High Speed Rail’s Strong Safety Record

Jul 27th, 2013 | Posted by

The driver of the Spanish train that derailed and killed 78 people in Santiago de Compostela this week is being formally accused of reckless homicide by the government:

Visiting Santiago, the Spanish interior minister, Jorge Fernández Díaz, said there were “reasonable grounds to think he [Garzón] may have a potential liability”.

“He has been detained since 7:40pm on Thursday for the alleged crime of reckless homicide,” Fernández said….

According to reports in the Spanish media, after realising the magnitude of the disaster Garzón said: “I fucked up, I want to die.”

Spain’s track operator, Adif, has confirmed that the section of the track where the derailment occurred did not have an automatic braking system, and that drivers are expected to slow the trains to 80 km/h. Numerous reports indicate that the train was traveling much faster than that, perhaps as fast as 190 km/h.

This is clearly a horrible tragedy. But it’s also a very uncommon one. Yonah Freemark, who runs The Transport Politic, wrote an op-ed for CNN yesterday that shows HSR’s strong record of safety:

Indeed, there have been a number of high-profile rail crashes over the years. In 1998, 101 riders were killed in Germany when a train crashed in Eschede. And in 2011, two of China’s newest, fastest trains slammed into one another, killing 40 people.

But those are exceptions to the rule, caused by poor maintenance and monitoring.
Rail, high speed or not, is one of the safest ways to get around. According to a National Safety Council review of 10 years of transportation fatalities, for every mile traveled, car drivers and passengers are more than 10 times as likely to die in accidents as passenger rail riders. In 21 years — between 1990 and 2011 — the Bureau of Transportation Statistics shows that nearly 900,000 people died in highway crashes, while fewer than 15,000 died in train collisions.

Other countries’ experience shows that high-speed rail can be even safer than the much slower U.S. trains. The bullet trains that zoom through France and Japan, for instance, testify to the astonishing safety offered by well-managed rail services. Each nation’s system has been in operation for more than 30 years and provided billions of rides.

Yet thanks to advanced safety systems and extensive maintenance, no passengers — zero — have died as a result of a high-speed train crash in either country. Improvements in the design of German trains and a review of maintenance operations in China have also prevented repeats of previous train accidents in those countries.

When you add together the Eschede, Wenzhou, and Santiago disasters, that’s still just three major crashes and less than 300 dead for high speed train travel. Air travel is still very safe and yet it has had many more crashes and a much greater death toll. Both the plane and the train are statistically far safer than the automobile, of course. It is estimated that 1.3 million people die each year in car crashes. That’s 4,000 times as many people than have died in those three HSR crashes.

On Thursday I laid out some important points for California to consider as it builds out its high speed rail system. All HSR operators should strive to build the safest infrastructure possible, rather than let austerity-minded politicians pinch pennies or let angry neighbors scale back a project’s scope. But even then, HSR still remains an extremely safe way to travel.

I’ve always wanted to visit Santiago de Compostela. And when I do, I’ll be taking the high speed train from Madrid to get there.

  1. Keith Saggers
    Jul 27th, 2013 at 10:43

    If the driver in the crash is going to be investigated then the operator Renfe should also be investigated.
    I have been on UK trains when the driver has mysteriously “made up” lost minutes caused by delays.

    Paul Dyson Reply:

    It’s called “recovery time.”

  2. n bluth
    Jul 27th, 2013 at 11:08

    The comment section on Yonah’s CNN opinion is toxic. No surprise there.

    Peter Reply:

    Haters gonna hate.

    VBobier Reply:

    Ain’t that the truth…

    D. P. Lubic Reply:

    What was shocking–though perhaps it shouldn’t have been–was the level of ignorance by the commentors. Highways a pure profit (yeah, we know better here, even John N. admits the road system is subsidized), no rail system makes money (tell that to the freight boys, not to mention some of the people with Shinkansen and TGV lines, not to mention the others that cover their operating costs if not capital costs), some people suggest a train should be a hovercraft affair, others think they’re totally obsolete even for freight, some are looking for Elon Musk’s tube train to make both trains and cars obsolete, anti-union comments, anti-government comments, and all that. Whew!

    At the same time, it’s still remarkable that we have the volume of comments. I’m not sure that would have been the case a few years ago, which means this is a high visibility thing nowadays.

