Why things change…..Hyperloop

This is another in a series of semi-regular pieces I do in which I try and set out what science and engineering can achieve and what it can’t do. In short, try and sort out science fact from science fantasy.

Figure 1: The proposed hyperloop system {Businessinsider.com, 2013 http://www.businessinsider.com/elon-musk-hyperloop-plan-2013-8 ]

Figure 1 – The proposed hyperloop system [Source: Businessinsider.com, 2013]

Last month, Elon Musk made a somewhat surprising announcement of a plan to build a 570 km Vactrain system from Los Angeles to San Francisco, for the bargain price of $ 6 billion. The media quickly reported it with great enthusiasm, however this has turned to gradual and increasing scepticism as they’ve inevitably asked engineers their opinion and they’ve generally taken the view that it borders on science fiction, if not science fantasy.

Certainly it has to be said, Elon Musk has a habit of proposing wacky ideas and proving his critics wrong. When he first launched an electric car business he was told that he was crazy and consumers wanted petrol powered cars. Of course he has since proven that there is a market for electric cars, much to the annoyance of some republicans who’ve been trying to use legislation to block the business.

Figure 2 – Tesla Motors [Credit: Tesla Motors Inc., 2011]

Figure 2 – Tesla Motors
[Source: Tesla Motors Inc., 2011]

Musk also took on the space launch establishment. When he founded SpaceX he was told that space was expensive because that’s just the way it is. Now while his current launch system is something of a cut down version of what he first proposed (the initially proposed five engine reusable first stage has ballooned into a nine engine expendable stage) and he’s now quoting costs considerably higher than originally forecasted (the original proposal worked out at around $2,800 per kg/LEO while SpaceX is now offering launches at around $4,100 per kg/LEO). But equally SpaceX are offering to launch payload for at least half (or less) the price of the established space industry and they have demonstrated that private industry can develop both a man-rated rocket and capsule for a fraction of the costs often quoted by the established space industry. So credit must be given where it is due.

Figure 3 – SpaceX, Falcon 9 with Dragon Capsule launch, In with the new, out with the old! [credit: Space.com, 2012 http://www.space.com/15751-spacex-dragon-capsule-launch-pictures.html ]

Figure 3 – SpaceX, Falcon 9 with Dragon Capsule launch, In with the new, out with the old! [Source: Space.com, 2012]

However the hyperloop proposal has the critics coming out of the woodwork for good reason. Firstly there is the sheer scale of the logistics. Hyperloop proposes a train line capable of over 1,000 kmph, 760 km’s long all for a cost of just $ 6 billion (about $7.9 million per km). By way of comparison Britain’s High Speed 2 (Phase 1) system will be 192 km’s long and is estimated at a cost of £20 billion (or about $35 billion dollars) or about $182 million per km (perhaps an extreme example given the high value of land over which the line will cross). The rival California High speed rail system proposes to build a conventional high speed rail for $ 68 billion (about $89 million per km, but it will involve upgrade of existing track where possible). Even the French quote a preposterously low figure of $15 million per km for their lines, and they often use various tricks to cut down on costs (such as long detours around difficult terrain, or simply upgrading existing lines to take HS trains, not an option for the hyperloop). So Hyperloop is proposing to be able to build a line capable of 4-5 times the speed for between 1/23rd and 1/11th the relative price.

Figure 4 – Hyperloop is intended to be solar powered and mounted on pylons [Credit: Dezeen, 2013 http://www.dezeen.com/2013/08/13/elon-musk-tesla-spacex-unveils-open-source-designs-700-mph-hyperloop/ ]

Figure 4 – Hyperloop is intended to be solar powered and mounted on pylons
[Source: Dezeen, 2013]

Indeed it is worth remembering that when it comes to the above quoted figures, those behind such projects are well aware that many rail, road and infrastructure projects have a nasty habit of coming in way over budget. Casing point, in the UK an upgrade programme of the railway lines into London saw them spend £850 million on upgrading Reading station, with a similar amount spent on Rugby Station (that’s 20% of Hyperloop’s budget blown just upgrading the track around one station!). Now before anyone starts going on and on about the excesses of public bodies squandering taxpayers money, the railway’s in Britain have been in private ownership since the 1990’s and both of these upgrades were performed by private contractors.

In part the reason for the $68 billion forecasted cost of the CHSR is the fact that local communities have demanded tunnels and viaducts at places where strictly speaking they aren’t needed. There’s also the issue of land purchases to consider as well as stations, car parks and public transport hubs at those stations (a large chunk of the cash for HS2 is to be spent on this) to service the line.

