It is a good question and I have no good answers. We must hope the answers will emerge to make such AWE safe enough.
I believe everything connected to the tether may quite easily be made redundant, if the initial design is simple. That is, the kite can itself detect any failures and land before the consequences are a full crash. To enable this one option is:
- redundant cabling
- redundant power storage for electronics
- 3 redundant processors
- 3 of each sensor
- 3 of each actuator, or at least an actuator combination to allow the loss of a single one
If you add this, the chances of a failure on these are very low.
There are still single points that cannot be made redundant
- the software to detect failure modes
- the tether
- the winch (if one is necessary for landing)
- the connection point to the tether
- structural failure, eg broken wing
- radio and GPS (if used)
I guess the probablility of one of these happening along with the probablility of a crash being fatal would have to be be small enough to warrant the usefullness of the plant.
Of these, the tether is the tricky one, as you want to optimize the dimensions as small as possible, while there is abrasion due to wear. If AWE succeeds, we will probably see some effort put into monitoring the condition of the tether.
The backup plan for tethet rupture for many rigs will be flight without tether, landing/crashing at a safe location.
Any power production will cause accidents. A coal power plant will on average have many fatalities. So do traditional windmills. I expect AWE would not necessarily be worst in class here… though the psycological «understood danger» may be quite big, much like nuclear power…
So to answer the question, I think the problem is not the computer, rather the issues with tether rupture for rigs with one or few tethers is the main culprit