    Keith Saggers Reply:

    Right on

  3. Keith Saggers
    Jul 27th, 2013 at 11:24

    •August 1, 2013 Board Meeting Agenda

    •Agenda Item 2: Proposal to Issue a RFQ for the Bakersfield to Palmdale Project Section RC Contract
    •Agenda Item 2: Draft RFQ Environmental/Engineering Services for the Bakersfield to Palmdale Section
    •Agenda Item 2:Draft Resolution #HSRA 13-15 Environmental/Engineering Services for the Bakersfield to Palmdale Section
    •Agenda Item 3:Proposal to Issue a RFQ for the Los Angeles to San Diego Section RC Contract
    • Agenda Item 3: Draft RFQ Environmental/Engineering Services for the Los Angeles to San Diego Section
    • Agenda Item 3:Draft Resolution #HSRA 13-16 Environmental/Engineering Services for the Los Angeles to San Diego Section
    •Agenda Item 4: Proposal to Issue a RFQ for the Sacramento to Merced Project Section RC Contract
    •Agenda Item 4: Draft RFQ Environmental/Engineering Services for the Sacramento to Merced Section
    •Agenda Item 4: Draft Resolution #HSRA 13-17 Environmental/Engineering Services for the Sacramento to Merced Section

    Paul Dyson Reply:

    Can’t understand why they are issuing these RFQs when the useless bastards haven’t even figured out how to get it past Wasco.

    synonymouse Reply:

    The collapse of leadership here is just appalling. I know it sounds like a broken record but it appears that Jerry Brown is shut down for the duration and we are locked on “stay the course”, led by utter dullards, your “useless bastards”. Do you know who fingered Van Ark and talked Brown into firing him? That move pretty much derailed successful hsr in California.

    With regard to the Santiago de Compostela disaster the most important questions relate to the very high level of damage to the cars and subsequent high number of casualties.

    1. There was no solid obstacle, such as a bridge abutment or other train, in the way of the train.

    2. Would the damage have been greater or less if the derailment had occurred on open surface ROW?

    3. Same question had it occurred on an aerial.

    4. What would have happened in a head-on with an obstacle?

    Of course everything possible has to be done to avoid derailments entirely but the residual key safety issue is are the body shells structurally robust and survivable. If they are easily squashed passenger rail is not that safe.

    Paul Dyson Reply:

    Doesn’t the concrete retaining wall count as a “solid obstacle”?

    synonymouse Reply:

    I would call that your lateral containment system, which probably would have worked better if you had jacobs bogies that make the cars hard to break apart.

    Imagine the damage in an open field. The jack-knifed cars ahead become your static obstacle since they are no longer moving as fast in your direction of travel. With no embankment the cars would roll farther and these ones don’t roll well obviously. The one car that went straight up into and landed at the top of the embankment would on an open field have fallen instead straight down, what, 20 or so feet to the grade.

    Max Wyss Reply:

    To some extent, the Jacobs bogie argument is propaganda. The Talgo system is also stiffening and reducing the chance of serious jacknifing.

    Something I noticed from the pictures. If you look which part got most damaged of the carbodies, it is the entrance parts, but not the passenger compartment. The entrance part (there is one per carbody in Talgos) seems to be designed as a crumple zone, absorbing quite a bit of energy. This is can be seen at the other end of the carbodies, which did not show that much damage.

    Jacknifing would most likely have happened with “conventional” cars, where the connection between the cars is not stiff at all.

    On the other hand, at this much overspeed, any vehicle would fail. Then, it depends on how well designed it is, whether passengers can be protected, and to which extent.

    Clem Reply:

    The train hit the exposed end of the retaining wall seen at 0:23 in this video shot from the cab.

    The driver is right, we are only human. He f***ed up, but the system should have been designed to catch his error. There is no way you rely entirely on one human to slow suddenly from a long stretch at 200 km/h to just 80 km/h without relying on some form of electronic supervision. From all indications, this was probably a wreck waiting to happen.

    John Nachtigall Reply:


    Joey Reply:

    Aerial: Wenzhou
    Head on: Chatsworth

    It’s usually better to implement the necessary safety measures to avoid these kinds of crashes anyway, rather than trying to make a tank on rails which will inevitably fail in a crash anyway.

    synonymouse Reply:

    Aluminum beercan cars are nowhere close to a tank.

    BART’s Rohr cars were so structurally deficient they went swayback from their own weight. You could looks from the Muni level and see the crinkles on the roofs of the “Cyclops” cars.

    Max Wyss Reply:

    Not being familiar with the design of the BART Rohr cars, I may be wrong. But there is a fundamental difference in designing a steel carbody and an aluminium carbody. If you use aluminium in “steel fashion”, things will fail. But if you design adequate to the material, you get a very good result. So, I presume that Rohr took a “classic” design and used aluminium… problems guaranteed (and they were not the first ones who made this experience…)

    synonymouse Reply:

    My understanding Rohr-BART-Bechtel went strictly monococque, ie. beercan, to achieve lightweight. Bechtel was deeply ashamed of building a railroad, thus coined the cute term “duorail’, and was indulging in gratuitous, weird stuff to pose innovative.

    Max Wyss Reply:

    Not wanting to do fingerpointing, and without knowing more, I may be completely wrong, but it looks as if they might have been looking for “best practices” in the US only, but not worldwide… Aluminium using pressed profiles, properly welded can not produce clearly visible wrinkles, for example.

    isgota Reply:

    Well regarding Talgo cars’ structures I can bring some first-hand info from my experience with their CAD/CAE models than I worked with near 2 years ago.

