 My colleague, Omar Siddiqui, and my name is Li-Amin from Stanford University. Welcome to attend the Digital Grade Summer Webinar Series. We're going to have the last panel for this summer, which will be focusing on transactive energy. Before we start the WebEx, a few housekeeping items. Everyone is automated when you join this webinar. There are the two ways you can ask a question. The first one is you will see a chat or a Q&A button down the screen. You can submit your questions through the chat or through the Q&A. Alternatively, you can raise your hand. We are recording this WebEx, so your participation is your consent. All the recordings and the presentations will be posted to both APRI and the Stanford website. A quick intro for the host perspective, if you are new to this webinar, and it hosted by APRI and the Stanford Based on What initiative. APRI is a non-for-profit organization and focus on the IND on generation, power delivery, and utilization for the public interest and public benefit. And the Stanford Based on What is an initiative at Stanford focused on the digital innovation for the 21st century electrical grid. Both of us come together with support from IT company VMware, have the same workshop goal, which is to convene utility, technology provider, academic leaders, government agencies, non-for-profit organizations to develop a research roadmap for standardized data platform to interface customer DR with a grid and inform an interdisciplinary and collaborative research initiative. So we cannot do this alone just by two organizations. In the last three months, we got a lot of support from more than 30 organizations that participate in this webinar service. More than 40 speakers that deliver the talk or participate in the conversation for this important topic. If you're interested in the last 11 web acts, you can find the recording from both Stanford and APRI website. Today, I'm very glad to have two distinguished panelists, Ron Melton and Ed Catlett to join us discuss the transacted energy. Also we are honored to have a special guest here, Terry Oliver, to help us to moderate this panel. Many of us, before COVID-19, many of us are frequent flyers of some airlines. And Terry is a frequent attendee for this webinar service. We haven't counted, but likely will be like 10 webinars in the past. And many of us know Terry, just a quick intro, Terry is the first chief technology innovation officer of the Bonneville Power Administration. He received the highest award from BPA, which is Meritorious Service Awards. He is retired, but not tied. And recently, he just awarded with the Great Forward Innovative Award. So today, he is going to help us to moderate the conversation between Ron and Ed. With that, I will hand this to Terry. Terry? Go ahead. Thanks. Transactive energy and Terry Oliver actually go back quite a ways, along with these two distinguished speakers. It got its start in a planning scheme for the Olympic Peninsula Grid-Wise Demonstration Project, a joint effort between Bonneville Power, the Pacific Northwest National Laboratory and a set of BPA customary utilities on the east side of the Olympic Peninsula in Washington State and graciously funded in large part by the Department of Energy. We began by imagining what if we could deliver prices or values directly to devices? And along the way, we discovered that that wasn't nearly enough and that much more was possible and beneficial. He now retired, every executive, Clark Gellings made a name for himself and every by imagining a specific set of utility benefits that could be halved by changing load shapes in a distinct way. Clipping peaks, filling valleys, moving load out of the peak into the valley, shifting loads, strategic load growth, and strategic conservation, something that Bonneville Power and the Pacific Northwest utilities took to heart and have made significant progress on, and flexible load shapes. The latter is kind of where Transactive Energy takes off. We wanted to go further, less in command and control and more into markets. At the same time, we were trying to avoid the downside of the approaches of load controls in use at that time by utilities and still primary use, angry customers and small impacts. The kind of thing that you got when, on the most extreme hot day, at the most extreme hot hour, you said, yeah, air conditioning off. In setting up the Olympic Peninsula grid-wise demonstration, we assembled a set of resources, home space heating, municipal water pumps, two elevated storage tanks, and backup generators. Each of these brought an interesting possibility or flexibility. Pumps could preheat and municipal water pumps could significantly vary when and how much they pumped into storage, and large enough backup generators could offset major loads and possibly actually feed into the local grid. We've set up a several-year experimental plan and a method for calculating local and regional values and a chain of resources aligned to deliver electricity to customers in the peninsula. Starting with the hydrogeneration values at the mid-Columbia generators and including transmission values and costs from the mid-seas to the peninsula, including that generation. When we had it all set up and in the dead of winter, we pulled the trigger on a big experiment. If we pretended that virtually half of the capacity to serve these loads simply disappeared, would the resources on the Olympic Peninsula keep the lights on while at the same time keeping consumers happy? And we did both. We thought that perhaps in a subsequent phase would be needed to deal with anticipated anticipatory load flexibility. It turned out that that worked right the first time. It worked because we also told the equipment not just the current value or cost of electricity, but that we gave them a future value. So that the equipment understood by the magic of programming that while electricity had a reasonable value at 6 a.m., it was going to be pretty expensive at 9. And we found that homes were choosing to preheat. No consumer engagement actually required to accomplish that. So one of the things that we knew but newly discovered was that well-insulated homes had a bunch of unrecognized value for as flexible loads. And you'll hear in Ron's presentation and in Ed's presentation more about how we advanced those ideas. So today we're going to hear about the work of the Pacific Northwest National Laboratory in systematically exploring how that could happen and the work of the California Energy Commission sponsored test of transactive energy actually applied in Southern California Edison via TMIX. So let me introduce our two panelists. The first speaker will be Ron Melton. Dr. Melton is the group leader of the Distributed Systems Group in the Electricity Infrastructure and Buildings Division, a senior power systems researcher and project manager at the Northwest National Laboratory. He's the principal investigator for the DOE Advanced Grid Research Project for an ADMS open source platform. He's a member of the core team for the Department of Energy Grid Architecture Project and administrator of the GridWise Architecture Council. He was the project director for the Pacific Northwest Smart Grid Demonstration that concluded in June 2015 and has 10 years of experience in cybersecurity for critical infrastructure systems and over 30 years of experience applying computer technology to a variety of engineering and scientific problems. Dr. Melton is a senior member of the Institute of Electrical and Electronics Engineering and a senior member of the Association for Computing Machinery. He chairs the IEEE Power and Energy Society Smart Buildings Loads and Customer Systems Technical Committee and the Smart Electric Power Alliance Grid Architecture Working Group. He got his Bachelor of Science in Electrical Engineering from the University of Washington and his Master of Science in PhD in Engineering Science from Caltech. Ed Casselette, also a PhD, is the founder and CEO of TMIX Inc., a transactive energy services company and a founder and VP of Megalot Storage Farms Inc., a grid storage advisory firm. He was previously a governor of the California Independent System Operator and also the founder and CEO of Automated Power Exchange, the first online wholesale power exchange. Dr. Casselette is a leader in the design of transaction services for electricity, the commercialization of electricity storage and the analysis of energy decisions. He has extensive experience in designing, building and operating high-speed, reliable transaction systems for electric power and then interfaced with the existing transaction systems and markets. Ed was also the chair of the Oasis Energy Market Information Exchange. He holds a PhD from Stanford, focused on economics, decision analysis and power systems and a