Here’s how it would work. Each substation would receive instantaneous capacity availability data from the grid. When it receives notice that an increment of capacity is available or if it senses a drop in its local load, it would transmit a distinctive pulse on the local power lines engineered to traverse the local interconnect with high reliability.

Current limiting meters having autos connected for recharging would react to each pulse by increasing their output by a very small increment. As the evening progresses, there will be a cascade of pulses ratcheting up the charging current on each of the autos until there is no further change in load - e.g. all the cars connected are charging at their maximum rate. At this point pulses would be sent continuously to catch up late connecting cars. Fluctuations from other loads would be handled in the conventional manner.

A system like this requires no separate communication channel, is single duplex, robust and will act effectively to limit load increases placed by autos to occur at a rate that the system can respond to without stress. Individual pulses can be obscured by transient phenomena but it should be impossible for a sufficient number to be dropped to keep a vehicle from being fully charged in the morning.

This would be practical if a low cost receiver was be engineered to receive the pulses reliably and increment the rate of charging accordingly. This may be a baby step compared to the grand digital grid solutions being promoted by wireless telcos, but I bet it’s a 90/10 shot at creating a scenario that is in range of other power management technology that’s coming on line now. Altair Nanotech makes high power-density lithium ion batteries. They delivered a 2Mw / 300kwh battery to the AES power company last December to use for power stabilization.

Power companies generate about 3% more power than is needed to react to instantaneous changes in load that gives them time to adjust generation. If they use a battery to react they save the fuel cost used to generate the mostly wasted “active reserve”. The payback for the battery is several months and it lasts for 20 years.

There’s no need for high precision, two way, broadband digital management to achieve high quality power management because a battery gets the fine structure power management done faster and better for a fraction of the cost.