interface rather than undergo deep integration.

The GWAC customer face addresses coordination of the demand side. Local agents representing smart buildings become participants in the smart grid. Initiatives like the Zero Energy Building foresee hybrid nodes, usually consumers of energy, but occasionally selling back site-stored or site-generated energy.

Kombikraftwerk and GWAC are compatible approaches that can easily build off one another. Because the defined interfaces of the GWAC are abstract and standard, new generating technologies can join the mix without extensive review. Easy recombination enables innovation by shortening time to market. More innovation enhances reliability by adding additional sources of diversity.

Money is the best, and most universally accepted, abstract interface for communicating scarcity and value. When we add money to each interface request, the natural target for the building-side of the interface is the enterprise and the tenant, not the systems. It may be that the best demand/response decision, when incentivised with pricing, is to shut down the office building, declare a telecommuting day, and run the building generators (whatever they may be) flat out. Such decisions would only further enhance Kombikraftwerk.

Looks like Kombikraftwerk and the GridWise Architectural Council could work well together…

References:

• Kombikraftwerk  - http://www.kombikraftwerk.de/index.php?id=27
• GridWise Architectural Council (http://www.gridwiseac.org/)
• Demand Response Automated Server (DRAS) http://www.akuacom.com/solutions/index.html