Potential Safety Improvements in Rebreather Design

[John C. Ratliff]

Odd,... Since it has been shown that nearly all rebreather fatalities have been shown to be the result of operator error, not machine design, whether, because of being a inexperienced or careless or such. Good training & having the discipline to follow accepted checklists would go a massive way in minimizing risks. For that rebreather to have been in the shape it was in for the accident to occur, there just about had to be red flags that were not noticed or ignored when/ if the pre- dive checks were done.

I've read about as far as I can without adding a comment (I'm about half way through this discussion). I have worked in product safety in the semiconductor industry, and worked as EHS Manager in the chemical industry (Process Safety Management). I also am a diver since 1959, although I have purposely stayed completely away from rebreathers. I dive mostly solo, in rivers, and do not feel that rebreather technology is where I want it at my age (almost 69, later this month). I do have some information to pass along which may help this discussion.

The above quote relies on training of the diver to ensure safety. Well, in the Hierarchy of Controls (what we in professional safety use to try to mitigate hazards), this is what we call an "Administrative Control." The hierarchy goes like this:

Elimination > Substitution > Engineering Controls > Administrative Controls > Personal Protective Equipment

Administrative controls (training) do work, but are subject to breakdowns due to a number of potential problems, such as fatigue, peer pressure, inattentiveness (distraction), sleep deprivation, "hurry up" sickness, etc. In short, while vital, training can only do so much. We see human errors occurring during training and for about 6 months after training, until the person is really attuned to what (s)he is doing. Then we also see human errors increase after about 6 years of familiarity with the job/task/situation.

Rebreathers are what we would term "safety critical" items, and because of this, single-point failures are normally engineered out of the system. This is a part of the profession of Reliability Engineering. This can be done by providing redundancy, sensors to sense an impending failure, or engineering so that the system cannot be improperly installed so as to fail. There are a number of tools to determine whether controls are necessary, and if so the types of controls. The analysis can be a Failure Modes and Effects Analysis, Fault Tree Analysis, a What-If Analysis, and a number of other techniques.

The point here is that if a rebreather does have a single-point failure potential, it is not safe as a life safety system.

SeaRat
John C. Ratliff, CSP, CIH, MSPH
CSP = Certified Safety Professional; CIH = Certified Industrial Hygienist.

---------- Post added December 2nd, 2014 at 11:20 PM ----------

The beauty of the checklist that doesn't work is that it can be easily modified by the manufacturer so that it does work. Look, this isn't teaching 17 year old sailors to run a reactor plant, this is teaching adults who presumably want to be there to dive a bit of scuba gear. So if what you infer by your reading of the coroners report is that the loop was assembled backwards, the counterlungs were put on the wrong side, and the unit still passed positive and negative, a concept I find a stretch, but then, I'm not certified on a Hollis, nor will I be, then a simple checklist modification may be in order, and Hollis should be able to do that by sending out a notice to all P2 users. Namely, step 6A would further identify the inhale counterlung with a description of what it looks like. Tammy says that the counterlungs don't look the same at all.

It's fine that you don't use a checklist. I've had a number of conversations with Leon regarding checklists, and our common ground is that they work if properly done, but most folks get too smart for them very quickly. Sadly, I had a very good friend and crewmember die on his meg. A checklist would have saved him, because he couldn't possibly have passed the negative pressure test with his particular problem. He installed his O2 selenoid incorrectly.(emphasis added, jcr)

This is an example of a single-point failure of a safety critical item leading to a fatality.

SeaRat

 

[Douglas Ward]

I believe what someone in this thread is advocating for is a rebreather assembly method/design that is foolproof. Such that the scenario goes like this: A completely disassembled rebreather is placed in front of a completely untrained diver. This diver should be able to completely assemble the unit and dive it without any instructions or training and live, because every single assembly and setup point is keyed to fit only its matching mating point. They would be able to assemble the rebreather simply by process of elimination by trying each connection point together until they found one that fit.

If this were the case training would only be necessary for the Instructor to give the diver points on buoyancy control and how to clean the rebreather.

Is this the end game?

I think in any sport such as diving, where there is always an inherent risk, the participant should understand their own limitations, understand and be completely confident in their training. If you're fresh out of training and aren't quite sure before you jump in on a dive, YOU DON'T DIVE. You bag the dive and take more time to figure out what you're doing, review your training, look at the manual again or call your instructor. During the course of training and at completion you have to sign indicating that you are fully confident with your training and completely understand the operation and how to dive your rebreather to the limits that you have been trained.

We are responsible for ourselves and no one else.

The diver died in this thread due to her failure to follow the checklist "CORRECTLY" and "LOOK" at her rebreather and make sure she assembled it correctly.

I don't see it as a possible resolution that a foolproof, anyone can assemble rebreather without training, is a possibility. If it did become a reality, then the argument goes to the rebreather must automatically control the dive and setpoint without any diver input. After that is accomplished then what's left? Buoyancy control? The rebreather must constantly control the loop volume and wing inflation to prevent uncontrolled descents/ascents??

Divers need to be well trained and take responsibility for their own actions. Bottom Line.

There are countless hours of logged dives on the PRISM2. One fatality due to one diver grossly neglecting the provided training does not indicate a catastrophic flaw in design. Does it bring to light a engineering point that could be improved? ABSOLUTELY WITH OUT A DOUBT!!!

