 Hi, everyone. Welcome to the SmartGrid seminar. This seminar is sponsored by the Stanford Bids and Wars Initiative. Today our speaker is Professor Ross Bonic from the University of Texas, Austin, and he's going to talk about the Texas Black Outs two months ago. It's an interesting presentation. I would like to everyone that our next seminar is next week, next Thursday at 2.30. The speaker will talk about ultra-high voltage development in China and the concept of global energy interconnection. And also please note that the start time for the last two presentations will be at 3 p.m. Instead of 2.30 p.m., the speakers are, as you can tell, they are in Australia, so you'll be 8 a.m. over there. Because today, Dr. Ross Bonic is Professor Emeritus in the Department of Electrical and Computer Engineering at the University of Texas, Austin. He received his bachelor's degrees from the University of Sydney, Australia, and his master and PhD is from UC Berkeley. His research includes optimization of economic theory, statistical analysis applied to power systems, and also the electrification of the transport industry. He has a book named Apply Optimization, which emphasizes optimization of applications in electric power systems. Dr. Bonic is a fellow of IEEE and the recipient of the 2015 IEEE PES Outstanding Power Engineering Educator Award. Okay, so without further delay, let's join me in welcoming Professor Bonic. So Ross, you can share your slides. Thank you so much, Chinwoon. It's a delight to be presenting. Chinwoon and I were graduate students at Berkeley together, and so I'm absolutely delighted to be presenting. So today I want to talk about our Texas Extreme Weather. I live at least some of the year in Texas. We had Extreme Weather in February, 2021. It had effects on the electricity system. For those of you who experienced blackouts in California last August, the blackouts in Texas were much more extensive. The rolling blackouts were much more extensive, both geographically and over time. So I want to give some flavor of the severity of that in this talk and also understand what happened. And at the end, I want to give some, perhaps warnings, you might say, of broader implications of what went wrong. So first I'll start with the weather context and that'll lead us to the supply side effects of the weather, but there are demand side effects of the weather as well. So we'll look at that. And then I'll present a little bit of a verbal description of the timeline up to and during the rolling blackouts. And then we'll cut to the chase of the reason for those outages in the blackouts, the implications for wholesale market prices in Texas or more precisely the Electric Reliability Council of Texas region or ERCOT region of Texas. And then as I said, I want to talk a little bit about wider relevance, even if the market design is different in California or somewhere else, what are perhaps some of the wider relevances. So I think probably most of the audience, even if they don't know about Texas or know much about Texas, even if they haven't been in Texas, surely know that Texas is best known for relentless high summer temperatures. And there are many days every year above 100 degrees Fahrenheit. And that's a normal year. Sometimes we have extremely hot summers where there are many tens of consecutive days above 100 Fahrenheit. And a real hot summer has maybe 50 days total above 100 degrees Fahrenheit. Highest temperature ever was 120 degrees Fahrenheit. Furthermore, Texas has had significant population growth about 1.7% per annum in the last decade. And I'll come back to that when I talk about the demand side, but I've just put that in there just to understand what's been happening recently in Texas. Correspondingly to that demand growth, our summer peak demand has grown on average by about 1% per annum in the last decade. And so the dates are not quite overlapping, but roughly speaking, we've had a demand, peak demand growth of 1% with a population growth of 1.7. So the peak demand is now about 75 gigawatts. Comparing that to California, California has a peak demand of about 45 or so. California has a significantly higher population. So Texans are energy hogs, particularly in the summer, because we have a lot of air conditioning. And indeed, most planning for extremes, most of what we do when we think about how are we gonna meet electricity demand within ERCOT, within the ERCOT region, is focused on summer. And indeed until this year, until 2021, ERCOT was summer peaking, meaning there is a peak in the winter and a peak in the summer, but the peak in the summer is the biggest, the bigger of the peak. And indeed capacity has been tight in some summers, just as last year in California capacity was tight for different reasons. One thing to observe is that because we face hot summers every year, people know that they need to get ready for hot summers every year, and the generation fleet, and indeed the whole infrastructure, the whole supply chain, is actually well weatherized for these conditions. And we might say that Texas generation or the Texas electricity system is well summarized. So I don't mean summarized in that there's a summary. I mean, it is well prepared for summer weather. And we said that occasionally Texas gets very cold. And I've got a list there of years where there was a low temperature record in one or more Texas cities. Lowest temperature ever in Texas was minus 23 Fahrenheit. So that range from minus 23 to plus 130 is a pretty big, pretty big range. That was in 1933. The most recent storm, the one in February, 2021, which is the topic of today, was reported by several media to be the third worst ever behind 1989, behind 1899. Now, I'm not sure that, I'm not sure that that was anything more than a media claim. So perhaps an anecdotal statement, but having lived through it and having lived through the last cold in 2011, I can say that this one was way colder and way longer, okay? Having said that, extreme cold for some length of time, whether that's a few hours or a day or multiple days has been a once a decade phenomenon since 1899. So we get it once a decade. The last time, which was 2011, coincidentally, not coincidentally, 10 years ago, there was a report into the extreme cold, the blackouts that ensued. The report was by the Federal Energy Regulatory Commission or FERC, and it was actually a joint study between FERC and the North American Electric Reliability Corporation. And they had a bunch of recommendations for weatherization. So various political reasons. And if you were to caricature Texas versus California, I would say Texas tends to be less a fair. California tends to take regulatory solutions. That's a caricature of both places, I should say. But both caricatures that have some truth and the recommendations from 2011 were not made mandatory with 2020 hindsight or 2021 hindsight, perhaps, that might have been a bad idea. Having said that, there's some evidence that some generation asset owners did actually follow the recommendations. Well, some of them did, but as an overall observation, the gas infrastructure, the natural gas infrastructure, about 50% of our capacity is natural