IT-based solutions to the energy/climate crisis

October 9, 2008 1:35 PM

What are some ideas for addressing the climate/energy crisis that are based on information technology and logistics, rather than environmental science, per se?

I am brainstorming for a topic for an academic project that involves detailing a method to address the climate/energy crisis. Most of the ideas my professor has suggested center on environmental/chemical engineering, but he is encouraging solutions based on other academic disciplines. My specialties are computer science and operations research; I know little about the geosciences. What are some proposed solutions to this crisis that center on the use of information technology, rather than environmental engineering?

posted by lunchbox to science & nature (12 comments total)

The first thing that comes to mind is the idea of the "smart power grid." Sadly, I don't know much about the concept, but according to the wiki page I just linked:

The term Smart power grid may best be defined as using communications and modern computing to upgrade the current electric power grid so that it can operate more efficiently and reliably and support additional services to consumers. Such an upgrade is equivalent to bringing the power of the Internet to the transmission, distribution and use of electricity - it will save consumers money and reduce CO2 emissions.

posted by General Malaise at 1:42 PM on October 9, 2008

I know I'm stating the obvious but: encourage the use of existing IT infrastructure to replacing commuting with telecommuting.
posted by -harlequin- at 1:55 PM on October 9, 2008

Somewhat related to operations research - there's enormous opportunities in waste heat recovery and industrial efficiency improvements.
posted by milkrate at 2:18 PM on October 9, 2008

You could look at telecommuting, and similar things (telepresence robots, video conferencing) that would reduce demand for fuel and office space.

You could talk about 'smart buildings' using sensors to accurately quantify running costs - i.e. so that you could get a power bill for every socket and lightbulb in the building. You could use this to support decision-making.

You could work with your university's computer services people to accurately quantify the running costs of leaving computer labs powered up over night. You could analyse usage patterns and systems like "energy star", sleep modes, wake-on-RTC, and similar things. In other words, how much power/money could you save with little or no negative impact on user experience?

You could beat the dead horse of thin client solutions, with a pinch of cloud computing: Why not locate data centres where power is cheap/power transmission losses will be low; then have efficient thin clients wherever they are needed?
posted by Mike1024 at 2:20 PM on October 9, 2008

Find out why the promise of paperless office wasn't delivered and try to use that to project how current promises of saving by IT and telecommuting can have unforeseen, wasteful consequences.

(For example, because of telecommuting it is easier to start distributed teams, but they still need f2f-meetings at some point. Do these few meetings make them more environmentally costly than having a whole team local?)
posted by Free word order! at 2:58 PM on October 9, 2008

The sustainability field (especially Clean Tech and logistics) has been growing like crazy, so "computer science" and "operations research" is a little vague.

Electrical Engineering has some strong possibilities. Semiconductors control much of what we do, they can have a huge impact in controlling data centers, industrial applications, HVAC systems, cars, alternative energy, appliances, anything with a chip. Battery storage is another major area of opportunity. Basically, there's huge opportunity in doing things more efficiently.

With your expertise, I would concentrate on data centers or supply chain.

I also reccomend reading The Clean Tech Revolution

Feel free to email me and I'll send you a gazillion links to dozens of emerging technologies that can impact energy.
posted by pokeedog at 3:08 PM on October 9, 2008

1. Relocation of data centers to areas with cheap/clean energy (see: Google in Oregon where there is lots of hydro power; also some crazy idea I read recently about putting data centers into a boat, anchoring it offshore, and using the tides to power it.)

2. Look at centralizing vs decentralizing. Probably more efficient to replace individual workstations with client/servers (the Sun Ray model).

3. Investigate more efficient power supplies in all IT equipment (I'm led to believe this is a major component of IT energy use)

4. Here's something original I just thought of. Energy profiling of different operating systems in different configurations. I have a theory that spyware and crapware keeps the CPU busy which increases power usage. On a related note, energy profiling of clean vs dusty computers, see how much energy is burned trying to keep them cool. Maybe companies can save a lot of money by vacuuming their computers once a year?

