All Episodes

[Episode #10] – Grid Architecture of the Future


What kind of grid architecture and markets will we need in order to actually operate the distributed, decentralized grid of the future? What sorts of regulatory models will be needed? And what does it all mean, from a philosophical point of view, about how human society is organized? How can mere mortals begin to understand these subjects? Never fear: We’ve got you covered, in this ultra-geeky yet accessible episode.

Guest: Lorenzo Kristov, Principal, Market and Infrastructure Policy with the California Independent System Operator (CAISO).

On the Web: Lorenzo’s profile on LinkedIn

Recording date: February 5, 2016

Air date: February 10, 2016

Geek rating: 11

Chris: So without further introduction let's bring Lorenzo into the conversation and learn a bit more about his vision for the grid, welcome Lorenzo to the Energy Transition Show.

Lorenzo: Well thank you very much Chris and I look forward to our conversation.

Chris: Great so you have published so much and offered so many interesting views of the grid both present future that it is sort of hard to know where to start with your material but I think I'd like to start with an interesting little thought experiment that you published in Fortnightly Magazine in May last year which we'll link to in the show notes. And in that you look back from the year 2050 to the year 2015 and wonder why the integrated decentralized electric system paradigm which had become dominant by 2030 was so hard to see coming in 2015. So before we really geek out and get into the nuts and bolts of your vision I think it would be helpful to start there. In general terms, what is this evolution that you see commencing and what is the integrated decentralized electric system paradigm.

Lorenzo: I think where I would start is by pointing out that for many decades, the electric power system has operated under a structure which featured large central station power plants. These are the big plants, anywhere from 50 to 100 megawatts and capacity, up to even thousands of megawatts for large nuclear or fossil fuel plants. And the system was designed and has operated such that these large plants, usually in locations that are not close to population centers, produce energy. It moves over the high voltage grid and then from there, out along the distribution systems to customers in one direction it's flowing from the center out to the edges of the grid to the end use customers.

Lorenzo: What is happening in the last few years and what I see as growing ever bigger as time goes on is that that one way power flow from the center out to the customers is no longer going to be the dominant way that electricity is moved and transacted, largely because of technological change. Change on the awareness of customers and the interests of customers and having more control over both how they use energy and of the impact of their energy use on the system and on the earth and the technological developments that have made it cheaper to have more powerful local small scale resources that provide energy. So we're seeing instead of this big central station paradigm one in which small scale, local, distributed, or decentralized resources, now can be the foundation of energy supply, can be customized to the needs of individual customers users whether they be commercial, industrial, residential, government, agricultural, and can also be customized to the locale in which these entities reside, in other words different climate zones different state, economic, and regulatory frameworks, different state environmental policies and so on. What that means for the grid however is that we need to rethink how it all works because we really had it down. We knew how to operate a system based on centralized power plants and one way power flow. Now we have to think, 'well if we don't have this one way power flow anymore' instead devices at the ends are putting power onto the grid and power flows are switching directions, and factors that we thought would be stable from one moment to the next are showing a lot greater volatility and variability. How do we operate in that regime? What kinds of communications do we need? What kinds of operating procedures, sensing devices, information sharing and how does that change how we plan the infrastructure of the system? Essentially this revolution to the small scale local decentralized and renewable resources is really affecting every aspect of how the electric power system is operated.

Chris: Great. That's the perfect transition because I next want to talk about your new paper with Paul De Martini and Jeffrey Taft titled two visions of a transactive electric system. And we'll link to that in the show notes as well. In that paper, you remind industry futurists, and I suppose I am one. You take us to task, suggesting that we might be getting ahead of ourselves with all this talk about integrating a large proportion of DERs and redesigning markets and utility business models and regulatory frameworks, to accommodate them and instead we need to start thinking about the operational side of things. How the grid can actually work from the retail customer, all the way up to the bouncing authorities. And you outlined two distinct visions here which I'll just take a quick minute to introduce before we get into the pros and cons of each one so the first architecture you call a grand central optimization. And in this architecture a TSO controls and optimizes a complete centralized system, on which it has total visibility from the top to the bottom. And on which it may also operate wholesale markets as an independent system operator ISO or regional transmission organization (RTO). And this vision would basically be an extension of the system we have today. Then there's the second and in some ways the opposite vision which is decentralized where as you put it the optimization at any given layer only requires visibility to the interface points within X layers above and below. And in this paradigm the TSO would only see a single virtual resource at the TD interface and would not have visibility below that level. So let's start there remind us what these layers are and what you mean by optimization at a given layer.

Lorenzo: OK if you don't mind I'd like to back up just a step because I think it's worth putting a little more context about the industry and how things change. And you started out by noting that I was taking industry futurists to task in a sense and that that task may be best epitomized by the famous economist joke that has as it's punchline, 'assume we have a can opener'. I'm sure everyone is familiar with that one right.