gas. The electric, and so it's supplied by the natural gas system. Some of the natural gas system, some of the electric generation system and the water system, all of them had failures. And I think it's fair to say that each of these infrastructures is less weatherized for once a decade events. In other words, I think it's fair to say that we are less winterized than summarized. Turning now to the demand side, and there's a Q&A about weatherization of or insulation of houses, and I'll come to that in a moment. Let me observe that electric demand is affected by low temperature. I'm sure everyone on the line, everyone watching understands that, primarily through heaters, such as residential heating, particularly. If you have a gas-fired furnace, then there'll be an electric blower to blow it around the house, and that might be only a kilowatt consumption. You might have a heat pump if you had installed an efficient form of electric heating, which would be up to several kilowatts. And you might, if you were not wanting to invest much in your heating infrastructure, just default to resistive heating, which might be up to over 10 kilowatts. For each of these, of course, what we'll find is as it gets colder, the duty cycle will increase, and the average energy consumption will increase. But here's a very important point that's relevant to our extreme peak. Even if you have heat pumps, when the temperature drops sufficiently, the heat pumps become inefficient, and in fact, automatically switch to resistive heating. So even if we use a heat pump that has, in many cases, is typically very efficient at moving heat into a house, for example, pumping heat into a house, the amount of energy consumed is much smaller than the amount of heat that's transferred. When it gets very cold and it got very cold in Texas, the heat pump becomes resistive heating. There's been a bunch of new housing stock added in Texas in recent years. Now, let me mention the Q&A that came up. There was a question that said, one of the typical insulation levels in homes, businesses, insulation can help with both summer and winter peaks. Yes, absolutely true. As a general observation, the building regulations in Texas are not as strong as in California. Having said that, as a general observation, newer houses, you could expect that they will just be installed with better insulation. A very old house, such as my house in central Austin that was built in 1907, didn't have much insulation in it when it was built. I've personally added some, but new houses at a matter of course, tend to have that. So we'd expect the thermal performance to be better than in the old houses. And since so much new housing stock has been added recently, we might observe that we should expect that thermal performance to be better on average in those new houses. But anecdotally, I don't have good statistics on this, but anecdotally, it appears that much of the heating is electric, not clear on whether it was heat pump or resistive. And it seems that at least some of the new construction was not well insulated, despite one's expectations. Well, the combination of widespread extreme cold and electric heating, including heat pumps switching to resistive heaters, resulted in a record electric demand in February, 2021. So I think it's fair to say that the housing stock in Texas is probably not extremely well insulated. And given that we have extremes of temperature, both hot and cold, we could probably observe that it could stand to be better insulated. So what happened? Well, we had a very big peak in February. And what was astonishing about that was that prior to 2021, I observe we've always been a summer peaking region. The summer peak has always been higher than the winter peak. So since 2015, our summer peak has been at least 70 gigawatts and it's currently 75. Winter peaks have been smaller than that and more variable. My bar chart here shows the winter peaks and the top of that chart is at 70. And you can see that recent years have been below and in most cases, well below 70 gigawatts. The only exception is that blue bar for this year. Top of that blue bar is at 69 gigawatts, which is in fact the peak demand that was served in or caught in February. Very soon after that peak demand was served, we had several generator outages and curtailment. And I will describe that in more detail later. But what then happened was that not all the demand could be served. There were estimates of what the peak demand would have been had it been able to be served. And those peak estimates are about 75 gigawatts. In other words, about as high as our summer peak. That red bar shows the top of that red bar shows where the peak would have been had there not been curtailment roughly speaking. It's of course an estimate because it was not served. It's an estimated load that would have occurred. And as you can see, that's winter peak demand that would have occurred had there not been curtailment was literally off the charts. It was literally far out of expectations. Well, was it completely out of expectations? We knew that there was gonna be extreme cold and the Urquhart system operator. So in California, you have Kaiso, the California independent system operator. In Urquhart, we have the Urquhart's independent system operator. They anticipated at least by February 8th, and there might have been earlier forecasts that there was going to be extreme cold from the 11th to the 16th. And then by the 11th, Urquhart was gonna anticipated a record winter peak. Few days later, February 14th, which was the Sunday going into that Sunday, there had already been more than 20 gigawatts of forced outages, meaning generator failures. I haven't gotten to the bottom of all of the causes. There are some lists of that and there are some several studies. I'm tangentially part of one that is trying to collate that information and put it all together. But exactly whatever the cause was, some of it might have been cold. There was already about 20 gigawatts of forced outages out of 107 nameplate. Now that nameplate capacity includes wind and solar. We wouldn't be expecting the solar to be at night and we weren't expecting much wind at this time. So that's not literally the total dispatchable capacity. The total dispatchable capacity is around about 80 gigawatts. February 14th morning, there was an appeal to the public to conserve. So things were already looking bad and a concern about the natural gas supply. And as I said, gas fire generation is about 50% of the total supply. So here's a graph over the days. It starts on the day of February the 14th, which was that very cold day. You can see that we hit that peak demand of 69 gigawatts. The demand fell a bit after that, but then very early on on Monday the 15th, there had been further outages. And I'll talk about that in a few more next slides. We found that generation outages then necessitated curtailing of demand because there was insufficient generation capacity to meet the desired demand. So we hit 69 gigawatts. That's three gigawatts more than the previous record in 2018. And the dashed curve here shows an estimate of what the uncurtail peak demand would have