Why research on smart energy grids?

For some reason it looks like the energy grid of the future is becoming more and more a focal point in my current research. Perhaps because it is truly an "Ultra large-scale System" in the very essence of the words. But it also raises considerable security concerns on all levels: personal security with respect to privacy and business models, corporative security with respect to access to devices, energy provisioning and billing and finally on a national and international level due to border crossing energy transports and things like global nuclear strategies and ecological questions. Peak oil, PV installations and technology, e-mobility all play some role in this huge system.

And it is exactly this system that is supposed to change substantially in the next couple of years. Political agendas, ecological forces and the fact that we have already passed peak oil detection and use all force nations, corporations and finally individual users to think about their (substantial) investments in energy systems and machinery.

This really comes down to very personal decisions and questions: should I install some form of PV system? Use sun energy to heat my water? Buy an electric car? Install a heat pump system? Large water tanks in the cellar?

As the owner of a historic vineyard those questions equate with possibly large investments with unclear ROI and lifetime. But there are even more reasons to do some research on the energy grid: it is also a very distributed system and as such it raises many questions about stability, efectivity and optimal topologies. It allows comparisons with the structure and resilience of the internet. But as soon as one enters into smart grid research some hard facts become visible: Being a computer science person does in now way guarantee a good understanding of the energy network: electric power, its generation, distribution and use seems to be quite different to regular IT networks with respect to mission criticality, realtime characteristics, persons and corporations involved and finally also the topology used. The following sections try to list some major characteristics of the smart grid - as far as my understanding goes at that point.

The smart grid of the future

It will combine different forms of energy into a reliable and ecologically sound form of provisioning and consumption using - at least currently - mostly private business models. It allows realtime data monitoring at private households and manipulation that goes far beyond simply turning off the supply remotely if a customer does not pay. Instead, intelligent household devices would be able to use energy when it is cheap and in ample supply, store energy when possible and even load the energy grid when energy is scarce and in high demand. A core element of the smart grid is provided by the smart meter and its associated gateway at customer premises.

Other elements of the smart grid are various forms of energy providers and consumers ranging from large scale wind parks in the North Sea to individual small scale installations of PV, wind and perhaps some form of water power stations on the supply side to huge steel corporations or electric cars. And the latter can turn around and join the supply side at any time, feeding energy into the grid when there is a high demand.

Ecologically the requirements for the grid are: avoid CO2 production as much as possible, avoid energy waste as much as possible.

Politically the requirements are very much heterogeneous: from billion-dollar off-shore installations controlled by a few corporations (the current situation with respect to energy production and distribution) to highly distributed ways to create, consume and balance energy use. There is currently no consistent model of this smart grid and its parts. Even worse: it seems to be impossible to even name the variables and the fixpoints in this concept because no one knows how core elements of the system will develop over the next couple of years. This seems due to two developments. First, the speed of technology and its introduction to real life is staggering. Just look at the mobile sector. At any time disruptive developments in the area of batteries, condensers, PV effectivity and so on can render long-term billion dollar investments into battery packs and PV installations obsolete. Second, many developments depend on politics controlling their environment and by that their rentability and lifetime. Private PV customers have experienced this volatility lately when politics changed the profit per kw considerably. Even worse: the rule for uploading energy seem to be volatile as well creating massive uncertainty in private households. This results - as a recent guest from a company serving the energy corporations expressed - into nothing substantial happening right know to foster the proclaimed change to an alternative energy system. And it is quite understandable: who would invest billions into infrastructure technology which might not last very long? But one could ask as well: why are those investments necessary? Do high-power cross-nation power lines really serve the purpose of a distributed grid? Or are they only needed for those billion dollar offshore parks to work? And on a deeper level: why doesn't the power grid scale as easily as the internet?

The elements of the smart grid

For consumers the most important element for sure is the so called smart meter and its associated connection to the world, the smart gateway. The smart meter collects energy consumption data and the gateway does the transfer to energy providers. But the gateway can also connect to a private power management system which controls intelligent devices within households. And it could possibly also allow energy providers to control energy consumption within the household to achieve some kind of optimum (and it is currently very much unclear which business model will define what "optimally" really means in this context and for whom). Let's assume some form of intelligent fridge in the household of the future. The fridge is able to charge its cooling batteries and last at least 24 hours without energy after doing so. Its insulation is very good. It can tell about its current state, temperature and when it will need some charge to maintain safety of food. When should the fridge recharge? Who should it ask to supply energy and when? (show ecological and business perspectives, discuss proprietary problems and security)

Other core elements of the smart grid are of course the power lines, the interchange points, meters and sensors on any level, the power factories and so on.