 

[tomfcrist]

I honestly can't believe certain people are still beating the same dead horse. It is impossible to design anything that is completely impossible to assemble incorrectly. It's not going to happen...ever. Not a rebreather, not an OC kit, not a firearm and hell, I'm a sailor...I've seen people that could assemble a damn hammer incorrectly. The CE standard that you keep referring to is unrealistic, kind of like the unrealistic student instructor ratios in most courses in ANY agency.

I challenge you gian to design and build anything that has more than 1 component...I will personally assemble it incorrectly.

 

[Storker]

I challenge you gian to design and build anything that has more than 1 component...I will personally assemble it incorrectly.

Corollaries to Murphy's law:

• It is impossible to make anything foolproof because fools are so ingenious.
• Nothing is foolproof to a sufficiently capable fool.
• Make something idiot-proof, and they will build a better idiot.

OTOH, don't rely on operator training and guidelines, because then even the least ingenious idiots will screw it up. Or, basically, what John C. Ratliff said.

 

[tomfcrist]

It is really no big deal for Hollis to improve on the design and have left and right hand thread to reduce the risk of user error and a repeat fatality.

Maybe something for Hollis Prism 3.

To deny that there is something to be learned from this fatality and something to be improved in the equipment and the training is reckless.

Regardless, even if they do design the 3 in such a way, when the next genius assembles that unit improperly, you will be right back here spewing the same BS all over again. The reality is that based off the strict adherence you seem to think the rebreather MFG's should adhere to in the CE standard, you will never be satisfied with ANY rebreather.

 

[Brad_Horn]

I believe what someone in this thread is advocating for is a rebreather assembly method/design that is foolproof. Such that the scenario goes like this: A completely disassembled rebreather is placed in front of a completely untrained diver. This diver should be able to completely assemble the unit and dive it without any instructions or training and live, because every single assembly and setup point is keyed to fit only its matching mating point. They would be able to assemble the rebreather simply by process of elimination by trying each connection point together until they found one that fit.

At least half a dozen rebreathers do achieve that. It is a reasonable requirement, to eliminate opportunity for an accidental misassembly that impairs the safety of the rebreather. The reason I say it is is reasonable, because it is achievable, and it would reduce accidents. Training in how to safely dive the rebreather is still required but not how to avoid assembly design errors.

NB: On your suggestion about controlling buoyancy .. have you seen DivewithSUBA.com?

 

[PRO-FOUND]

Simple, yet complex, at the same time.

Before I got my rebreather, My instructor constantly drilled me on Nitrox formulas. Being that I am also one of his OW instructors & his repair technician, he would point out my mistakes, even the small ones, & say, "if that was your rebreather,... you would be dead". It changed my way of thinking & analyzing things. Once the unit arrived, I completely built & broke down the unit, by the check lists & ran the pre-dive tests more than a dozen times, before I even took the unit to the pool. Now,... being that I am staff & work for my instructor, I was in no hurry to get the course done & we took a much more thorough & complete route. From the time I got my unit until I completed the course (MOD 1) was about 10 months (Dec.- Oct.). I spent the winter in the pool getting used to the buoyancy characteristics & the little nuances of the unit. By the time I took it to Open water in June, I had over 40 hrs in the pool alone. My instructor & I made 10 or more trips to the local quarry for my course. I was able to get in a couple post training dives before the weather started turning cold & went back to the pool to further refine my techniques. It took me most of this past summer to gather the necessary hours & experience to begin my CCR Cave training. I got a little taste of that about a month ago. Hopefully, in the spring, I'll be going back & gaining much more. Learning to dive a rebreather has taught me to be much more critical of my attention to detail,... a valuable asset. I fully realize,... I'm human, I make mistakes (comes with being human,... unfortunately), Yes, my rebreather can kill me,... But I am thankful that my instructor cared enough to slow me down & to be as thorough as anyone can be.

One little change to what you wrote

"Yes, my rebreather can kill me,... "

Should read

Yes, I can kill myself with my rebreather

 

[Omisson]

Including safety improvements is obviously a Good Thing, and in this particular case, I don't see any downside to Hollis using different threads or fittings to keep divers from putting the counterlungs on backwards. The real issue, though, is that it's just one more example of a diver not using the safety equipment provided--there's no way that she could have built that unit the way she did had she just used the supplied checklists. The surprising thing to me is that she was a recent student--you usually read about divers taking a little more time to become that complacent.

Looks like complacency can start very early on!

 

[victorzamora]

At least half a dozen rebreathers do achieve that.

Let's do this: let's go diving one day. Bring each of those rebreathers you believe to be foolproof. Let me assemble them, you dive them. I'd bet i could kill you with at least half of them. If i had more rebreather training, i could guarantee i could kill you with all of them. Nothing is sufficiently fool proof to a sufficiently exceptional fool.

 

[Akimbo]

I'm in the camp that rebreathers can never be made 100% foolproof...

That’s a pretty high bar given that we haven’t achieved that level of reliability on open circuit Scuba in the last 70 years.

… Personally I'd rather see the manufacturers develop units with more obvious and immediate user feedback when a failure mode is entered...

A lot of work needs to be done on the whole sensor, electronics, and software system to make a significant impact. That represents a considerable amount of R&D which is hard to amortize over so few units.