been. And the peak of that over the days would have been about 75. In fact, due to the supply failures, there were rolling outages. And again, I'll supply them, I'll describe them in a few more moments that forced the served demand down closer to about 50 with an approximately 20 gigawatts of curtailment. So what caused that? Well, just to reiterate, going into the early morning of February 15th, there were further forced outages compared to what had already occurred. In other words, generators were broken. There was less wind production than expected. Although that turns out to have been a fairly minor part of the whole game, the biggest effect was reduction in thermal capacity due to failures of thermal capacity and reduction in gas supply to natural gas generation. Morning by about 1.30 in the morning, there was additional generation outages and the rolling blackouts began. And at about 1.30 that morning, I woke up because the power went out at my house. As I said, the uncritailed demand would have been 75. That would be an annual average growth rate of 4% from 2018, far outstripping the population growth. So in fact, that is really representing not population growth, but truly an unexpectedly high demand because of large amount of heating load, okay? We had about 35 gigawatts of forced generator outages and D rates. So by a D rate, I mean a generator that had a capacity of 100 megawatts, let's say, could only produce 60 megawatts for some particular reason due to a failure. Then there were some generator failures outright, totaling up to about 35 gigawatts and up to about 20 gigawatts of load shed. There were outages in all technologies, nuclear gas, coal and wind. And as I already mentioned, the gas supply to gas fire generation is also reduced. Now in California's problems last August, California often relies on imports from other regions. And in that extreme time, that it was hot throughout the Western United States and the ability to import power into California was limited because everybody else needed the power. And a somewhat analogous situation happened to Urquhart. Urquhart is famously asynchronous with the rest of the U.S. that has about a gigawatt or so of import capacity on high voltage DC and other asynchronous ties, both to the Eastern Interconnection and to Mexico. As it happens, it might not have been too cold in California, but most of the U.S. experience cold weather and high electric demand. And in fact, the surrounding regions around Urquhart also had rolling blackouts, meaning that the imports, they were not planning to, they weren't going to export any power to Urquhart because they were already curtailing their own demand. So in fact, during much of the blackout, the imports to Urquhart were reduced. So even if we'd had 10 gigawatts of import capacity, we would not necessarily have been able to import much power. They were rolling blackouts, but here's the rub. Rolling blackouts deliberately not imposed on two main classes of consumers. The first is that there are distribution feeders in Urquhart. So distribution is the wires typically radial that do the last few miles of delivery to consumers. Distribution feeders have so-called under-frequency load shed relays. What are they, what is the idea there? The idea is if there were multiple uncontrolled generator outages, the electrical frequency would drop as supply drop compared to demand, frequency would drop and these under-frequency relays would trip. So 25% of Urquhart feeder load is on under-frequency relays. That's our last ditch defense against a cascade that would cause a complete system blackout. So we're not gonna impose rolling blackouts on those loads because we need them just in case. Additionally, there are a number of feeders in California and everywhere that have critical load. A good example of that is a hospital. So if you happen to be fed from a distribution feeder that also fed a hospital or some other critical load, you are not gonna be switched off. It turns out the capabilities or flexibility to switch off the consumers. Although there is some additional flexibility in principle, for the most part, it was controlled during this blackout by all or nothing on particular feeders. So if you had a feeder, if you were on a feeder with a hospital, you didn't get blacked out. Unfortunately, the required load shed from Urquhart was so large that many distribution feeders were outaged almost continuously for several days. And again, I haven't gotten to the bottom of this but I think in cities such as Austin, Houston and Dallas that have a lot of hospitals, what basically happened was that when you added up the feeder load on under-frequency relays and the feeder load supplying hospitals, loads that were fed on feeders that also fed hospitals, we found that the feeders that were left over represented much less than 50% of the available total load. So many distribution feeders were outaged almost continuously for several days. My own home was blacked out for all but 90 minutes over 59 hours. I went off, as I said, at about 1.30 a.m. on Monday. It came on briefly at two. I jumped up and switched my gas-fired heater on. It's, the blower is electrical, so I needed a little bit of electricity. We cranked that as hard as we could because we didn't know how long the electricity was gonna be on. It was only on for a couple of hours. And so then lost power again, all of two, rest of Tuesday, we didn't have power. Power was finally restored Wednesday afternoon and my house at its lowest temperature went to 43 degrees inside, went to 43 degrees. It went as low as 10 degrees outside by the way. So what caused this? Well, I think it's pretty obvious that's extreme cold. And we should observe that this is a common mode cause of both the high demand and the failures in the electricity, natural gas and water infrastructure. So we have a double whammy here, right? We have a common mode. It's not just causing multiple generator failures. It's not just causing increased demand. It's doing both at the same time. And so that posed a real problem for us. And I think to the extent that heat pump load played a role in this, it probably had a very non-linear effect. In other words, as the temperature dropped, each additional degree of temperature drop meant that more and more heat pumps turned into resistive heaters, jacking up the demand rather more significantly than you might have expected from the heat pump load. So it was very inconvenient. I hope I've conveyed that it was inconvenient to me. And there was, but far worse, there was deaths of both directly and indirectly due to the cold. And I would say electricity outages indirectly cause some deaths due to the cold. There were also apparently various interacting issues. So it turns out that to move gas along a gas pipeline, we need to pump the gas. We need to pressurize it. And some of those pumps are natural gas fired. In other words, they use some of the natural gas that is being pumped, but some of them are electrically powered. And it turns out that that infrastructure should have been nominated as critical infrastructure to the transmission distribution companies, but was not. And so