Furthermore, its worth considering that the only existing Maglev of comparable scale is the Shanghai system, which is only 30 km’s long , has a top speed of 430 kmph, yet it took two and a half years to build and cost $1.33 Billion or about $43 million per km (probably a vast underestimate, the Chinese government has a habit of hiding cost overuns seen as embarrassing to the regime). So Hyperloop proposes a system 25 times the distance at twice the speed, built 5 times faster and at least 1/6th the relative costs, even though the labour costs in China are vastly lower than they are in the US.

And the chances are that if anything the costs of the Hyperloop proposal will be a lot higher than those for a conventional railway, or indeed even a conventional maglev system. A normal railway line after all doesn’t need an enclosed, pressurized cylinder around the train. Much of the line will have to be cut and cover tunnels or viaducts, with fewer options to “go around” obstacles. This is in part enforced by the fact that maintaining an air tight seal is going to be difficult (particularly so if the tube has to follow lots of contours as that means more joints and opportunities for leaks).

While opting away from a complete vacuum does make things a little easier, it still means you’re going to have problems with air leakage into the tunnels. Consider that Brunel trialled a similar idea, his atmospheric railway, back in the 1840’s. He only enclosed a small “pusher” element rather than the entire train and even this proved too much for technology at the time. In all probability you’d need pumping stations every couple of km’s along the route to regulate the air pressure and pump out excess air.

Figure 5, Brunel's Atmospheric Railway [Credit: ikbrunel.org.uk http://www.ikbrunel.org.uk/atmospheric-railway ]

Figure 5 – Brunel’s Atmospheric Railway [Source: ikbrunel.org.uk]

Then there are other issues, such as the problem of choked flow and transonic turbulence around the vehicle. This has always been a problem for high speed flows (aircraft, gas turbine engines, engine manifolds, gun barrels, etc.). While the Hyperloop team claimed to have resolved it, I’ve not seen anything that indicates rigours scientific study of this problem. This is significant because the sort of buffering forces such a train would be exposed to are enormous, and the slightest instability could lead to enormous friction forces on the vehicle (which would require a lot of energy to overcome) not to mention discomfort for passengers (NVH related issues) or even catastrophic failure.

Also there is some debate as to the energy efficiency of maglev’s. Although they are potentially a lot faster than conventional railways. Mc Kay (2005) suggests Maglev’s are slightly more efficient than conventional high speed rail, but of course the figures for high-speed rail is backed up by decades of experience & thousands of miles of track, while the data set for maglev’s is considerably smaller (and under fairly tame conditions).

Maglev’s are also a lot less fault tolerant, vastly more expensive to install and harder to maintain (this is what brought down the Birmingham Airport Maglev). There are also issues such as noise to consider. And putting a maglev in a tube will if anything make these problems worse not better.

The Hyperloop also seems to consist of lots of small pods rather than a multi-carriage train. One has to question the wisdom of this. The whole reason why trains, planes or buses tend to be as large as economically possible on any particular route is to maximise economies of scale (i.e. it takes 2 pilots to fly a ATR-42 with 50 passengers, same number as it takes to fly an A380 with 500 on board), as well as to reduce the number of vehicles in service. The logistics of having a hundred small pods in transit, against a half a dozen multi-carriage units is obviously going to be a lot more complicated.

Remember in the real world things go wrong. Trains break down, faults develop with the track (there’s a huge problem in the UK with thieves nicking signal cable), “passenger incidents” (suicides, heart attacks, bomb alerts, people being slow to board/get off), all of which serve to slow down trains. On the highly congested UK West coast line, a single incident can have a ripple effect such that it adds multiples of delays to other trains (5 minute delay to one train can lead to a 30 minute delay to a following train) and can lead to whole services having to be cancelled. And accidents involving maglev’s have proven to be a lot more serious, largely due to the combination of high speed and lightweight construction. Also, returning to the transonic flow issue, any fast moving object through a narrow tube is going to produce considerable wake turbulence, even in a near vacuum. You would have to leave a considerable distance between pods to even out these effects.

