On September 4, I met with Professor Antonello Monti at RWTH Aachen University. Prof. Monti directs RWTH’s Institute for Automation of Complex Power Systems, working on topics such as monitoring and distributed intelligence for Smart Grids. He was also the technical manager for FINESCE, an EU-funded project running from 2013-15 that developed an open-IT infrastructure for Smart Grid-related applications. Prof. Monti and I sat down for an interview, and he also gave me a tour of his lab.
(Note: The following interview was recorded, and has been edited and condensed for length and clarity. The interviewee reviewed this writeup before its publication.)
Priya: What is your definition of Smart Grid?
Prof. Monti: I’ve encountered many definitions. My preferred definition is that it’s the transition from a grid operating as planned to a grid operated in real time. In the current power system, a lot is calculated and planned the day before, and we assume we’ll have to take only minimal actions to correct for deviations from this planning. In the future, as we imagine a higher penetration of renewables and resultant high volatility, we’ll need a system that can make significant decisions in real time.
Priya: In one of your papers, you discussed three types of control architectures — centralized, decentralized, and distributed — and outlined their pros and cons in terms of privacy and ease of control. In which direction do you think Germany is headed?
Prof. Monti: Today we have a top-down system, with centralized intelligence distributing energy to the distribution grids, which send energy passively to the consumer. However, the European Commission plans to create what they call a “customer-centric grid” in Germany and Europe. A truly customer-centric grid makes customers the real kernel of the system, which means the grid has roughly 40-50 million points of control. Such a grid cannot be controlled by a fully centralized system; the work has to move in the other direction. The system will then certainly need to be more distributed, which makes the real open question what this distributed system will look like. I think it’s logical to imagine some sort of bottom-up aggregation approach, in which city quarters coordinate within themselves, these quarters then aggregate at the city level, cities then aggregate at the district level, etc. In the end our system will be a mix, with different centralized layers in an overall highly distributed intelligent system.
Priya: The same paper talks about demand-side management of electricity in buildings, in particular storing electricity using thermal devices such as heat pumps. How does this thermal device strategy interact with people’s need to heat their homes?
Prof. Monti: We don’t have too many heating or gas grids in Germany, so when we think about demand-side management with heating, we imagine that buildings will eventually use electrically-driven heating systems such as heat pumps. This transition makes sense because 40% of Europe’s energy consumption is related to buildings, and heat pumps will give us a high level of flexibility for demand-side management. Heating system storage is also extremely cheap and uses existing technologies, so we can implement it today to significantly increase grid flexibility without significant cost. Obviously, no technology is the silver bullet solution to all our storage problems. Electricity-based heating puts more strain on our electric grid in the winter, when people need to heat their homes. This seasonal dependence may change, however, since central Europe is experiencing hotter and hotter summers during which heat pumps can be used for cooling. Another disadvantage of heat-based storage is that heat is lower-quality energy than electricity. Electricity is flexible, reversible, and usable as any other form of energy, whereas heat is not. Heat-based storage will therefore be used in conjunction with battery-based storage, and the question is in what proportions these storages will be used.
Priya: Are there any other interesting connections between electricity and heat for demand-side management?
Prof. Monti: Absolutely. Thermal processes are everywhere. For instance, large commercial buildings such as supermarkets and malls have a large cooling capacity, which means you can store or remove a large amount of heat without changing the temperature too much. Such buildings have a high flexibility in their demand. Data centers are another interesting case. Computers generate heat, so from an energy perspective, data centers are big locations where electricity is converted to heat. You can explore a lot of interesting relationships between heating, cooling, and electricity in the context of data centers. My lab actually has a new project dedicated to optimizing data centers in the context of Smart Cities.
Priya: You brought up the issue of seasonal dependence. What approaches are being taken to store energy across seasons?
Prof. Monti: At a large system level, people are looking at the gas grid for long-term energy storage. The existing gas grid has been over-designed, which means it has a lot of excess storage capacity. The German Association of Gas Companies also calculated that it’s possible to insert forms of gas other than natural gas into the gas grid. Once power-to-gas techniques mature, we can use them to convert electricity into hydrogen gas for storage in the gas grid.
Priya: Could you tell me about some of the grid simulation models your lab is working on?
Prof. Monti: One area in which we’re quite strong is real-time simulation, where we can replace our grid models with the reality of the grid. This type of work is critical in the energy system. We can’t play directly with the real grid, because the wrong move might result in a blackout; the laboratory substitution allows us to test real systems without real repercussions. The goal is to make these tests bigger and more sophisticated over time to cover a larger grid system. My lab contains the second-largest real-time digital simulation (RTDS) in Europe, with 8 racks, but the German grid is still much bigger.
Priya: As you just mentioned, we can’t test Smart Grid technologies on the real grid system because there may be blackouts. Are there any other reasons you think Smart Grid simulations haven’t been tested on more realistic systems?
Prof. Monti: One definite reason is that operators are very conservative. One mistake can cost them a lot, so they don’t like to take risks. Another reason is that the European Smart Grid transition mostly affects the distribution grid, which used to be completely passive. It’s a huge jump for utilities to think about adding intelligence and completely monitoring the grid. However, state-estimation in distribution grids is now a normal topic; people start realizing over time.
