Makes it very clear how hard it is to make it practical; grid, even a small one, is just better.
An interesting insight further down: "It turns out that most remote islands are powered like ships are"
An interesting insight further down: "It turns out that most remote islands are powered like ships are"
On the mainland, investment in that infrastructure has gone in a cycle, I think we're coming off a decade of very large investment after a decade of 5x less investment in that infrastructure.
What's going to replace the expansion of transmission and distribution will be storage. It's often cheaper to buy storage at today's (relatively high) prices than to continue to size the transmission infrastructure for peak demand, rather than something closer to average demand. It's that peak demand that is the most costly to serve, batteries help with this immensely. In NYC, Con ED is delaying $1.1B in substation upgrades through the use of storage:
http://www.utilitydive.com/news/coned-seeks-efficiency-distr...
And on the smaller scale, in Arizona, a community that's growing is getting batteries rather than upgrading the transmission lines:
http://www.utilitydive.com/news/aps-to-deploy-8-mwh-of-batte...
Electricity billing is fundamentally flawed: Industrial users will pay a demand charge according to their peak usage, when they should really pay their contribution toward the local/regional/national utility's peak usage.
Residential customers may pay a time-of-day generation rate, but still pay a flat per-kwh rate for transmission and distribution. All 3 should be time-of-day, at the minimum. Wires sag more when you use them during the hottest part of the day and when they're already running hot. Upgrades need to happen to handle peak usage, not off-peak usage.
In open energy markets like those in California and Texas there's things called "Locational Marginal Pricing" and "Congestion Revenue Rights" to deal with this.