Bachelor of Science in Massive Science, Degrees in Engineering from the University of Washington. Welcome to our speakers. Let's go on to Ron's presentation. Well, thanks, Terry. Good morning, everybody. I have to say, Terry mentioned sort of the genesis of the work here in the Northwest back with the Olympic Peninsula Grid-Wise Demonstration Project. I think we should also recognize that much of the work you're going to hear about today, certainly the work that we do at PNL, actually has its roots back in the late 70s, early 80s, and worked at Fred Shweppy and Richard Tabor and Jim Curtley were doing at MIT on something that they called homeostatic control. The techniques for engaging customers and the flexibility of loads of customers were greatly laid out by that work that they did. There's an MIT Technical Report, MIT-ELB1-033, or maybe that's 81-033, that is searchable. You can find that online if you're interested in going back into the history of the field. What I want to talk to you about today are sort of four things. I'm going to go fairly quickly because I have a lot of slides to cover, but we'll give you a little bit of background and context, some of the motivations for transactive energy. Terry alluded to this when he talked about the flexibility aspect of loads. I talked about some of the sort of foundations for laying out framework for discussing transactive energy systems to have a common way of referring to things, nomenclature, taxonomy if you will, and a roadmap for the development of transactive energy system technologies at scale that were coming from the GoodWise Architecture Council. Then a little bit of a great architecture perspective coming out of our good architecture work, and finally a quick overview of the various things we're doing at PNNL and transactive systems research. So motivations, I'm not going to dwell on this because I think you're all pretty well familiar with this, but basically the energy system is changing and changing dramatically with the introduction of renewable resources at large scales. And as we get more and more of those resources, distributed energy resources in particular, and we also get an increasing amount of distribution automation and other intelligent devices spread throughout distribution systems, the old techniques are just not going to work anymore. And so we need new techniques that are able to harness the flexibility that's brought by the distributed energy resources and be able to take advantage of edge computing and other advances in terms of the technology that's laying in front of us in the electric power system to take advantage of those changes. So we have a number of ideas here about how we might engage these different devices. First, taking into account owner's interests. We sometimes refer to this also as boundary difference. And this calls us to think about not necessarily how to control. As Terry mentioned, it's not about controlling the devices on the hottest day at the hottest hour. It's about coordination and incentivizing behaviors. As we do that, we need to have some sense of what will in fact happen. And so we need some sense of performance agreement so that we can have services requested and responses to those services offered and know something about what's actually going to happen. And one of the ways we do that is of course with markets. And the markets both give us a chance for some market based optimization and efficiency. And they also give us a chance to use those incentives that I mentioned. We need to provide access to this to all interested parties. And we need to make it simple. We need to have it to the greatest extent possible, some form of plug and play integration. Now, Terry mentioned the work on the Olympic Peninsula. And we used a double auction market. Dave Towson was a part of that. I see Dave is on the on the call today. And the market, as I recall, was designed in consultation with Lynn Keesling. And basically, the slide here gives you a general idea here. There's automated price responsive devices. They have some curve that tells how they will respond to a elasticity of price. The various systems in a home are aggregated together. They give you the overall price flexibility curve for that home. The service provider aggregates the curve from the customers that say what they will do at what price. They clear the market. They send out the clearing price. The devices behave according to what they said they would do at what price. Either they get to play or they don't. And you just repeat that loop over and over again in time on the Olympic Peninsula project that Terry mentioned. That was a five minute time interval. And this work and the foundations of that work, the things we've learned sort of led to many of the discussions that have taken place in the grid wise architecture council. And we spent a great deal of time debating over about a six month period definition for transactive energy that you see on this slide. A set of economic and control mechanisms. And so we recognize the use of the economic techniques, the markets, but also from a reliability and resilience point of view, the need to have some sense of treating this also as a control system to have the dynamic balance of supply and demand coordinated through the use of value as an operational parameter and otherwise economic signals. We don't use value of precisely in the economic terms here, but value in the sense of we want to figure out how to monetize things. How do we monetize, for example, comfort so that one can transact on comfort with the smart thermostat and somebody's house. We also generally agreed that these are going to be distributed systems and that there's various various things that they might do for people. This was all pulled together in the Transactive Energy Framework document, the second version of which version 1.1 was published in July 2009. It's available at gridwiseac.org. And this both includes definitions that also list a number of attributes of the TE system. For example, who are the transacting parties? What is the transaction mechanism? What is the extent of the system? Is it just in a distribution system? Is it from end to end, from transmission through to end use, et cetera? And also we identified in the second version six TE principles that will go over in just a minute. And we then discussed in more detail the context for deploying TE systems from a perspective of policy and market design, business models and value realization, provide some conceptual architecture guidelines and then some discussion of cyber physical infrastructure related to the implementation and deployment of TE systems. Let's see. There we go. My screen just went blank, but we're back. Okay, so the Six Transactive Energy Principles that we came up with, these were discussed at a meeting at PJM about about 2010 or 2011. And we had agreement across a number of participants at that meeting on these six ideas. First, that these systems require highly automated approaches and that's a form of coordinated self optimization. We'll talk often about greedy local optimization. They need to provide the nondiscriminatory participation by anybody that's qualified to participate, has the necessary devices and systems. And you saw that just a few minutes ago that there's some sense of accountability that has to be in place that is applied to the transacting parties so they can't say they'll do something and then not do it without any repercussions. But observability and auditability at all interfaces is required. One has to be able to sort of book the transactions, if you will. And so I'll ask you what's going on to do that. And also there's some sense of situational awareness that may be needed. There's a need to consider reliability and control while doing this optimal integration and coordination of the DER. And that, for practical purposes, these systems need to be scalable, adaptable, extensible, et cetera, across a variety of different DER and variety of different participants and geographic extents. So these form, if you will, a set of high level requirements for the systems. The work of the council further sort of builds on the GWAC stack through the stack of interoperability that comes from the council's interoperability context setting framework. And we can recognize those four sections that I mentioned from the TE framework also correspond to breaking the eight levels of the GWAC stack into four paired levels from the very basic technical levels below up to the organizational considerations of regulatory and policy and business objectives. Taking that particular view and combining it with the IT road map stages that we got from the work that Paul DeMartini and Lorenzo Christoff have done for the future utility electric utility regulators project at