….One sad reality about power systems is that we always need a big blackout to make people aware that there is a problem. Even if it’s possible that the system will collapse, people don’t care until it actually does. Historically, all major power system changes have occurred only after a catastrophic blackout. For instance, SCADA grid monitoring systems were only invented after a big 1965 blackout in the northeast of the US. In Europe, the last newsworthy blackout was 10 years ago in Italy. This blackout was due to human factors, a miscommunication between Switzerland and Italy in which the operator made an incorrect decision.
Priya: I’ve been told that Smart Grid technological problems are “solved,” and now we just need to think about market tariffs and communication. What’s your perspective?
Prof. Monti: I’ve heard this statement many times, and it depends on exactly what you mean. I agree that we already have the technology that would allow us to run a Smart Grid. However, I would disagree that we know how to really use that technology to run a Smart Grid, even at a fundamental level. Nobody has any clue how to build a safe Smart Grid infrastructure, from a cybersecurity perspective. If we were to build a truly customer-centric system now, it would be down in five minutes. We don’t know how to do it, and we don’t know how much research is needed to do it.
….The European Commission wants to have a 100% renewables-based grid by 2050. To me, one of the most intriguing questions is how we can run this grid. Right now, our grid works because we have a significant number of gigawatt-sized power plants that provide huge energy storage in the form of coal and oil. These plants can automatically and perfectly enter into action to balance the grid. In a 100% renewables-based grid, with a huge amount of PV and wind and no significant quantity of conventional energy, who will provide this balance? It’s technically possible with inverters and controllers, but we’re talking about millions of controllers working together. We’ve never built something that complex, and I haven’t yet seen any technical architecture that would guarantee a 100% renewables-based grid to be as reliable as it is today. As a result, our energy group at RWTH wants to start a project researching how a completely electronics-based system can work at the country or continent level while still being as good as our system today.
….In general, much Smart Grid technological research lacks system-level thinking. A lot of things have been tested in reasonable, good field tests, but only in the context of an otherwise large and robust grid. Even if your field test isn’t that good, nothing really happens; the overall grid still balances things out. The question is, if I develop a “fantastic” microgrid model, will it still work if everyone is using that model? We don’t know. We don’t have the appropriate system-level picture to understand how a weak, large infrastructure — as opposed to our current robust infrastructure — will accept the technologies we’re developing.
Priya: Some of my interviewees in Germany have called Smart Meters rubbish, in particular saying that they’re not useful for people understanding their own consumption. One of your papers, however, talks about Smart Meters as being useful for state estimation on the electric grid. What’s your take on Smart Meters?
Prof. Monti: First we have to agree on a definition for “Smart Meter.” Some people think of a Smart Meter as a device that provides billing information. However, an emerging idea is that a Smart Meter is a gateway of information between the customer and the utility. That gateway is fundamental in a customer-centric grid, since we need a way to bidirectionally talk to customers. Smart Meters are customer-level measurement points that can tremendously increase our knowledge about what’s going on in the grid, which is vital to avoid stray investment in how we develop the grid.
….If you want to be provocative, you could ask me whether we need meters in the traditional sense. In the future, maybe we don’t. Why? Currently, our cost of electricity depends on the cost of fuel used to generate electricity, so customers are charged accordingly for how many kilowatt hours of energy they used. In other words, the system is OPEX-driven. In a 100% renewables scenario, the system works differently. Renewable energy generation has a close-to-zero operating cost but a huge capital cost, so the system is CAPEX-driven. In any economy, a CAPEX-driven system charges for intensity, not flow. To understand this concept, think about telecommunication bills. We used to pay for the amount of mobile data we used, but now we don’t pay for the amount; we pay for the speed.* Similarly, we may charge for energy services instead of energy quantity, negating the need for a meter.
Priya: Could you talk more about these “energy services?”
Prof. Monti: Currently, the utility is interested in selling as much electricity as it can. However, that’s not sustainable. We’ve been discussing a new vision with utilities in which they don’t sell electricity to heat a house, but instead sell the guarantee to a warm house. The utility can then manage the system in the most efficient and low-cost way. Such solutions are attractive to customers since they save money and time, shifting management and maintenance responsibilities to the utility. Such solutions are also necessary for utilities to stay competitive. Europe has a free supplier market, which means customers can opt for the energy provider that offers the best solution. A similar transition happened in the telecom industry. When I was growing up in Italy, what is now Telecom Italia was a monopoly that didn’t care about customers. Now, the company fights like crazy to get customers, calling them with attractive offers. The same thing will happen with electric utilities.
Priya: How well do you think technical and social realms of knowledge are being connected in Smart Grids?
Prof. Monti: This topic is a very critical one. Traditional engineers tend to imagine that their technology is perfect and people should use it, but the reality is different. I think the Smart Home era has been a clear lesson for everyone because the market hasn’t taken off so far; people get excited playing with their real-time energy consumption data for a few days, but then they get bored. Nobody’s been able to create long-term engagement, but in a customer-centric grid, we need to have the customer aboard. We do have some good experiences in that area. Within RWTH, my institute cooperates with the Institute for Future Energy Consumer Needs and Behavior. We’re also conducting a large project called Forschungscampus, which involves our humanities institute on technical communications. This institute focuses on how to communicate technical facts to people.
