Sunday, January 9, 2011

114> In Search of a Cheap Lunch

One of the dirty little secrets of “green” electric cars is that the batteries have consumed a lot of energy and created considerable greenhouse gases before they are even installed in the vehicle. The real environmental cost of batteries goes all of the way back to the mines, where diesel fuel is used in large quantities to extract ore. Fossil fuel is an ingredient in the plastic battery cases. Refining the ore into metals and useful compounds often is extremely energy intensive. It takes fossil fuel to ship the materials to the battery maker and still more energy to assemble them. Of course then they need to be shipped to distributors or to the vehicle manufacturers.  More energy lost. The real energy costs should probably even include the energy budgets of all of the employees of all the companies involved insofar as those expenses are directly tied to the manufacturing process. (A miner’s gasoline costs getting to the mine, for example) Then there is the energy to move the electric vehicle’s extra battery weight, and eventually everything involved in the steps of removal and recycling. Then there is the fact that fossil fuel is burned by utilities to generate electricity to recharge the batteries, but let’s leave that one aside for the moment.

It is a reasonable to ask, “How much energy is actually saved over simply fueling vehicles with gasoline directly?” After all, gasoline has one thing going for it. The pipeline between the well and your car is very efficient. This is something to consider with other supposedly “green” products as well. Solar cells, for example, are notoriously energy intensive to make and, likewise, do not last forever. It’s like the oil used to make the fertilizer for the corn to make cleaner burning ethanol fuel. There is no such a thing as a “free lunch…”

I am certainly not saying that this stuff is a waste of time to pursue, but that it should be considered in the design considerations of nascent technologies like PRT. This applies to all design choices, not just whether to use batteries. In particular, I would point out that my call for a minimalist track profile is not purely for aesthetic reasons. We ought to be asking ourselves, “What is the greenest possible medium for moving from point A to point B within the urban/suburban environment?” This, as luck would have it, will also probably be the cheapest, and least objectionable to look at.

I submit that a power-carrying micro-monorail system is the greenest alternative, all things considered, unless we can invent a way to make ski lifts have branching routes and off-line stations. It should be as close to invisible as possible and use minimal materials. It should allow very flexible routing options including tight turns, steep slopes, etc. If it can’t be run somewhere, then people can’t use it. I further submit that it should be thick enough to be a “workhorse” that can take fast vehicles and span wide streets without shaking or sagging. Being too thin mandates closely spaced supports, which can also be a disadvantage. On balance, this trade-off puts me squarely in the Ed Anderson camp, size wise, of about a meter high and about two thirds of that in width. Long-time readers of this blog know how much I have agonized over these dimensions. One advantage to a self-leveling suspended vehicle, I would note, is that it can transition in elevation easily, so that the main routing need not be on the same level as the stations, enabling track that can be higher and more out-of-the-way, if that is what the community demands. We don’t want to cut trees to put PRT in.
PRT has been caught up in kind of a “Gee-wiz, I’m so futuristic!” mindset, even though there is nothing, in this age, futuristic about it. But it is still about being green. My last post was about how free-roaming robocars had co-opted the PRT moniker, and we’ve been having a lively debate on better names. I would just like to add this thought to that debate. If PRT is the physical equivalent of the internet, then the track is the equivalent of telephone wires or fiber optic lines. I say, “Let’s go broadband from the start!” Furthermore, let’s make that infrastructure as green as it can be. That means not being designed to be scrapped, but rather being modular, so it can be moved and reused rather than melted down; It should contain zero fodder for the landfill.  PRT, of the powered rail variety, isn’t just another green transportation alternative. It is the ultimate green alternative, bar none. (I’m not counting open-air or human powered vehicles) So maybe it should be presented that way, by the infrastructure, and not the vehicles or the difficult-to-explain operational characteristics.

After all, if you are promoting “elevated microrail transit,” then the whole rest of the PRT paradigm becomes implied. 
“Automated or involving lots and lots of drivers?”  - Automated.
“Make everyone wait behind a stopped vehicle or have off-line stations?” - Off-line. You get the idea.