Lawrence Berkeley lab, where they defined three stages of adoption of distributed energy resources, a low adoption rate where prices are basically coming from administrative mandates of things like feed-in tariffs to a moderate level of adoption where you actually have enough DER in a system that you can aggregate it and potentially interact with an organized market and establish value that way up to very high adoption, where in fact local value can be established through interactions within a distribution system, for example. So using those principles and those three stages, we created in the TE road map a structure that you see like this, where we consider the four areas, we consider the three stages and within each of those four areas and across the three stages, we look at what are the benefits that would accrue from transactive energy systems, what are the results are actually working our way down? What's the vision of what could be done? What are the things that are required to enable that vision? What are the results and if one is able to do that and what are the benefits to all the interests and stakeholders when one does that? We don't have time to go into details on this today, but you can take a look at the road map yourself. It's intended to answer the question, what's required to actually be able to deploy and operate these types of systems at scale and utilities? So switching gears now. Mission that coordination is extremely important as opposed to control that one of the reasons for going to a transactive energy approach is to be able to do coordination. And we spend a little bit of time in the grid architecture work considering that as well. And you see here this notion of coordination sort of looking at the various new interfaces that come into play and also at some of the structural considerations as we move to things like distribution system operator concepts and have the ability to coordinate what's happening in a distribution system with the operation of the distribution system avoiding bypassing that you see, for example, in the figure in the upper left, where the microgrids are communicating directly with the TSO. As we look at that need for coordination and the roles and responsibilities, one of the things that we did in the grid architecture work, just one, there's a lot of other work coming out of that project. It's also relevant. But we took a look at in particular at the mathematics of distributed optimization. And the mathematics of distributed operate optimization basically breaks problems down into master problem, sub problem, sub problem and so forth. And it gives us a naturally layered structure that we chose to label the laminar coordination framework. And by considering that laminar coordination framework, we actually get a set of practical architectural guidelines about the types of information that needs to be exchanged between the layers and in a layered approach. And you can see that maps onto the grid, we can actually consider the physical topology of the grid relative to such structures, we can also address things like boundary difference and so forth. Going further in that mapping, you see these coordinator nodes are basically the interactions within the levels. And one of the reasons I wanted to show this this morning was with a high high number of devices, especially edge devices with computing capabilities and other smart devices spread throughout the system. In order to be able to have these distributed coordination techniques and distributed system, there's a high degree of standardization required at all at all layers of the system. And this is sort of a new requirement, I think that we need to step up to as an industry, don't have a lot of answers for that. But that is the kind of thing we're working on. And Terry mentioned, when introduced me, the open source ADMS platform grid app Steve that we're working on, which is in part trying to address some of those problems. These laminar coordination frameworks and the layered approach give us give us a number of features within that standardized construct to be an extensible supporting boundary difference, the greedy selfish optimization, and the ability to be scalable and secure. So that's our sort of quick drive by the grid architecture work. As I mentioned, there's there's a lot more work on that, including a new concept we come up with with something we call a logical energy network, which is a virtualized construct in the electric power system. And that that those reports and efforts are all available out on grid architecture dot PNL.gov. So now quick run through the PNL transactive systems research. So so Terry started with the Olympic Peninsula grid wise project. So I'm not going to dwell on that other than just to remind you that was a double auction market that was used to show that we could manage a constrained distribution feeder without without going above whatever boundary had been set. That technology was further deployed using the in the AEP grid smart project, which was America Electric Powers, American Reinvestment Recovery Act project in the PGM territory using PGM, Locational Marginal Prices is one of the bases for establishing the double auction market prices. Also, it was able to do that with constraints in there and introduced in their system. That also led us to the Pacific Northwest Smart Grid demonstration project where we took a different approach with a forward forecast of cost of power delivered in the forward forecast of load. And through an iteration consensus process between those two and arrival at an agreed to price at a future point in time for the power that would be delivered. This was a nodal approach. You see the little nodal diagram showing the interaction of any given node with its electrically connected neighbors to exchange the type of information I was just talking about. And post Pacific Northwest Smart Grid demonstration project in AEP grid smart, then we began a set of transactive campus projects, which continues today and I'll say a little bit more about that in a moment. This looks really behind the meter now, if you will, at a campus and looks at what can be done to coordinate energy consumption across the campus and within the and within the individual buildings in the campus using transactive techniques. And that gives them the ability to manage both the energy consumption within the campus, which has been great interest to the building technology office in the DOE energy efficiency and renewable energy. But it also gives us the opportunity to, in fact, and bring that ability to coordinate flexibility to the meter and work then with the utility on the other side of the meter to offer, offer response to needs expressed by the utility. We're also doing a project as a part of the DOE transactive systems program, we call it DSOT distribution system operator plus transmission. The idea here is to do a large scale simulation to assess the benefits at large scale within a within a transmission system of the distribution system operator offering offering flexibility and other services into that transmission system operator. That project is based on ERCOT and it's actually kind of a mashup of ERCOT and PGM, not using ERCOT markets, but using the size of ERCOT using the roughly the transmission network of ERCOT, and then the distribution systems of ERCOT to be able to be able to do this simulation, which as you can see, is going to include, you know, a huge number of huge number of buildings. And then within the buildings, you know, a huge number, total number of devices, including over 20,000 electric vehicle chargers, 27,000 rooftop PV installation installations and so forth. And this gives us this gives us the ability to really do some detailed analysis of the benefits of the transactive energy system at scale. So for more information about all of this, I've only really been able to give you the 30,000 foot view here, because there was a lot of different material I wanted to cover. But you can find the information about the GridWise Architecture Council work at GridWiseAC.org There's the TE framework, the TE roadmap, the decision makers checklist, a recent report, we just finished Smart Buildings as a Transactive Hub, another recent report reliability and resilience considerations for transactive energy systems. Also, I invite you all to join us for the seventh International Conference of Workshop on Transactive Energy Systems, which is now a combined GridWise Architecture Council and IEEE Power and Energy Society conference. So you can go to the IEEE-TESC.org website, that'll be a virtual conference the week of December 7th. You can also find out more about the transactive energy system simulation platform, which we're using to do the DSOT work. I just mentioned that there's a read the