Failure Combinatorics

[lamont]

Given a team of 3 with 6 functioning posts to begin with, what are the likely failure scenarios? An unfixable post or side of a manifold will take out a single post. An OOG will take out both posts on a diver. The team must have 3 posts to breathe off of (well, maybe not, but lets keep this tech1-level and simple).

So what happens? Lets start with a very simple one beginning with an OOG:

1. OOG (2 posts down, 4 left, one diver out of it completely).
2. Valve shutdown (this is the only possibility left, which leaves 3 functioning posts).

Now, which diver is going to get the valve drill in a GUE tech1 course? Giving it to the diver not involved in the OOG is not terribly interesting. Much more likely (from a training perspective) that the donating diver will get a valve failure and need to switch the OOG diver off of them and manage to shut down the correct post without shutting off anyone's air. If an OOG is the first failure, the subsequent valve drill is almost entirely predictable. This failure scenario can repeat step #2 for as long as the post is fixable with the OOG victim needing to swap each time. Once the valve failure is unfixable the failures are complete.

Next one, starting with a valve shutdown:

1. right side unfixable (1 post down, 5 left, one diver cannot donate)
2. OOG on different diver (3 posts down, which leaves 3 functioning posts)

Since the diver that cannot donate is in the middle, the OOG diver needs to bypass that diver to get gas. Once a valve failure occurs it should be expected that an OOG is going to happen to another diver. This scenario is slightly forgiving in that if the OOG diver forgets and goes to the diver with the right post failure that diver may reflexively attempt to donate the backup and notice the problem immediately.

Next one:

1. left side unfixable (1 post down, 5 left, one diver cannot donate)
2. OOG on different diver (3 posts down, which leaves 3 functioning posts)

Since the diver that cannot donate is in the middle, the OOG diver needs to bypass that diver to get gas. Again it should be expected that an OOG will probably occur to one of the fully functioning divers. This scenario is less forgiving in that the middle diver has a donatable long hose, but no backup. Since tech1 instructors turn valves back on before doing additional failures this is a scenario where its likely that the middle diver will get a note handed to them saying "you've been breathing off a failed post for the last X minutes".

Next one:

1. either side unfixable (1 post down, 5 left)
2. OOG to affected diver (2 posts down, 4 left)
3. valve shutdown (probably on donating diver, 3 left)

This one is actually somewhat more realistic. So there's a right or left manifold unfixable (tank o-ring), followed by that diver going OOG, followed by someone else needing to fix a post. It is actually the first scenario mentioned above, but you take a valve drill and an OOG to get there. It should be easily handled by the team since the OOG diver is now in the middle and either side is the correct choice so the only difficulty becomes swapping the OOG diver around during the last valve shutdown.

1. either side unfixable (1 post down, 5 left)
2. the other side valve drill (2 posts down, 4 left)

Here's a test of the guy with the valve failure to see if they remember that if they're down one post they need to go OOG on someone. If its unfixable then the diver is really OOG and if this was the first failure you can expect a valve drill on the donating diver (per the first scenario) or if there was already a valve failure then you're done after this one.

1. either side unfixable (1 post down, 5 left, 2 donatable divers)
2. other diver either side unfixable (2 posts down, 4 left, 1 donable diver)
3. OOG on one of the victims, valve shutdown drill on one of the victims.

This is probably the worst test I can think of for remembering what has failed on the team. If either diver with an unfixable side fails to go OOG during a valve shutdown then they've failed to mange team resources. If both valve failures are left side unfixables then its easy for an OOG diver to wind up taking the long hose of someone who cannot donate. This ends when one of the divers with the unfixable sides gets another unfixable side, or they go OOA or the third diver gets another unfixable side.

In general all the scenarios are over after one diver is OOG (either OOG or unfixable-OOG or unfixable-unfixable) and another diver is unfixable, except for one permutation of the last scenario where all three divers have unfixable sides. But typically you wind up with a diver with no posts, breathing off a diver with all posts, with the other diver having a failed post. Then you're done.

Did I miss any combinations? I think anything remaining will reduce the team to 2 working regulators and at that point you are buddy breathing, or using the LP inflator as a regulator, or feathering a valve, or hopping on your deco gas and hoping you get lucky...

 

[lamont]

And as a followup...

People say that you need to "learn to think better" underwater, and I sort of disagree. There's only a very finite amount of fundamental failures that can occur. It is simply important to remember who is down a post and who can donate (of the independent divers). The divers who are OOG should be obvious by the long hose hanging out of their mouth. If you run through the failures and practice them mentally it should be possible to react correctly, but without thinking as the problems occur.

Stop thinking. You know if you've got one OOG diver and two functioning ones that all the team gets after that which is easily done is a valve drill. You know that in a class its more interesting to have it be the donator. Expect it.

You know that after a valve shutdown you sandwich the victim. Then either side can have an OOG and need to run to the diver farthest from them. If you visualize that enough it should become reflexive that after you sandwich a diver (for gas/valve reasons rather than just a light failure) you run to the other guy.

The memory part is just remembering who has valve failures. If you find yourself running for gas to a person with a valve failure you need to stop yourself and find your other buddy. If you see a buddy with a valve failure doing another shutdown you need to donate to them (assuming you can donate, if not get your other buddy to donate). If you have a valve failure you cannot donate and cannot do a valve shutdown without getting gas. If you are breathing off someone's long hose and that person gets bubbles you need to switch. If you have donated and you get bubbles you need to shutdown that post without cutting off anyone's air (act accordingly to free up that post).

