In backup generators we trust?

Image: Photo: Jemimus shot this photo of a new backup generator (cc-licensed).

Backup generators exist anonymously. They are metal boxes, squirreled away on a roof or near a loading dock. You are meant to not see them. The point is that they are there when you need them and, the rest of the time, they do their best to be unobtrusive.

The problem is that this very job description makes it more likely that your emergency generator won't work in an emergency.

On Monday, New York University's Langone Medical Center lost power during Hurricane Sandy, and ended up having to evacuate 215 patients when the generator that was supposed to keep its charges alive and its critical systems running failed to turn on. Across the United States there are about 12 million backup generators. Most only operate during blackouts — times when a hospital, or a laboratory, or a bank, needs electricity and can't get it from the larger electric grid.

But backup generators aren't 100% reliable. In fact, they won't work something like 20%-to-30% of the time, said Arshad Mansoor, Senior Vice President for Research & Development with the Electric Power Research Institute. The bad news is that there's only so much you can do to improve on that failure rate. The good news: There are solutions that could help keep a hospital up and running in an emergency, even if the emergency power system doesn't work.

So why do backup generators fail? The short version is that we only use them, you know, for backup. Most of the time, these generators just sit around, doing nothing. It might seem like you're keeping them safe, but it's actually a pretty rough way to treat a mechanical system.

If, like me, you've ever owned a scooter in a cold climate, you'll be familiar with this paradox. You store the vehicle away, nice and safe for the winter, and when you get it out in the spring it actually runs worse than it did back in October. Maybe the battery is dead. Or the oil needs drained and replaced. Whatever happens to be specifically wrong, leaving mechanical systems to sit around for long periods of time isn't really good for them. This is why the spring tune-up exists.

"It's not an issue with the actual quality of the generators," said Dan Zimmerle, assistant research professor at Colorado State University's Engines and Energy Conversion laboratory. "It's maintenance related. For instance, if you don't burn deisel fuel sitting in the tank, it will start to degrade and clog the fuel filters. Things that don't get used tend to fail."

Together, the combination of nationwide electric grid and backup generator means you'll most likely have the lights on at any given point in time. But it's not a guarantee against blackouts.

Instead, Zimmerle and Mansoor say we need other lines of defense, if we want to avoid hospital evacuations in the future. And microgrids are one possible solution.

Today, we rely on an electric grid that stretches coast to coast, throughout the United States and Canada. Large areas of that grid can be managed independently — everything east of the Rocky Mountains, everything west of the Rocky Mountains, Texas, and Quebec — but for the most part, small hyper-local bits of the grid can't really break off and do their own thing in an emergency.

There's a few reasons for that. First, most of our workaday electric generation is done in bulk. By which, I mean that it happens at very large power plants, which each serve millions of customers, and those power plants are located relatively far away from the people who use the power. The second issue has to do with the way the grid operates. Electric grids have to maintain a constant, almost perfect balance between electric supply and electric demand, and that means maintaining a constant voltage and frequency.

Right now, your neighborhood gets that voltage and frequency signal from the larger grid as a whole. If you're suddenly cut off from the signal, your neighborhood will cease to have a working electric system — even if there are sources of generation right there down the block.

In an emergency situation, we do suddenly have lots of hyper-local generation sources — those 12 million backup generators. What we don't have is the infrastructure in place to take advantage of that. A backup generator can power a building, but, in general, it can't share resources with the building next door.

A microgrid would change that, enabling areas the size of neighborhoods to operate independently in the event of an emergency. "Your backup generators are tied together and then you can redirect power from where it's available … say at a bank … to a hospital, or a fire station, or someplace more critical," Zimmerle said.

Doing that means updating technology, but it also means changing the way we think about legal and regulatory frameworks. In particular, Zimmerle pointed to power purchase agreements — contracts between the people who get electricity to your house and the people who generate it. In some places, those two jobs are done by the same people. But where they aren't, power purchase agreements usually limit the amount of electricity that can be generated locally. That cap can be as low as 5% of total and it includes everything from college campuses that make their own steam and electricity, to the solar panels on your neighbor's roof. The contracts aren't evil. But they do make it difficult to set up microgrids.

The other big problem is cost. Updating infrastructure is expensive. And it's been hard to convince utilities to spend billions on a system that they're only going to use in rare emergencies. Even when, in one case, you're spending billions in small doses over a long period of time, as opposed to having to spend billions (and possibly a greater amount of billions) all at once to deal with storm damage and shutdowns. Basically, we don't have microgrids now because utilities have looked at the balance between cost and benefit and didn't see a big enough benefit.

But, as 100-year storms become more frequent, the outcome of that analysis could start to change.