some of these compressors, electrically powered compressors were actually interrupted, which further reduced the gas supply. I think it's easy to overstate that. I don't think it happened on a lot of feeders, but it's certainly contributed to the reduction in natural gas supply. And let me say this, at least from a public perspective, the relative significance of various specific issues is really not fully clear as of today. At least not clear to me and not clear in the public domain. And indeed, there are several things we don't know. We don't know how many generation and other asset owners paid attention to the 2011 recommendations. Some did, some didn't. We need to do some forensics on that. We also don't know whether the things that broke in 2011 was the same equipment that failed this time. So a large thermal generating system, generating station has a large number of piece of equipment and sensors in the 2011 blackout. In many cases, it was sensors that failed, which you would think would be relatively inexpensive to weatherize. It's not clear whether the things that failed in 2011 also failed in 2021, or whether there were different things that broke. And furthermore, we don't really know whether the failures of generators or the failures of the natural gas infrastructure was more significant. Although I have to say it's beginning to appear that the natural gas infrastructure might have been the most important failure. I'd state that tentatively because I think we're still waiting for a good deal of evidence about that to become public. There are probably various private entities that have a good idea about this. And I have my own private opinions, I should say. But in terms of public information, I think we're still waiting on some of that. So where does that put us? Well, I think it remains true that expectations about annual high temperatures in Urquhart, high demand is sufficient to motivate summarization. Generation asset owners in the electricity system generally has a pretty good handle on how to be able to meet high air conditioning demand, which drives our Urquhart's peak in the summer. But in contrast, extremely high levels of winter peak has not been really well thought out. Either in planning studies, if you look historically at what Urquhart's paid attention to, it's paid attention much more to summer than to winter. That kind of makes sense because the summer peak so much higher historically. And similarly, the market participants have not paid so much attention. Even despite the fact that we have occasional low temperatures, they just haven't been sufficient to motivate winterization. And something that should be born in mind here is that summarization and winterization are somewhat antithetical. If you wanna prepare for summer, you typically want your sensors, your assets to be able to get rid of heat, to not get too hot in the hot summer weather. If you put a bunch of insulation around it, that might cause it to get to, to, excuse me, on the other hand, if you put a bunch of insulation around it to cope with the winter, you might find it in the summer, it gets too hot. So summarization and winterization are somewhat antithetical. And that partially explains why Northern US and Canada, which goes regularly through much colder, more extreme temperatures, is able to have its generation stock handle these colds, okay? Those sort of cold weather. The issue is in Texas, we're experiencing both extreme heat and extreme cold. Our cold is different to most markets. And that's gonna have some implications for, has had implications for our prices that came out of this big event. It's a so-called energy-only market, or you might better call it an energy and ancillary services-only market. It's similar to the Australian market as it happens and to the Alberta market, but it's somewhat different to other markets. And the difference is that capital formation, that is decisions to invest in a new generation assets specifically, and contracting around those assets is on the basis about expectations of occasional high prices when supply is tight or even when demand is curtailed. Like every other ISO market, including Kaiso, Rekard has a short-term forward day head market. But the principle difference is that we don't have a mechanism to ensure capacity is built, nor to ensure that that capacity is reliable and available when needed. Nevertheless, that's been enough, I would argue, to get summarization to work. But I'd say apparently, or maybe even obviously not sufficient for winterization of the gas infrastructure, the electric infrastructure, and the water infrastructure. So as I said, we have no capacity market. Other US ISO markets have either a capacity market and the idea of a capacity market is the ISO forecasts about three years out, what they think the peak demand would be, and then procures enough generation to get that. Another alternative that applies in California is an obligation on retailers to have contracts to have with generation assets, again, based on something like an ISO forecast of demand out a year or two years. People have argued in the context of ERCOT that the problem was our lack of a capacity market. But the problem is our winter peak in 2021 was so much higher than anyone would have might have anticipated, in retrospect with 2020 hindsight, or these days I call it 2021 hindsight, we might have expected it, but I don't believe a capacity market would have had requirements set high enough to meet this peak. So it's not clear to me that a capacity market would have procured enough generation capacity to comfortably meet the peak. Now, on the other hand, it's also argued that a capacity market is designed to make sure generation is available. It includes penalties for not being available, for not turning up. And that motivates, for example, for a gas fire generator, that the gas fire generator agrees to a firm gas supply contract. So a firm gas supply contract with a gas supplier is basically saying, I want the gas, make sure you've got it and make sure you can get it to me. Another thing that it can motivate is a so-called dual fuel capability where a generator can both be fired on natural gas, but also on diesel, basically. Having said all that, it's not clear that that would be sufficient for a multi-day extreme event. Dual fuel capability, the tank is likely to be only big enough to survive a few hours. And furthermore, there were multiple firm gas contracts that were curtailed in our big event under force majeure clauses. That is to say the gas supplier basically were negged on the contract. So it's not clear how one would deal with it, would have dealt with that in a capacity market. And it's not clear that a capacity market would help. When we have scarcity, meaning curtailment or nitric curtailment situations, the Urquhart-Holstein market has another feature which is a little different to QISO, but there are similar arrangements in place in PJM, for example, Pennsylvania, New Jersey, Maryland. The Urquhart-Holstein prices are set administratively, automatically, we might say, to high levels when there's scarcity. This is called Operating Reserve Demand