Preliminary lessons available to be learned from Feb 2021 extended cold spell
A large number of “hot takes” are appearing now that the cold wave that began arriving on Feb 11, 2021 has moved into areas where sub-freezing temperatures in Feb are normal.
If the politically charged nature of the takes could be harnessed, the energy released would be able to keep quite a few homes supplied with power. But, no one has found a way to capture and convert words and hot air into electricity – at least not yet.
That doesn’t stop writers from writing. I plead guilty to the charge of adding to the pile of non-electricity producing words.
It’s a necessary endeavor because so many of the hot takes have been produced by people whose agendas are different from mine.
Though no power source worked perfectly throughout the five day period when the Electricity Reliability Council of Texas (ERCOT) declared the grid condition to be at Emergency Energy Alert 3 (EEA 3), some power sources worked better than others.
While almost every generating source in the system has room for improvement, some have limitations that cap their performance no matter how perfectly they live up to their maximum potential.
The storm revealed severe weaknesses in the current grid resource management model that are worth discussing in an informed, responsible way. Lessons learned discussions are just entering into the early stages, but with an open minded, questioning attitude, it’s not too early to produce some recommended short term actions.
Aside: The title of this piece is intended to indicate recognition that performing “lessons learned” analysis and creating action plans is no guarantee of better future performance if responsible people choose not to take recommended actions. End Aside.
Why did the grid get so stressed?
About a week before the cold weather arrived, it was evident to those who pay close attention to electricity grid supply and management issues that the ERCOT service territory was going to experience an extremely challenging period. Wholesale prices were going to increase by many multiples and would challenge the existing cap of $9,000/MWh (300 times the usual price of $30/MWh).
Every generator in the system would be motivated – incentivized by high prices – to produce as much electricity as it could possibly produce and even then, there probably would not be enough power available to serve every customer as much as they wanted or needed to buy.
Despite the very high prices, almost no help would be supplied from outside the ERCOT system. In order to maintain its fiercely protected independence from the Federal Energy Regulator Commission, ERCOT has kept connections to other US grids at a bare minimum. That ensures that its electricity falls outside of the Constitution’s interstate commerce clause used to justify federal regulation. There are some significant cross-border connections into Mexico.
Few commercial or retail customers would know how much their power was costing at the time they decided to use it, so the system would not receive much help customers making informed choices about timing or limiting their use.
Most of those customers would not even receive a sharply higher bill at the end of the month, because their rates would be adjusted over time to repay the costs of a sustained period of high spot market prices.
That is how the system in place is designed to respond to severe weather or other stressed on the power system. ERCOT has chosen an “energy-only” resource management model where competing generators bid their capability into wholesale electricity markets that are settled and priced every 5 minutes. There is no other source of revenue.
Why choose an “energy-only” resource planning model?
An “energy-only” model keeps wholesale prices low during fair weather. Low prices encourage customers to add devices and equipment. On a larger, longer term scale, it encourages businesses and even residents to migrate to take advantage of having low cost electricity available.
But it doesn’t provide sufficient predictable revenue to encourage investment in durable generating sources or long term, guaranteed delivery fuel supply contracts.
Because all energy sources have different cost structures and different bidding strategies, wholesale prices have wide and rapid swings. By design, the system provides massive returns during periods where demand exceeds supply. This characteristic is supposed to provide all of the necessary incentives for generators to be ready at all times to provide their full capability.
But even with almost a week’s notice that an historic weather event was on its way, there were limited actions available for most generators to prepare to maximize their returns from the coming period of scarcity. If they did not already own reserve fuel tanks or winterized generator packages, it was too late to make arrangements for installation.
Some generators might have been able to get a rush liquid fuel shipment to top off any existing tanks – at an ever increasing price – or they might have been able to make careful inspections and fix obvious system weaknesses. If they discovered some missing insulation or a non functioning heat trace system, they might have had enough time to make repairs.
But most would have had to simply hope for the best and do whatever they could to keep producing power.
Most customers were blissfully unaware of the decisions that had created a system where participants depend on scarcity pricing to make a profit in the business of supplying electricity to the ERCOT market. They didn’t know that many of the generators in the market knew they would only be able to produce limited amounts of power, even with sustained prices at or near the cap of 300 times the usual grid price.
They didn’t know that most generators in the market would be richly rewarded even if they were only able to produce 10, 20, or 50% of their expected capacity during the several day-long deep freeze.
Few frozen wind turbines
The 2021 cold weather event began February 11, 2021 with freezing rain, one of the most impactful kinds of winter precipitation. The nation began paying attention to the incoming cold weather as a result of news reports of icing roads in the Dallas-For Worth area that led to a massive, 130 car pile up on I-35.