In the end, being green, being efficient, and being prosperous are all one-in-the-same. Battery powered electric vehicles, though not a complete red herring, do start with substantial energy deficits that should not be ignored, so environmentalists should be made aware of the fact that powered rails are a much more efficient option.

Elevated, line-powered, mini-monorail transit: To me, it’s a no-brainer. There should be non-profits promoting it, universities developing vehicles for it, the works. It’s where we need to go. Delay in doing so is simply squandering resources, including our land, our raw materials, our fuel, our time, (spent in traffic) our time (spent building and unbuilding stuff) our (still not totally carbonated) atmosphere, and of course, our money. 

PS - If there’s anyone who can find a link to actual studies on the energy used in the life cycle of batteries I would be grateful if you would share… I have only found this paper, which is so outdated that it doesn’t even have figures for Lithium-based types. Finally, I would like to share this video, listed as “300 years of fossil fuels in 300 seconds”.

19 comments:

akauppi said...

This article is touching the subject (of Li-Ion batteries and Lithium mining).

Vinod Khosla On Why Lithium-ion Batteries Are Overhyped

Also see the comments - many good ones there.

akauppi said...

Car battery's carbon footprint

The Swiss study referred to in that article.

Now, would someone have the time to read the study (I didn't).

Lars Endre said...

“How much energy is actually saved over simply fueling vehicles with gasoline directly?”

None. Since You are clever and ask specifically about "energy"!

The loss of energy is higher when charging a battery, than with the fueling of a gasoline tank. Much higher; and won't weigh up for the fact that electric-to-motion-conversion may be more efficient than gasoline-burn-to-motion.

Gasoline is in fact so far the most energy efficient way for mobile solutions.

Now, talking about carbon emission is something "else". If (but only if) the electric power you use to charge Your batteries come from "green tech" somewhere, is it better - BUT then, to make matters complicated again; you should remember that the electric grid is "fuelled" by all kinds of generators, and that "one kWh gone to charge an electric vehicle" will have to be taken from somewhere.

I have two powerful arguments for the PRT to take it's energy right from the grid; 1) NO conversion of electricity from grid-to-battery (and thus also no battery weight in the pod). 2) "Green energy production" are much faster and better put to use directly onto the grid, than if it were to be implemented on the pods.

Lars Endre said...

...and a link for You; http://www.lowtechmagazine.com/2008/01/bumper-cars-o-1.html

Andrew F said...

I seem to recall that gasoline vehicles are also shockingly inefficient from a wells-to-wheels perspective. Something like less than 5% and quite possibly less than 1% of the raw crude in the ground makes it into energy to turn the wheels. Not that batteries are necessarily more efficient. It is just to say that there are enormous opportunities for improvements in efficiency along the whole system.

I agree that eliminating the battery charge/discharge from the system helps greatly with efficiency. You have to compare the track/vehicle as a cost as well in comparing to batteries, though.

Dan said...

Dan the Blogger Responds –

Thanks for the links, guys. Let me a couple. I am still a big believer in wheel motors (hub motors) …and I DO mean the DIRECT DRIVE type, where the middle stays put and the motor/wheel turns around it. (Last I looked even Wikipedia couldn’t be trusted to make the distinction) Check this out. (they also give some automobile figures for comparison) I have the spec sheet for this motor and it lists it at 97.4% efficient, not 98. It’s about 2.5 horsepower, to use the old fashioned terminology, so 4 wheels would only give 10 hp, which is a little shy for multiple passengers. By the way, it has 40 poles! (if you’re into that kind of thing) There is also this youtube video of the motor in action and opened up.

Lars Endre said...

Cool motor! Compared to combustion engines You save the following energy conversions / power transfers:
Potential energy in the fuel converted to kinetic energy of expanding gas via combustion
Kinetic energy of expanding gas converted to linear piston movement
Linear piston movement converted to rotary crankshaft movement
Rotary crankshaft movement passed into transmission assembly
Rotary movement passed out of transmission assembly
Rotary movement passed through differential
Rotary movement passed out of differential to drive wheels
Rotary movement of drive wheels converted to linear motion of the vehicle.