Lorenzo: So my point is that because the change that I just described to this de-centralized structure is really so dramatic and affecting every aspect of the industry. There really is a tendency to try to address all the issues at once, and parties who are interested in regulatory structure focus on the regulation and parties who are interested in the markets, focus on the markets and parties who want to develop resources. Focus on the business cases, how they're going to earn revenues for their products and services and so on. And it's natural to have those kind of siloed focuses. But what gets overlooked is the fact that the laws of physics are not changing. Electricity is still going to move through wires. That part is not changing and it can be tempting to develop really slick market designs, regulatory procedures, rate structures and so on, and forget about the connection to the physical system on which it all has to work. Ultimately electricity is a large scale complex system that obeys the laws of physics. And if our policies or our market designs or our regulatory frameworks forget about that connection then you have problems. Because the markets need to come up with results that say I'm going to move energy from Point A to Point B. The underlying physical system has to be able to accommodate that. So that's why I'm taking the emphasis and with the colleagues I'm working with. You mentioned Paul diMartini and Geoffrey Taft. We've been really trying to focus our discussion our papers on the actual operation of a physical complex large scale system, and start from there, and let the market design come out of the physical requirements for reliable operation and let the rate structures come out of that. Let all these other things build upon having a solid operational framework so that when I get into those two visions that you just summarized which was a very good summary those really are asking the question, how is this physical system, this ultra large scale complex system that has to obey the laws of physics. How is it going to work? Because we can come up with the greatest market designs, but if the underlying physical system isn't compatible with it. Then it is not going to work very well. And then there's another piece that I need to fill in here that I think will be helpful for listeners. You went through the acronyms at the beginning between TSO - transmission system operator, ISO - independent system operator, RTO - regional transmission organization. All of these are entities that are concerned with the high voltage transmission system. In contrast, you get two D.S.O. distribution system operator or the distribution utility and the DER, the distributed resources that are connected to the distribution system. That's another entity and typically the entity who's running the transmission grid is a separate entity from the entity who's running the distribution system. Even in a large vertically integrated utility, where the utility owns transmission and distribution and power plants and all those things, the departments are typically separate and don't really talk to each other very much and in the old world where power was just flowing one way they didn't need to, and it didn't matter a whole lot. In the new world, it will matter a whole lot. And so starting to create new relationships between how do the people who run the transmission grid interact with the people who run the distribution system. That interaction now is a crucial focal point of the redesign of the whole system that we need to think about. And so that's why I was focusing on that and that's what you captured in the acronym of the transmission distribution interface or the TD interface. Because there are separate entities on both sides of that interface and because traditionally there's not been a whole lot of emphasis on coordination between them. Now there needs to be and so we need to think about designing how that interface is going to work in the future.

Chris: Gotcha. OK so, when you talk about these different layers, are we just talking about sort of a high voltage layer and sort of the low voltage lawyer or are you looking at more layers than that in this layer cake.

Lorenzo: Well we can start there I think that's the most fundamental layering that we need to talk about. And so probably the best way to get into that is to think about the fact that something which is a large complex system has mechanisms that coordinate the activities of everything that's part of that system. So on the transmission system there are large generating plants. In California, we have hundreds of them. Seven, eight hundred, of different sizes at different locations and there are points. These TD interfaces where there are lines leading out and emanating towards the end use consumers. Somehow that whole system has to operate in balance. In what we call real time instantaneously supply and demand have to be in balance because if they get out of balance then you start creating operational problems on the grid. Everyone's aware for example that AC current runs at 60 cycles. Well if you get an imbalance such that you have excess supply and not enough demand, it starts to push that frequency up. And there's a little tolerance with which it can move up and down and it's not a problem but you get outside the tolerance and you start have operational problems. The other direction, if demand starts to exceed supply then but frequency starts to go down which is another kind of problem. So it's important then for the operator of that system to have procedures in place, control mechanisms, sensing devices etc. that enable visibility into what's going on and mechanisms to detect little perturbations and make adjustments in the operation. Some of them through human intervention and some of them through automatic controls that will keep the system in balance and avoid any kind of significant disruption. Now those kinds of systems are different on transmission, which is high voltage which is an integrated network which means the lines all crisscross and interconnect in different ways. As compared to the distribution system, where the lines tend to be what we call radial. The line comes out from the TD interface and it just goes out in a straight line out to customers and you have more of these radial lines coming out and also they're at lower voltage, so a lot of the operational procedures, a lot of the response times in which you need to address a perturbation are very different. So now we get to the point where these distributed resources that are out on the grid and that could include rooftop solar, but also two, three four, five, megawatt solar farms that are like community installations or on the tops of warehouses or in industrial parks or subdivisions. They could be large battery installations that are within say an industrial park or university campus or a major medical center and we could have vehicle charging stations that are workplace stations where you might have facilities to charge 50 automobiles, or a fleet of government vehicles. So you now have installations that can be of significant size, not the huge power plants that you find on the transmission grid but of enough size on the distribution system, that they have a real impact on how that system is operating. So the question is, how can those devices out on the distribution grid, essentially not only use the grid to provide what they need but also provide services back to the grid. How can they engage in transactions. If I have a battery facility, and it can discharge energy to help relieve peak load on the system. Well, that device can now participate in the market and actually provide services and earn a revenue stream from doing that or aggregations say of 500 electric vehicle charging stations, could provide a service by having fast response time frequency regulation so that if the frequency changes a little bit it responds quickly. That's a service that helps out the grid and it can earn revenues from that. But that means then, we need a control system that's going to be able to tell those devices what to do, be able to monitor what they're doing, be able to see what is needed at different locations on the grid, et cetera. So this is part of the system redesign that I was talking about.

Chris: So that kind of coordination that you're talking about there that's what you mean by sort of optimization within a given layer.