That wave of the storm did not actually produce massive quantities of ice; the slick road conditions resulted from less than a tenth of an inch of ice on a road where drivers didn’t exercise sufficient speed restraint.
While freezing rain can accumulate on almost any surface exposed to the weather, including aircraft wings and wind turbine blades, there is no available evidence suggesting that a major portion of Texas’s installed base of >30,000 MW of wind turbine generators experienced icing sufficient to have much impact on their generating ability.
There were some reports that claimed iced or frozen turbines were a significant cause of lost wind power generation, but the real culprit was a relatively common, and predicted winter storm weather pattern that included long periods where high pressure covered a huge land mass. When atmospheric pressure is the same over a large area, there is no driving force that creates wind.
Many people that strongly support the continued rapid development of wind and solar power generating systems declared that their favored power sources performed at or above their day-ahead predictions. They wrote lengthy defenses of wind generation and declared that the historical performance of turbines in areas that regularly experience colder weather than what Texas experienced in Feb 2021 proved that there was nothing inherent about wind that made it especially vulnerable to severe weather.
Midwestern utility company MidAmerican Energy Company has shown that wind energy is highly reliable, even in harsh Iowa conditions. In 2020, 80 percent of the utility’s electricity was generated by renewable energy — the majority of which comes from its 3,300 wind turbines, said Geoff Greenwood, a spokesperson for MidAmerican Energy.“Wind turbines can handle the cold just fine. Just look at Iowa.” Vox Feb 19, 2021
From the wind and solar advocates’ point of view, better-than-expected performance means that wind and solar should bear little or no responsibility for low generation during an electricity supply shortage.
Even in a supply crunch severe enough to cause an Emergency Energy Alert stage 3 (EEA-3) – the highest available response level – there is nothing that wind and solar generators can do to make their systems supply power. They must wait until the wind starts blowing or the sun comes up. Low sunlight inevitably affects areas spanning more than half of the planet all the time. Wind is more localized, but there are times when entire continents can be still for many hours several days at a time.
Vast majority of wind and solar advocates are observant enough to know these facts. They even take offense when they are introduced into energy discussions. If challenged about the value of continued strong support and mandates for increasing wind and solar penetration, one of their arguments is that using the wind and the sun to supply energy when it is available allows fossil fuel generating sources to burn less fuel.
That would be a reasonable response if the only competitor to wind and solar was fossil fuel. It’s even a reasonable response in systems where large hydroelectric dams are part of the generating mix because it allows the water to remain behind the dam, ready to be used when wind and solar generation falls off.
But opportunistically displacing other sources of power can lead to unproductive consequences like eliminating enough revenue from nuclear plants to make them struggle financially. Right now, there are firm plans in place to close five operating nuclear plants in the US during 2021.
Though some industry leaders have vociferously denied that wind and solar power can be blamed for those closure decisions, the financial evidence is clear. Low grid prices and grid congestion fees in regions where there is abundant wind or solar power available create a “missing money” situation that stresses large steady-running generators that serve base load very well.
There is a correlation and a causation between the location of Exelon’s Byron and Dresden power plants in high wind areas and their financial performance. The same holds true for Diablo Canyon, but the culprit in California is a massive quantity of solar power generation that can create negative pricing during the middle of the day.
Large numbers of gas-fired generators could not produce power
One of the design features of the “energy-only” market model in Texas is that it rewards low capital cost equipment that can burn natural gas. For the past 13 years, natural gas has been abundantly available in many parts of the US, especially in Texas.
The Permian Basin, much of which is under Texas soil, is one of the world’s most prolific oil and gas reservoirs, but it isn’t the only major source of gas in the state. There are other shale formations and there are large gas reservoirs in the Gulf of Mexico off of the Texas shore.
Natural gas, which is more accurately called methane, burns cleanly enough so that a stream of its combustion byproducts can be directly used to spin turbines in a Brayton cycle. Those machines are simple and cheap compared to the huge Rankine (steam) cycle plants that are needed to burn dirtier fuels like coal or lignite. Brayton cycles work well in combination with simple, relatively small steam plants to produce highly efficient Combined Cycle Gas Turbines (CCGT) power plants.
The “energy-only” market structure has helped gas to push most coal and lignite off of the Texas grid, producing significant air pollution reduction and a reduction in greenhouse gas emissions. Using more natural gas in power production has been beneficial to the Texas economy as well, since most of the gas burned in the state is extracted in the state.
But a known challenge related to natural gas is that it is more difficult to store materials that are vapors (gaseous) than it is to store solids or liquids. Gas can be compressed and it can be liquified by cooling it to extremely low temperatures, but both of those processes add costs and consume power.