... and You're only left with a transformation from electric energy to rotary motion!

Still, the power has to come from somewhere, and the question of whether this is an onboard battery or the grid/rail is still open. That's where I stress many points on taking it from the grid/rail; You skip the energy transfromation: Grid to battery. (And You save the weight and room of the battery).

You also get to implement any kind of improvement in electricity production directly to the grid powering stations with immediate effect, rather than having to take xx thousand pods in for a refit (however modular and standardized their design might be).

akauppi said...

I am a proponent of using batteries.

While there are some benefits to "grid" (power from the rail) operation, which one is optimal differs in my opinion by the geography and market segment.

Batteries allow the rail to be simplest, and safest. They allow more independence between the vehicles since one central place of malfunction (losing the power delivered to all vehicles on a certain track section) is eliminated.

But most of all, consider that in India up to 50% of the power fed to the general grid is actually stolen by illegal wiring. That alone is enough reason for me to go for battery instead. More robust.

It may be different in the US. Electricity from the rails may be the right approach out there. But even so, battery technology is supposed to develop hugely in the 5-10 years to come, and power from the rails - well - will not develop. Thus a proper comparison would not only need the geographics, the market segment but also the time of application to be considered. In my opinion, battery powering leaves more options open.

India electrical wiring (Google)

Well, this story is from US. And we all know the US grid badly needs repair. So maybe the stealing applies there too. Stealing electricity is on the rise

Andrew F said...

The problem for any reasonably large and successful system using batteries is that a substantial portion of vehicles will, at any given time, be out of commission as their batteries are required to recharge. Unless battery technology substantially improves or you go with a switchable battery, this will necessitate a larger number of vehicles for a given level of daily throughput.

akauppi said...

Switchable batteries, of course. But as you are stating yourself, this is only necessary because of the charging times of most current batteries. If that changes (to say 2-3 minutes) then maybe fixed batteries are okay.

Most matters with PRT seem to be such multidimensional optimisation cases, with time and technology changing the "best case". This is the problem with ULTra and others, who's design was essentially made with 1995 technology options. Time has changed. Optimal is elsewhere. Any product and company must remain flexible enough to embrace such changes, and change their offerings accordingly. Optimal is not a static thing, at least not in the 2010's.

Why I like the battery approach?

A big stack of switchable batteries (in a warehouse / pitstop place) solves many issues. Grid black-outs. Not a problem. Using solar or wind power for charging. Not (such) a problem. We could even help balance the grid at large, like people are expecting parked electrical vehicles to do (which I think will not succeed, for many reasons). With lots of charge capacity at PRT pit, doing this kind of grid load control will be easy. Of course, we'd charge the grid for providing this buffering capability (charge when electricity is cheap, push back when it's more expensive).

So - remind me again why batteries are a bad idea? :)

What I really would like is for us to move on. Take real challenges preventing PRT from getting further on the hype curve. Because I see people who'd like to buy these things - only they don't really exist. How to change this?

Crowdsourcing, maybe?

How about just listing the top N reasons keeping PRT in the dream world. Here's mine.

1. financing. Lack of it.

2. overly belief on automated cars (which does not solve everything, and betting on just one horse is always bad)

3. financers don't ride public - thus they don't really see the problems?

4. too slow change of infrastructure. Traditionally we've been in 10-20 year cycles at least in Finland. This must change to 2-5 cycles (akin to what computers did to sluggish office work)

5. Lack of imagination.

Andrew F said...

Even if batteries do get better and can be rapidly charged and/or swapped, the battery paradigm necessarily is less energy efficient (charge/discharge efficiency loss), as well as necessitating the carrying of heavy/expensive batteries and/or swapping mechanisms. I don't doubt that batteries will get a lot better, but I'm doubtful that it makes sense to carry them around in the vehicles when there's a more straightforward alternative. If you're concerned about power failures, put the 'batteries' (or fuel cells, generators, etc.) in distributed backup power centres along the guideway. As an example of such a system, look at the landline telephone system, which is very resilient, even during blackouts.

akauppi said...