Lorenzo: Yeah, I think that's a good way to think about that. It's being able to have a view of all the facilities that are connected, and be able to utilize them in the best possible way to meet the needs of all the customers who are depending on the grid, and also to maintain reliable operation and be able to respond to disturbances. That's in a sense what the optimization is all about. So the blessing I was going to say, just to tie back to the two visions that you started out your question with, would be to say the two ways of looking at this are do we want a future system where the transmission operator who operates the high voltage grid essentially extends its optimization, or its controller, its management of the system all the way down to these distributed devices, and sees all of them, and is making decisions about which ones to use and how to use them for the entire system as a whole. That's the Grand Central optimization paradigm that I called in paper. Or, as an alternative, create a really clear boundary at the TD interface and say, the transmission system operator is only concerned with what's going on above that boundary, and what's going on across that boundary, and below that, the distribution system operator is now the entity who's in charge of managing all of the coordination and optimizing at the local level in that area. And that would be a very different structural arrangement in terms of the companies that are doing things and what their responsibilities are and also the regulatory frameworks, in other words the relationship between federal regulation which governs high voltage transmission and wholesale markets versus local or state regulation that governs distribution and retail markets. So all of these things are part of the evolutionary change process but I think the question that I wanted to pose with that paper, was to say let's think about which of these futures is more desirable from an operational perspective. Which one will give us a more stable system a more reliable system and ultimately a more efficient system that is more secure against possible disturbances and achieves the bigger goals of the society, which include you know equity and reduce greenhouse gas emissions, cleaner energy and so on.

Chris: Right and I do want to get into some of those questions but I just want to. Sort of get the architectural vision clear for our listeners who may or may not of have read your paper so. So just to sort of try to rephrase this a little bit. You say the Grand Central optimization could actually have as much DER participating in it as any other design on both the customer and the utility side of the meter, individually, or in aggregations, including markets for unbundled services that DER can provide and peer to peer transactions so. So in either architecture, we could enable all of this DER growth, but the TSO would see the DER either as a single resource, located at the TD substation in what you call the minimal DSO model, or it would be able to see the DER all the way down to the actual locations on the distribution circuits in the total model. So, in the minimal D.S.O. model for example the TSO might see 1000 EVs connected to the grid as a single resource, whereas in the total ISO model it would know the physical address of every electric car plug into the grid. I'm on the right track so far.

Lorenzo: I think I need to correct that a little bit.

Lorenzo: The total DSO is the one in which the ISO sees a single resource at the substation. And the DSO is managing everything below it. That's the layered version. Ok ok in the minimal DSO, that's where the ISO is seeing thousands and thousands of distributed resources and the DSOs role I call it minimal because it's playing a coordination function but it's not really making any decisions about optimizing the resources and and what services they can provide and so on.

Chris: Okay then moving over to this layer decentralized optimization the other architecture, instead of having all of these do you are bidding directly into the wholesale market and being dispatched by the ISO, a D.S.O. would aggregate all the DER in a local distribution area. Another lovely acronym here, LDA located below a single TD interface substation or LNP pricing node. So here the total D.S.O. would bid all of the DER under its control into the wholesale market. At the TD interface as a single bid and the TSO would only need to concern itself with balancing the interchange between the TD interfaces without needing to know all the detail going on below that level and a D.S.O. would have the responsibility for balancing supply and demand from the distribution level all the way up to the LDA, is that correct?

Lorenzo: Yes that's right.

Chris: OK so now with all that out of the way. I tell you or as I'm really struggling to mentally track all this stuff I hope our listeners are not struggling as much as I am for the three people in the world who are still listening, let's discuss the pros and cons, so you say of the second architecture of the layer decentralized optimisation. If your objective is to have a high penetration of DER, that's the preferable architecture and that's because of two important issues that you talk about. Two are bypassing and scaleability. So should we elaborate on those a little bit?

Lorenzo: Sure, those are examples but I think behind them all is the question of what's going to be a more stable system from an operational perspective and stable in a sense means the ability to maintain reliable operation in the face of different kinds of disturbances, and the layered is really much stronger in that regard.

Lorenzo: The analogy that I like to use is to think about natural ecosystems and biological models. And, I use the example of say a cell in your body. Which is an example of what I would call the layered composition. You know each cell has to look out for what's going on inside its cell membrane. It is maintaining homeostasis. It's got little energy plants in there and the mitochondria, it's got a membrane through which it interacts with its outside world. But in a sense it's responsible for its own dynamic equilibrium. Now at the same time, that cell in your body maybe it's the cell in your liver. So it's part of your liver and as such, there is a function that the liver performs that that cell has to contribute. So while it is doing its own thing and maintaining its own stability and health and well-being, it is also contributing to the function of that organ. Similarly, your liver as an organ has to perform what its functional responsibilities are. And yet it has interfaces with other parts of your body. It's getting information from your nervous system from your bloodstream from your endocrine system and so on. And part of its responsibility is that it has a function within the complex structure that's your body, that it needs to perform so it's looking out for its own health and well-being in a sense. But it still has a function in the greater whole.

Lorenzo: Take that another step forward. You know you are an individual I'm an individual and yet we have a place inside our families inside the neighborhoods that we live in and our circle of friends and relatives in the places where we work and our relationships with our coworkers. So we need to look out for ourselves, for our own health and well-being. And at the same time the dynamics of our interrelations within our different overlapping social networks is also a crucial part of our health and well-being. And then you can take that up any level you want we're part of say a city and we participate in the city government more or less we vote in elections we respond to laws. We follow traffic lights when they give us signals on the road about what to do. So there is this kind of layering that happens where there is in a sense an optimization happening at each level. And. going back to my body. My brain does not need to see what's going on in every cell. It doesn't need that information because it says, cell you take care of yourself and stay healthy. I'm sending general instructions for what I want your liver to do. So it's that kind of thing that we can imagine for the power grid. You can take a building and it's not too far in the future that it will be very cheap for a house to have rooftop solar electric vehicle charging station perhaps thermal storage where it's chilling water during the day when energy is cheap for use later on for air conditioning or battery storage and electronic control systems that control the devices in your house so that your house becomes a little energy system on its own.