Without any source of revenues for power generations other than selling electricity, there are no reasons why any generator would spend money to store fuel on site to use in the rare case where there are interruptions in the fuel supply.
Even in fair weather, only a portion of the methane that is extracted gets burned to produce electricity. Some of it gets used as a raw material for petrochemicals and plastics. Another portion gets used in cooking – both residential and commercial – while another is used in industrial process heat and to heat water in both homes and large buildings.
During cold weather events, heating buildings quickly grows and can become larger than all of the other uses combined. But natural gas production rarely increases when the weather gets cold. During the event that lasted from Feb 11-Feb 18 2021, daily gas extraction fell by nearly 20% due to various issues in the system.
The predictable, though not often publicly predicted, effect of a high dependence on natural gas to supply its usual amounts of electricity, to expand its production to make up for low wind and solar production, and to supply building heating systems is a system where some needs are not met.
Under the low cost, just-in-time, fuel supply model that is an inherent result of an “energy-only” market scheme, there is simply not enough fuel in the system to supply all demands all of the time. When the fuel that supplies the majority of the power generators in a system is stressed, all generators that burn that fuel can be affected.
In the lingo I learned as an operating power plant engineer and as a participant in a power plant design project, insufficient gas during a cold weather event is a predictable “common cause failure” for an electricity supply system.
As designed, the market uses pricing signals to balance demands with supply. But those price signals have to be dramatic to change behavior because both supply and demand have a large amount of inertia and cannot be easily changed.
When price signals aren’t sufficient to change behavior fast enough, the only option the grid operator has left is to balance demand with available supply by turning off the power to some customers.
What about the nuclear plants?
At 0537 on Monday, February 15 South Texas Project unit 1 tripped off line. (That link includes far more details about the event than can be fit into this post.) Other than that single event, all 93 of the 96 nuclear plants in the US that were operating before the cold weather event began continued producing as much power as they were asked to produce.
The only nuclear power station that did not produce as much power throughout the event as it possibly could have produced was Arkansas Nuclear One. For part of the week, the regional transmission operator asked the plant operators to supply less than their plant’s design power in order to keep the system in balance.
Here is a quote from an Entergy Arkansas spokesman explaining that period of less than 100% power.
Arkansas Nuclear One’s dual units continue to operate safely and securely throughout the weather event, with essential functions staffed by Entergy’s team members. Both units currently are operating at reduced power at the request of the independent grid operator.
The Midcontinent Independent System Operator is an not-for-profit organization that works to ensure reliable power supplies in part of Canada and 15 U.S. states, including Entergy’s service territory in Arkansas, Louisiana, Mississippi and portions of Texas.
In doing so, MISO often asks generation facilities to change power levels during times of potential grid instability. Entergy’s regional generation facilities are coordinating closely with the grid operator, and power levels at our plants may continue to rise or fall as the dispatcher works to keep transmission functions stable.Direct message from Entergy Arkansas (@EnteryArk)
Approximately 60 hours later, STP 1 returned to service and increased its output to its maximum capacity. That lengthy shut down exposes one of the reasons why there have been few new nuclear plants built in the US during the past 30 years.
The large, light water nuclear power plants that were selected to be built commercially in the 1960s-1990s work best if run steadily. If they are taken off line for any reason, power restoration can take many hours to several days. While it might be possible to improve that situation for existing reactors, it is best done via a meticulous, methodical, time consuming path.
Under our current construct as refined by many decades of continuous operational improvements, nuclear plant shutdowns and start ups are rare events. They happen less than once per year at each unit.
If the population size for nuclear reactors is restricted to the four units physically located in Texas, the operational score for nuclear during the 5 days of rotating outages turns out to be about 86%, which should be a solid B in most grading systems. (That is calculated on power produced compared to the power that could have been produced if all four units operated perfectly throughout the event.)
But given the widespread nature of cold fronts and the impacts of stresses in the nationwide natural gas delivery system, it might be fair to include the performance of a larger population of nuclear plants. 92 out of 96 operating at or near 100% produces an A in almost every known grading system.
On the scale of producing as much power as expected by grid planners, nuclear did about as well as it was expected to do. It’s not a perfect power source; there are numerous ways for systems to fail to produce at 100% of rated output 100% of the time. But nuclear met or exceeded some pretty high expectations.
It is important for systems planners or people who influence system planning actions to recognize that nuclear plants offer several important features. Among its strengths is its independence from the common cause failures of fuel supply constraints and direct dependence on wind and sun availability.
It is also a clean power source, with life cycle CO2 emissions that rival onshore wind and beat both offshore wind and most solar systems. It produces virtually zero air or water pollution. Its ‘waste’ heat could become a valuable resource if systems were properly designed to use it beneficially.