docs site and a GitHub site for that if you're interested. And last but not least, we're going to be having three transactive theory workshops in September on September 14, September 16 and September 17. Those are three, the meat of the discussion broken into three different times. So it's not the same thing done three times. It's three, you know, a continuing discussion across the three workshops. Those are open to anybody that's interested. We've already got the speakers selected, but their participation is open to anybody who's interested. If you want to participate, please contact Susan McGuire at PNNL and register with her. So I will call it good with that. Thanks for your attention, and I will turn it back over to Terry. Thanks, Ron. You caught me up. And now we get to hear from Ed Caslett on a real world experiment conducted with Southern California Edison. There you go, Ed. Thank you, Terry. Let me see if I can there. Okay. Well, thank you. I'm going to be talking about a pilot of transactive energy that we completed a while back. The report is on the CEC website at the end of this presentation. There's a list of references. They include the links to that report. And I'm also going to be talking about the current situation in California and what motivates this type of approach in California. And I'm delighted to be on this panel along with Ron and Terry who I interact with on transactive energy for probably a couple of decades right now. So everybody's aware of the rolling blackouts just last week in California. That's just only the latest of a series of crises in California that call into question how we design our markets. And it goes back to the design of the markets back in 1990 and 1995 led to the first energy crisis and various crises along the way. So we keep learning from these crises. And I think there's an increase in realization by all of the major entities in California ISO CPC and CEC and many other parties that the supply side approaches to retail electricity must give a way to demand side approaches. That is trying to fit in end customer modes and distributed assets into a centralized centralized market dispatch market run by ISO is too hard and really pretty inefficient. So we're seeing the concept of capacity or RA in our areas really failed evidenced by the rolling blackouts prevent shortages and is an expensive might be an expensive and overly complex way to do especially now that we have solar storage and EVs in the market. What's the capacity of a contribution of a solar device or even a storage device depending on this duration. And EVs are you never know when they're going to come on. So and then the other observation not just mine but others that distributed energy resource aggregators are we're thought to be the solution but they're they're struggling and struggling with low margins and disappearing they're pivoting to the software. Good operators have little trust in these aggregations incremental incrementality are we double paying for RA? Are we paying once through RA incentives another one through say something like demand response. And then there's the multiple use problem when you have a behind the meter asset and you want it to increase its output or decrease its output. Who gets your control whether increases or decreases that the customer the distribution operator or the ISO and then who gets the benefit of that just too complicated to think about. So after demand side approaches seems pretty attractive for this is again as focused on demand side resources such as behind the meter storage etc. And after 25 years of trying what do we have here in California. Lots of interval meters but we don't use them for much. To use just only recently time to use rates only just start your rollout now and I think you can say that because of their block structure and flexibility they're too little too late especially to deal with things like went on last week in the rolling blackouts. Real-time pricing is viewed as a solution but it's still years away if ever just because of the politics of it. Goal I think is a tariff and I think the key part of the design transact assistant starts with the tariffs. We need tariffs that charge or pay the customer the granular marginal costs or when we don't have a marginal cost it's really marginal willingness to pay. We may not have marginal costs because at the margin most of our assets are becoming fixed cost assets such as solar storage, wires, transformers etc. And it's very little going forward in California where we're going to 100% renewables or clean energy. There's very little marginal cost so it's really which customer is willing to pay more or less for electricity. We need retail tariffs that recover all the allowed fixed and variable costs and can just do a pure marginal cost structure. And we need to think not only about the real-time self-dispatch devices but the incentives that the tariffs provide for self-investment. So we have two visions that approach. One division is development of distribution system operators, interface and Cal ISO that operate like retail ISOs and that's more or less what one says exactly like what Ron has talked about. The other one is the retail automated transactive energy system approach that I'm about to describe. So in this context I define T transactive energy, TE very simply. So it's simply a way to create a decentralized retail electricity market. In that market all retail parties can buy and sell with their load-serving entities or the distribution operators or their peers peer-to-peer transactions if they're permitted. It is primarily a forward market hours, days, months or year ahead and then with sub- hourly transactions for fine tuning. A key element is we define two basic products energy and transport on the distribution grid. It's really in my opinion important to separate those two products because once you separate those you find out you can work with a distribution operator that's much like the structure of the current distribution operators and not have to create a DSO that simultaneously has to manage energy as well as the transmission of the distribution grid itself and big simplifications. Other products in this framework you only have the real energy but reactive energy or transact particular flavors of energy such as green energy, solar energy. In this framework derivative products such as capacity are not needed which addresses the resource advocacy problem I just described about. We get by without the derivative product such as capacity by lots of forward contracts and highly dynamic in real-time prices. Next slide. There's a very small arrow I got to find here. This is the overall retail system that we piloted in large part in our pilot we just completed. At the top you see the customers and prosumers and distributed generation. In the middle you see the transactive energy platforms and the volume sensors in the supply side which is the distribution operator, the mode serving entity, the California ISO and the forward wholesale markets on the right at the bottom. And so the idea is in general offers to buy a cell I call those tenders and the forward and spot flow up at various levels of granularity certainly locational but also hourly 15 minute and five minute through the transactive energy platform to the end devices where through a standard interface optimizing agents control individual devices. All the optimization in this framework occurs at the edges so it's the retail customer having a smart agent is helping him schedule his storage devices scheduled to be charging in his car or HVAC. Now the violence optimization going on within the California ISO etc. In the middle the transactive energy platforms are primarily ledgers. They're ledgers that record the transactions and provide settlement functions and and standard APIs to allow all the parties to interact. We standard APIs the protocols for that were developed through the smart grid interoperability panel and the energy interoperational and e-mix panel, e-mix panel energy market information exchange Terry described earlier. T-Mix is energy market information exchange with a T in front of this transactive energy market information which is a profile of the standards we developed back with the smart grid interoperability panel. So the transactive energy platforms can manage the transact record the transactions for the retail transport the distributed the energy and the wholesale energy. Typically the wholesale energy interface might be operated by the California ISO the retail energy operated by LSEs which include community choice aggregators in California and the distribution operator might run the distribution transport project. The distribution operator in this case as I mentioned doesn't have to think about dispatching energy. They