    Those are not monococque aluminum “beercan”, far from it. They are extruded aluminum made of 2 shells welded together by a pretty complex web of stiffeners.

    Those Talgo VII cars, like the ones that crashed in Santiago, are pretty strong and meet European norms, even if they are very light. What’s more, Talgo apparently has been able to make them FRA compilant in their Series 8 cars with not huge modifications, not many European train buiders can say something like this.

    This pdf article says the modifications only added 1 extra ton per carbody, to a total of 18.7 tons, still they can resist 800,000 pounds of compression-force.

    Best regards.

    synonymouse Reply:

    @ Joey

    These cars seem to be particularly weak. Chatsworth, I think, would have been much worse. Like a wooden passenger car hit by a steam locomotive.

    morris brown Reply:

    @Paul Dyson:

    You write…

    “Can’t figure out why they …”

    Simple, funny paper simply. Its all about the Money . Spend as much of Prop 1A funds as you can as fast as you can.

    How much more ridiculous can it be to be spending money on Sac to Merced and LA to San Diego when any possibility of their being construction on these routes is at least 2 decades away.

    Travis D Reply:

    The precise routing is to be worked out but what seems certain is that it will proceed to Palmdale once east of Bakersfield. As such it makes sense to start on the engineering now so that it can be built as soon as money is available.

    And the Tejon Boosters just need to live with that fact.

  4. John Nachtigall
    Jul 27th, 2013 at 12:01

    How does a line built in 2011 not have automatic safety features? I could understand it if it was an old legacy line, but all the information I have read about this accident indicates this line was built within the last 2-3 years.

    This is a severe failure in management. The speed of the trains can be measured in a variety of free or almost free ways. You then weed out the drivers not following the rules so this never happens. It not complex, its not expensive, and it prevents this problem. Long haul truckers are monitored for speed by their companies all the time using a variety of methods, how is it that one of these most advanced and the largest HSR network in the world does not have this in place. Disgraceful.

    That said I think it is an object lesson for CAHSR to learn and understand, but not necessarily a precursor of things to come if they learn the lessons here.

    PS. Saying air travel has had “many more deaths” and comparing raw numbers on cars is just misleading. If you want to compare deaths per passenger mile then pull the stats. Cars are not 4000X less safe than trains.


    Paul Druce Reply:

    The line uses legacy portions in certain areas, including where the crash occurred, and that was built back in the 1970s I believe.

    Joey Reply:

    The segment in question is a transition between the new HSL and legacy track leading to the station.

    Joey Reply:

    Richard informs us that the crash occurred about 4 km beyond the end of ERTMS territory.

    Joey Reply:

    Okay, so reading Richard’s next comment, it appears that the train in question was not under ERTMS supervision. That will start when it’s converted to standard gauge, connected to the rest of the high speed network, and its top speed raised from 220 km/h to 350 km/h.

    Richard Mlynarik Reply:

    Not exactly: ETCS is active on the high-speed parts of the new cutoff line between Ourense and Santiago de Campostela. Again see the schematics shown here and watch a cab ride here. ETCS Level 1 is between PK 1.857 and PK 80.149 on the relevant direction and track.

    ASFA, the legacy national train stop enforcement (but not speed control, other than requiring trains operate at under 200kmh) signal system is also active on the entire line, including though the ETCS/ERTMS-equipped section.

    (As an irrelevant aside, you may not be aware that this is an Iberian broad gauge (1668mm) line, unlike all the other standard gauge new HS lines in Spain. But this does not affect the signalling systems, except in that no 1435mm-only ETCS-only train could every run there.)

    There’s more official ADIF information on national ETCS than I can possibly manage to understand or read in Consigna C Experimenta No 17 of 2012-02-16.

    The matter of which trains communicate with ETCS/ERTMS is a different one. All (as far as I know, but I am not certain of this) contemporary mainline Spanish passenger trains talk ASFA, and all new ones are ETCS equipped, but not all new trains have had their ETCS mode formally approved for life safety use, and it is possible to an ETCS-approved train to operate (at reduced, sub-200kmh) speed on the dual-signalled line in ASFA mode as a fallback.

    The train in question did not run in ETCS/ERTMS mode. Even if it had, the site of the accident was 4km beyond the end of ETCS territory and hence not under ETCS speed control.

    (One thing I do not know and have never see mentioned anywhere is what static upcoming civil speed limiits static Eurobalises towards the end of the ETCS section might have sent to an ETCS-communicating train. The system allows a balise to say “300kmh limit at 0m; 80kmh limit at 4081m”, and that might affect the operation of a train exiting the section, but I have no idea. ASFA/ETCS transition must be at under 200kmh by definition, but that’s irrelevant.)

