Heat waves and violent storms recently knocked out power to millions of Americans from the Midwest to the Mid-Atlantic, and for many the power was slow to come back on.

These weather-related power outages highlight one of the U.S. power grid’s key areas of vulnerability – others include susceptibility to conventional or cyber attack, and man-made or solar electromagnetic pulses – and the heat-related deaths and lack of economic productivity due to the outages show just how dependent on the grid we are.

Many suggest that “smart grid” technologies would enable electric operators to diagnose failures and reroute power more efficiently, making the grid more reliable. However, centralized power generalization would still be the norm and it’s conceivable that in some instances, for example a conventional attack on a power plant, a smart grid would be no more reliable.

A recent paper coauthored by M. Granger Morgan and Anu Narayanan of Carnegie Mellon University titled, “Sustaining Critical Social Services During Extended Regional Power Blackouts” advocates the strategic deployment of distributed generation systems, which would create electricity “islands” to support critical services like hospitals, markets, and schools running in the event of an outage. However, the authors point out that “There are currently a few obstacles to implementing such a strategy, including state laws that prevent the deployment of cost-effective combined heat and power (CHP) ‘microgrids,’ and the lack of incentive for power companies to invest in such a system. We have the technology to make our critical services less vulnerable to large blackouts. What we need now are the right policy initiatives to make it happen.”

What investments in grid reliability are needed? Is distributed generation a solution? What policy initiatives would most effectively support the deployment of a distributed generation infrastructure?