simply need to price the services on their each distribution feeder at the locations that the customers receive power. There's a concept I'll describe in a minute of the subscription. We can't just do this with all spot pricing real-time pricing already 50 minutes because of the volatility that's required to fully represent the schedule. So the only forward contract is one concept for a forward subscription which is just a series of hourly quantities of electricity at a fixed monthly price that provides a basis for long-term transactions among the parties. So the each importance each of the agents is not just looking at the current spot price for prices but looking forward over whatever horizon of tenders is available. Typically in the spot market at least 24 and 36 hours here in California. So I'm going to the next slide because it's having trouble advancing it. I can help please. There you go. Okay so how does rates work? Well the LSEs frequently offer forward tenders five minute hourly 15 minute even longer depending on what's available via the TE platform to customers and the LSEs that are essentially acting as market makers continuously forced to buy and sell tenders go bid and ask if you like that get set up all the way to end customers and their devices and so these tender prices are designed to recover costs all costs and they're typically low when solar wind is surplus and the circuit is heavily loaded and high when the opposite is true solar wind is scarce or distribution circuit is lightly loaded. On the automated end agents in the serve the end customers they they're tuned to the customer preferences and they schedule the operation of the of the individual devices. You can do this almost completely automatically within the occasional customer input required so this is very important in terms of customer participation. Typically the devices include heat pumps, conventional AC, electric vehicles, pumps, water heaters, battery stores, large appliance, data centers, refrigerators, warehouses, behind the team next service interface we don't care what's there they just have to respond to the tenders and respond and come back with transactions. Devices typically operate or charge more when tenders prices are low, operate less or discharge when tender prices are high. They're simultaneously doing that in each hour but they're also shifting and shaping the operations devices across the hour so in California if we got surplus of solar in the middle of the day Tim he is going to attempt to charge all the stories and EVs and do as much air conditioning that period of point in time and then in the evening when prices may go way up it's going to tend to ride through there using the thermal inertia or the storage that Terry mentioned. What's in the for customers they save money better follows available wind and solar supply and available distribution customer which reduces investment in those features thereby saving further for customers. I'll mention subscriptions that reduce the bill volatility both through the benefit of the customers and to the benefit of the suppliers DO, DSO and generators who also for revenue stability and TTE platforms records the transactions of distributed ledgers and computes payments among the parties and energy and transport and the design of this platform is the scale infinitely you have many TTE platforms that you want use a customer typically would transact with a single platform but you have the option of connecting up to two platforms in case you want to buy or sell power outside of your local area. Next slide. So this is getting the pilot. The pilot was funded by the California Energy Commission epic program. There's a $3.2 million award as proof of concept pilot. It's done jointly between TMIX and universal devices with about 100 homes on an S&E circuit, the Moore Park circuit in Los Angeles. The TMIX did the transactive energy side of universal devices, did the devices and the customer interfaces. This is kind of a cartoon and the same did larger diagram and went through where you've got the California ISO interacting with Edison interacting with the TTE platform flowing tenders up the tenders go across into the house that received in the TTE service interface. There's a set of devices that are communicated through the universal devices ISY box, the black box that talks to each of the devices. Also talks to the meteor. The meteor in Southern California Edison produces 15 minute readings that we only get after a day or two delay through the green button system. So we augmented that with five minute readings directed by the ZigBee Hon interface. So we have five minute readings, so we had a five minute transactions that interface with the five minute markets in the California ISO. So there's a set of devices in here. A typical operation would be say, are you plugging your car and it knows the forward tender prices in the next five minutes through the next hour, maybe 15 intervals, and hours each hour through the rest of the day. And it figures out when the chief's time is to charge it, it starts, it reserves, transacts for all the power required to meet your charge schedule. And then if in the middle of the night, in the middle of the day, there's significant changes on the grid and prices go up or prices go down. It will recontract whether it's sell backward or might have purchased at one point and then purchase in another hour that might be cheaper to the benefit of this consumer and to the benefit of the grid because you had a price response. So let's go to the next slide. So this is how the subscriptions come about. There's many wonderful ways of doing it. What we did in this study in the top of the screen is we said, let's give the customers an amount of power each hour of the next year that's based on their historical, their technical consumption. So if your solar costs were in January, you had a particular 24-hour shape to your load. So you get that amount of power in every day in January at a fixed price, fixed monthly cost for January. Not a price, but a cost. Same for the other months. So that gives you a fixed schedule of costs and revenues to the utility. And then if in any interval you use more than that, you can purchase it or you automatically purchase it at the tender prices. If you use less, you get paid at the tender prices. And the two tender prices are close together. So and how do we get the tender prices? Well, what we did in the SCE pilot is this working closely with some of the rate makers, the marginal cost expert at SCE who said, okay, for the delivery portion, that's distribution portion on the circuit, that we want to recover so many dollars per year, say $400 a kilowatt year on the circuit. Megawatt year. And so let's recover more of that when circuits were heavily loaded and less when it's lightly loaded. And if it's very lightly loaded, they don't want the circuit to go below a given level so they want the price to go negative. So there's a curve that's created and we create that curve using their judgment and my judgment. And then we scale that curve based on the forecast and loads on that circuit to recover approximately the annual fixed cost of that circuit and then update that from year to year. We also add to that the losses on the circuit based on the marginal cost, marginal losses and the Cal ISO energy price. For the fixed cost that are not included in the NLMP from the ISO, they broke it into two parts. One was the base generation cost so the similar concept to the fixed cost recovery for delivery where the recovery increased with a function of total load on the grid and for the flex portion of that, so we looked at the 3R ramp and they said say 40% of their costs are fixed for flexing 60% per generation. That was the rate making assumption. We covered more of that 40% of fixed cost per flex when the 3R ramps are highest. So we put all that together plus the LMP from the ISO and that's what the tender price is. The buy price and sell price are slightly different. So this recovers all the costs and reflects essentially marginal cost on the grid. Next slide please. So now I'm going to go through some of the results from the project here. This is what the tender prices look like for first 31 days of March 2019. You can see big spikes in around 6 p.m., slight morning spike. Sometimes the prices went down near zero and these are retail prices. So including the distribution, the distribution operator cost and the S&E fixed costs. This is what scarcity pricing does for you. Very dynamic pricing. Why that's why we need the price, the subscription component because you're getting paid this price. If you use less when the less that you subscribe to and you get paid, you pay this price if you use more than what you subscribe to. Next slide please. So this is