Were lessons available from the Texas cold weather event of 2011 (Super Freeze) actually learned?
It’s not accurate for people to claim that the freeze of 2021 was a complete surprise or had no precedent in history. Galveston Bay has frozen solid several times in the past 50 years. In 2011, there was a cold weather event that brought temperatures just as cold and just as widespread as those experienced this year, though that event was shorter.
Many of the recommendations from 2011 post event reports were not implemented. The state persisted in pursuing its fierce grid independence. A substantial increase in wind power generation was accompanied by a growing boom in solar energy and major new transmission lines to move their power. Natural gas dependence has increased by double digit portions.
As much as it pains me to admit this, if the nuclear plant construction plans that were announced in 2007-2009 had been pursued, they would not have helped avoid the issues that appeared.
It’s difficult to prove a counterfactual historical point, but it’s easy to point to the only US nuclear project that survived from that brief period of excitement about new nuclear power plant construction. Vogtle units 3 & 4 will not enter service until sometime in the next two years. They were the leader projects from the Nuclear Renaissance and they are still not complete.
The next time we revive the nuclear plant construction industry, we must do a better job. We must achieve better cost and schedule performance and we must make design choices that recognize the importance of flexibility and responsiveness. That might include implementing some of the speedy recovery capabilities that have long been a part of military nuclear power plant design and operations.
If society determines that it is unacceptable to have a power grid that cuts off customers for many hours at a time during a period when being without power can be deadly, it must accept the fact that markets cannot be the decision makers.
Cheapness on a short duration scale – like 5-minute settlement markets – cannot be the sole criteria for selecting power sources.
Well done Rob. Would love the opportunity to share our approach at SO on advanced reactor R&D.
Other than a required ancillary services market and stated reserves, ERCOT runs an energy-only market. There is no market for capacity.
For details, see
This has worked for the expected summer high load. Since adopting the energy-only rules for bidding by generators there have been no rolling blackouts. Until this February.
I don’t particularly blame ERCOT or even the insufficient winterizing by the generating companies. I do blame, in retrospect, the Texas Railroad Commission for failing to insist that natural gas pumping and piping companies be capable of JIT delivery despite the unprecedented cold.
As a footnote, I point out that the nuclear power stations in ERCOT Texas compete quite well in the energy-only market; there’s no need for a capacity market.
@David B. Benson
I don’t think capacity markets are a solution. As you noted, nuclear plants can comfortably serve in energy-only markets (they produce massive amounts of energy over extended periods) as long as energy prices are high enough for enough hours every year. But if energy prices are too low or too unpredictable, new cones won’t be built.
Current nuclear plant owners in Texas are rewarded handsomely and more frequently under conditions where the system approaches instability. Scarcity pricing is good for their revenue.
Question I am posing in this piece is whether or not that situation is acceptable to electricity customers. Does it produce the results that they want from electricity supply system?
ERCOT Texas is also unusual in enabling a multiplicity of retail plans. Some do dollar cost averaging over a year. Customers with such plans won’t notice, much, this fiasco in their electricity bills.
Others choose wholesale+ plans. They are the ones stuck with very large electricity bills.
Many were seriously affected by the inability of the rolling blackouts to actually roll. The lack of generation was too severe.
Capacity markets provide money for new power plants to be built. These are (surprise) inexpensive gas-fired plants…that’s all the money capacity markets can supply. So now you have an extra power plant. What about the fuel supply for that plant?
My recent book, “Shorting the Grid,” describes how capacity markets do nothing, nada, zero—for winter fuel security.
In a short form, this November blog post updates “Shorting,” and gives somewhat of an overview of earlier material. In “Shorting,” the section on winter reliability starts with the chapter: Winter Lights. Meanwhile, here’s the blog post.
People constantly think that a new auction (capacity! pay for performance!) will solve the reliability problem. New auctions are bandaids, and they are kludges. They are part of the problem, and not the solution.
While I am a staunch free-enterprise, the whole capacity market scheme (aka price gaming) is Enron all over again. No reinvestment of windfall gains in better capital or utility debt management. No consumer choices to upgrade to efficient and cost effective arrangements. No real wisdom of utility choices in energy mixes and technology. Certainly no investment in a truly cost effective national grid. Just price gouging when consumers least need it, and hidden utility cosys.And where do you think Enron decided to build it’s new corporate campus before they went belly up?
Will the lessons be learned? It seems to survive an event like this in the future natural gas storage at some number of strategically located gas plants would be the solution. Texas relies so much on natural gas that you have to be able to heat and generate through the duration of the vent. Not sure this is practical.