Well, clearly opinions differ on this - as they do on many other PRT design criteria.

Have you got practical experience on vehicles that take current from an electrical wiring, s.a. trams? I used to travel them daily, and electric arcs on the overhead wires are a common sight. While this does not really trouble tram usage, it will be a different scenario with smaller vehicles where passengers are closer to the power rails.

Oh, but I forgot for a while about the "enclosed rail" design that Dan's idea is using. Sure - in such a power rail is probably best. It's anyways safely contained within the enclosure so any visual sparks are contained as well as the power rail not being exposed to weather etc. Sure.

Dan tried in one stage to make a categorisation of "all things PRT". That would be helpful, since we're really talking of a family of things. Not like frogs, but more like amphibians. The different PRT's might not share much in looks, though they have common philosophy and ancestory. I think this is even confusing ourselves, as evident of my forgetting the enclosed rail thingy.

Dan said...

Dan the Blogger... took to long to post his comment and now it's two comments later...

I'll post it anyway, it's almost midnight..

Power is stolen mostly because it is so darn easy to do. I could bypass my home’s meter with a knife and pair of jumper cables in 2 minutes. If the voltage or modulation is exotic, however, (not 12v or house current) it would be fairly useless without a lot of conversion gear.

About vehicle independence… most powered rail designs call for a backup battery that can get a vehicle to a station, if not to the destination. There is a lot of difference between a battery that can move a vehicle for 10 minutes and one that can work for many hours.

Akauppi, it seems to me that you may be glazing over some difficulties with the battery approach, although I agree that these difficulties are liable to be alleviated somewhat in upcoming years. Have you figured the (weight/cost/power output) requirements of these batteries? My best guess is that they would weigh at least 70 kg, and cost at least $2500. That’s $5000 for two, one for the vehicle and one to swap.

Batteries don’t last forever. Lithium Ion batteries can only be charged 1200 times at most, Ni MH maybe 1500. What kind of distance and fare receipts do you anticipate per charging cycle? 1200 charges is a lot for a car or scooter, (they anticipate an hour or two of driving per day, tops) but not for the continual use of a transit system, especially one that you want to be lightweight. Could a battery even last half of a day on a single charge? I’m guessing 3 charges on a busy day. And then there is the extra labor involved in swapping the batteries, the time/energy to get there, the real estate for the charging stations, with their special cooling, backup and security needs.

Of course adding an electric “third rail” is no piece of cake either. It is one of the many reasons I went for expense of an enclosed track, and I STILL have not worked out all of the details. That is why it has never appeared in any of my drawings. Either way, as they say, “There’s no such thing as a free lunch!”

Dan said...

Dan the blogger continues, almost endlessly....

As for reasons why PRT has not been adopted…under financing,… I’ll just say this:

Financing requires a very detailed cost/benefit analysis, which pretty much means that a system must be completely engineered using available technologies and parts. Then manufacturing/construction/operating/maintenance/depreciation costs can, at least, be estimated. Only then can a comparison be made between projected earnings and those costs, to see if there is sufficient profit potential to make financing worthwhile.

That’s a hard nut to crack for anyone who’s business model requires unique machinery. I know, because I once got caught having to personally design and build a piece of factory equipment for making a product I had patented because I could not nail down and substantiate such figures. How do you know what something costs that has never been built? The money got bored and moved on and I was left with the completed machinery and no money for raw materials. Heck, the machine would spit out almost s liter of sawdust per second but I couldn’t even afford a dust collector for the beast. Sad but true.

Andrew F said...

This is a bit off topic but I was thinking about this as I was falling asleep a few days ago and thought I might as well share it.