Curve, replaced previous tolerance of exercise of market power by small market participants under tight supply, which was called a scarcity pricing mechanism. And prices have occasionally risen towards or to the offer cap of $9,000 a megawatt hour. There is a so-called circuit breaker to change the cap after a threshold. But it's fair to say that when all of this was set up, when this Operating Reserve Demand Curve was set up, no one imagined more than a few hours here or there that capacity would be tight or there would be curtailment. So extended contiguous periods that the cap would not contemplate it. Unfortunately, that's what happened. We had extended periods at the cap. I see there is another Q&A, and it's a really interesting point. It says, sorry BS on the insulated sensor crap. And I would defer to someone who knows some more about sensors on that. And I fully admit, Terry, that we don't know publicly what was broken this time and we still don't know whether it's the same as what was broken last time. You've got another question, Terry, about how does Texas propose to deal with a range of wider weather things. Let's discuss that at the end, because I think it's fair to say that there's a lot of proposals on the table and it's not clear exactly what the direction will be. So I mentioned that the prices occasionally get very high, but there is a circuit breaker that changes the cap. How does that work? The circuit breaker is based on a so-called margin, the accumulated margin between the price and the peak or marginal cost. It's meant to allow for enough revenue in the market for generation asset owners to recover their capital cost, even though the energy prices most of the time only reflect energy costs. That circuit breaker is designed that when enough hours of high prices have accumulated, the offer cap will be lowered to avoid excessive wealth transfers. It was triggered, but unfortunately it was ineffective. It was triggered because our wholesale prices were at the offer cap for several days during the blackouts and enough hours accumulated at high prices that the peak or net margin trigger was triggered. But as I said, it was ineffective since the lower offer cap is based on a large multiple of gas prices. And those gas prices were also quite high. And in fact, ironically, the lower offer cap turned out to be higher than the higher offer cap. So it was not effective. There were pretty large wealth transfers at least between certain market participants. Indeed, high electricity prices persisted even after the end of the blackouts, which in my opinion was just basically a mistake somehow. There's current Texas legislation that's gonna possibly force repricing, although there's been a lot of back and forth about that. Having said that, even though the prices were very high, $9,000 a megawatt hour for basically four or five days, which is about 100 times, 50 times higher than what they typically are in Urquhart, very few end-use residential customers were exposed to these prices. Now, you might have seen some media reports discussing Gritty customers. Gritty has about a few tens of thousands of customers out of a total of 10 million. Gritty and a couple of other retailers have real-time rates. They sell energy, they price energy to their retail customers at the real-time wholesale price, which went to $9,000 a megawatt hour. And so some particular customers that didn't pay attention to Gritty's texts, asking them to conserve or find another retailer ended up with pretty big bills. But as a general rule, very few residential customers went through that. Many commercial industrial customers were at least partly exposed. Many of them in Texas have purchased contracts that are indexed to gas prices and gas prices were quite high. So that would have led them to have fairly elevated ultimate costs of purchasing electricity. So who lost? There were significant financial losses incurred by, for example, generators with forward positions that had outages or supply limitations. If you were a generator, even without a forward contract position and you had an outage, you forwent the opportunity to sell power at a very high price. So there was significant opportunity costs to generators that were out. Retailers that had obligations through fixed tariffs that weren't fully hedged were also on the losing end of this. But having said that, the rolling blackouts meant that less than the normal amount of electricity was being provided in many cases. However, at least one and I think several retailers have gone bankrupt and consumers that were not hedged. So in particular, some commercial industrial customers that purchased at wholesale or indexed wholesale rates and those relatively few retail customers on real-time wholesale prices. As I said, the price continued to be at the offer cap for two days after the blackouts ended. In my opinion, that's an error in administrative price setting. There's current controversy over whether to reprice because of the implications for other forward contracts. We've had several bankruptcies already announced. One retailer, a rural cooperative as it happened has declared bankruptcy and there are several others. There've been multiple resignations and firings, multiple investigations proceeding, multiple bills proceeding at the state level. And we expect further inquiries. And because of the significant role of gas here, unlike in California where, for example, the California Energy Commission looks after all energy and the California Public Utility Commission, as I understand it, regulates retail gas as well as electricity. In Texas, we have a very strange situation where electricity comes under one regulator and gas comes under another. And I think a lack of coordination there is something that we probably need to investigate. So is there wider relevance? And I think some of the Q&A will explore some different dimensions of wider relevance, but wider climatic relevance. This extreme cold phenomenon in the Northern Hemisphere is a jet stream phenomenon, as I understand it. I should point out, I'm definitely not a weather or climate change expert, but it does seem to be a very peculiar to the Northern Hemisphere, less likely, less severe, for example, in the Southern Hemisphere, not relevant to the tropics. But having said that, similar analogous effects could happen in the Southern Hemisphere. For example, recent heavy rain, some of you may be aware that there was very heavy rain in Australia. And that was in fact due to a combination of a La Nina and a jet stream. Whatever the exact cause, I'll observe to you the common mode phenomena are relevant to tight supply conditions everywhere. So for example, we might have a combination of widespread high temperatures, for example, California and Western US, low wind conditions on the coast, particularly after sundown, if you didn't get an evening breeze, coastal areas would be hot, air conditioning load would be high throughout the West. And that again could be a common mode phenomenon. So my point is whether it's hot or whether it's cold, I think we're particularly concerned about these common mode phenomena, which as a rule have not