Comment by TotesAThrowAway | Hacker News Reader

Comment by TotesAThrowAway | Hacker News Reader

Parent Full thread

TotesAThrowAway

Friend of BillinghamJ here.

We do build mock facilities and we have recently started working on the Fukushima Daiichi inspections, and there is only so much I can actually say but I'll share what I can. I'll explain what caused the failure in this part and then I'll move onto the fun stuff in the next post (robots!).

In the past, the mock facilities we have made in the past were a quarter of a full scale reactor, rather than a quarter scale reactor, if that makes any sense at all.

Anyway back to Japan. I'm assuming people have a basic understanding of how fission reactors work (Boiling Water Reactors if you are interested in doing further reading).

To break the situation down, the cooling failed (believe it or not, diesel generators don't work too well on water!) on reactors 1, 2 and 3, causing a complete meltdown of the fuel rods. When cooling failed, all of the cooling water was turned into steam, which in turn reacted with the released radioactive isotopes creating hydrogen. I'm assuming people know mixing hydrogen with oxygen is basically a recipe for an explosion, and that is important for what happened next. They tried to vent the gases out to the atmosphere to prevent the pressure vessel from exploding, but the hydrogen went the wrong way and caused reactors 1 through to 3 to explode in various places. I can't remember correctly (I think its reactor 3?) but the explosion happened within the pressure vessel which is why there was a large contamination breach. Because of this complete loss of control of the reactors, and the meltdown currently happening, they flooded the whole system with sea water and pumped as much out into storage as they can, but they lose a lot of it out to the sea, hence the Americans whining about the radiation in the Pacific. Now, the reactors are stable (ish) and they are continuing to pump water in to stop them going critical again.

I'll move onto the robots when I get home in part two, but right now I need to go have an argument with a lawn mower as my hair is getting unruly. The video linked below explains more about the actual failure of the reactor:- https://www.youtube.com/watch?v=JMaEjEWL6PU

Brb.

I should probably also mention there's 6 reactors on site, 4, 5 and 6 were relatively unharmed by the tsunami and 5 and 6 could actually be turned back on. However for obvious reasons they have been decommissioned and wont be turned on ever again.

Are there real reasons not to turn them back on once the cleanup is complete? Or is it just the matter of politics/popular perception?

AFAIK, fukushima was actually a very small "disaster", with negligigle human life loss (if I recall correctly, no life loss, except due to the evacuation panic)

There were some design flaws that were exposed with this incident.

Would you, in conscience, turn back on a series of nuclear reactors that had their emergency backup generators located in an area that can easily be flooded?

There are a lot of other human/social aspects involved as well.

It's better to apply the lessons learned from Fukushima to a safer type of reactor, built better, designed safer, and operated with a different set of procedures and by a different company (not TEPCO, for example).

^that.

Does "...in the past ... quarter of a full scale reactor, rather than a quarter scale reactor" mean that

1) in the past the mocks was just a subset portion of the reactor (ie the quarter that is relevant to operations)

Hopefully I didn't mangle your words, but this is how I understood your phrasing.

The quarter rig (imagine cutting a 90 degree cross section out of the reactor and plonking it into a rig hall) we made was for a remote solution we did for EDF for some of the nuclear fleet in the UK. As much as it would have been fun to make a full size to scale rig, it would have cost a small dortune and we are a business afterall and profit is our end goal. So we only made enough of it to be able to train the operators to deploy the kit on site.

In terms of Fukushima, the manipulator we are developing will eventually have a testing rig built for it, however we have only just been involved on the project in the last month and as such haven't reached a stage where we need to build the rig yet. With the Japanese government funding though I imagine it will be a similar job with a scale rig.

Ah thanks for the clarification

PART 2

So, ROVs.

We probably aren't going to use a ROV for our solution as it doesn't suit and they're a pain in the arse quite frankly.

To start, I'll address the issue with the 'wires' failing. An American company built a ROV that was heavily shielded and was driven via an umbilical. Wireless is hard to use and autonomy is too hard to use as the reactor conditions are unknown. The ROV was a good design and could survive the radiation for a reasonable amount of time, however for whatever reason they used PVC wire sheathes that break down under heavy radiation, and as a result the wires touched and shorted the electronics out, rendering the ROV useless and 'dead'.

In terms of why it takes so long and why we can't use an off the shelf version, basically radiation is a bitch. At the base of the reactor vessel, just above the corium, the radiation output is estimated to be around 3000 microsieverts per hour, that translates roughly to a human life expectancy of around 6 seconds, give or take. This amount of radiation causes electronics to fail (transistors commonly), and materials to break down. The breaking down of materials caused the American ROV to die, and another example would be that it can causes greases to harden, which stops motors working.

Reactors aren't big spacious areas either, so it's not like we can just deploy a lead (lead weighs a metric shit'tonne) shielded tank to have a look, it's just too big and cumbersome. We decided against using a ROV as it had to be 30kg or less, which is absolutely nothing once you bring in drilling packages and the likes.

Also quite often you will be deploying through a hole between the size of your fist up to just smaller than the diameter of your head, so that restricts you hugely as well.

You also have material compatability. If you get something stuck you have to be sure that it won't react and cause the reactor to become critical again, which could happen in one of the reactors (can't go into more detail sorry).

One last major consideration as too why it takes so long to build and test a ROV to suit. The reactors are under immense thermal stress, and metal likes to bend and warp when it's heating/cooling. You have to build your solution around the worst case scenario. An example would be we went into a boiler tube trying to plug a 1 inch hole from 18m above it using a manipulator arm. That's already hard on its own, but then we discovered the originally 7mm gap we were aiming for was actually now as small as 3.5mm. Trying to develop ROVs and remote solutions is really not easy, the best way I can put it is that this line of work is an art, not a science. That's why it will probably take the best part of a century I reckon to fix this problem.

   That's why it will probably take the best part of a century I reckon to fix 
   this problem.

Just wow.

But a question -- you say you won't use a ROV. But you say wireless is hard and autonomy is too hard. So that leaves what?

And thanks for taking the time to share.

The alternative and what is most commonly used for control, in the nuclear industry, is a very long 'umbilical' (wire with loads of cores). A lot of nuclear providers are against people using wireless as if you lose connectivity and it gets stuck you are in trouble.

Also we are going to use a manipulator arm more than likely. Manipulator arms are cheaper and more widely used in our industry.

So its like a very long colonoscope that has camera on its end and can be manipulated to turn around a corner? https://duckduckgo.com/?q=colonoscope&iax=1&ia=images

More like this minus the electronics:- http://www.lasersnake.co.uk/images/galleries/2b8q6hcombined-...

Your idea would be perfect if we were just doing an inspection, however we have other work we will need to do down there which will require various packages so it has to be somewhat bulkier.

I assume you meant 3000 milisieverts (3 sieverts) or even 3000 sieverts.

3000 microsieverts (3 milisieverts) is most definitely not fatal, as it is similar to the yearly exposure.

> and the meltdown currently happening

Sorry?

That should read 'was happening' sorry. I was in a rush and didn't proof read properly.

Ok, thank you for the update. You had me worried for a second there, being so close to the proverbial fire that was a scary thing to read.

Reply on news.ycombinator.com