    As for the question “How does a line built in 2011 not have automatic safety features?”: it’s important to understand (if one wishes to understand, not just score points, or not just to suck down $250 million of public funds) that railways are a system with decades-old legacy “installed base” and muti-decadal “software upgrade” paths. The new cutoff high-speed line is glued into low-speed tracks and low-speed stations at both ends, and those tracks and stations are used by older trains that never run on the high speed line. Similarly, the trains that run on the high-speed line must operate on older tracks. There’s no way to have a “big bang” “flag day” in which the network of the entire country converts to a new signalling system, no more than it is possible to change the track gauge from 1668mm to 1435mm, or to change all American roads to be left-hand running. (Hell, it’s even been a challenged in Spain and everybody else to upgrade from older versions of the ETCS specifications.)

    It’s also possible — not just possible but likely — that attempting to interface, or to aggressively replace, existing legacy systems with something “newer and better” will lead to worse problems than are being “solved”. (That’s why one would never ever ever ever choose to introduce multiple life-safety-critical systems, and why everybody in any way involved in Caltrain’s $250m CBOSS systematic fraud and scam needs to be lined up against a wall.) Putting the shiny new safety system in place as well on the shiny new tracks is hard enough; but requiring it to be deployed throughout the entire network, including on every track of the old lines and stations is a route not just to economic ruin but also an invitation to technical and safety disaster.

    There are certainly many lessons to be learned from this terrible accident, and it’s abundantly clear, world-wide, that the old regime of train drivers relying on “route knowledge” of maximum line speeds isn’t appropriate, but it’s not as simple as declaring that tomorrow all trains on all lines will limit their own speeds. Would that it were. No doubt large systemic changes will be recommended and made, just as they are after air accidents, but just as with aviation, some things are immediately implementable, some take years, some take decades; some things are cultural, some operational, some physical; some are negative cost, some cheap, some cost tens of billions.

    John Nachtigall Reply:

    With GPS and microprocessors why not control the train, not the track. The GPS knows where you are (and can be enhanced to provide location with inches look at modern transit equipment) and the software tells you how fast you should be going.

    So you can update the trains 1 by 1 instead of the system.

    This is not an impossible problem to solve. The real problem is the rail industry worldwide is as about as open to new and different technology as the Pope is to contraception. They acknowledge its existance and wish it would go away.

    Richard Mlynarik Reply:

    * “GPS” is a bad knee-jerk solution. (Unsurprisingly, the local jerks are all over it.) Knowing location safety-critically and to sub-metre resolution (which is what one needs to know which of several adjacent tracks is occupied) and continuously is a big hairy deal.
    The bad seemingly old-fashioned one of the tracks communicating some information to passing trains isn’t as bad and costly and old-fashioned as you might imagine.

    * The safety-critical database distribution logistics problem of keeping all trains up to date is far from trivial in the real world.

    * Static information about approaching grades and curves (that might be solved by GPS plus on-board track database) is a small part of it. Knowing dynamic information about the location of obstructions (other trains, signals, missing bridges) is a bigger piece.

    * Temporary speed restrictions (newly maintained track, sub-maintained track, workers in the vicinity, extraordinary temperature or wind, etc etc etc) are a huge and rapidly-changing database which is hard to reliably distribute among all vehicles.

    * OK, you know exactly where you are (using GPS MAGIC!) and exactly where other trains are and how they’re moving. Who decides who makes what move to avoid conflicts? (Even achieving a distributed quorum to say “EVERYBODY EMERGENCY BRAKE RIGHT NOW!” isn’t trivial.)

    * Even translating static information into do-not-exceed train speed parameters is hard in the real world. For just one thing, just “knowing” what the train being controlled is (How long am I? How much do I weigh? How is the weight distributed along me? How is braking capacity distributed along my length?) is much harder than one might think at first, though generally simpler with contemporary unit-sized passenger trains.

    There are huge problems with the “rail industry worldwide”, sure. But a magical inability to make overnight continent-spanning changes in life safety critical systems (hardware and software) distributed over hundreds of thousands of miles and hundreds of thousands of pieces of equipment isn’t a unique problem, or the worst problem. (How’s the US Air Traffic Control modernization doing again? What, controllers with pieces of paper? What, aural communication of critical safety data? What, no Heavens!)

    Yeah, it sucks. That’s engineering. There are better and worse engineers and better and worse engineering solutions. The global rail industry doesn’t have a monopoly on the worst — though the US Passenger rail sheltered workshop of egregious incompetence and unemployable rent-seekign time-servers certainly tries and generally succeeds at scraping beyond the bottom of the barrel.

    Paul Dyson Reply:

    Richard M: Thank you for these last two posts, a dose of reality. Maybe, just maybe, something positive will come from this re Caltrain signalling.