another chart I'll kind of indicate the scales on the side because it's hard to read. So this is just for this past week when we had the problems on the California grid. And you can see the top case is the LMP for the ISO grid. And you see the first couple of days of two spikes are very low. That's typical operation. And then the last few days the spikes from the LMPs are fairly low. Then in the middle we had some bigger spikes that went up to $1,000 then to $1,500. These are for the P-node, the location in the Southern California War Park circuit. Then went up one day all the way to about $1,500 megawatt hour. And that's way higher than what you typically see. Next component below that is the base generation cost of recovery the fixed generation cost which vary a lot with load. You see in the middle there much higher than at the ends. Next we have the the next we have the the flex generation cost and you see it was high at the end but in the middle of the summer what happens is you don't have big ramps because it's just high load every day. So it's just continually increasing so the flex becomes less important in the summer it's more of the base generation cost recovery than the LMPs. Below there you have the that's the total LSE price adding up to three curves above. You add in the distribution operator transport cost and that's going up to the limit because of the high cost of losses on the grid because you're paying $1,500 or $7,000 megawatt hour for the losses. Finally down the water you have the totals. And so they're very different every day try to fit a time of use price to that price. So you know it's really different every day it's chosen to benefit and every month of the year since things are very interesting. Next slide please. So this shows the operation the pool pump the red is when the pool pumps on across the top you have the five days one two three six days in January of 2019 down the side of 24 hours it's red when it's on. So every day the operations this pump is different operating optimizing for the pool if you save the money for the custody. Next slide. This is the operation behind the meter storage battery with an agent that optimized the operation of that. The blue case this is for two days. The blue is where indicates the discharge the orange indicates the charge and you can see the gray is the stated charge going up and down through the day. Below that are the price tender prices that produce that result. So you virtually always get substantial discharge and warning in the evening in this case and charging in between at low prices. On the left you see the battery specifications and the red first day and secondary revenues. Next slide. This shows the operation with a heat pump in a heating mode and the up the top where the temperatures and then we have the kilowatts of power used by the heat pump and below we have the tender prices that motivated this operation. The you can see that if you look at the kilowatts it avoided all the high price times pretty much. It's different each day because of the outside temperature being different and you can see the orange up the top shows the deviation from the customer's preferred point of 74 degrees in this and it shows that this deviation is controlled by essentially a slider preference that the customer sets from time to time for more comfort or more savings. Next slide. We add a quick time track. Can you wrap this up in about two minutes? Yes, I can. Thank you. So this this shows the operation of a EV, Tesla Model S and the middle of the top of the tender price in every five minutes. Now when you start charging it's you're plugging it at 630 at night or something. And by 730 next morning I want it 75 percent charge. It shows you the middle of the five five-minute charges finding the cheapest places adjusting it through the night and the state of charges at the bottom goes up. Next slide. This shows the same EV charging over a period of about 10 days. It was unplugged. It goes down through the load and so off and then you'll see it getting charged and you get charged when you look down the bottom at various where you see the prices and you literally see the charges. So after the operates that charging. Next slide. So in summary you know the race race project achieved its goal of a proof of concept and work to design. We have lots of implementation challenges and cumulative delays to overcome but we got through it and we got it done. Our goals in California 100 percent cleaning energy and electrification huge need for flexibility. Demand side price response will work as evidenced by this this project. CEC 2020 load management rulemaking is is working to require often dynamic real-time pricing tariffs by the utilities by 2022 or so. However, I believe it needs tenders and transactions be effective. You can't just do this sort of peer pricing approach at scale. Renewable storage of wires and mostly fixed costs best to recover such costs when you scarcity pricing concept. Resource adequate and bill stability need the forward subscription process. That's going to work a lot better in the arbitrary resource adequate rules we have to come up with. And we need end-to-end automation for customer convenience savings and simplified billing and operations. So thank you. There's two or three additional slides. One is a bunch of references. Others is kind of a roadmap for getting California summary benefits. So I'll stop there. Thank you. Thanks Ed. Boy, we've got a bunch of questions. We also have a panelist joining us. Paul Hines is an associate professor of engineering at the University of Vermont. And wears a second hat as the CEO of Packetized Energy a company that is also working on transactive style integration of end uses end use flexibility. I'll give him a little bit to figure out if he has a question and I will pick a question from the many that have been submitted so far. There's a question about I think last session or two sessions ago on this webinar series there was a series of presentations on the application of blockchain and other sort of open source approaches to racking up all of this information. And I wonder how the panel thinks about the application of blockchain to the kinds of things you've described. I can go first. The blockchain think of that as an application that sits on that is a essentially distributed database just like the T-MICS platform is a distributed database. You could substitute blockchain for that distributed database but you'd still have a T-E application labor layer. Sometimes they want to use smart contracts. I find blockchain and I tried it too slow, too expensive, too hard to work with in its current state. And so it offers no advantage over the highly secure distributed commercial databases I can use in transactive energy that are available from any number of Microsoft and other vendors. So for now you've got to distinguish T-E from blockchain. Blockchain essentially is a distributed database that could enable T-E. Yeah, I got a pile on with Ed here, you know, blockchain's got to distribute a secure ledger and again possibly as Ed mentioned the smart contracts may be a way to trigger an event but there's a lot more to successful T-E systems and technologies than a secure ledger and being able to trigger a contract to execute. So the real work of T-E is in the logic and the functionality to make the decisions about what to do and what not to do in the marketplace. Yeah, thanks. Paul, you got a question? I love it. First, they were I mean bigger is the green it's on that one. I think that we've looked at blockchain and just haven't found it to be a cost effective solution. I think with anything that we do in this space it really have to focus on figuring out how to make a cost effective relative to other solutions. So yeah, we haven't found it to be better than just using a good database. I do have a question though and one of the things that we're finding is really important for making these demand side flexibility programs cost effective is making the experience for the consumer really good so that that everyone wants to join. And one of the concerns I've had about some of the transactive energy things is just really hard for people to understand what's going on. So I guess, you know, maybe Ed or others, how do you make this easy for people to understand and join? Well, yeah, they don't need to understand it. They're on a current tariff. Maybe it's flat, maybe it's time to use. This is just from their point of view. A new tariff and the prices vary and they get they can go out and buy a certain smart thermostat if they like. They don't have to and that thermostat can be pre-programmed to communicate to pick up the the tenders and communicate with the platform the operation of that thermostat. They have one control over