The headway concern expressed by many seems to only really be a serious control issue for merges. Diverges seem pretty straightforward to me and are not particularly risky (short of some mechanical failure that sends the first bogie one way and the second another way on a multi-bogie system. Such a failure would be like hitting a brick wall.) Following seems straightforward enough that some feel comfortable enough doing it with cars/trucks on highways with sub-second headways. (I think this is at least a little bit nuts--one blowout and you have an instant unavoidable pileup.) Merges are tricky from a timing perspective, and are always the part that critics point to to say that large headways are required for safety. 'Imagine 0.5s headways with vehicles merging like the teeth on a zipper'. I guess the obvious answer, and one that would allow systems to scale without significant capacity problems, would be to vary the headways depending on the operation being undertaken.

Visualizations of PRT usually show vehicles evenly and randomly spaced over the guideway. It strikes me that platooning is what would happen in practice, with platoons breaking apart and reforming whenever sorts are required (at diverge points, generally). Within these platoons, the headways could be subsecond, and at merges, we ought to keep the headways between platoons in the several second range. This would require some negotiation for the traffic passing through a merge over the next 30 seconds, say, to ensure that oncoming platoons speed up/slow down to meet headway requirements and leave room for each other. You could even have insertions in the middle of a platoon by adding a 6 or 7 second gap in the middle of it in advance of a merge. These insertions could actually make a decent amount of sense, in terms of grouping pods with similar destinations to simplify future merges?

As I write this, I'm not sure if this is exceedingly obvious. I haven't seen this discussed really in the literature.

Dan said...

There’s probably more discussion of these kind of control issues in this blog than can be found anywhere else. The best overview I can think of is from Anderson, and even there, he doesn’t go into all that much detail. The system is, after all, a trade secret. He particularly references sub-second headways, at least. The system, or an off-spring of it, is presumably the one used in Skyweb Express.

There are so many issues that come up when it comes to maximizing efficiency of extremely dense PRT traffic that I have concluded that some form of AI will be what is required, shared between affected vehicles. In such a system, various advantageous and disadvantageous factors would be weighted differently, and those weightings could be tinkered with in simulations to arrive at the best result.

I think nobody has really thought that far ahead in much detail because it’s a classic “cross that bridge when we come to it” sort of situation. It does touch on the autonomy vs. central or regional control issue, in that it is using destination information. Personally, I prefer local but shared decision making with information “from above.” Your comment on the bunching, if not platooning, of vehicles reminds me of issues raised in the “Anti-Gravity” post back a few months.

akauppi said...

( A comment of mine got gulped up by Blogger. Here are the highlights. )

The Wired magazine has an article on what happened to AI. That it's already all around us, but not in the 1980's form. Stock market, automobiles, automated warehouses. Algorithms that are alien in the way that they don't even try to mimic the human. Yes, I agree.

And fuzzy logic is the weight-based decision-making mechanism that Dan refers to. It's not really AI, but just a way to consider multiple "fuzzy" variables and try to pick an optimal. Dead simple, really.

Dan said...

You are right, Akauppi. I used the term AI way too loosely. It's really not AI at all. I really meant that the decision making would be done locally between vehicles, rather than by some all-seeing central computer, so in the most unscientific sort of way, the vehicles would be "thinking" for themselves and reaching a consensus, based on rules that could be weighted. AI suggests that they could learn for themselves, which isn't what I meant at all. Interesting article..
I always buy a "Wired" at the airport when I fly.

Andrew F said...

AI is a term that is used pretty loosely, Dan, so you are forgiven. Let me tell you, I was disappointed when I peeked behind the curtain on AI and found out it is mostly things like search, rudimentary decision classification techniques, etc. Not much in the way of actual intelligence at all, really. More the illusion of it. I'd say describing an agent-based control system as 'AI' is fair game. It wouldn't be the first time it's happened. You have to think pretty hard about what that actually means though. Do we just use a greedy algorithm, where each agent tries to locally optimize, and see if that leads to a reasonable behaviour? I'm sure there's plenty to learn from how internet routing is done, but with one important difference: packets don't collide and kill people.