If necessary, legislate the presence of the frame.
That's why we don't let politicians design technical systems. They're bound to come up with solutions that sound good but that will not work in practice.
Some problems with your approach:
- partial loss of signal -> false positive -> power surge
- loss of signal -> false negative -> grid overload
- how do you signal the need for a little bit more power?
The British solution is typical: it's simple, and it 'just works', the speed with which hydro comes online it is just not enough to be reactive but even a little bit of forewarning and they'll be fine. Heck if the TV fails he can call home to mum, or some other place.
Solution of last resort: a slight brown out due to lack of information, but as long as this system has been in place you'd think if that were a statistically frequent thing that it would have happened by now.
Compare it with the auto-traffic-info system present in most car radios.
http://en.wikipedia.org/wiki/Radio_Data_System
The alternative frequencies system works along the lines you suggest (though technically it's not a 'frame' but an entire channel with its own carrier wave). This fails so frequently that I usually disable it. Either you get switched to some totally unrelated station (without traffic info to boot) and then never switched back or you're listening to a CD or so and then you'll suddenly be in the midst of some radio broadcast and again, you're never switched back.
Easy to propose, using an unreliable medium such as a TV or radio channel apparently very hard to get 100% right.
And if you're going to change major parameters in the electricity generation capacity of a nationwide grid I assume you'd put the bar for reliability a bit higher than I do for my car radio.
In practice, FM radio signals travel so haphazardly that in most areas, you're likely to be interrupted several times an hour with travel news for towns and cities 40, 50, 60 miles away. I regularly get the bulletins for places as far afield as Shropshire, Humberside, Merseyside and Lincolnshire. If, while driving, the signal becomes too weak for the 'traffic off' flag to be received, the radio will simply stay tuned to the FM frequency until the listener manually cancels it.
In addition to this, illegal 'pirate' radio stations have started to take advantage of this feature - some of these transmitters have the traffic flag set to 'on' constantly, with the obvious effect of switching listeners' radios to the illegal station regularly.
Most people have it turned off!
First, and the biggie, the BBC would be responsible for including Cue Dots/Marks/Frames which signal the program's TTL not that "more power is needed in UK". That'd be irrational and dangerous. It's up to the power guys to interrogate the cue information and then react to it. That's just basic separation of responsibilities.
Second, you'd still want a centralized component to interrogate the cue information and then tell power utilities to turn on/off, just like the current model. Of course, they will still need people in the loop - the computer would be there to propose which assets to activate and to provide a countdown to when a decision will need to be made - a guided decision making system. (Computers can do this sort of cost optimization faster then people.)
Third, their man in the power HQ is already relying on a TV set to predict when the spike is about to occur, so any sort of system has to be at least that reliable (i.e. not very).
Fourth, what if the power boffins lose the signal to the BBC? Calling "mum" is a terrible idea, but having automatic failover and/or redundant reporting stations are the type of precautions a basic safety review should uncover.
TLDR: Not suggesting that the BBC tell the Power Co that they should add more power to the grid. Just suggesting a system to augment decision making based on Program TTL.
http://en.wikipedia.org/wiki/Cue_mark#Cue_dots_in_television