that. It could be in the thermostat or you can say to the smart speaker, increase my comfort level or decrease my comfort level. Increase my savings level decrease my savings. That's all. They might do that once or a couple times either. That's it for that. The rest of the stuff like operating a pool pump, you say I'm in the hours per day. You need, you might change that seasoning through a voice interface and your car you plug it in and say when you want it to what charge level and that interface is probably in the car interface today when it operates that. Very little change from the customer. Yeah and I like going all the way back to the Olympic Peninsula grid-wise demonstration project where we offered the customer a smart thermostat that basically the setting they got to choose was a slider, more comfort, more savings. They moved the slider for turbo direction they wanted to at any point in time and that picked a point on the price elasticity curve that was embedded in the smart smart thermostat functionality. You know the other Paul the other comment I have about your question you know I hear especially within the electric vehicle charging sort of discussions talk about time of use prices and sort of expecting customers then to you know either set a timer on their car or something like that to take advantage of a time of use price which I think is a lot more complicated than the type of thing Ed was just talking about which is hey I want the car to be fully charged by 6 a.m. I don't care how it gets charged as long as it gets charged by 6 a.m. That's the parameter I set and then you know automated automated systems take care of the rest and can take care of the you know the potential benefits between the car charging and the utilities needs and any any you know financial side of those benefits that accrue to the customer or accrued and the customer doesn't have to worry about it. It's just a statement objective and then step back and let the system do its job. Yeah good point so I've got another question here it's sort of a theory it's it's a sort of nested version of a question and it relates to micro grids and peer to peer energy trading how do you see that working in the Transactive Energy System? I like to go take it on but I'm sure Ron's got some views so I'd say in terms of peer to peer if you take the rate system where it's set up and if you have a the market makers the LLC or the and the DO continually making a market others having buying and selling prices and if that's true say in the real-time market then there's not not much advantage for peer to peer transactions because you can sell it to your labor for about the same price you get it yourself where the advantage of peer to peer you come in terms of long-term investments I want to put up a solar system my house the neighbor doesn't have roof space we share the costs and we write a subscription the contract says shares the output of those between the two on the peer to peer system and so one of the advantages of the rate structures we built into rates is that which you could still do those peer to peer transactions and the utility will still be recovering its costs under the Transactive subscription Transactive terror so they wouldn't have any objection to that process as they do now under current rates Michael gets it in the same thing maybe the I'll add one thought to that that's kind of related is we you know we have a distributed algorithm that we do in our system but we haven't really done it using prices as the primary signal we're using a variety of different signals and we think of our Transactive energy system as a contracts to devices approach rather than prices to devices that we're not always pushing prices and we like that approach because you know we don't have as much complicated peer to peer stuff and and it allows us to do that you know give a similar test for experience where they're choosing like high medium low flexibility but they have a kind of a known amount of value that they get out of that so that they can choose high flexibility that's you know the 10 bucks a month of savings and low flexibility is two dollars a month of savings and they just can really you know easily make those trade-offs so but you know I know there's lots of different ways to solve this problem but just if we're going to do peer to peer we have to really figure out how to make it super simple for the customer yeah I'll be right up front I'm not a big fan of peer to peer I think I don't actually think the economics will pencil out for individual consumers in the long run and it may maybe something that sort of is early adopter interesting sort of stuff but that's for most people it would probably never really pencil out well and and quite frankly it also may obfuscate the real needs in our energy economies and that is to be able to support you know deep decarbonization if somebody can make make an argument why you know peer to peer truly supports reliable and secure system that's deeply decarbonized then I'm all for it but I really I'm skeptical that such argument can be made I've got another set of questions about sort of differentiating electric vehicle integration versus battery storage versus sort of thermal mass benefits of productive energy well so the transactive energy systems are not an end into themselves I've been waiting for a chance to say that I think this is the point so the question with storage is you know is the coordination that one can achieve through a transactive energy system approach to coordinate the behavior of the storage system relative to other operational needs is that can be stated and monetized to provide the basis for price signals to incentivize the behavior of the storage system is that the most optimum and efficient way to get those behaviors that are that are beneficial to the owner of the storage system to the operator the grid and society overall to come into play and I tend to think that TE is a great way to do that but but one should keep in mind that you do really have to start with what's the what are the objectives of having the storage system in the first place especially by whoever has made the investment in the storage system and then due to TE systems offer a an approach to achieve those objectives and I generally agree the key things say take take storage or behind the meter storage okay there you got lots of different technologies everyone's a different state of charge all the customers have different reasons for having a storage device and try and aggregate those and bid them in as a virtual storage device perhaps plurian demand response and you know HVAC shifting into kind of the virtual storage plant and you're bidding it and tell ISO for example is really really really hard to do we lose a lot of granularity in the information that the customers are concerned about and on the other hand if you take the demand side price responsive approach you can have a very very fine-grained operation taking account of each individual storage devices situation of the customer preferences and still get I think much higher responsiveness to meet the grid needs from the catalyze so Vegas agreement and I really like what Ron said about we need to think about what the ends are I mean carried quote I don't know if you should it's one today but like we need wiggly load because we've got wiggly generation coming is key and the you know the whether it's transactive energy prices devices quantities whatever it is whatever the structure of the goal is to figure out how do we get enormous amounts of flexibility into the system so that we can solve the grid problems and make it affordable and get to lots of renewable energy so you know I I don't think that we should try lots of different ways to get there and figure out what works and do those and unfortunately I have to run from there so thank you for the invitation to join the quick good to see you Paul thanks everybody so when we began exploring what transactive energy could do and we didn't really have an idea of how wiggly generation was going to be and it turns out to be really important but we also thought that this is something that was so obvious that utilities would pretty well adopt it within you know 10 years or so and it's been a long time so the question came up which which utilities in the US or globally would be most interested in this and and I got an answer from Portland General who's launched set of three smart grid pilot projects Larry Becadal there says yeah it is the future we are going there but we're not going to call it that anyway to the remaining panelists well I I can I can report that I don't know that I don't know how public the information is I'm not going to name them but there is a large east coast utility a very large east coast utility that is