Hacker News new | comments | ask | show | jobs | submit login

Its a nice write up but I don't believe it. It focusses to much on capacity factor, which is an interesting easy measure. But not nearly as important as one might expect. The key measure is actually dispatch-ability. i.e. how quick can you turn it on and off again. In markets with energy trading this is key for profitability.

The CSP plants with molten salt storage do really well on the market because they can turn on and off as fast as gas peaker plants. (For as long as they have storage of course). They will off course have a low capacity utilisation as they are peakers in all but name (aimed at daily evening peak load).

Nuclear has a high capacity factor, not because they are base load. But because once they are on they are on. And being on they will sell all they produce for as long as possible. Driving all other producers of the market. i.e. in the current market with 20% nuclear in the US they are rarely forced off the grid because demand is always greater than what they can supply. Once we hit 50% peak load supply by nuclear this high capacity factor will drop. For the simple fact that while a Nuclear power plant can produce, no one is buying.

So for a whole system using just nuclear capacity factor will be around 60% not 90% due to market realities on the supply side.

Wind is at 40% capacity factor, not because of technical limits (one could build 90% capacity factor windmills if you where crazy, you just derate the generator but keep the same blades, e.g. put an 1mw generator in a turbine designed for 3mw and you will get your 1mw most of the time, but never capture the 3mw you could some of the time)

Wind power plants are repowered at the moment after 20 years, not due to limitations of the tech. But because turbines made today are so much better than those of 20 years ago. Most of these old windmills are actually resold on the second market, and its quite difficult to get enough 500kw mills at a good price. Also upto 2/3 of the value of a wind mill is in its steel tower, value that exists in 20 years as much as it does today (little rust and easy to recycle)

He also takes 2.5mw turbines from GE as an example but those are rather small these days. They are heading into the 8mw territory today and we will see 12MW plus mills in the coming years.

Critically, looking at it on an economic perspective. Once you start building a wind farm, it takes 18 days to build on site a turbine, from base to grid connection. That means you know if its going to work in a month. This leads to easy financing compared to nuclear, where the average best case build time is 4 years, which often descends into decades of building. Financing wise that is a completely different game.

Also wind and solar power are added at 100Mw order of magnitude to the grid in single year projects. This is better for project financing and risk management. Panel in project phase1 bad buy a different one for phase2, same for wind turbines.

All in all 500, 1000APs could not be build in 20 years even if we lined them up one after the other. The infra in forging and assembly is just not there, not even if we went all in on it.

Wind and solar can be build at that scale. Because its distributed manufacturing. Best case nuclear numbers barely meet offshore wind numbers today. Wind which has manufacturing on scale benefits that nuclear does not, will reduce costs much faster than nuclear.

PV efficiency will go up, e.g. look at the first solar efficiency roadmap.

If you are buying a power plant, then you will quickly see that Nuclear is not a cheap or easy to finance option. A 1 billion windfarm not working out, you cancel midway in construction with 500 million of pain. A 4 billion nuclear power plant not working out, you are decade down stream with a 10 billion bill. 4 billion is a sum few electricity companies can gather, 500 million borrowing for 18 days during construction after which a wind farm is sold on a power purchase agreement made is much easier for many more electricity companies. Think: if you owe 1 million the bank owns you, if you owe 100 million you own the bank...

Nuclear should work, and if it did it would sell like cup cakes. The problem really is that these cup cakes cost as much as a trip to the moon... You might wonder why I am bringing up financing so often. In a energy market like the EU or US where there is enough current generation this is key for a plant/farm to be build or not. Often you can only finance if you can drive an other generator off the market, by being cheaper.

Solar and wind benefit from the law of large numbers, nuclear does not. If a nuclear plant goes into maintenance you lose 1gw. If a turbine goes down you lose 10mw, a substation blows up you lose 100mw. A nuclear power plant transformer fire, and you lose 1gw in a minute, is a major grid issue. Design issue in your plant, 4gw offline for regulatory reasons, major grid issue.

Nuclear power plants are to big and to expensive, small modular nuclear could work. But no serious market players are in this field.

Energy storage is currently still to expensive, but even there prices are dropping utility scale power storage is not just pumped storage. (Funnily enough in the UK this was build because of the nuclear investment). Hydrogen, heat, batteries, compressed air, fly wheels are all being investigated and each has deployments in the market. Currently still rather specialised depending on local market needs, but getting closer to taking on gas peakers.

This is a robust and cogent analysis of the case for nuclear. Advocates, often motivated by global warming concern, do not look at the "grid that will be", and rather look at "the grid today" or "the grid of the past". That nuclear gets more expensive (lower capacity factor) the more of it there is should really give people pause.

Thanks for such a good note.

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact