4 ms·

Your first statement is clearly wrong, or at least you should define what you call "lot of latency". Third generation nuclear reactors in France do load follow

Your first statement is clearly wrong, or at least you should define what you call "lot of latency". Third generation nuclear reactors in France do load following all year round, with great success: https://en.wikipedia.org/wiki/Load_following_power_plant#Pressurized_water_reactors https://en.wikipedia.org/wiki/Load_following_power_plant#Pre...

rndgermandude 5y ago
Load-following in nuclear works to a very limited degree, sure, but >These reactors have the capability to regularly vary their output between 30–100% of rated power, to maneuver power up or down by 2–5%/minute during load following activities And even the BWR can maneuver only at 10% per minute (60-100%). So first, you cannot just shut them (nearly) off when you don't need them. You will have to run them at 30% at least essentially all the time. Not great if your goal is to use solar/wind as the primary mode of generation. Then 2% to 5% per minute in changes is high latency compared to alternatives. If you need 10%-20% more NOW because there e.g. was an unscheduled break in a football (soccer) game and everybody got up, turned on the lights and the coffee maker and whatever then you're out of luck because the latency is too big to cover those additional 20% in the time you need them. 20% would take up to 10 minutes for nuclear to adjust, 2 minutes at least, realistically something like 5 minutes. And by the time it adjusts your grid will have broken down and everybody will be sitting in the dark. Load-following with nuclear plants only works well enough if you have a pretty good idea of when the load is going to change in advance. Which in many cases historically you did, and you had non-nuclear fallbacks to compensate for when you were a little or even a lot off. You had a pretty good idea because you could estimate both the demand as it progresses during a day and the generation too. E.g. you knew what a "normal work day in November with this weather forecast (cloudy average temp 8°C, here is how temps will likely develop during the day), sunrise 07:07, sunset 16:22" looks like with enough historical detail from similar past days to plan for it. With solar and wind, you still know that about the demand side, but you can be a lot less sure about the generation side. That's why you usually use gas (and historically coal) "peakers"[1]. They were and are in use for when your estimates were off, even if your base load comes from the best load-following nuclear. They can be maneuvered a lot faster with demand, and can be (almost) shut down when not needed. And what you need in a wind/solar world in particular is true peakers not base load plants with limited load following capabilities. Does this clarify what I meant? [1] https://en.wikipedia.org/wiki/Peaking_power_plant https://en.wikipedia.org/wiki/Peaking_power_plant

rndgermandude 5y ago

Load-following in nuclear works to a very limited degree, sure, but >These reactors have the capability to regularly vary their output between 30–100% of rated power, to maneuver power up or down by 2–5%/minute during load following activities And even the BWR can maneuver only at 10% per minute (60-100%). So first, you cannot just shut them (nearly) off when you don't need them. You will have to run them at 30% at least essentially all the time. Not great if your goal is to use solar/wind as the primary mode of generation. Then 2% to 5% per minute in changes is high latency compared to alternatives. If you need 10%-20% more NOW because there e.g. was an unscheduled break in a football (soccer) game and everybody got up, turned on the lights and the coffee maker and whatever then you're out of luck because the latency is too big to cover those additional 20% in the time you need them. 20% would take up to 10 minutes for nuclear to adjust, 2 minutes at least, realistically something like 5 minutes. And by the time it adjusts your grid will have broken down and everybody will be sitting in the dark. Load-following with nuclear plants only works well enough if you have a pretty good idea of when the load is going to change in advance. Which in many cases historically you did, and you had non-nuclear fallbacks to compensate for when you were a little or even a lot off. You had a pretty good idea because you could estimate both the demand as it progresses during a day and the generation too. E.g. you knew what a "normal work day in November with this weather forecast (cloudy average temp 8°C, here is how temps will likely develop during the day), sunrise 07:07, sunset 16:22" looks like with enough historical detail from similar past days to plan for it. With solar and wind, you still know that about the demand side, but you can be a lot less sure about the generation side. That's why you usually use gas (and historically coal) "peakers"[1]. They were and are in use for when your estimates were off, even if your base load comes from the best load-following nuclear. They can be maneuvered a lot faster with demand, and can be (almost) shut down when not needed. And what you need in a wind/solar world in particular is true peakers not base load plants with limited load following capabilities. Does this clarify what I meant? [1] https://en.wikipedia.org/wiki/Peaking_power_plant https://en.wikipedia.org/wiki/Peaking_power_plant