Lorenzo: That does a lot for itself because with solar panels it's generating electricity and with batteries it's storing it and providing for your own needs. And yet that house or smart building a commercial building can be part of a larger structure which is also optimizing at the next level up. The commercial building might be part of a commercial or industrial park or it might be part of a major medical center or university campus. The house may be part of a subdivision of 500 houses that has also some community solar and community storage and the electronic control systems that enabled that subdivision to operate as an electrical unit and then similarly that electrical unit the subdivision or the industrial park can be part of the next layer up which is the local distribution area that the utility operates that may have you know five thousand customers and a whole lot of commercial buildings and different kinds of merchant renewable power plants and batteries and so on and it's operating that system.

Lorenzo: The key thing is that when a smart building is embedded in a micro-grid the micro-grid knows the buildings there but it doesn't have to know everything that's inside that building. It just needs to see what's the interface what's my interaction with that building. Similarly the D.S.O. or the distribution utility that's managing that local distribution area doesn't need to see what's inside the micro-grid. It just needs to know that it's got an interface with the micro-grid and we have rules for how we coordinate around that interface. But meanwhile underneath it, inside the micro-grid, their taking care of business for themselves in their own way. So you bring that up now to the level of the ISO and the transmission system, and that interface with the distribution grid.

Lorenzo: The ISO doesn't need to see what's going on inside that local distribution area but it knows that it's got a relationship with the distribution utility that's maintaining a certain stability and certain protocols for managing the flows across that interface. So this is what we call scaleable. That the same concept of a smaller unit doing something, managing its interface with the next larger unit up, can scale at different levels, as the technology enables us to do that and its that same scaleability that I would say mimics what we see in biological systems in ecosystems, and on that basis it's a very resilient system. You know right now there's a lot of discussion for example about cyber security or physical security threats to the grid or or even climate events, weather events like big hurricanes that take out portions of the grid and you have huge areas that don't have electricity for perhaps weeks at a time. Well with this scaleable system, each of these smaller units could disconnect and be a self-contained power system that could operate for a day or two days or perhaps even weeks, and the possibility of taking out a huge area of the grid through one of these major disturbances becomes really diminished once you have this scaleable structure.

Chris: It's really interesting that you describe it that way, because it really kind of brought to mind for me a fractal you know a structure that's replicated over and over from a tiny scale all the way up to a large scale. It also made me think actually back to my days as a software engineer. Because when I first started coding we were all using procedural coding. That's what we were taught and then eventually we moved to object oriented coding where you have a self-contained object with well-defined interfaces so you can communicate with objects but you don't really need to know what's going on inside of each of those objects and you don't need a master procedure that understands the functioning of the whole system. You just have to have enough coordination that the objects can work together. Are those also apt metaphors?

Lorenzo: Yeah, I think that's a very good analogy. In terms of the whole system, this is where the architecture comes in, because you know as we're thinking about this transformation that's happening in the industry, we want to take an architectural perspective. In other words, let's look at the whole system, all the way from the level of the Western interconnection and you know all the western states and what parts of Canada and Mexico are all part of one interconnected grid. And that whole system, all the way down to all of the end use customers. It's one really physically interconnected electrical system. Let's look at that and then start to think about the architecture of the operational structure, the coordination and control structure in a way that maximizes the stability and efficiency of that whole system. So we're looking at it from a design perspective as a whole system. But then we're creating these units. At different levels and these layers at different levels such that the control and the management of variability and volatility can be de-centralized at different levels in the way that makes the most sense for the local needs.

Chris: So your paper moves on to discussing how grid power futurists are thinking about applying the locational marginal pricing concept to the distribution grid instead of just the transmission grid aka the LMP plus D, where D is for distribution. And we're thinking about this because it offers a way to unbundle the services the DER can provide. Where and when they exist on the distribution grids. And this is something that we've talked about in previous episodes about how we really need to unbundle some of these things so that for example storage can get properly valued for all the different services that it can provide. On the grid. So what are some of the issues wrapped up in this question?

Lorenzo: I think the first one is that the distribution utilities are not accustomed to thinking of devices on the distribution system as being able to provide services to them. They largely view the interconnections just as they view customers, they are entities that are receiving service. The grid is providing a service. They're delivering electricity. The traditional one way electricity flow, it comes off the big power plants we deliver it to the customers and they are then centrally the recipients of the grid services. It's been I think just kind of an automatic response which makes perfect sense that as rooftop solar has been coming along, and electric vehicles and charging stations, to continue to see these as facilities or devices that are taking service from the grid and the change in thinking that's beginning now to take hold is to realise that many of these devices at the edges of the grid can provide service, as the behavior of distributed resources raises new operational challenges much of it having to do with volatility of say, voltage and frequency. Some of the devices on there could be under the control of the distribution utility and can help manage the operational challenges. And so now we create new kinds of relationships where the distributed resources, instead of just taking the service of the grid, they're now providing a service that they can get compensated for, and the distribution utility sees, well I've got these five, six, ten different entities around on the grid that I can call upon that can help me manage voltage fluctuations or frequency fluctuations or power quality or phase balancing. These are just different aspects of the operational things that distribution utilities have to deal with, but the distributed resource has now become a source of service to do that. So the first step after recognizing that those services are possibility is to actually define what the services are. And this is where we're at an early stage I think in the industry defining what are the services that we want distributed resources to provide, and define them in a pretty specific manner, in other words, What are the performance characteristics that this resource needs to have? It needs to respond in four seconds say or it needs to receive a signal and do something in response to that signal. It needs to be available 24 hours a day, whatever. But to start to spell out the performance requirements for each service, and then to begin to think about what is the value of that service. How would the utility procure the service. How would it compensate for the service. So these are all, I think, at a very early stage of consideration, the distribution resources plan proceeding in California under the California Public Utilities Commission is taking up this topic. But it's still at a pretty early phase.