    Clem Reply:

    Maybe, just maybe, pigs will fly

    Paul Dyson Reply:

    Clem, I’ve spent 50 years fighting for lost causes, over the top one more time, lads.

    John Nachtigall Reply:

    So GPS is imperfect…

    So the engineering is hard…

    So the prefered solution is nothing???

    The rail industry is mortibound when it comes to technological advancement. We are on the cusp of self driving cars which is a problem with 2 orders of magnitude harder given all the variables. I am not saying it is solved, but its close.

    I am not saying it is a trivial problem, but to claim it is beyond current technology is laughable. It just requires capital investment and embracing technology which is something the rail industry has never done so why should we be surprised. I mean how many miles of US track are still “dark territory” in a day and age when UPS can track millions of packages a day.

    wdobner Reply:

    You’re conflating two very different operating regimes. Nobody is talking about running HSR or even anything more than the most rudimentary passenger services (such as those operated in Alaska) through dark territory. Most passenger trains are going to be operated on double track lines currently equipped with signalling systems, and GPS does not reliably provide the requisite accuracy to differentiate between adjacent tracks on multiple track alignments. Thus geolocated positioning beacons, such as the balises used by ETCS, provide a superior solution which can provide greater accuracy at key points but is still relatively low cost.

    Those areas which are currently non-signalled are quite likely to be single track, and thus don’t need sub-10 meter accuracy with such absolute reliability. A few positioning beacons around passing sidings would be a good idea, but aren’t necessary.

    John Nachtigall Reply:

    Actually I am making a point about a few things.

    1. It is incredible to me that Spain would build a HSR rail system and not have automatic controls in place at all times. They had nothing on this stretch of the line, and they did that on purpose?!?!?


    the money quote

    On the approach to the crash site, the train had switched to Spain’s own safety system, Automatic Braking and Signals Announcement. The Spanish system also features automatic braking, but it depends on electronic beacons to monitor the train’s speed and none had been installed on the stretch of track heading into Santiago de Compostela, Mr. Alfambra said.

    The train crashed as it was making “a transition from a system that is secure, reliable and automatic to…a system that is in a certain way manual,” Mr. Alfambra said.

    A spokeswoman for ADIF, the company that manages the tracks, said ERTMS systems aren’t typically used in stretches where a line enters urban areas, like the one in Santiago de Compostela.

    That is because motormen are trained to slow down in such situations, so there is typically a reduced risk of accident, and the more complex ERTMS system is considered superfluous, she said.

    However, Mr. Alfambra said the accident indicated that such policies should be reviewed. “There’s political interest in opening new lines, which is great, but we should do it with up-to-date technology,” he said.

    Exactly what kind of logic is “motormen are trained” If you wanted to 100% rely only on humans you would not have an automatic system in the first place…much less rely on them for a “tricky” part of the line. Its a system that reduces speed and your excuse for not installing it is that they are “supposed” to slow down. That is the 1st place you should have it.

    2. Railroads around the world (not just the US) fail to embrace technology. i.e. dark territory in this day and age.

    3. Some system is better than no system, which goes back to point 2.

    4. The Spanish system, which has been pointed out many times on this blog as an example to follow, got at least part of their low cost from cutting corners like this.

    5. And most important. If you believe Richard, and installing advanced signaling systems on legacy track actually increases the chance of something going wrong then the choice seems to be install and get an accident or don’t install and get an accident. I don’t like that choice.

    Joey Reply:

    John – The problem with that particular segment of track, as has been noted, is that it it is a transition between the high speed line and the conventional line leading into the Santiago de Compostela station. Adding a new signaling system there would require adding it to all trains that use that station, even ones which aren’t ever going to approach high speed.

    Note that the place where the crash occurred is right before the local Oruense-Santigo line merges. Pushing the changeover point any further would have created another safety issue.

    Also note that the high speed line is currently broad gauge (1668mm). Once the Olmedo-Zamora HSL is complete, it will be converted to Standard Gauge (1435mm). Once that happens there will be no more shared track with local trains and the line can be fully ERTMS supervised.

    John Nachtigall Reply:

    I get that, but they were not even using the standard system that the track was built with??

    I would say it was a poor decision not to upgrade, but it is un-defendable not to install the existing system

    Richard Mlynarik Reply:

    John, the were using “the standard system that the track was built with”.

    The train wasn’t strictly speaking on “old” track but was on the old right of way on tracks that connected the existing station to a new cut-off line, approaching a junction point where the old wobbly line and the new straighter line merge, and as such under the control of the common (important!!!!) legacy signal system.

    That system, like most (not all, but most) rail safety systems of the 20th century, doesn’t enforce speed limits. It does enforce stops after blowing through red lights.

    The old standard system has huge flaws, but is standard (in Spain) and does prevent most common accidents. That flaws and its older technical base (essentially one-bit communication: “stop” or “go”) are why there’s a new system. The new system isn’t perfect, but it’s OK engineering and does, among many other very desirable things, enforce speed limits. The new system has to co-exist with the old system for decades, including the years 2012 and 2013.