been very well treated in many reliability studies of the electricity industry. Unfortunately, as observed in a recent EPRI report, impactful weather events, and I'll quote, are increasing in frequency and intensity and geographical expanse and duration. I think it's very likely that some of those are climate change induced, whether the cold event in Texas was climate change, I'm not sure, I've had some discussions with some climatologists and they're a little skeptical of that, but of course we should expect that the hotter weather extremes are related to climate change. And I'll finish with a couple of observations about the implications. I think it's fair to say that as a society, for example, Texas as a community finds the levels of outages that occurred in February to be unacceptable. In other words, the societal risk tolerance for that sort of common mode phenomenon is lower than the apparent risk tolerance and exposure of the market participants. Roughly speaking, you might say that the market participants are gonna prepare for what they think on average is gonna be problematic. And that includes summer, but not so much for winter. And to me, that means that if we have these occasional extreme events, we're going to need an important role for regulations and standards to handle such phenomena. These are not new to the electricity industry. There are low voltage ride-through requirements that are aimed at making sure that generators stay connected when a common mode issue such as a fault that produces a low voltage occurs on the electricity system. So I think it's very important to think about regulations and standards for this. And a complicating factor, of course, is that we need to track potentially increasing occurrence, increasing probability of such events, not just historical averages. Now, in the case of cold, maybe we should just expect that it's gonna happen once a decade, but the summer extremes, it would seem due to climate change are gonna occur more and more often. And we're certainly gonna need to deal with that, with tracking that increased frequency as we think about our tolerance for exposure to the risks of such events. So I'll conclude by saying that extreme weather affected electric demand in supply in Texas. We had extended blackouts that were incredibly disruptive and I think unacceptable to the Texas community. Common mode cold was the common mode cause, I should say, of those multiple failures. And although that level of cold may not be relevant to California, let's say, may not be relevant to Australia, let's say, might not be relevant to the tropics, let's say, common mode causes are definitely relevant to other regions and we should be concerned and thinking carefully about them. So with that, I'm just gonna open up the Q&A again, just to have a look at it. Terry's question, first observation was BS on the insulated sensor crap. I'd definitely like to know more about that. I'm not sure whether this is the right forum for that. So I'd like to perhaps take that offline but I would like to address Terry Oliver's question that says, how does Texas propose to deal with a much wider range of weather and extremes arising from climate change? Let me say that there still are quite a few officials in Texas who at least affect to be skeptical of climate change. I think that's changing a lot and it has changed a lot. There are currently quite a few bills in place. I think there is going to be some legislation that requires weatherization of critical infrastructure. One of the things that I'm concerned about there, and Terry, I would like to take up the sensors. I'd really like to know what you know about this and it's reflected in part in a blog post of mine that is entitled ready, let's not ready fire aim part of the problem is that we have not got to the bottom of what exactly were the real causes. So I can certainly believe that improved weatherization somewhere is going to be important. And it might be important to improve weatherization in homes, in electricity infrastructure, in gas infrastructure, in water infrastructure. But we really need to understand a little bit better before we start deploying our resources. And one concern I have is that the political process going on at the moment has been excited to get on with it, but without I think investigating what's really needed. So in answer to your question, how does Texas propose to deal? I think at the moment, we're not dealing very well with it. We're not paying attention to the forensics before we make decisions, but I hope that in the course of time, we find that some of the crazier bills get winnowed out of the legislation. And now turn to a question from Octavi Semenin, I hope I pronounced that correctly. Aren't lots of utilities increasing their rates today to cover the losses they incurred during the event? So it's natural that if you were not hedged as an entity and that you might increase your rates. But for example, I'm an Austin Energy customer and it turns out that Austin Energy did pretty well weatherize its generation fleet and might have even made a profit from this event. So indeed, you can't expect people to be perhaps traditional utilities wanting to recover their costs, but bear in mind some entities didn't lose money and others that are competitive retailers are not really in a position to recover their losses if they were not fully hedged. Why is that? Because those competitive retailers have tariffs that they publish and that they have to stick to. Of course, if I consume more electricity, I'm gonna be paying more, but they won't be able to retroactively change those specified rates down the track. They might try to increase their prices, but bear in mind in the competitive retail areas, if retailer A starts to increase its prices, it's gonna lose customers to retailer B to some degree. So I think that in a retail restructured world, that puts a lot of downward pressure on those retailers and basically forces them to take the losses themselves, which by the way is the important incentive in a retail restructured world to get them to hedge against their exposure. So Octavia, I'm not sure if I fully answered your question, please feel free to ask some more if you have a follow-up on that, but I'm now gonna turn to Travis Nix. What are the effects of intermittent renewables increasing in percentage of capacity, having on ancillary services in moments of crisis where load spikes? So there's a couple of elements of this. One thing to think about is we had a large amount of thermal generation that wasn't working. In a future world where we're trying to meet demand without thermals, that's kind of like we were in February. And the difficulty is that some of the time the wind is not going to be blowing very well and it's gonna be cloudy. So I think what we experienced in February is a wake-up call about the implications when we ultimately have much higher levels of renewables and we get climate or weather conditions where the renewables aren't performing. So of course, as intermittent renewables increase as a percentage, we might expect that ancillary services increase. And in fact, there is a Texas bill