Rebreather World - View Single Post - Point of no return on-the-fly calculation

AD_ward9 · 15th November 2006, 19:05

Quote: (Originally Posted by Faceless)

It may sounds weirdo, but after playing with your simulation a little bit more, an interesting idea came to my mind.
As Bob has explained to me, deco obligation is being calculated on the fly ( i.e. there is no static dive tables), and given that O.R. rebreather has full gas sensing including gas pressure in both diluent and oxygen tanks, and in addition it can predict state of scrubber, is it possible with these data to derive a so called point of no return in real time as well ?

If you're using finite state machine then I guess, using just plain dull voting logic for O2 cells is silly, probably you have buit some mathematical model of O2 cell falure modes, then hardcoded it into set of states or trends, right ?
So, let say we have some dive/tank pressure/deco obligation/etc. trends and there are junctions of these trends which may lead to inability to perform a required deco obligation. If we can hit this junction i.e. when user is near that point or has reached it due to gas leak or smth else not so fatal but pretty dangerous, let's warn user about it.
Kinda a piece of common sense.
Sorry in advance if my idea is stupid

Offtopic: Alex, I wish you smooth and nice trip to DEMA and hope your demonstrations will go as you planned. Keep up the good work.

Thanks for your good wishes. It was an excellent show.

As well as the photos I will put into the gallery next week, when I have some time, we will do a video of the unit underwater.

We predict the turning points based on the third rule, that is, when 3 * (1 - O2 in bar / starting O2 pressure in bar) >= 1. This might do what you want. Time to surface is displayed on the handset (top left), then ceiling time and ceiling depth.

On your question on O2 sensors, we eliminated all modes where the sensor fails high (by getting the sensor vendor to change the sensor design). We use 4 sensors, and we need only take the highest reading. This means 3 out of 4 can fail. When there are depth changes, we predict the O2 change, and if it is outside the expected range (which is tiny), then we check the O2 sensor and injector by widening the spread of PPO2 control. The PPO2 control can maitain PPO2 for just under 30 minutes even if all PPO2 sensors are out, including for depth changes.

Cheers,

Alex

15th November 2006, 19:05	#2 (permalink)
AD_ward9 So much more to learn Current Rebreather/s: Other CCR Other Rebreather/s: Other CCR Join Date: Jun 2005 Location: Scotland Posts: 1,862	Re: Point of no return on-the-fly calculation Quote: (Originally Posted by Faceless) It may sounds weirdo, but after playing with your simulation a little bit more, an interesting idea came to my mind. As Bob has explained to me, deco obligation is being calculated on the fly ( i.e. there is no static dive tables), and given that O.R. rebreather has full gas sensing including gas pressure in both diluent and oxygen tanks, and in addition it can predict state of scrubber, is it possible with these data to derive a so called point of no return in real time as well ? If you're using finite state machine then I guess, using just plain dull voting logic for O2 cells is silly, probably you have buit some mathematical model of O2 cell falure modes, then hardcoded it into set of states or trends, right ? So, let say we have some dive/tank pressure/deco obligation/etc. trends and there are junctions of these trends which may lead to inability to perform a required deco obligation. If we can hit this junction i.e. when user is near that point or has reached it due to gas leak or smth else not so fatal but pretty dangerous, let's warn user about it. Kinda a piece of common sense. Sorry in advance if my idea is stupid Offtopic: Alex, I wish you smooth and nice trip to DEMA and hope your demonstrations will go as you planned. Keep up the good work. Thanks for your good wishes. It was an excellent show. As well as the photos I will put into the gallery next week, when I have some time, we will do a video of the unit underwater. We predict the turning points based on the third rule, that is, when 3 * (1 - O2 in bar / starting O2 pressure in bar) >= 1. This might do what you want. Time to surface is displayed on the handset (top left), then ceiling time and ceiling depth. On your question on O2 sensors, we eliminated all modes where the sensor fails high (by getting the sensor vendor to change the sensor design). We use 4 sensors, and we need only take the highest reading. This means 3 out of 4 can fail. When there are depth changes, we predict the O2 change, and if it is outside the expected range (which is tiny), then we check the O2 sensor and injector by widening the spread of PPO2 control. The PPO2 control can maitain PPO2 for just under 30 minutes even if all PPO2 sensors are out, including for depth changes. Cheers, Alex Last edited by AD_ward9 : 15th November 2006 at 19:15.
(Offline)