Lorenzo: The other thing that I would mention and that was in the paper we've been talking about. Is this notion of markets versus control. And the simple way of thinking about markets and controls is that when you have a market there are prices in which say a supplier is willing to sell something and a buyer who is willing to buy something, but it's voluntary. The price, for the same price. Some suppliers may be willing to supply and some buyers may be willing to buy and others may not. It's very much of a voluntary decision and on any given day or any given moment, somebody who bought at that price yesterday might not buy it at same price today. So it has this voluntary aspect to it. And markets on the electricity system have a lot of that aspect to it as well. Controls on the other hand are things that are automated. They are hard wired so that a certain signal goes out and the device that receives that signal is going to do very specific action in response to that signal. Assuming that the system works, of course with a malfunction that all bets are off but assuming that it works the way is intended. There isn't that voluntary aspect in it. When you get to managing a complex system that is physically based on electrical systems, then you can't rely completely on one of the other. Well, let me modify that a little bit. You probably could rely on a complete control system but that may not be the most desirable way to do things. As we found with electricity, the idea in the 1990s to create markets in the transmission system for wholesale trading of electricity. We had a more purely controlled type of system with the utilities that made all the decisions about what resources they were going to dispatch, what resources they were going to procure. They built a lot of their own resources and the regulators oversaw that. But starting in the 80s and then moving with restructuring in the 90s, it was realized that getting competition when you are getting investment coming from profit seeking investors rather than just the utilities was going to be economically beneficial, so let's create competitive markets. So in many ways those wholesale markets have been very successful and they've been functioning well for 20 some years now. So you take the ISOs in the East, New York, New England, PJM, Midwest California, those wholesale markets work pretty well.

Lorenzo: As we get down into the distribution system, there's a tendency to say, 'well let's just do markets there too' it's markets all the way down because they've clearly shown themselves to be good.

Lorenzo: And part of what I wanted to do in the paper was just offer a cautionary note to that, that there are limits to how well markets can operate. When you get into control of physical systems. Because there are certain things that happen on very short timeframes where that voluntary aspect of a market response may not get you what you want.

Chris: And you also point out that the market approaches that LMP+D won't necessarily support the kind of long term capital commitment as necessary when you're trying to build out a high DER system.

Lorenzo: Well that's true as well and I think we've seen that in the wholesale markets as well, that locational prices are very good for providing price signals in the short term, day ahead hour ahead, real time, to get generation units and demand response say, to operate in a way that's compatible with the needs of the grid. But what we're seeing in most places is that those spot prices aren't necessarily sufficient to stimulate investment because they fluctuate too much. You know you need more longer term contracts that provide price certainty that make investment in large infrastructure more bankable. So there is that aspect as well but that's not particularly new to just the distribution system. I think the part that he knew to the distribution system is that we're operating more and more on very short timeframes, short response times, and short refresh times of the parameters that characterize the operation so that we need controls that can operate really quickly in those short response times. So while I I and my co-author started to sketch out in that paper is the notion that what we will need is certain controls that can act very fast but they can be controls that are part of integrated market hyphen control structures. And the example I like to give is for second regulation service that we have on the transmission grid. So regulation service gets a four second signal every four seconds from a centralized system that's looking at frequency and is sending out signal to generating plants every four seconds to very slightly adjust their output levels so that they're maintaining this dynamic supply and demand balance.

Lorenzo: Now the market part of that, is that not every generator on the grid has to follow those signals. We just need a certain amount of it so that we have a sufficient amount to deal with the expected fluctuations that we are going to have to deal with back can be really well estimated statistically. So now we go out a day ahead of time and we say well tomorrow for this hour we need 300 megawatts of capacity and that is able to be hooked up to the four second response and perform accordingly. Well, the market part of that is that day ahead solicitation and we do that in the ISOs market in the day ahead market time frame. We put in a constraint that says for this hour by 300 megawatts of regulation of service and that is generators that can move in the upward direction in response to the signals. That's a market to procure those 300 megawatts. We have thousands of megawatts of generation that's on the grid, and maybe four, five, six, eight hundred of it might be capable. So, we look at who wants to offer to sell it to us. And it's the market that procures the megawatts of capacity that will provide the service, once they sell off that service, they get paid for providing the capacity. But then, in order to deliver the service they're hooked up to an automatic control system. They don't have a choice once they're on that automatic generation control and they're getting that four second signal. They don't think about well do I want to respond right now or not maybe I don't feel like it. Or what's the price is the price high enough.

Lorenzo: All of that is over with. Once they sold us the capacity in the day ahead market, now in real time it's simply a control. They get a signal. They respond accordingly. So that's an example I think for folks familiar with great operations of a service that's pretty well known but you can characterize it as a market control structure because it's got both a market aspect to it and a control aspect to it.

Lorenzo: When it gets to real time operation, it's that control aspect thats governing the behavior.