These are more about the problems of IT rather than how IT can be used, though.... if you're asking a question like "what can I accomplish by throwing computing power at the problem", I don't really have any good ideas.
posted by PercussivePaul at 3:17 PM on October 9, 2008

1) A kind of smart shipping solution, so that, for example, people ordering stuff in the same neighborhood would _receive_ it at the same time (assuming they weren't too picky about when they received it), thus eliminating the need for 25 trips by the UPS truck into my neighborhood on the same day?

I don't know everything that would be involved, and it could get complex - trying to build in customer-selectable schedule flexibility, coordinate across vendors, etc. I expect that UPS, for example, is doing some of this already, but still - if it's some kind of overnight documents shipment, probably that could be conveyed on a motorbike, scooter, or little wagon-y car. A bunch of clothes from Land's End? I'd be fine getting that next week as my neighbors' Amazon orders and carpet samples.


---

2) Another idea: make public transportation _work better_ for the people who use it. Our system (Chapel Hill) participates in some kind of real-time tracking thing, so you can see on a computer where your bus is at any given moment. But having schedules that work right is important. If you don't know exactly when the bus is going to arrive at a given stop, then everyone who rides that bus -- especially at times when there are fewer bus trips per route -- has to allow a big block of time to be sure to get the bus whenever it does arrive.

One might have to allocate an extra _two hours_ per excursion (for the bus I take at ~1:00 weekdays for a 2:00 class on a campus literally five minutes away by car). This is both an environmental issue and a social justice issue - people who ride public transport because they can't afford cars are literally wasting time that they could be spending with their kids, or improving their education, or exercising, or cooking non-fast food.

Why can't our bus follow its published schedule? It's not the driver's fault -- the estimated route duration is imprecise, and probably calculated at a time when there wasn't as much traffic or something.

---

3) Yet another idea: figure out a way to make food (I'm thinking produce OR prepared meals) be delivered more efficiently, reducing the amount of shipping distance food has to travel. It might actually involve recipes and/or coordination with local tiny prepared-food vendors to make sure people would _want_ to eat what was prepared locally -- the information would need to flow both ways, preparers <> producers. Then you get into a similar delivery logistics problem as in the first idea above.

A simpler version might involve some kind of very-easy-to-update inventory system for local farmers (as in farmers' market vendors), customers being able to place orders with multiple vendors through one system, and a hyper-efficient delivery system. Just ordering produce locally from farmers means less shipping trans-continentally. Even for this, though, I'd seriously recommend some carefully-selected recipes as part of this system, so that I'd know what to do with my locally grown kale without having to go out to the store and get some exotic thing to go with it, and so that I could get dependably yummy results without pining for a lettuce-based salad when it's just plain too hot to get good lettuce here.
posted by amtho at 3:30 PM on October 9, 2008

PS - Where are you studying? What's the course? Sounds interesting. Please tell us more about people who are actively learning useful stuff!
posted by amtho at 3:35 PM on October 9, 2008

You're almost certain to find something useful at the Rocky Mountain Institute.
posted by flabdablet at 3:38 AM on October 10, 2008

Thanks for the great ideas, guys. I'm going to consider a lot of these.
posted by lunchbox at 1:19 PM on October 10, 2008

Adding to what General Malaise said in the first reply, the application of It to changing the way we manage electricity grids is likely to become increasingly significant as we move from the traditional model of centrlaised generation into a model with increasing volumes of distributed generation such as renewables and other small scale generators. Denmark is probably the most advanced down this path, this paper should provide a fairly detailed introduction to the ideas (its from a project I worked on a few years ago, though I'm not the author). you might also look up active network management as a useful search term (as opposed to the passive we mostly have now).

The UK has also seen some advances in the area, you could look up innovation Funding incentives and Registered Power Zones at the UK electricity regulator, Ofgem to find out more. It's quite interesting (to me anyway) but the idea of both was to stimuate technical solutions but the general upshot has been more management type solutions.
posted by biffa at 2:26 AM on October 13, 2008