Figure 6 – Technology readiness levels http://en.wikipedia.org/wiki/Technology_readiness_level , its important to know where you stand! [Credit: NASA, 2004 http://as.nasa.gov/aboutus/trl-introduction.html ]

Figure 6 – Technology readiness levels, its important to know where you stand!
[Source: NASA, 2004]

However, perhaps the big problem I can see is the background engineering and research. I’m reluctant to argue either way as to whether or not the Hyperloop is technically feasible, as there simply isn’t enough research data in the field to give a reliable answer. Completing all of that research to fill in these gaps would take both time and money. The German and Japanese Maglev systems (both considerably less ambitious) consumed many tens of billions to develop. And no bank or investor is going to commit money at this sort of scale until such research has confirmed the feasibility.

Don’t get me wrong, I’d love to say that a solar powered mass transit system was a great idea. However there are some serious holes in this proposal. Not insurmountable problems, but certainly some major question marks that need answers. I certainly would be in favour of research on this area, but as things stand, if you want to move people rapidly between the Californian coastal cities, I’d stick to conventional high-speed rail for the time being.

About daryan12

Engineer, expertise: Energy, Sustainablity, Computer Aided Engineering, Renewables technology
This entry was posted in economics, efficiency, energy, future, transport. Bookmark the permalink.

12 Responses to Why things change…..Hyperloop

  1. Your comments about the impracticalities inherent in the Hyperloop design are spot-on, in my opinion, but your casual listing of conventional maglev’s faults really misses the mark.

    When you say, for instance, that maglevs “…are also a lot less fault tolerant,” you offer no proof because, I suspect, there isn’t any. Both the German and Japanese high-speed systems are at least doubly fault-tolerant on the things that matter, as in the levitation/guidance magnets in the German case. And when you say maglevs are “…vastly more expensive to install…” the only case that can be made seems to be the Shanghai line, for which reliable numbers are hard to verify. Further, there’s no evidence that maglevs are harder to maintain than comparable high-speed rail systems (forget the Birmingham Airport Maglev — it was so reliable that by the time replacement parts were needed the suppliers were gone). And there really are no “…issues such as noise to consider…” since the study you cite (not studies, plural, since I have found only the one, done so poorly as to suggest such an issue exists) has been roundly criticized by experts.

    Oh. One other thing. I’d really like to see the evidence supporting your statement that the German and Japanese high-speed maglevs “… consumed many tens of billions to develop.” That’s a real eye-opener of a contention, one that I’ll chalk up to youthful exuberance after your having built up such a head of steam on the potential failings of Hyperloop.

    • daryan12 says:

      Fault tolerance is pretty simple to understand if you look at how a Maglev works v’s a normal train, lots more systems, which often have a very narrow optimal range of operation, needing careful control, more things that can go wrong, etc.

      In a similar vein this is why I tend to be skeptical of the Fuel Cell car. Sounds like a nice idea, until you realise just how sensitive FC’s are and how hard they are to manage and install (e.g.. a few degrees of temp or Volts out is the difference between optimal performance and catastrophic failure). Another example, because all the different FC support systems run at different voltages you often end up with 3 or 4 different DC to DC converters, indeed the electronics and support systems often take up more room than the FC!

      Oh, and while the suppliers often claim efficiencies of 50-60%, I’ve found that the balance of plant performance is closer to 40-45%, which isn’t far off the performance of a GT or high performance diesel engine (typically about 30-40%).

      And that’s even before we consider the thorny issue of hydrogen. Now while I reckon long term we will develop a renewable H2 grid, we’d likely use it for the same things we use NG for….and generally that doesn’t include cars! Indeed, given that most H2 comes from NG it would seem more sensible to me in the short term to just build a series hybrid car and fit it with CNG tanks and avoid the whole mucking around with H2 and FC’s.

      Don’t get me wrong, I don’t think there’s any reason we should stop research into FC’s or maglevs. Indeed I’d point out that given the trillions spend globally on roads, cars, trains, planes, etc. a couple of billion spent on alternatives is practically a rounding error. Both technologies are gradually improving, in particular as new material technologies mature.

      But equally, if and when the time comes to commercial deploy such technologies, the performance has to be there and they have to be cost effective. And extraordinary claims of performance requires extraordinary proof. Otherwise the public won’t buy it.

      • I agree with you: fault tolerance is pretty simple to understand, because I do. And if you look at how a Maglev works compared to a conventional high-speed train, maglev effectively has no moving parts — no wheels, gears, drives, pantographs, etc. — and is doubly-fault tolerant in key subsystems, which often have a very wide optimal range of operation, needing careful control only in monitoring dynamic air gaps and vehicle positioning. There are fewer things that can go wrong with maglev than with a conventional high-speed train and, further, the things that can go wrong are accounted for in the basic production design.

