* https://www.youtube.com/watch?v=bGySq7QBRiY
* https://www.youtube.com/watch?v=CfkoMUmmUCI
* https://www.youtube.com/watch?v=rbKf_TbI000
It mentions a few 'macro' issues with Vogtle: the first being the choice of using an AP-1000 design (Westinghouse Toshiba). Not because it is/was necessarily 'inherently' a bad design (the guest, James Krellenstein, describes himself as an "AP-100 fanboy"), but rather because it was (at the time) never built before. That, combined with the fact that the US had little/no institutional knowledge of building nuclear plants means that a whole lot of learning needed to be done.
Contrast this with another available (at the time) option, the ABWR, which GE Hitachi Nuclear Energy had building been for over a decade in Japan. Between ABWR build beginning and commercial operations it took about 4-5 years on average:
* https://en.wikipedia.org/wiki/List_of_commercial_nuclear_rea...
They had a ready-to-go supply chain, with plenty of experience.
Tied into the fact that the AP1000 was a new design, they decided to start building before 'detailed' drawings were only ~40% done. So they had the high-level view design done, but the drawings for machinists and pipe fitters needed were not ready, and so that added to delays, as they had to first create them, and then there was all sorts of back-and-forth to refine them. These issues were called out in the article:
> When Vogtle’s Units 3 and 4 were approved by Georgia regulators in 2009, the reactor model, known as an AP1000, had never before been built. (It was Westinghouse’s flagship model, combining massive generation capacity with new “passive safety” features, which allow reactors to remain cooled and safe without human intervention, external power, or emergency generators in the case of an accident.) It later emerged that the reactor’s developer, Westinghouse, had not even fully completed the design before starting construction, causing a significant share of the project’s costly setbacks. While that was bad news for Georgians, it could mean a smoother path ahead for future reactors.
> “In the course of building Vogtle,” Kozeracki told Grist, “we have now addressed three of the biggest challenges: the incomplete design, the immature supply chain, and the untrained workforce.”
Somewhere in the podcast they also go over some interesting 'bureaucratic minutiae': the plant was originally approved under a part of the regulations where everything is submitted 'all at once' and approved 'all at once', but any later changes to the design may take longer. As opposed another part where the design is submitted in parts, which may have been better as it would have allowed for gradual refine of the plants and submission when they were 'actually' finalized.
It also doesn't help that the US has so many 'one-off' nuclear plants, so that experience in one cannot be directly transferred between them. Contrast that with France, which has a fairly cookie-cutter fleet, or even (closer to home) the province of Ontario which has CANDUs everywhere (which were generally constructed in pairs and were often ordered in four- and eight-packs).
Some CANDUs are currently undergoing refurbishment, and the project is on-/under-budget and ahead of schedule. Interview with the fellow in charge of the refurb program:
> Gary Rose, VP New Nuclear Growth at Ontario Power Generation, walks us through the CANDU refurbishment program whose stunning successes have laid the foundation for deployment of the west’s 1st grid scale SMR, the GE Hitachi BWRX-300.
* https://www.youtube.com/watch?v=LHTtthahwjs
Of course now that the AP1000 has been built twice (Vogtle Units 3 and 4), plans exist and there's experience. The question is whether any new plants that are going to be built will use that experience, or choose yet another design.