During the extreme polar vortex , I think it was 2014 a lot of gas plants in the NE were dual fuel and switched to oil. There were also reserve oil burners used.
And if gas was available for the generators would that have been enough with near zero wind? How much does Texas credit wind in their planning?
But the heavily politicized finger pointing won’t solve anything
Apparently their LNG plants were disabled too. We hear that Boston again had to import it from Russia despite the much vaunted capacity of TX to export cheap gas via fracking. Since the EROIA of LNG is very low compared to modern nuclear or even wind, I suggest using nuclear with cogen to process the gas would have been very shrewd. Better yet ethane, which is ordinarily burned off. It is easier to liquify than methane and contains more heat content per tanker, and turbines have now been designed to burn it. Much propane and butane would have been obtained for auto gas, a clean gasoline substitute, well within the reach of Texans.
Thank you Rod; this was the most cogent analysis I’ve seen yet. Excellent work!
Excellent point about the waste heat. It’s a much-overlooked resource which really ought to be included in integrated resource planning.
The Chinese and Russians are more forward-looking than we’ve been. The AP1000s at Haiyang are being connected to a district heating system, and the floating NPP Akademik Lomonosov is supplying heat as well as electric power.
Nuclear district heating systems would go a long, long way to displace natural gas. I ran the numbers once, and an all-nuclear USA would have more than enough waste heat to replace every bit of NG burned in the country even in January. As in, several times as much.
Don’t need nuclear to follow your idea. A CAES concept developed in Macintosh AL was ready in the 90’s but was deemed too thermally inefficient to replicate. It now turns out that seperating the heat from the compressed air brings the thermal efficiency up to modern gas turbines, and certainly warrants underutilized off peak electricity to compress the air. Salt deposits from Texas to Alabama provide endless opportunity to site CAES adjacent toelectris and gas grids. Would have saved a $60 bil effort to create a Sourhern gas grid in the late 90’s, which was stranded when Bush II decided to grant coal a moratorium on clean air regs for 15 years! Texas of course, rejected investment in CAES, preferring dodgy wind, for the pleasure of subsidizing Texas folk to use their washing machines and TV’s at night when wind was strongest.
Agree with everything you say, Rob. Here’s my historical take from the possibly weird perspective of a Texan who reported on the California crisis in 2000-20001
Will – are you addressing me? Did you mean Rod and not Rob?
Sorry, Rod. I’ll join blame game by blaming the cold for my fingers not typing correctly!
When in doubt, always blame the autoconnect!
So basically, wind and solar were not to blame because we already knew beforehand they’d be useless in a blizzard. Got it. I guess it takes a Ph.D to explain that. They are useless but its ok because we knew that already.
Here in Holland the capacity factor of solar PV averaged around 2% the last 3 months. No blizzards, maybe 4 days of snow. Quite a mild winter otherwise.
Hey its ok because we already knew solar PV wouldn’t be there 98% of the time. Cue the Ph.D.
The news media is reporting that the two-unit Comanche Peak nuclear power plant was as little as 3 minutes away from automatic reactor shutdown, and the entire grid as little as 4 minutes and 37 seconds away from total collapse, according to ERCOT. These statements raise more questions than they answer.
There are a number of potential reactor trips that could have been experienced, e.g. loss of load/turbine trip or loss of offsite power leading to any of a number of trip functions. During the 2003 Northeast Blackout event, nine US nuclear power units experienced rapid shutdowns (reactor trips) as a consequence of the power outage while six nuclear units experienced significant electrical disturbances but were able to continue generating electricity. Indian Point Units 2 and 3 tripped on under-frequency to the reactor coolant pump buses at around 58 Hz. This trip is a design feature to protect against unacceptably low core flow and to prevent the nuclear fuel rods from experiencing localized fuel failures (departure from nucleate boiling). It is public knowledge.
It is also public knowledge that ISO New England has emergency operating procedures and training to shed up to 50% of the load in 10 minutes to respond to catastrophic natural gas pipeline events. It is not clear under what set of circumstances and what combination of equipment failures and human errors that frequencies as low as 58 Hz might be reached, resulting in the automatic trip of the Seabrook and two operating Millstone Units. Whatever analyses might have been performed by national labs remain FOR OFFICIAL USE ONLY.
Suffice to say that the key lesson to be learned from all of these experiences is that intermittent sources of electricity such as wind and solar AND just-in-time fuel sources such as natural gas are unreliable and/or not resilient. Ironically, FERC just closed the docket on resilience. A half-dozen nuclear units may close this year, and up to one dozen units by 2025. You be the judge on the wisdom or lack thereof regarding our nation’s grid.
Note: the commenter is a former manager of safety analysis, previously responsible for LOCA analysis, non-LOCA analysis, containment analysis, and risk assessment for the Connecticut Yankee and three Millstone units. Opinions expressed are those of the author and do not reflect the views of past or current entities.
The wind and solar enthusiasts defence is rather hilarious. Oh yes, our energy sources, which are dependent on the WEATHER, and are by definition unreliable (ZERO dispatchability on demand) are in no way responsible for issues during extreme WEATHER. They need Ph. D.s to explain this.
The nuclear plants seem to me a classic case of insufficient design basis – equipment, insulation, heat tracing, HVAC etc. were simply not designed for such temperatures/wind speed combination, so unsurprisingly there will be failures when one goes far outside the design range. Anything can be designed for – 100 year blizzard, 1000 year blizzard, 10000 year blizzard. Just costs ever more money. At some point one has to decide on a residual risk level. If solar panels and wind turbines must be designed for a 1 in 10000 year blizzard, rest assured they wouldn’t be cheap. Looking at some photos of damaged solar panels from hailstones, its pretty clear they aren’t even designed for a 1 in 20 year hailstorm.
But the reactor DNB thing – are you saying that a few percent drop in the coolant flow rate causes issues with critical heat flux?!
Surely these plants aren’t designed with such poor thermalhydraulic margins?
Also, instead of tripping the reactor, wouldn’t it be better to decouple from the grid and run on islanding mode with the reactor throttled down automatically?
Throwing out your main energy source (reactor) as soon as external energy (grid) is not available seems to me a poor design philosophy. Similar to the Japanese plants automatically tripping on seismic – not smart as large seismic likely trips the grid, now you lose 2 sources of energy and are dependent on the plant diesels.
Nuclear power plants are designed to have 95% confidence or greater of 5% probability or less that an anticipated design basis transient such as a loss of coolant flow would result in DNB. And that is typically to the worst location of the hottest fuel rod in the hottest fuel assembly.
Some plants are designed to ride out a loss of grid event, but for how long, and to what purpose if they are not generating power? Plants will typically disconnect from the grid on under/degraded voltage to protect safety equipment. I can not describe 60 years of nuclear reactor safety philosophy in 250 words or less. Not tripping the reactor when safety limits are challenged is against all good principles of safe operation for fuel, equipment and the public.
Don’t get me wrong, I’m not saying don’t trip if there’s a risk of core damage, rather the limits being (apparently) rather fragile.
This seems like a poor design to me – this shouldn’t be a twitchy formula 1 racing car where everything mechanical and thermal is designed at 95% of the limit. This should be more like a reliable long haul diesel truck with big margins on engine mechanics and cooling.
Don’t get me started on the probabilistic stuff, claiming a 1 in 10000 year core damage frequency and then saying there’s a 1 in 20 we’re wrong due to uncertainties. Not a convincing argument to me. Your core damage frequency is now determined by the uncertainty.
I’m a thermal-hydraulics guy. When I see a 5% margin to critical heat flux I nearly get a heart attack. Not what I want to see in a design. If my car has a slight cooling flow blockage and gets 95% of the normal cooling flow and that overheats the engine I’m not going to be a happy customer.
These PWRs have very high power densities and seem to be a glass giant operating right up to the margins. Given uncertainties in local power production (changing fuel isotopics with burnup, control rod movements etc.) and local flow (complex fuel assembly design, fixed orificing) this is clearly an area where we want to use big margins.
The more I learn about PWR the more I think they are formula 1 racing cars operating close to various limits, rather than the gentle giant long range diesel trucks they should be.
“To what purpose?”. Don, wouldn’t a progressive power-down allow the operator to burn away any accumulation of I-135? Tripping the reactor from full power and staying down for several hours would (allow accumulated I-135 to decay to the neutron poison Xe-135 and) remove its capacity to generate for several days.
Exactly – avoid iodine pit. Also a power operating plant (even just on house load supply) can get back on the line faster.
From a safety viewpoint I think this is much superior to just a knee-jerk trip of the reactor as a “precaution” to every upset condition.
Islanding with throttle back on power will increase thermal margins, not decrease them, as coolant flow per unit thermal output is increased. For PWR it would be likely on constant coolant flow so a low core dT and very high T/H margins. For a BWR you can eventually just trip the recirc pumps and run on natural circulation (need small feedwater supply of course).
Secondly, islanding allows another layer of defence in depth for both the heat sink and plant power supply. That is a huge deal in safety terms. The islanding/throttle back can be fully automatic and can be a non-safety system. If it fails (or if an upset condition is so severe as to exceed trip points) then the safety systems could still trip the reactor as a safety backup.
