Not able to adequately handle peak load demand this afternoon, Con Edison is -- at the moment -- trying to avoid a blackout by resorting to emergency measures.

The company has resorted to calling customers multiple times (automated calls were placed at 4:28 p.m. and again at 6:39 p.m.) asking customers to reduce demand (i.e., power consumption) by turning off all non-essential electrical appliances.

Apparently "because of problems on electrical equipment supplying power to several neighborhoods in Brooklyn, Con Edison has [also] reduced voltage by 8 percent to the area."  While it may be true that the problem was due to "electrical equipment," you have to wonder if reducing customers' level of service actually was a cheaper solution for the investor-backed utility (relative to buying power on the spot market at peak)? For every 1-volt reduction, the effect is a reduction of total system demand by approximately 1-1.5%. So in this case, Con Edison just reduced system load by approximately 8% to 12% (for the neighborhoods affected). 

Utilities in the U.S. are required to deliver power to consumers at 120 volts plus or minus 5%, which yields a range of 114 to 126 volts.  Con Edison's website did not mention whether the utility had dropped service below the 114-volt threshold.  It's worth noting that many consumer electronic devices and other digital equipment start to malfunction, and can become damaged, when they are "fed" power at a voltage level below 114V.

An alternative to "gaming" the voltage and/or buying incredibly expensive power at peak on the wholesale market, would be using a demand response system (a.k.a. "demand side management"), whereby end-users are incentivized to turn down their power through a financial rewards program.  (In fact, the premise of having "smarter homes" outfitted with home area networks and smart appliances is largely based on this idea of consumers selling "negawatts" back to the utility during high demand periods.)

Beyond demand response, simply having a more dynamic distribution grid -- one, for example, that is capable of integrating information such as weather data into its modeling -- would reduce the need for the kind of rudimentary response that Con Ed, and many other utilities, are currently forced to take in the middle of heat waves.

Obviously, this all leads directly into the case for smarter grids, but I'll spare everyone the lecture... for now.