        I realize that the above could be seen as a nuisance, since your expertise is more in line with energy and other-than-maglev pursuits, but your blog seems to be less run-of-the-mill than most I come across, so I wanted to at least make the points for the maglev side of things.

      • daryan12 says:

        “no moving parts”

        This is what the FC people are always saying to me. And my response is, what’s wrong with moving parts? The engine of my car averages 3000 rpm when in cruise, so that’s a couple hundred million rev’s its done in its life time and its still in perfect working condition. By contrast, I’ve seen FC’s written off after a few hundred hours of operating, often under conditions nowhere near as brutal as to what I put my car engine thro.

    • daryan12 says:

      Oh and my “tens of billions” comment probably relates back to a figure banded around during debates I heard in Scotland back in 2007, when there was a brief consideration of installing a maglev between Glasgow and Edinburgh.

      Given that Transrapid existed for at least 30 years, if we assume a couple of thousand employees, apply the normal costs of doing research, throw on a couple of research grants over that time, assume a similar figure for Japan and all in all I suspect you’ll wind up with a figure, maybe not tens of billions on reflection, but certainly one with nine zeros behind it.

      Indeed, its worth noting that the Glasgo-embra link was eventually shelved as it was argued that it would be far more cost effective to spend the money upgrading public transport in both cities or improving existing rail links (the maglev might improve city to city journey times, but these would be rendered moot without quicker front door to city centre journey times within Glasgo/Embra).

      Similar arguments are being made as regards HS2. That we’d be better off upgrading existing track, improving public transport and fixing various bottlenecks in the UK train network than building an entire new railway line.

      I would note that I don’t entirely buy this line of reasoning above. The whole point of HS2 is that the West coast line is at breaking point and the UK needs another trainline. However, I suspect if you proposed a maglev instead of HS2, that would be an even harder sell. Very fast yes, but is it value for money?

      • Again, not to beat a dead horse, but calculating the sunk costs for German and Japanese maglevs would include at least 30 years of a several hundreds — not thousands — of employees, and I suspect you’ll wind up with a figure of one or two with nine zeros behind it.

        And I do remember the UK Ultraspeed debates from several years ago. Dr. Alan James and his staff, with support from the German technology supplier, compiled some of the most compelling business cases for maglev vs. high-speed rail that I’ve seen, allowing for the capital cost for UK Ultraspeed to come in 50% lower that HS2, with better operational performance, a smaller fleet and wider system coverage.

        This really isn’t the place to revive the debate, but in the end, as you suggest, we can certainly agree that the successful sale of maglev would be even harder to pull off than the sale of HS2, which has lately come into real dispute.

  2. John Harding says:

    Daryan, you have made a very important contribution to this subject. All these issues need considerable investigation. However I disagree with some of your points regarding maglev- Shanghai reliability and safety has been exceptional good. With regard to energy efficiency, especially compared with modern trains like Japan’s Nozomi*, actual energy consumption for the Shanghai maglev is 132 Wh/seat-km (although this is for a 30 km trip at a top speed of 430 km/h). However energy intensity for the proposed 420 km Anaheim to Las Vegas project with a top speed of 500 km/h is estimated by Transrapid to be 95 Wh/seat-km. The Japanese superconducting maglev system, presently abuilding, is estimated to consume 70 Wh/seaat-km with a maximum speed of 500 km/h. All these trains operate at atmospheric pressure, which is responsible for much of the energy used. However in the case of the Transrapid examples there appears to be an enormous amount of excess dissipation, presumably caused by inaappropriate power invertors, which should be replaced with equipment that can produce pure sine waves. No doubt this will add to the capital cost but only in segments where large thrusts are required.

    * (18Wh/km-seat is the value calculated from actual performance of the N700 series on the Tokaido Shinkansen at 270km/hr (average of 210-220km/hr).

  3. John Harding says:

    Of course there is much more to it than the number of moving parts-how about the total number of parts, period, Computers have billions if you count every element. Marchant calculators had hundreds of moving parts which often failed-a repairman’s delight. Each system has to be evaluated on its own merits. New systems will have start-up problems but it will be hard to predict eventual reliability.

  4. Pingback: The Hyperloop hyperbole | daryanenergyblog

  5. Pingback: The Hyperloop hyperbole | daryanblog

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.