Chris: Right, and so then as we move toward this more layer decentralized architecture, you say that we're going to have to sort of start leaving that sort of control market behind as you were just describing and move toward more of a transactive energy concept. Am I right so far?

Lorenzo: Well, no I wouldn't say we're leaving it behind and in fact I think the caution that I would raise. So this concept of transactive energy which is being talked about very widely, has been for a number of years, in the industry, based on the idea that these distributed resources out at the edges of the grid can engage and all kinds of transactions on the system, with each other or with the distribution company or with the ISO and so on. But I believe that we cannot abandon the need for a certain amount of control structures in place to maintain grid reliability. Now here is where I think that we would differ probably with some of the other people who are engaged in the industry discussions I've heard expressed the concept that if we get price signals correct, we can rely on transactive markets to deliver everything we need to maintain reliable operations. And that's where I think that's perhaps theoretically interesting, conceptually interesting, but in fact I don't know that systems as complex as the power system can forego control structures where you're actually sending signals and things respond automatically to those signals because we need that predictability not in a statistical sense but in a very precise sense at the location where it's needed. So you know I think the notion that we can abandon centralized dispatch, centralized optimization completely, centralized control structures. To me that doesn't seem very feasible, and I kind of view that as going too far in a sense. But that doesn't necessarily foreclose the opportunities for distributed resources to engage in transactions and revenues from that and provide services.

Lorenzo: I think all those things are possible, but it's complemented by the architecture of a control system that's going to make sure reliability is maintained in these really short timeframe response times that are needed to respond to perturbations.

Chris: So you are clearly making the point that we really need to understand our policy objectives here and be clear eyed about the implications of the architectural choices we make. If we go too far towards this transactive energy concept, for example I mean what could go wrong?

Lorenzo: Well, you know I don't want to say that I'm rejecting the transactive energy concept I think it's just a matter of maintaining the complementarity with the centralized control in a layered sense. I'm not saying, you know the ISO controls everything either. It's just recognizing that these are all different parts of the system that play a role just like individual cells in the body do their own thing. We wouldn't expect individual cells to do the right thing to keep our body alive in the absence of a central nervous system and a circulatory system. And you know an endocrine system that all have to work together. So you have these centralized systems that are sending instructions that are moving information from one point to another.

Lorenzo: And you still have your liver doing what it's supposed to do as a liver and your heart doing what it's supposed to do. So it's really the system is comprised of all of these things and they all have roles. It's just not to go to one complete extreme or the other.

Chris: So how do we get from the architecture that we have today more toward this layer decentralized grid of the future I mean, I sort of wonder like what are the mechanics of getting from here to there. Will the common sort of integrated resource planning the utilities are doing today will that get us there.

Lorenzo: No, not by itself I think it requires more than that, which is a lot of the reason why we're writing these papers. I think the first step is to begin to envision a future that is like along the lines of what I described. That is, these decentralized systems at different layers. We see some examples of them in existence today, we see smart buildings some of them exist, we see campuses that have become micro-grids. They exist today. We see communities that are local jurisdictions that are starting to develop diverse resources and control systems that optimize the use of those resources. There's experiments going on to demonstrate that aspect of it. So that's sort of the first step for utilities, distribution utilities, and state regulators, in many states the expansion of distributed resources is moving very slowly at the very preliminary stage, so they're looking at it and they they go why do I need to worry about that. You know there's really nothing happening here but maybe there isn't much happening just yet. But, I think we will see with the declining costs and increasing power of small scale distributed resources and the electronics that enable automation then and decentralized control, these things will start to grow elsewhere. So part of what the distribution utilities and their regulators need to be asking is, how do we start to modernize the distribution grid to be able to sustain this more de-centralized future. What do we do first? So think about a system that was designed for one way power flows from the transmission system out to the customers.

Lorenzo: Now, it's going to have devices all over the place that will change the direction of power flows and change some of the operating requirements. Let's start to, Number One really have good sensor information about what's happening at various places of the grid. My understanding is that in most places distribution utilities don't really have good information that you might call situational awareness. What is the status of all of the feeders and substations that are part of the local distribution area? A starting place is to just think about that. What would we need to put in place as a distribution operator in order to have a really good up to date instantaneous snapshot every minute or every five minutes. And in some critical locations maybe every few seconds, as to what's happening on the grid. And then secondly, if we want to encourage this transition into more decentralized options, recognizing some of that is going to be driven from the bottom up anyway. So I think this is another mind shift, that especially policymakers and regulators need to realize. We're used to thinking about policies being made at the top and filtering down. And we're in a situation where a lot of the driving force for change is going to be from the bottom up. Customers are simply going to adopt things. There is this autonomous expansion of distributed resources that happens through individual customers making decisions. And I'm using customer in a big sense a customer could be a resident it could be a commercial building it could be a campus. It could be a municipality that wants to meet its climate action plan.

Lorenzo: Or that wants to customize resources and and do its own energy efficiency program to assist the low income community. And you know on the subject of municipalities I think it's worth mentioning that the distributed resources and the electronics and control systems that go with them are enabling new ways of thinking about municipal services so that you know we're used to having a Department of Water Supply and a department of wastewater treatment and a department of solid waste management and a Department of Transportation, that deals with local transportation. Well, how do we view these things how can we begin to view these things as comprising a whole system where municipal services utilize the electricity grid? But it's integrated with how we do water supply and how we do wastewater treatment and it's extracting energy value from the waste stream and extracting nutrient value that then becomes fertilizer for local gardens and agriculture. And in many ways these kinds of convergences of services and whole system thinking I think will be tremendously beneficial for cities and counties that continually struggle with budget problems and can afford to do everything partially because an awful lot of what they do is wasted or is uncoordinated are siloed.