    You’re trying to set up a straw man and have me defend 80 deaths in a horrible accident. That’s not going to happen. I don’t — unlike some of the weak brained types here — gloat over tule fog freeway pileups, bus plunges or airline CFITs either. Learn and improve.

    jonathan Reply:

    John, the “standard system” is ASFA. ASFA is an Automatic Train Protection system which enforces that trains stop at signals. ASFA does not enforce speed limitations, or provide in-cab display of braking curves and target speeds. IIRC, ASFA works by placing a balise, or beacon, in advance of each signal.
    The balise tells the train whether it needs to stop for the signal; if the train needs to stop, AFSA will slow down and stop the train.

    There’s no speed-control (excluding enforcing an ASFA system-wide upper bound of 200 km/hr).

    John Nachtigall Reply:

    I am not trying to get you to defend the existing system. I get you are just explaining the facts hand doing it quite well. But you have a person saying that beckons for the existing system are routinely not installed close to stations. You have a huge use change not he line in 2011 with no corresponding change to the safety system.

    It just astounding to me, actually astounding, that supposedly modern rail systems would rely on NOTHING to enforce critical speed limits. That is just a concept I can’t get my head around

    Joey Reply:

    It certainly does bring into question the safety of the design of the whole setup, though finding a solution isn’t easy. Would it make sense for the line to be ERTMS supervised in the interim even though it would require at least 3 signaling changeover points for all trains using the line? Would it make sense to push the signaling changeover closer to the station at the expense of having to build quite a bit of additional infrastructure. Would it make sense to put a speed limit at the end of the supervised section just to make sure trains are going slow enough when they hit the curve, at the expense of a bit of travel time?

    Max Wyss Reply:

    Would be nice. However, it is not sufficient that the train knows where it is; it is even more important for the control center to know where the train actually is. Therefore, GPS can only be used as a supportive system to a communication-based signalling system.

    Another aspect is that the railroad environment is extremely harsh, and does require very high reliability standards of the components. Particularly in electric (or diesel-electric) locomotives, the electromagnetic environment is extreme. That means that “industry level” components can not be used for safety-relevant applications. (in many cases, GPS is used for non-safety-relevant systems, such as passenger information.

    wdobner Reply:

    GPS has a way of losing accuracy under power cables. Lots of fun to use a surveying DGPS setup and watch the GDOP and PDOP skyrocket, have it start beeping at you, and then look up and realize you’ve wandered under some residential power lines. GPS, as it exists today, is *never* going to provide safety critical information regarding track occupancy.

    And IMHO GPS isn’t particularly great for passenger information, especially on lines which feature extensive tunnels. Balises, even purely passive positioning beacons, are far superior to GPS for that sort of thing.

    Useless Reply:

    Any system using GPS for positioning must have a backup system in case the GPS goes down, because the GPS doesn’t guarantee service at all times, and I am not talking the European, Russian, or Chinese global positioning systems as backup systems.

    jonathan Reply:

    Would be nice. However, it is not sufficient that the train knows where it is; it is even more important for the control center to know where the train actually is. Therefore, GPS can only be used as a supportive system to a communication-based signalling system.

    Max, that is not strictly true. Wabtec’s ETMS uses on-train GPS as an integral and essential part of ETMS. You can argue whether ETMS meets your definition of a “communicaiton-based” signalling system. Where does ERTMS-Regional fit in your taxonomy?

    Max Wyss Reply:

    Jonathan, after I wrote my comments, that system came to my mind (used for very low-density traffic lines, executed exclusively with (D)MUs). The “communication-based” part would come in by using GSM-R for the communication with a control center. This system can be made working by either excluding passing loops (and using local control), or by using splittable switches, assuring that the train will always have its reserved track (with that, the lack of precision of GPS can be compensated). The vehicles must be self-contained, assuring the integrity of the train (when a GTW loses a part, it will definitely not get unnoticed…).

    Jonathan Reply:

    hi Max,

    Would you be surprised to learn that Wabtec’s ETMS — intended to be used for humungous US freight trains — does in fact use on=board GPS to locate trains; and uses communications to send location-information from trains to the “back office”, and to send movement authorities to trains?

    Joey Reply:

    Automatic train operation has caught on for some metro systems, but not for mainline rail. Possibly because those systems usually have to mix with legacy trains and/or grade crossings. But then again, setups where some segments of track are automatic and some are manual do exist.

    Ted Judah Reply:

    No one has mentioned so far that under the current design, PB has CAHSR blowing through at full speed through intermediate stations between the end points for express travel.

    This crash bolsters the argument that automatic control is preferable even if legacy track is used for parts or through stations. The good news is that if the Spanish come up with a cost effective solution to this sort of a problem, CAHSR will benefit.