Costs of this option are typically low since systems are required for startup and shutdown anyway (like aux/startup FW/boilers and so on).
For LWRs, loss of plant power and loss of heat sink are proven to be much bigger risks than reactor criticality control. Can’t shut off decay heat… or run important emergency systems on fairy dust imaginary electrons.
I will not perpetually argue with you except to say that I have 40 years of nuclear reactor safety analysis including conventional design basis safety analysis and risk assessment at a nuclear utility and regulatory body. The trade-off in risk between reactor trip and loss of heat sink is minor and is not a reason for not tripping the reactor or separating from the grid on degraded/undervoltage. The risk of permanently damaging safety equipment due to undervoltage conditions, and the equipment being damaged and not being available should the transient progress, far outweighs what benefit of islanding might afford. The reaction is not knee-jerk. I have been talking about an automatic reactor trip at 58 Hz that occurs in minutes due to a degraded grid. You can not manually override it. These trips have been established on sound engineering basis by the NSSS vendors, and approved by the US NRC and its predecessor. You’ve nit-picked on some minor point of my comment, while missing the forest for the trees “…intermittent sources of electricity such as wind and solar AND just-in-time fuel sources such as natural gas are unreliable and/or not resilient…” compared to nuclear power plants. If you want to argue that point, fine.
Please avoid making judgmental comments that include phrases like “knee jerk trip of the reactor as a precaution to every upset condition.” That statement cannot be true, given the decade+ record of a US fleet-wide average of about 0.5-1.0 unplanned scrams/year.
There are plenty of upset conditions that do not initiate a shut down signal. Many have time delays that allow temporary conditions to be cleared without scram.
But when you have a very large, high value unit, you work hard to make sure that it doesn’t suffer permanent damage. Professionals like Don have played a huge role in creating nuclear energy’s incredible record of safety and reliability. Their work and success is what allows us to be ready to evolve into slightly less restrictive paradigm.
Let me describe a success story that, while not a station islanding situation per se, perhaps gets to the point that several commenters wish to make.
During the 2003 Northeast Blackout, power flows went back and forth between New England, New York, and New Brunswick. The grid frequency was oscillating. Interties with New York were cut by relay action. Most of New England and the Maritime Provinces islanded.
I next refer to publicly available testimony of the September 11, 2003 Energy & Technology Committee of the Connecticut State Legislature to describe the response of Millstone operators.
The first indication in the Millstone Units 2 and 3 control room occurred at 4:11 pm when numerous internal indicators and alarms began showing significant grid instability on the 345 kV system. Both units noted large swings in grid frequency and voltage. When the operators saw the grid frequency changing the way it did, they knew that they were one of the few generators left on the grid in the area. To be able to change grid frequency, the operators have to change turbine speed by changing the control valve position. These are situations that the control room operators train on and they responded extremely well in that critical first couple of minutes. As part of the plant’s stabilization, Units 2 and 3 operators took manual control and reduced power respectively to 98 and 83 percent. This reduction in power helped to match station generation with system load and thus helped to reduce grid oscillations all across New England.
The Chairman of the Department of Public Utilities Control, Donald Downes, states that “Millstone stood like a rock and was probably one of the main reasons that the system here survived as well as it did under the circumstances.”
So, yes, there is capability to affect the course of a Blackout. No automatic reactor trip setpoints were reached in New England that day. Seabrook stood. Setpoints were not reached to separate power to emergency buses from the grid. But run a thousand different scenarios in different parts of the country under different circumstances would there be the same outcome? Yes, it would be great if the next generation of nuclear plants were specifically designed and licensed for islanding.
So, my apologies to Roger Clifton and Cyril R. if I did not listen closely enough to your comments.
Full disclosure: I left Northeast Utilities and my affiliation with Millstone prior to the 2003 Blackout
Thanks Don. That’s an encouraging story.
You’re right we were talking past one another, actually seem to agree on just about anything really.
Islanding mode seems like a great capability to have as an add-on to more safety-related/limited systems, and to prevent such systems from being called on for the more common and expected transients. It won’t deal with all scenarios but will help a lot in the more common ones.
The newly proposed plants all seem to claim such capabilities – though it is of course a case of “so they may claim” at the moment.
How does NuScale or BWRX300 do in terms of islanding or self standing operations? Black start capabilities?
Gen IV seems well suited to a more robust and tolerant architecture. There’s typically a lot more margins on the fuel, with either high temperature capable fuel or liquid fuel design, the latter being my own speciality. Getting rid of fuel thermal and critical heat flux limits altogether is a big deal, I’ve found. The reactor core becomes just a glorified heat exchanger with internal heat generation in the liquid, a well understood though bit peculiar science. Such reactors are much more amenable to rapid load changes and frequent stop/starts, part of the reason for their focus during the aircraft reactor program… back when people were still pioneering and innovation wasn’t just miniturizing an old school PWR and call that novel.