that is wanting to assign the cost of ancillary services caused, so to speak, by renewables to those renewables. And this is a little strange compared to our typical history because in all markets in the US, we currently assign the cost of ancillary services to consumers. So it's kind of strange to now be saying, okay, we're just gonna pick out one category of generation and we're going to assign the costs of procuring ancillary services to them. That's a Texas bill that's winding its way through the Texas House and Senate at the moment. As it happens, it's not clear to me that we're needing currently different situation if we get to much, much higher levels of renewables, but at least at current levels, we haven't needed to procure a lot of additional ancillary services. And there's a couple of reasons for that. One of them is that the short time scales of frequency regulation services, frequency regulation services, it turns out there's been a number of changes in the ERCOP market. And my PhD student, Juan Andrade and I, you could Google Juan Andrade, A-N-D-R-A-D-E, looked at the amount of procured frequency regulation over time as we increase the amount of renewables. Well, it turns out that indeed with increasing wind, wind is still the majority renewable in Texas, you need more frequency regulation, but there have been a number of changes to the ERCOP market that have more than compensated for the amount of wind. So that's meant that we really have not had great increases in, we haven't had great increases in the amount of frequency regulation. Now, what other effects might there be? Well, the California duck curve, which I hope all of you folks are familiar with is a phenomenon that's associated with a lot of solar and particularly a lot of renewable, excuse me, a lot of rooftop solar, I should say. And we hear the two have not had that much solar, but more to the point, the solar that we've had is large scale and therefore dispatchable and therefore the ISO can smooth out its ramps, for example, just by dispatching the solar down. So I would say as a general rule, we have, ERCOP has been able to integrate a remarkable amount of renewable energy without incurring significantly additional and solar services costs. Now, there may be a few people that might dispute that a little bit, but that seems, I think that's a fairly solid observation. We've hear the two had enough thermal capacity that the load following, keeping up with changes in renewables has been easily done by the thermal capacity. I'll emphasize though that as we go to much higher levels of renewables, that's gonna be a lot harder, right? So not clear that we can stay like that forever. There's a comment, heat pumps that use CO2 can operate at quite low temperatures versus heat pumps that use refrigerants, such as HFCs. Ole, I don't dispute that. I'll just observe that there are the commercial heat pumps. I recently bought a heat pump for my house and the ones that are commercially available are currently not gonna operate down to minus 20 Fahrenheit. But certainly conceded that you could redesign it and you could use specialty heat pumps to do that. And I commend that you have a look at some of the writings of, for example, Amory Lovens, you might look up our conference that I was involved in in April, the Austin Electricity Conference, check out the keynote speech by Amory Lovens and check out more generally Amory's comments on this. And he definitely has made this observation, but commercially available for residences currently, I believe that they've improved in the last 20 years, but they're not necessarily able to deal with a 10 Fahrenheit. So I'm glad to see Terry is asking. We have eight or nine more questions. Maybe if you can try to answer five or four. Yes, yeah, let me get to a couple more. Is there any that you would particularly like me to answer? Let me try a couple here. How should markets think about balancing efficiency versus resilience? I think this is a really good open question. Again, I think I wanna emphasize to everybody that we don't really know at least publicly what broke is increasing feeling that there is, and I observed this earlier that the natural gas system was problematic. There may be particular points where we can very cheaply add a lot of resilience to the network just by investing in a little bit of weatherization here or there. So let me observe that markets are only gonna think about more or less risk-neutral behavior. They're gonna think about what is the level of resilience or redundancy that I can justify the cost on the basis of the expected value to me of being able to, for example, sell my energy when supply is tight. And I don't think we can rely on that for, for example, the extreme cold, okay? Someone said, would a heat pump which uses gas as a backup be something to do? And yeah, I think that would be a possibility. And in fact, next time I buy a new air conditioner or a new heat pump, I'm definitely gonna try and see whether that's a possibility. I think that makes sense. I think part of the problem is that a lot of these consumer products, even though they're quite expensive, you know, they compete on price and being able to provide for the backup situation in extreme may not be their main selling point. And so again, this might be something that needs to be done from a regulatory perspective rather than expected to play well in the marketplace. I'll also mention that from a California perspective, I believe that California is intending to be basically only adding houses that are electric only sometime into the future. And so naturally gas backup would not be relevant there. Carlos Nacimente, Nacimente, excuse me for my poor pronunciation there. If you have robust smart grid, we will know where is the trouble in the network. Yes, I think Carlos, it's a good point. I think more broadly there are some smart grid concepts that could have helped a lot here. So I mentioned that theaters stayed off for the whole 60 hours or so because we were not able to more finely control things. One technological fix to that that would come under the smart grid rubric is automated or automatic, excuse me, automated remote sectionalization where you could disconnect the load on part of a feeder and connect it to another feeder that might enable you to modulate things better. Furthermore, as it happens, every retail customer in ERCOT is directly remotely switchable off. They can be disconnected remotely. Unfortunately, the bandwidth to do that can only support a few thousand or a few tens of thousands of connections and disconnections a day. It's supposed to, it's intended to facilitate retail choice. I think what we're gonna find is as well as additional remote sectionalization, we will see more bandwidth to be able to control those meters. Now, I'm not sure Carlos, if you were meaning out on the distribution level or at the transmission and generation level, I think already at the transmission and generation level, we already know where the troubles were. And for example, they were not transmission problems, they were predominantly generation problems. How much, next question is how much storage capacity