    Indeed, in some sense, the situation occurring in Spain is a blessing in disguise.

    Joey Reply:

    Well, at least CalTrain and Metrolink will have some PTC on the ends of the system. Though I don’t know if they will have speed enforcement. And changeover points are just another potential failure point, particularly for PTC solutions which are home-grown and untested.

    Jonathan Reply:

    well, actually they _have_, and at some depth; just not in this thread.
    See Clem’s analysis of CHSRA’s contractor’s simulation speed-curves, on his blog.
    At least for the Peninsula “blended” portion of the route.

    I find your “blessing in disguise” utterly appalling.

    Max Wyss Reply:

    Good explanation. However, ETCS Level 1 is essentially classic signalling, using ETCS components (which includes cab signalling). As it is stated that the section is ASFA equipped. I may be wrong, but it looks to me that it is ASFA, using ETCS components either as an overlay, or as the equipment for the actual signalling.

    What is normally understood as “ETCS” in high-speed applications, is ETCS Level 2, which is communication-based, and backed with Balises to synchronize the actual position of the train. I am aware of one place where ETCS Level 2 and another signalling system are used concurrently, and that’s the LGV Est Européenne. It is a certain challenge operating with two systems, but it seems that they have it under control. In this case, ETCS Level 2 has precedence.

    Good point with the legacy systems and the long upgrade cycles. Signalling upgrades are expensive, and do require good planning; an upgrade project can take several years between the go-ahead and the activation. There were several (fortunately not so serious) accidents at places where the upgrade would have prevented it, and the upgrade was scheduled to be begun within months after the accident.

    However, one can wonder why potentially dangerous places (such as sharp curves after stretches of high speed) are not treated with a higher priority. On the other hand, because of its high inherent safety of the “railroad” system, the “safety culture” may not be as evolved as with other transportation systems (such as airways).

    Max Wyss Reply:

    oops, read the last remark as (“such as aviation”)

    … a kingdom for an edit function …

    Richard Mlynarik Reply:

    Max, the relevant ETCS L-1 feature not provided by classic lineside signalling is speed enforcement.

    As for the priority of speed-enforcing sharp curves after stretches of high speed: well, it does seem so in retrospect. Interestingly enough, ETCS is in operation in the opposite direction (but on both tracks) from 42m within the 80km speed restriction. So if the cutover approaching Santiago de Compostela were at the same location as that leaving, and if the train had been operating under ETCS/ERTMS, then the normal ETCS brake profile enforcement would have prevented this accident. If if if.

    My guess (but I do not know!) is that the directional asymmetry of ETCS cutover would have been motivated by safety: it is important that a train entering a congested station area have a working signal system, so the failure-prone cutover from one system to another is placed on open track well outside the location where other non-ETCS trains and signals are located. So ending speed ETCS-linked speed enforcement early (and remember, this train was not operating under ETCS!) may have been a conscious safety-related decision, rather than a deliberately short-sighted failure to protect a tight radius curve. Crashes between trains that overrun stop signals at stations are, after all, hugely more common than overspeed derailments. Trade-offs are seldom simple or unambiguous, and the data upon which to make them are sparse and poorly sampled.

    wdobner Reply:

    it is important that a train entering a congested station area have a working signal system, so the failure-prone cutover from one system to another is placed on open track well outside the location where other non-ETCS trains and signals are located.

    Oh joy, so we’re gonna see CBOSS to Los Banos!

    Maybe there is something to those real estate speculation conspiracy theories. I can hear it now: “Well, the Caltrain signalling system goes there, the Caltrain trains should too!”

    Max Wyss Reply:

    Thanks for the clarification, Richard. I could have gotten a bit confused by ETCS Level 1 LS (Limited Supervision), which is supposed to become network wide at the SBB network within the next few years.

    The explanation for the asymmetrical switchover does indeed sound plausible, as switchovers are always a source of disturbance.

    Derek Reply:

    PS. Saying air travel has had “many more deaths” and comparing raw numbers on cars is just misleading. If you want to compare deaths per passenger mile then pull the stats.

    That would also be misleading. People tend to fly longer distances than they take trains. A better statistic for knowing whether you would reach your destination safely would be fatalities per x passenger trips.

    Alon Levy Reply:

    Ugh. Safety per passenger is only relevant when the transportation system itself determines distances. This is the case for local urban travel, but for intercity travel, building high-speed rail isn’t going to make Los Angeles and San Francisco magically farther away. Likewise with aviation.

  5. Keith Saggers
    Jul 27th, 2013 at 12:10

    ASFA – Anuncio de Senales y Frenado Automatico or signal notification and automatic braking – relies on a series of beacons to communicate with the driver’s cab – so does not have the constant communication of ETCS.

    The system gives audio and visual warnings to the driver if speed limits are surpassed, and will step in and brake the train if there is no response from the cab.

