Given that not everyone here knows how to perform an elementary calculation of the risk due to flying through volcanic ash, I put a crude one up at
http://www.abnormaldistribution.org/...cial-aviation/
It is part of what I presented in my Risk course this morning. I thought it would be cool to illustrate the things we had been talking about with a topical example.
PBL
The calculation may be elementary, but (as ever) the devil is in the inputs.
On your item 4, you seem to have taken pCatastrophe = pIFSD - this isn't correct. Ok, you mention dead-stick landing, and lets say we accept that even with a Sully/Burkill you are still looking at a hull loss so catastrophe impact applies. Still it's not right though:
Firstly, you need
all engines shutdown - and the documented incidents show asymmetric damage. Correlated, but not 100%.
Secondly, you need a failure to relight. We know that descending to clean air will cool, solidify and shatter the glass, typically allowing a relight (every time so far, I believe).
So, pCatastrophe = pIFSD * (correlationIFSD ^ no.engines) * (1 - pRelight). Couple of orders of magnitude there ?
Then on item (2) - you are probably going to have the increased inspection costs
anyway as a regulatory cost of flying. Adjust your business model - or go bust.
On (3), your impact is probably going to vary much more widely. Engines have a limited life anyway, in some cases you will have effectively just lost some remaining hours use. Notably, the NASA report states that the most damaged engine was coming up for major work anyway - and it isn't clear if the 3m cost included that engine, or the work that was going to be done anyway. Conversely you could hit Sunfish's problem and have the cost of a fleet grounded due to lack of engine repair capacity. I think the impact range on (3) is more like 10^5 - 10^8 (Sunfish scenario, or bust).
But that's all minor nitpicking really. 
Then we get to
the cost of not flying.
You've used
gross ticket revenue as the loss! Seriously, did no one in your audience call you on that ? [ Probably a good job I'm not on your course, I don't think I'd be a good quiet student these days
]
In the EU, at least, the cost of
not flying is very little to do with ticket price. Much as some airline bosses would
love to say "flight cancelled, here's your 50p back (less £5 admin charge)",
they can't.
The main cost of not flying is the cost of getting your pax back.
Options, roughly:
1. your pax agrees to take ticket refund
2. you pay for them to get on another airline
3. you put them on one of your flights later (+hotels/meals etc. for 50% who are on return leg)
Now, (1) you'd love, but you can't choose, (2) ain't going to happen with everyone grounded, so you are stuck with (3). Cost of putting pax on your own flights - zero (more or less)
if you use your spare capacity. Unfortunately, if you are running daily at say 80% load factor, a 1 day shutdown means a four day backlog (5 days total, 3 days ave pax. delay). A six day shutdown would be a 24day backlog, 30 days total, 18 days ave pax delay. At, say, 100 per day per pax for hotels (who aren't exactly going to be handing out discounts...) etc., that's 1800 per 50% of pax, or approx 150k
(+ some for the outward pax) for your example flight. Reality is that it may be a bit less than that because you will give seats to these pax that you would normally have sold - but those will be last minute seats that would have gone for premium prices, so you still lose several times the average ticket price paid by the stranded pax. May be cheaper than the hotels though...
And then of course if the shutdown is holiday season when your load factors are much higher than average...
Feed those numbers (150k) into your risk equation and (by day five or six) you need pMajorDamage around 1 in 5, and pIFSD about 1 in 20 (say pRelight is 0.95) to be worth not flying. Now, if you've done a few test flights looking for ash, and seen no sign of a (1) event let alone a (2) or more, I think you'd be legitimately screaming for airspace to be opened.
Same equation, even same risk events and impacts, ...
very different results. All in the assumptions and inputs. [so who's right ? Well, I am not an aviation safety expert like PBL - but that isn't where we differ on our figures. Reality is that neither of us is an accountant for an airline. ]