Equally important elements of the smart grid are the corporations dealing with the production, distribution, provisioning and maintenance of energy and the grid. As unbelievable as it may sound there are seven!! different business contracts concealed within one power meter in Germany. This is due to the so called "un-bundling" of the energy sector, i.e. destruction of vertically organized energy companies. The way the corporations are organized and the way the consumers are tied into this system will largely define economic power and dependencies: will consumers have access to the energy markets or will they depend on energy providers? Can consumers spontaneously form "virtual power stations" to consume or provide energy by bundling thousands of households into megawatt providers? And how do we prevent damage to the grid when this happens?

The old (current) grid

The "old" power grid has some distinctive features which pose a challenge for the use of alternative energies. Those features are in particular:

  1. the lack of power measurement devices below the 20kv level

  2. a strictly hierarchical, one-way organization going from large power providers to individual consumers

  3. a separate and proprietary control network based on rather old hard- and software

  4. a lack of intelligent devices on the consumer premises

  5. a highly "unbundled", vertically de-regulated business organization

  6. A lack of switching and routing fabrics (like e.g. the Internet Exchange Points)

Once the one-way power highway becomes a two-way street there is the danger that surprise changes in supply or demand from customers overwhelm the 20kv switches. Network providers have no way to realize those situations before they become a threat to the stability of the network. Much less a way to control the supply and demand behavior of customers e.g. through financial incentives.

There are some important parameters related to the old grid which are hard to get, e.g. the percentage of power lost during transmission (around 10-15) or the percentage lost due to standby features of electronic devices (again around 10-15). The numbers were taken from Demirel.

Transformation

The transformation of the traditional power grid into a network that allows the integration of alternative energy forms could happen in various forms. One way would try to keep the large corporations currently providing the energy in the lead by requiring large, centralized wind or solar parks. Those are very costly investments only possible for large corporations and because those parks could be built almost anywhere require new high-voltage cross-country transmission lines.

Another way would be to focus on decentralized, small and cheap installations and let the sheer number of those do the work. This will require more intelligence at the customer premises as well as the network level. It would allow virtual power stations by combining the demand or supply power of customers and possibly avoid the need for new and expensive cross-country lines. It would require better switching and routing stations within the network to deal e.g. with highly dynamic changes in demand or supply.

Stability and Reliability of those two architectures are highly controversial. More components raise failure numbers as well as reduce failure severity, in the best case to a level that is no longer recognizable. Centralized installations are easier to control and administrate but offer critical surfaces for attacks and in case of an error have a potentially much larger influence. The centralized vs. de-centralized debate finally is also a very political question as it influences business models considerably. The centralized model basically replicates the current dependency on large corporations while the decentralized model distributes control over energy across a larger population.

Alternative energies are kind of "unreliable" from the traditional point of view: the sun may shine or not, the wind may blow or not and on top of that there are seasonal changes as well. Power demand on the other side is much more stable and predictable with e.g. the typical peak during daytime hours. The dynamics of alternative energies could be fit to the (currently) rather stable demand side in various ways.

  1. Build storage capabilities at the production or network side and leave the demand side unchanged

  2. Build storage capabilities at the demand side and leave the supply side patterns largely unchanged

  3. Improve the dynamics of energy exchange across larger regions or countries assuming that shortages are kind of local or regional events.

  4. Start controlling customer demand by dynamically turning customer devices on and off and providing financial incentives for customers to spend money on intelligent devices which can be controlled remotely

  5. Do like in 4. but try to balance supply and demand locally in a peer-to-peer manner without going through providers.

  6. Do all of the above

The role of energy producers, distributors etc. change considerably with every alternative. And the distributed solution (5) suffers from a lack of intelligence and sensing equipment in the 400v range typically used and produced at the household border.

Acceptance

Some future business models assume that people will actively optimize energy consumption when the price is right (high or low enough to force behavioral change). We should not forget that "optimize" lies in the eyes of the beholder. In that case it is the energy provider who defines the optimal timing for energy consumption. People are creatures of habit and on top of that the daily work and familiy routine does not allow a lot of flexibility with respect to energy consumption. Two factors have a major influence on the acceptance of alternative energy handling: convenience and safety. A solution is convenient if it fits into my daily routine and at the same time does not require active intervention. This means automation at its core. And a solution is safe when I am protected against financial disaster. This finally means a flat-rate business model.

Physics, Business and the Role of IT

I did already mention the difficulty of modeling the smart grid of the future because there seem to be only variables and nothing fixed. This is not entirely true as electricity has some physical and ecological properties which work independent of business and political agendas (but certainly can worked around to some degree).