would have to be implemented in Texas to avoid the blackout? So if we say 20 gigawatts times 60 hours, we would have needed about a hundred and 1200 gigawatt hours, so about a terawatt hour. That comes to a couple of billion, sorry, I think it comes to a couple of trillion dollars. So it's probably not viable. We need to figure out other ways to do it, okay? Travis Nick says, is there a max percent capacity that can be filled by renewables before ancillary services start being affected? Well, let me say, Travis, in my earlier comment, I was not saying that the wind doesn't contribute to increased ancillary services. But what I was saying was that we had figured out other ways to better utilize at least the frequency regulation ancillary services, at least that part, so that the improvements in the market enabled so much better utilization of the generation capacity providing those ancillary services that we've been able to integrate a huge amount of wind without greatly increasing the procured capacity. But it's true that the increasing wind and indeed the increasing solar tends to, all other things equal, increase the amount of ancillary services that we need. There's no doubt about that. I think the difficulty here in the answer is that as you start to increase additional capacity, it's not so much the ancillary services as defined in ERCOT that will be affected. It's the ability of the market to follow net load ramps, okay? Now in other markets, MISO and KISO, there are already ramping products. So I would say already in both MISO and KISO, they're the particular mix of renewables and generation and thermal generation means that they've already had to define new ancillary services, right? That means we're defining new things that are affected by the amount of renewables. So depending on how you calculate things, you might observe that California, might estimate the renewable level in California. Something to bear in mind there is that California balancing area now through the, it's really throughout the West. So the amount of renewables throughout the West is much smaller than the amount in California. So I'm not trying to prevaricate here, but I'm observing that it's not an on-off issue. It's the more renewables, the more you're gonna affect and at some point, you're not gonna be able to follow the net load with the action of the market, a five minute by five minute market. And you're gonna need to define new ancillary services that hasn't happened yet in ERCOT. And I don't think it'll happen anytime soon in part because we still have a significant gas thermal fleet. But once we start to get integration levels that are significantly higher, I think we are gonna need some additional ancillary services. I haven't tried to figure that out specifically. And I can't point you to good research that's estimated on that. And one of the reasons why I think it's hard to get a good estimate, the lesson I learned from looking at the frequency regulation material is that very often by changing the market, tweaking the market rules, you can fix what would otherwise appear to be a problem. So it might look like something's gonna be a problem next year or in five years time, but if you have a market design that's susceptible of being changed and there are people of goodwill who wanna change it to make it work better, that can forestall a lot of these issues. So I put it to you that ERCOT's ability to integrate so much wind, which is truly amazing in fact, has been in part due to various market design changes. And that's down to ingenuity and it's hard to predict ingenuity. Okay, are hospitals in Texas required to have backup generating systems? Yes, essentially all hospitals, I don't know exactly the requirements, but every hospital that I've been to in Texas has a backup generator in their parking lot or somewhere around. The thing is you wouldn't typically want to rely on that. So the backup generator is if the hospital gets outaged, I wouldn't wanna deliberately outage the hospital because that would be kind of a scary thing to do, but here's another place where I think your point is a better one or perhaps more relevant. There's a supermarket here in Texas, it's called HEB. And it turns out that they have backup generators because if there's a blackout, their fridges will warm up and they will lose all their produce. So I think they are a really good candidate. They have backup generators, they need backup generators anyway, and in tight supply conditions, they can switch on those generators. And so I think that's perhaps a better example where you know what? If they got interrupted, they'd still be running their generator and if the generator failed, it would hurt their produce, but it wouldn't kill people, right? So I think that that's a better example of where the additional flexibility can come in. Okay. There's one more. One more, it says, comment, it's from Karl Lennox. The Texas Legislative Bill is not allocating ancillary cost caused by renewables. It allocates all ancillary costs as I understand it. And I guess that means all of them to renewables, right? This is inconsistent with basic principles of cost causation because the level of AS needed is a function of the overall system architecture and generation mix. And there's no fair way to allocate cost back, because that's what I've heard. So the bill is very limited. There's just one sentence. And I don't have it in front of me, but I thought that it was specific to renewables, but it might've changed since I last looked at it. Let me say this. I think it's a very ill-advised bill. So essentially, no matter what's in it, I think it's gonna be a very ill-advised bill. Whether it tries to identify what's caused by renewables or otherwise just allocates everything to renewables, I think this is not gonna lead to more efficient outcomes, but perhaps more to the point, it's not addressing the fundamental causation of our blackout in February, which is apparently what's motivated all of the bills that are running around in the state. So I just don't see it as relevant to the issue at hand, which is how do we deal with the next cold event? So with that, I think that's the last question. That was a great set of questions. I hope I gave some reasonable answers. If anybody wants to challenge me on anything, and I certainly appreciate the correction on the weatherization of the sensors, please get in contact with me, because I'd love to carry on this conversation further. Ross, thank you very much for the wonderful presentation. My pleasure. So yeah, I guess if the audience have any more questions, they can send you an email. That'd be delightful. Yeah, yeah. OK, with that, I'm the normal questions, and I'm all talked out. Thanks very much for inviting me, Chin Woo. I really enjoy presenting this. I wish I could get back to live presentations because I'd sure like to see the faces of folks asking the questions and see whether they're smiling or frowning when I answer. And but maybe that's next year. Yeah, I'm sure we'll see you in a very near future. Super. Yeah, OK, good. Thank you very much again. My pleasure. Bye-bye, everybody. Thanks for coming along.