I had remembered Professor Knox using one in his anchor tests, but did not mention him as I did not recall that he had specifically designed and offered one as an alarm - though I thought someone had. Thanks again for the compliment - and even more for the clarification! 
I meant problematic to get any useful data from a strain gauge because this is a clamped (i.e. pre-stressed) joint so the change in load in the bolt varies very little for large changes in the load on the joint.
I think it would either tell you that the bolt had failed, or that the layup had compressed, or that the washer under the bolt head had sunk into whatever it butts up against, or that the strain gauges had come unstuck. I've only ever used tham in pairs because otherwise, you can't tell the difference between tension and bending. I've never used them on the shank of a bolt, so I don't know what the twist down the length of the bolt would do to them as the bolt is tightened.
I can't picture that in my mind's eye? Surely there is a tapped hole in the keel, and a plain hole in the bottom of the boat? The keel bolt passes through the boat and into the keel and its head bears against the bottom of the boat? I'm picturing a bolt (say) 20mm in diameter and a hole (say) 21mm in diameter, with the underside of the head of the bolt bearing against a washer and the washer bearing against the inside of the hull. If so, I can't see (a) that there would be enough clearance to put a strain gauge and wires on to the bolt shank and (b) bring the wires back up the hole to the inside of the boat? (Although, as the (now somewhat smug!) owner of an encapsulated keel, I am unfamiliar with the concept of worrying about keel bolts)!
Yes, I could see that working, in theory, but in practice, it's asking an awful lot of a strain gauge to stay stuck to a bolt shank for the life of the boat (or at least, the life of the keel bolt)! Hell, I used to have enough trouble getting them to stay stuck to flat plates in a warm, dry lab for several months rather than years! We used to re-calibrate these things once a month, but I don't know that you could reasonably expect the shole system (gauges, wire, connections, amplifier and signal conditioning / readout unit) to stay in calibration for such a long time, but there may be other applications that use load cells for long periods. Surely there are places selling off-the-shelf "load washers" that you can just trap between the head of the bolt and the laminate? (Don't know whether they'd be up to givign a constant reading for (say) 20 years though).
As for the hull. This would be much more tricky and less useful. I did notice, in my research, another boat which had this done but it was for a university research project. I was thinking that as I develop my layup I would make test panels and then subject them to known loads to get an idea of maximum loads and also of fatigue effects. The fatigue would be more difficult as it would require building a machine to make repeated cycles. There would be no particular use other than if we took a significant impact from, for example an ice berg, we would have some indication of whether the hull had reached a critical load. I currently have an idea that I might have an aramid layer on the inside of the layup so it might not be immedicately obvious to the eye that the structure has been weakened.
As I said earlier, geeky stuff and with little overall value other than to satisfy my boundless curiosity.
A properly designed fastener system installed and tightened correctly will not come loose or fail unless it is abused by say a collision. In those circumstances I can't really think why you wouldn't check out the relevant bits.