It doesn't really take a whole lot of thought, IMO. And visualization of the possibilities beforehand should cut down on the amount of thinking required.

 

[TSandM]

This works, because the number of possible combinations is small. But I will warn anyone reading this that this kind of thinking, trying to outsmart the class, causes a great deal of trouble when you start mixing in other failures and issues. I've made precisely the wrong decision several times because I was overthinking and trying to solve the problem I thought I was being given (because I expected that one) rather than the actual problem I was facing.

But the basic principle that the guy who's been sandwiched is probably not in a position to help anybody else is a nice, solid, ought-to-be-gut-level insight.

 

[lamont]

This works, because the number of possible combinations is small. But I will warn anyone reading this that this kind of thinking, trying to outsmart the class, causes a great deal of trouble when you start mixing in other failures and issues. I've made precisely the wrong decision several times because I was overthinking and trying to solve the problem I thought I was being given (because I expected that one) rather than the actual problem I was facing.

But the basic principle that the guy who's been sandwiched is probably not in a position to help anybody else is a nice, solid, ought-to-be-gut-level insight.

That is sort of the idea though. Stop thinking.

If you're in the middle of the class trying to figure out how to solve a theoretical issue you've invented in your head, that indicates that you're not sufficiently mentally prepared and you're inventing issues for yourself. Be prepared for what happens and then take of whatever is directly in front of you without thinking. Do all your thinking beforehand.

What do you do if you're shutting down your right post and become convinced that you've been actually given a left post? Nothing. Just shut down the right post. If bubbles keep coming, shut down the isolator. Call in your buddy. If you "messed up" then you just gave your buddy a left-thought-it-was-right to debug instead of a left. Overthinking the problem can lead to bizzare reactions to a valve issue. The sequence for the affected diver is just to signal, listen, guess, shutdown, listen, maybe isolate, call in buddies. That's it. No thinking involved other than the first guess and there's no punishment for getting that wrong. It isn't your job to guess the correct post.

And similarly for multiple failures, without overthinking and attempting to game the system, there are some very obvious reactions that you should have. You need to be tracking who has failed posts. That isn't gaming the system, that is just real life. In general the person with a failed post is going to be in the middle which makes it easy. Act accordingly and go for the diver on the other side in an OOA. This isn't actually gaming or overthinking, its also just real life -- that's the diver with the donatable gas over there, the diver in the middle can't.

 

[rjack321]

I agree that there's a finite number of problems. In class the problems "may" be presented in a manner which leads to the need to remember who is down a post and that you need to hopscotch over them. Certainly a good idea to always be aware of what equipment is working and what isn't.

But that leapfrogging combination is quite remote compared to the likelihood that the post failure drained a bunch of gas and that weak link in the middle is more likely than the front or the back to go OOA.

Class failures tend for force the team to work like one (or not). These types of problem cascades are rather far fetched compared to one person having a problem then hoovering through what little gas remains. So #2 goes from being a weak link with a post failure to being OOG. Weak points beget bigger problems. Otherwise why put them in the middle eh?

 

[lamont]

I agree that there's a finite number of problems. In class the problems "may" be presented in a manner which leads to the need to remember who is down a post and that you need to hopscotch over them. Certainly a good idea to always be aware of what equipment is working and what isn't.

But that leapfrogging combination is quite remote compared to the likelihood that the post failure drained a bunch of gas and that weak link in the middle is more likely than the front or the back to go OOA.

Class failures tend for force the team to work like one (or not). These types of problem cascades are rather far fetched compared to one person having a problem then hoovering through what little gas remains. So #2 goes from being a weak link with a post failure to being OOG. Weak points beget bigger problems. Otherwise why put them in the middle eh?

Yeah, the leapfrogging situation is really more of a class situation, to test awareness. As i noted the unfixable valve failure followed by the OOG in the middle diver is the most realistic one to deal with.

Although its not totally unrealistic. More likely way to see it in real life is that after an unfixable failure you thumb the dive and on a gas switch someone else goes OOG due to not having access to gas rather than not having gas. I've seen that before with a diver switching to a deco regulator which failed to deliver gas and their impulse was to OOG a buddy rather than to go onto their necklace. Now if the diver throwing the OOG does it to the buddy with an unfixable valve failure you've got a cascading set of mistakes which could end badly.

So, for defense in depth, both practice going OOG off of a bottle onto your backgas and practice going OOG onto the buddy who doesn't have the valve failure.

 

[SparticleBrane]

I can't say I ever thought I'd read a post in here advocating that someone doesn't think...

 

[KMD]

whoa... someone just went through T1.

 

[KMD]

So the whole manuscript of thinking could be boiled down to:

You hear bubbles:
1) If there is a diver on your long hose, switch him off
2) trouble shoot like normal


You are out of gas
1) go to the diver who has donatable gas

Communication:
Make sure to communicate all failures to entire team.

 

[lamont]

well, there's the little details, such as sandwiching the diver with the worst problem, etc...

generally, though, i agree, its very simple. just fix the problem right in front of you (correctly) and move on. thinking about stuff should be fairly minimal...

( is 'minimal thinking' better than 'dont think'? )