laying laying the groundwork for implementing transactive energy solutions in their in their system yeah I think here as every utility you look at there's different silos of interest you know and and it it takes time in the large utility so I think many of the smaller utilities are going to find transactive energy much more interesting particularly you know if we can make it simple and cheap to implement because they can more quickly get around internal politics I think the at least in California I think the move to transit active energy will be driven by utility and just but by regulated interest both in telephony energy commission and the CPC because they're struggling to keep the system operational as pretty evidence by the problems in the beginning of August here so Terry you know I mentioned I mentioned in my presentation the report by Paul DeMartini on Lorenzo Kristoff on the future electric utility regulators project that you find the three stages of deployment of DER and and I really think until you get to high high levels of deployment DER that are in fact creating operationally a distribution system that's very difficult to operate without new approaches I don't think you have necessarily a huge motivation by many utilities to adopt a TE type approach because they can get by with traditional central optimization and voltage management and things like that that so they don't have a compelling economic reason to change their operational strategy but once you get to high levels of deployment of DER that that are causing some sort of operational pain and suffering that can't be alleviated with just more of the same I think then there'll be a lot more motivation on the utility smart to adopt a highly highly distributed approaches that then include use of economic signals so that's a point where I debated Paul DeMartini on that a few times I have a completely opposite view of that curve that the right way to start is to fix the tariffs if you get the tariffs and get at least some progress towards a highly dynamic terror hopefully based on transactive energy then you'll start making the right investments both on the customer side as they'll see the value of stories value of solar etc EVs and on the on the supply side because they won't have to expand circuits because people are charging their cars at the wrong time or they're not putting not pairing their solar with storage so unless we get both sides making good decisions on early on both the investment in the operational side and we only do that if we get the prices right and that's the terror and the tariffs they're politically hard to change but they're cheap to change because you don't have to put steel in the ground to fix the terror well I guess Ed I agree with at least some of what you said but the challenge I think is sort of that changing the tariff part in which I think you really need to be able to start with some operational objectives and then be able to create the tariffs to support achieving the operational objectives and so you get kind of in a you know chicken chicken or the egg which came first sort of a situation so I tend to think that operational requirements are going to drive it and I think you tend to come at it from the other side but the two are two have to be taken together at the end and with that thanks to this excellent set of presentations what a fascinating topic for me anyway and apparently for 40 some odd participants so thank you very much and I'll turn it back to Omar and Lynn great first of all thank you Terry and to Ron Ed and Paul that was a fantastic panel I think we had I think we broke a record for most questions asked in fact I know we did so Terry that's obviously a testament to your your great moderating skills and this was a great session a great way to round out this webinar series so again thank you to our panelists here today just want to take a step back and just through this digital webinar series starting with our workshops in June through this summer and speaking on on behalf of you know Liang and all the great folks that stand for bits and watch initiative and us at every you know uh you know with us as as the host and I want to give a special shout out to VMware for their contributions for the initial webinars and and throughout that they have been an active participant so I wanted to make sure that I appreciate we've spent our appreciation to to VMware but through these 12 webinars over 42 speakers 800 plus attendees if we count you know the attendance of each one which you know we had a lot of common people that have been attending regularly but I think you're just a testament to the interest level in this in this area and again I just want to thank thank everyone all of the attendees all of our distinguished speakers and moderators throughout these 12 sessions I don't want to give a special special shout out to Orelia Bear from Epri and Wahila Wilkie from Stanford they are the ones that have been making sure that everything works well behind the scenes orally for making sure that these webinars actually function and we could not do this without you and Wahila for doing so much with the organization and reach out to the speakers and everything else so if everyone could just give a virtual hand to Orelia and Wahila they did a great job and it's been it's been a wonderful wonderful experience absolutely I just want to just say okay we've laid this great foundation we've we've covered a lot of ground where do we go from here and I think we definitely intend to continue this this momentum and in a few ways first we're going to maintain this repository presentations and webinar recordings and they are accessible from both the EPRI site and with links with the Stanford site also having links to the to those sites as well we will migrate that to a more permanent home and we'll make sure that everyone has that that permanent link but they will remain accessible because it's a it's a real library that that that we've developed together we also intend to launch interest groups to sustain this engagement that we've developed through these webinar series and what we're envisioning is a an interest group for the utilities to to join and participate in and to provide some backing in so that we can keep this as self-sustaining activity and for a stakeholder group because we have really through these webinars we have a whole ecosystem of experts that have been brought together including academic leaders researchers industry technology leaders from the large companies to startups and everything in between to government agencies to thought leaders in various related disciplines so we want to have a home for that group to continue to meet and engage close amongst themselves and also with with with the utilities so we're going to be working you know every and Stanford together to to build on the foundation that we've set here to enable this sharing of ideas and best practices to continue in a structured way so that we're advancing a research roadmap around digital grid development specifically enabling the enabling data platforms that are needed to allow the types of you know transactive energy architectures that have been talked about here today part of those interest groups are intended to incubate projects pilots demonstrations that say are born out of the recommendations from these groups so that we can again push the ball further down the field towards that end and ultimately set to transform the industry this is a big lift and it's bigger than anyone company and anyone perspective so this these webinars have really been a testament to that fact so again I want to express my appreciation we will be sending out a survey that'll be forthcoming we're going to be sending it to everyone who has participated in any one of these these webinars to get your feedback on them and recommendations on these next steps including what we laid out here but in a little bit more detail so there will be further information on that with that if we hang if you want to put your video back on I want to personally thank you and and bits and wads for for all that you've done and again thanks to all the all the panelists and Leanne would you like to say any final words before we before we close just very quick thank you wrong add and carry again for contributing to this webinar and thank you all of you who attended the webinar series in the past really appreciate your support as almost that well it's a time for us without the conversation it's a time for us to loading up our sleep and the real do some work and we will do it in an open and the collaborative way because both April and Stanford University we have the mission which is do something good for the public interest thank you again for your support thank you everybody thank you thanks thank you