Production and consumption of electric power follows certain physical rules. Solar panels produce energy when the sun shines and this energy gets fed into the power network. This causes a raise in energy level within the network which starts to balance the energy across the network (if possible). When devices start consuming energy the opposite process starts and again a balancing happens to smooth the surge of energy caused by consumption. This seems to be a somewhat automated process which works until certain high or low ernergy marks are not crossed. The importance of this balancing shows in the need for power destruction in real-time to avoid damage to the grid. If one is able e.g. to run huge aluminum ovens just to consume power a lot of money can be made today. The same goes for installations that can supply power in real-time to avoid grid collapse.

If the balancing would be perfect it would only require one flexible source of additional power at one point in the network to achieve a dynamically stable level of energy throughout the whole network. And only one installation that could destroy unneeded power in real-time and only in case there is absolutely more power generated than consumes at that time.

Another "physical" constraint is power generation precedence. "Alternative" power seems to be free with respect to resource consumption. From an ecological perspective this should mean that alternative energies like wind and solar should always get precedence. In other words: no oil or gas should be used to produce electrical energy when alternative energies are available. This again requires balancing throughout the distribution network. And perhaps - in the case of large scale alternative power stations e.g. in the North Sea, more and larger power lines. Of course there are also physical constraints behind nuclear power plants and other forms of energy production requiring sometimes a constant demand for energy due to the cost of starting the plant.

Now, I do not know if there are regular power stations which can provide such a dynamically changing level of energy. And second, I do not know whether there are range limits for the real-time balancing of power across larger distances. But the question really is: do we need to balance the complete network or do we assume islands of balancing within a larger network?

Let's just assume that we have a household that produces energy through solar panels and the sun is shining. The household feeds the energy into the power grid and a counter (smart meter) within the customer premises starts counting the kw supplied. Next to this household another household starts the dishwasher and draws energy from the grid. The question now is: are these events related? And: Should/could those events be related?

On a physical level they are already related as the grid balances supply and demand automatically. It can do so because both, supply and demand are close and need not pass exchange points. But the energy counters in both households are not related. Indeed, both can belong to different energy providers. Notice that from a physical point of view those providers are only needed when the balancing of supply and demand within an area or across the grid does not work due to one-sided production or consumption. And also notice that the consumer of power in this case does provide a physically important service to the grid as well.

Let's take a look at the business level for a moment: In Germany the household which generates energy will report this to its network operator, NOT the energy provider. The household which uses energy from the grid will report this usage to its energy provider, NOT the network operator. This leaves the question whether this asymmetric architecture is helpful for the grid and how balancing will happen on higher layers and what kind of latencies and losses are involved.

But we are not done with the physical perspective yet: should the events be related across households? This would mean that we start changing the patterns of consumption by relating it to production of energy. When should the dishwasher start? Should it start because someones solar panels produce energy in the neighbourhood? But what if the energy produced within this area is still not enough to balance the grid and therefore external energy needs to be fed into the grid? This leads us to realize that the way we define "area" to balance decides whether the dishwasher should start because there is an abundance of energy or it should wait because of a lack of energy in the grid. Finally, if we are willing to give precedence to locally generated alternative energy it has to be the energy situation within the network that needs to decide about energy use and direct our appliances at home.

The Game

Last term students at the computer science and media faculty of the Media University Stuttgart have designed and implemented a game with alternative energy at its core. The idea for this game was born after the first Energy Day at HdM last June and Ludwig Karg of Baum Group played a major role in it. But we were not ready to enter into a serious research or development contract due to several missing links on our side. And in my opinion it is always better to have a quick prototype to help further planning.

The game served several purposes. It was used as a prototype for a multi-level serious game on alternative energies. It was difficult to design a game plan for this topic. And little did they know about the energy business in general. The second use of the game was as a platform to learn HTML5. This turned out to be quite tedious: the canvas is little more than a frame buffer and my students spent time e.g. figuring out how to detect a pixel position (within or outside of a polygon - routine algorithms for all current graphics packages). They decided on a 2D solution which was the right way to go as they had no experience in writing a 3D-game engine and 3D in general would have taken too much time away from the game.

Almost more important than the current state of the Smart City Game is its potential as a simulation and collaboration engine which can be used to test different user behaviors. We plan to add roles, a catastrophy module simulating various weather conditions, a collaboration engine which would allow neighbours to combine forces or groups to act as power stations consuming or delivering mega-watts of electic power. Mobile interfaces could be added and finally some form of hardware-in-the-loop to allow mobile users to control and monitor their installations remotely.