Chris: Yeah, and I've thought a lot about that because you know it really sort of brings up this whole concept that the world is just moving faster than the sort of regulatory paradigms that we've attached to them. And it really makes me think about the need for regulatory reform in particular. In order to value the services of DERs more completely and accurately and to maximize that value and to eliminate some of the waste that we're talking about, we really need a whole variety of regulatory reforms it seems to me, and on the show we've talked about how storage is disadvantaged as I've mentioned earlier relative to thermal generators because of the services that it provides can't be properly valued in an unbundled fashion on the wholesale grid. And we've talked about how wholesale prices fall as more generation with zero marginal cost comes onto the grid. And rate design probably has to change in order to keep wind and solar from destroying their own value as they take over the generation role on the grid. I look at all of the issues at the regulatory level and I just sometimes wonder if it wouldn't be better to just get rid of concepts like wholesale and retail and qualifying facility and just shake off the whole calcified conceptual framework of grid power markets and start over with new goals and new architectures designed for the kinds of equipment that weren't really even a glimmer in anyone's eye when the current architecture was developed.

Chris: Or maybe we need to find a path from here to there incrementally I really don't know I mean what kinds of regulatory reforms do you think are going to be needed to support the kind of evolution that you're talking about here.

Lorenzo: Well I think you're on the right track I definitely agree with that. There are lots of regulatory changes that are needed. I tend to shy away from the idea of throwing everything out and starting over because when you do that then everybody's going to argue about what the new thing is and it could get infinitely bogged down. So, I try to look for ways that we can make significant changes that move us in the right direction. One thing comes to mind. I'll mention a couple of things that I've been thinking about that seem perhaps doable within the current framework. One of them is is bringing local jurisdictions local government agencies fully into the conversation. When we have discussions in the industry about the power system, it's you know the ISOs, the distribution utilities, the state regulators, the federal regulators, the big power companies, the DER developers, all of these enterprises and entities that have had traditional roles and then the end use customer at the bottom which is an individual customer. Well, what is missing from that is groups of customers, communities of customers, local jurisdictions, local agencies, that do water supply, the entities that are really grappling with climate change, with environmental impacts with budget constraints, with real things on the ground where they care about their environmental footprint the cities and counties have climate action plans. They care about collecting the energy value out of their waste stream and collecting the nutrient value out of their waste stream and materials, because to develop a really sustainable way of living on the earth as humans, we not only need to change how we do energy we need to change how we do waste. We can't just keep throwing stuff away because an awful lot of what we throw away is perfectly usable. In fact I heard a really cool song the other day that your friends on the extraenvironmentalist used as a tune in one of their podcasts, where the refrain said 'there's no such thing as waste there's just stuff in the wrong place.'

Chris: Or as the old saying goes, you can't throw anything away, because there is no way.

Lorenzo: Right. And we don't need to because there is value there. So I think part of what needs to happen then is in all of these industry discussions let's find ways to bring cities and counties into that discussion, the local jurisdictions the local government associations and make them real players in this, recognizing that this is a bottom up change that some real innovation is happening at the local level and let it be part of this conversation. So that's one thing that I think could be done simply by thinking bigger about who the customers are and who the agents of change are.

Chris: And taking the whole system approach, yeah.

Lorenzo: So the second thing I would say is that we have a concept of electricity which is excessively focused on the commodity. And you know, go back to the 90s we're creating wholesale markets and commodity, commodity, commodity was thrown around as the buzzword and we were making analogies to commodity markets, and you know the trading of our wholesale megawatt hours. You know so it was all very much a commodity basis. Now in the old paradigm where you have big central station generators and you have customers out at the end. Well yeah, it's basically a commodity that's flowing from one place out to the other. But once you start getting into high volumes of distributed resources, there are a couple of different things to think about. One of them is, the services are more important than the commodity. And I mean we've known this for a long time when all the energy efficiency discussions going back into the 70s and 80s recognized that, oh yes, customers don't really care about kilowatt hours, they care about comfort and other services they get. But I think that needs to take on new meaning now. It really needs to expand, because just take the idea of local resilience. If you're creating a local power system and through this layered structure that power system, if there's a disturbance on the grid, they can disconnect and they have power in that local area. Well, what's the commodity value of that. It's not measurable in terms of a commodity. It's really measured in terms of the value of being able to have continuous operation and to maintain, not just electricity kilowatt hours that flow to customers, but all of the integrated systems that depend on it which may be your water supply at your wastewater treatment and so on. So let's start thinking about services instead of commodities. And then a compliment to that when you move on to the system that right now at the ISO even and on the distribution system, we're looking at small time intervals every five minutes. How much does this generator produce in five minutes. How much does this local load take out point load node consume in five minutes. And we're looking at essentially the commodity that's flowing one way or the other in a small time interval and then we charge based on that all of our charges are essentially based on that. Even when you get down to the retail level, we're looking at okay how many kilowatt hours that you consume in a month or in an hour or in a certain time period. So, as we move into a more renewable and distributed future, and I'm saying both because we do have large scale renewables on the grid we have big central station solar farms and wind farms. Those will continue to exist but they bring a different characteristic to the grid, they bring a certain kind of volatility that makes them more difficult to manage operationally because what they're doing is dependent on how much wind is blowing how much sun is shining whether there's clouds etc. and then the same volatility is happening at the distribution end even more so because now you've got vehicles and sometimes they're charging and plugged in and sometimes they're not plugged in and sometimes they could be discharging when they're plugged in. And then you have micro-grids that are doing their own behavior and you have customers some have just plain rooftop solar others may have rooftop solar plus a battery. So what I'm getting at is, instead of just how many kilowatt hours are being consumed or produced in an interval. Let's start looking at the volatility impact of all of the things that are connected to the grid and look at volatility as a driver of the cost of providing reliable service at the transmission system. Part of providing reliable transmission grid is being able to manage volatility and that has a cost because we have to procure resources that are very responsive that can ramp over a couple of hours to meet severe changes if you lose a whole bunch of solar due to cloud coming over. And similarly on the distribution level there is a lot more volatility. The distribution company is going to have costs associated with managing the volatility in the form of say paying certain resources to be able to operate on command to maintain voltage frequency et cetera. So let's think about if a resource is attached somewhere to the system. Part of what it's going to pay for distribution service is a measure of its impact on volatility. If it adds volatility to the grid it pays more for service. Same thing on the transmission system. If a facility adds more volatility to the grid it should pay more service because it's that the cost causation model. It's driving a cost of managing volatility. Similarly, resources that can help manage volatility get paid for providing the service of managing volatility so that now the cost of transmission and distribution service are not just attached to the commodity, but are looking intemporally, not what did I do in this five minute interval but let's look at a sequence of 12 five minute intervals and look at the variance of what your consumption or production was. Or maybe the variance of your voltage over those series of intervals. That's a measure of volatility. And, now if you are very volatile and you're exploiting that volatility onto the system you're creating a cost that you're going to have to pay for. So that's a new different way of thinking about how we're making the rates associated with distribution service and transmission service.