….Communicating technical facts is extremely important, since technology acceptance is majorly affected by how you tell people the story. There are some good results in this area. As part of the FINESCE project, we created a local microgrid in a small town in Denmark, equipping 25 houses with solar PV panels, heat pumps, and electric cars. The team leading the Denmark test site employs human factors experts, who approach the people in a nice way to get them enthusiastic about the project. One day, we organized an open house in the village to show how the system works. When one of my project colleagues arrived, he saw someone he didn’t recognize giving a PowerPoint presentation and demo in their house, explaining how the FINESCE project works. It turned out that the presenter wasn’t a project team member, but actually one of the customers from the village! The customer was the best seller of our project because he was excited and proud to be a part of it.
Priya: What policy changes do you think are necessary to support 100% renewables?
Prof. Monti: Europe needs to do a better job operating as a union, in every department including energy. The European Commission recently wrote a white paper called Energy Union that addresses this point. Right now, the situation is a mess. Each grid operator has its own rules, and rules differ even between operators in the same country. Companies need to build a different models of renewable energy technologies for each operator because each operator has different requirements. We need more harmonization to build a real European energy market, so that products can be used all over the place without costly modifications. A bill or legislation can help with this harmonization.
….I also think it will be important to introduce local markets. Right now, the energy system operates inefficiently because all power has to go through the centralized transmission grid. Renewables are installed on the decentralized distribution grids, and pushing them into higher and higher levels of grids creates losses. The best interest is for renewables to be used locally, and it’s important to think of a market that encourages energy to stay where it’s produced as much as possible. That’s not happening right now. Particularly in Germany, the solar panels on people’s houses produce energy when nobody’s at home, so this energy is injected into higher and higher levels of grids. A local market might encourage more battery installations to optimize grid operation and increase grid flexibility. Along this vein, we conducted another trial within FINESCE to match commercial energy consumption with the energy produced by a PV-based power plant, to reduce losses and make best use of the energy.
Priya: Do you have any observations about Smart Grids in Europe as compared to other countries?
Prof. Monti: I often tell my students that the European and American Smart Grid situations are completely different. I lived in the US for about 9 years, and the power system’s quality and reliability is very bad compared to in Europe. Therefore, the US is trying to modernize its infrastructure and generally make its infrastructure better. In Europe, our interest is in increasing the penetration of renewables. Since a renewables-based grid is hard to manage, we’ll be happy if the future European Smart Grid is as good as the grid we have today. There are also other differences, such as the fact that nuclear is acceptable in the US but not in Europe, that completely change the story.
….An interesting global commonality is the topic of microgrids. In the US and India, people want to build their own grids because the main grid isn’t that good. In Europe, companies and other large entities are trying to optimize their premises for energy efficiency. To me, microgrids are the most international topic, where you can really exchange experiences.
Priya: Is there anything else you’d like to add?
Prof. Monti: As I mentioned earlier, there’s no silver bullet technology. To restructure the energy system, we need different technologies for different purposes. Batteries are the only option for providing quick support to the grid. Heating-based storage works well in the 1-2 day range, since (for instance) a well-insulated building could shut its heating system down for 24 hours and use stored heat energy instead. The gas grid is likely the only option for dealing with the seasonal effect.
….One last point is that we’re never able to fully predict technologies. I think that the grid as of when I retire will look technologically similar to today’s grid. However, I think the grid my child will see will look completely different. We as humans are never really able to see the new big thing that comes and completely changes the world. For instance, when the first international GSM phones came out and you could travel the world without a new phone, we thought we had everything. But then the iPhone came out, rapidly and quickly changing the world of mobile. On the other hand, the utility sector has been pretty much the same for 100 years, since many people in this sector are very conservative. We’ll see what happens.
After the interview, Dr. Monti took me on a tour of his lab. Here are some pictures!
* Older German cell phone plans used to charge customers for the amount of mobile data they used. Today, many German cell phone plans have unlimited low-speed data with the option to purchase some amount of high-speed data. In other words, you’re not paying for the number of bytes of mobile data (i.e. the flow), but for the bandwidth of your data (i.e. the intensity). [back]
** Technical specifications: RWTH Aachen’s RTDS has 8 racks, with 4 processing cards per rack and 2 parallel processors per processing card. In total, this makes 64 processors. The RTDS normally runs simulation models in real time with a time step of 50 microseconds. Typically (but depending on the complexity of the model), each rack can solve 80 bus bars in real time. It is also possible to run real-time simulations using a so-called mini-step of 2 microseconds, but then the size of the model is significantly reduced. In addition to the computational boards, RWTH’s system has digital and analog I/O channels; digital cards that can generate digital messages emulating automation communication messages (e.g. IEC 61850 messages) using data from the simulation; and digital cards allowing synchronization with an external clock typically generated by a GPS signal. [back]