I’m forward looking and often lament the paths not taken. The smartest and talented people I know are nuclear engineers, but we are stuck in a path dependency and saddled with baggage like a mule, stuck in a system that neither wants nor accomodates real change, and rewards process and red tape over real innovation and progress.
I’m sure you’ll agree, wind and solar offer no real alternative… with zero dispatchability, low capacity factors and high use of land and nonrenewable resources, these are not practical energy sources for a modern industrialized civilization. Increasing our dependence on the weather is backwards, running contrary to the most major of achievements of humanity the last millennia.
Cyril R. —- NuScale Power states that their SMR has cold start capability using a small on-site startup unit.
From what you describe, it reads like the islanded part of the grid was dealing with an over-frequency. What would be the operators’ procedure for dealing with an underfrequecy, which is generally more common event?
After New England and New Brunswick islanded during the 2003 Blackout, frequency oscillated between 60.4 and 59.6 Hz as measured at the Northfield Mountain pumped-storage facility in western Massachusetts. The Connecticut transmission voltage reached high levels: more than 385 kV on the 345 kV system, and 130 kV on the 115 kV systems. Some minutes later in the Blackout, when other equipment was lost and loads came back on, the voltage on the 115 kV system fell to approximately 100 to 105 kV. The Connecticut Valley Electric Exchange (CONVEX) eventually ordered manual load shedding to bring generation and load into balance.
In direct response to your question, I do not currently have access to nuclear power plant abnormal/emergency operating procedures for under-frequency/degraded grid events. Available options to plant operators are much more limited since the concern becomes primarily one of mitigation and protecting plant equipment.
My concern goes back to the original February 25th posting regarding catastrophic gas pipeline events in New England. The just-in-time aspect of natural gas generation, combined with the fact that gas now represents about 50% of the electricity production here, but at times as high as 70%, creates an inherently unstable situation against catastrophic gas pipeline events. That was the main message.
I am curious about the “islanding” comments. Are there new equipment designs that make it easier for Operators to battle Xe transients? BOL vs EOL significantly different and probably not trained on these days due to the lack of transients experienced now.
Last training I was aware of on Reactivity control you would never get a supervisor to buy in on what it took to manage a large Xe transient. Mid cycle to EOL, Go from 100 to 30% rapidly and maintain it? Put on your hat and spurs and giddy up.
Is life better now with new designs which can be retro fitted?
Thank you for that Don.
I agree with you about plant not realising their risks, but many don’t know about things until it happens. Fukishima and the tsunami is a case in point, though that is more a black swan event. I can relate a lot smaller one from my own experience. I had hooked up an atmospheric pressure transmitter at a remote site. This was to monitor a minor piece of equipment. The readings back was by a radio link. When they did an instrumentation on pumps upgrade a year later, they needed atmospheric pressure. They just used the signal from my transmitter without looking where it came from. Then the radio link went down. All hell broke loose (fortunately nothing tripped, until they got everything back under manual control. Now there are two hardwired transmitters with voting.
I agree that nuclear is caught in a mindset. For one thing, liquid air energy storage can increase the effective thermal efficiency of PWR and meet rapid demand response, in a way to anticipate ‘islanding’, scrams, load shedding, and the inconvenience of ramping instead of utilizing the reactor at maximum capacity factor. I have considered reviving the famous Oyster Creek reactor in NJ on this basis, which was shuttered for unwillingness to build a long overdue cooling tower. Your thoughts?
Exelon filed a permanent cessation of operations letter for Oyster Creek. This is a one-way door; it is not possible to “revive” an operating license for a plant that has filed a cessation letter. They could conceivably apply for a new license but considering the plant began commercial operation in 1969 there is no way such would be approved.
Exelon no longer owns the Oyster Creek site. It sold the site to Holtec. Holtec received the plant decommissioning fund and it using it to promptly remove the old plant.
It’s also preparing the site to be reused.
IMO, Holtec has some very smart people working on developing its various businesses.
Recent Comments from our Readers
I spent some time to listen to this podcast, and I still have two questions about module size and its…
Will I agree with your theory. Expensive designs are going to be expensive to build, even with practice. I would…
“And since we are seeing it in the West but not in the East (UAE, S Korea, Russia, China) is…
@Cyril R What was Tesla’s learning rate starting at the first Roadster? How much do you think that first unit…
A new engine or turbine product line doesn’t just cost triple a unit. That’d make it pointless. Yet this is…