Chris: Yeah, I can see very clearly how that would lead to really a totally different kind of regulatory paradigm and different kinds of rate design and all that. Yeah.

Lorenzo: And it's not that far fetched I think in terms of where we are, you know that's not overturning the entire system. That's just introducing a new element.

Chris: So actually on a very related note I wanted to ask you if you'd seen this new paper from Christopher Clack and his colleagues at NOAA and the University of Colorado, Future Cost Competitive Electricity Systems and Their Impact on US CO2 Emissions. So they found in this paper that carbon dioxide emissions from the U.S. electricity sector can be reduced by up to 80 percent relative 1990 levels without an increase in the levelized cost of electricity using only current technologies and without storage. If we build more HVDC transmission lines and switch to a national grid architecture, instead of remaining with the current regional system so. So here we are, where we're talking about a whole system view of the entire U.S. grid. Do you think that the national grid is even a practical concept here?

Lorenzo: Well, that's a good question. I haven't really looked into that a lot so I wouldn't really have an expert opinion on that. I guess when I hear though about really ambitious transmission projects I wonder if the advocates of that have thought about the potential for high penetration of distributed resources, because it seems to me we need to at least consider how cost effective would such a build out of transmission be if we moved to say 50 percent of the electricity consumed is being produced locally. You know, do we still need all those facilities. Right. And so I would just say that that as you think about the costs and benefits of the construction of such a system. Well, let's build in a scenario that's a very high penetration of distributed resources because my sense is that an additional benefit of the localization and the layering structure, will be reduced loads on the existing grid, reduced congestion on the existing grid, and as a result, the ability to have higher capacity factors on the existing grid because where the distributed resources are also having the effect with storage of smoothing out the load profiles. And you know a lot of infrastructure is being driven by peak demand. So if you start to smooth out the load profiles then you have less drivers of new infrastructure. So, you know I question really is that needed if we enable this more decentralized structure to move forward.

Chris: That's a great, great point. So, one final question and I'd like to maybe just get a little more philosophical, maybe a little less technical here. Why do you think it is that decentralized decision making just seems to work better in complex systems?

Lorenzo: Well, that is a good deep philosophical question. Let me shift your question a little bit to talk about human systems, people in community because I feel like a whole lot of the predicament that we're in as a species has to do with, in a sense, breakdowns of communities and relationships. We have an economic system that for 100 some years has been tremendously successful at producing lots of new stuff. But part of the cost of that is this sense of ourselves as individuals that have to have all our own stuff. And the breakdown of communities, structuring neighborhoods and ways that people hardly ever come out of their houses and get to know each other, and commuting long distances to work, getting in a car and driving or getting on public transport and going places. And I think, as I think about a future that I expect to be a lot more volatile in general whether it's climate disruption, extreme weather events, or just the instability of our economic system, which is also growing increasingly volatile and our social structures which are growing increasingly volatile, look at the quality of some of our national debates that are happening in the presidential campaigns. And a lot of it has nothing to do with the real problems that we're facing. Very, very little substance. And as I think about surviving for the future as I think about my children and my grandchildren and subsequent generations what I think we really desperately need as people is to rediscover what community means. To begin to come out of our houses, to begin to have gatherings and talk to our neighbors about what's really important not just about who won the Super Bowl or you know that kind of stuff but really talk about what is it that we really care about and how can we create resilient communities. In other words, how can we be less dependent on what happens in Washington or what happens in Sacramento. Because, in my neighborhood we're growing some of our own food we have permaculture plots on some people's yards we have chicken coops. We have community dinners. And we're talking about stuff, we're creating art projects, we're educating each other. We have litle cultural events we don't have to go to the arena with 10,000 or 20,000 people to see a concert. We could. Not bad, but we also have music happening right in our neighborhood. You know so to me decentralized is the essence of creating a human culture and human society that's going to move into a more sustainable 21st