    Route knowledge

    Rail expert Christian Wolmar, author of Blood, Iron and Gold, says that there will have been some sort of warning system in place.

    “We don’t know whether this failed or whether the driver didn’t heed a warning,” he said. It’s also possible that the driver was confused by moving between the two different systems, he added.

    “However the driver would have route knowledge,” he said, pointing out that all drivers would have been trained extensively on a section of track before taking charge of a train carrying passengers along it.

    Philippa Oldham from the Institution of Mechanical Engineers said accident investigators are likely to be looking at a number of factors which could have caused the accident, including the role of the signalling and speed advice system, as well as the role of the driver.

    They would also look at whether breakages or vandalism damage to the track or the train contributed to the derailment


    PS Contradicts what Adif said above about driver expectations.
    PPS Last sentence adds to Syn comment about presence of graffiti indicating insecure location.

    Paul Dyson Reply:

    Very few railroad locations are truly secure. My last train trip in Europe; Koln, Bruxelles, London, was so depressing because of the graffiti everywhere.

    synonymouse Reply:

    Apparently this train was relying on ASFA only and throughout.

  6. synonymouse
    Jul 27th, 2013 at 15:50

    Link thanks to Altamont site to an IRJ article which states this type of equipment was using ASFA exclusively:


  7. Amanda in the South Bay
    Jul 27th, 2013 at 17:13

    What is this Altamont site youve mentioned before?

    synonymouse Reply:


    Not as active as this site but you have engineers and other rr people posting including one fellow who owns and operates his own shortline.

    I pissed off one guy by critiquing SMART, but it is in my backyard, and now I sense others are catching up and picking up on its shortcomings. He says I hate Thomas the Tank Engine. Naah, but I still think they should have retained the British narrator.

    Neville Snark Reply:

    Hey Syn – can you just list your main criticisms of SMART? (One: it stops short of the Larkspur ferry!).

    synonymouse Reply:


    1. Turn the page of history on the NWP. Nostalgia is poignant but it is transit’s time now. Adios freight and AAR requirements.

    2. Electric light rail – 750vdc OC standard gauge, compatible with Muni.

    4. To the south PB-BART style scorched earth cutting a swath down 101 to a station at Marin City, connecting cross platform to buses to the City, including conceivably some Muni service.

    5. Essentially SMART replaces the trunk route #80 with the #101 remaining and now serving directly the Graton Casino in RoPo. Hypothetically 24/7 service on that route.

    6. To the north service as far as Willits would be nice but practically connecting MCI coaches from Healdsburg or Cloverdale north would do in the interim.

    7. Infrastructure. Yes, and that means some of those damnable stilts. Viaducts in San Rafael for sure, and maybe Petaluma and other locales. Underpasses, overpasses, anything to separate those expensive and fragile trains from auto traffic. I had the wild idea of using one of the 101 viaducts over the Petaluma River, which Caltran plans apparently to tear down, for SMART to cross at high level. Would have been worth looking into, but the insiders are clueless and bereft of imagination.

    8. I’d have GGT take over SMART. Novato bus yard is adjacent to the NWP and could store cars. I’d like to see tracks in Bellam Blvd. in San Rafael and streetcars at GGT’s big central maintenance facility there. The District extends all the way to Del Norte Co., so I could see GGT buses taking over from Amtrak serving that corridor all the way to the Oregon border.

    Neville Snark Reply:

    The points 1-2 are well-taken and 4 exciting but won’t 6 and 7 be expensive?

    synonymouse Reply:

    They could have saved a bundle by purchasing San Diego’s Siemens that went instead to Argentina. Good enough for starters and spend the money on infrastructure.

    Peter Reply:

    You are of course referring to the clapped-out 30+ year-old Duewag cars that San Diego justifiably replaced with new equipment (again, why would we want them?). Still not sure how the cost savings of not having to purchase rolling stock would balance out the significant additional capital expenditure to install OCS to begin with, the significantly increased costs of vehicle maintenance (the vehicles are ancient, and would need more maintenance), plus increased ROW maintenance costs (must now also pay to maintain the OCS).

    synonymouse Reply:


    Jonathan Reply:

    Peter, surely you have figured out by now that facts (at least post-1959 facts) are water off a duck’s back to Synonymouse?

  8. John Nachtigall
    Jul 27th, 2013 at 17:33

    I thought France had no fatalities on HSR?


    This was 2 weeks before he wrote the OpEd.

    synonymouse Reply:

    Legacy electrified rail network and I do not believe the verdict is in whether the loose rail bracket jammed into the frog of the switch likely causing the derailment came undone by itself or was sabotage.

    Richard Mlynarik Reply:

    Regular old boring train on regular old boring track, as in Spain.
    Tragic. Avoidable in a perfect world … just as with every other of the thousands of daily human fatalities.

    Fatalities on a high speed train on high speed track are inevitable over time, just as hull losses of A380s are inevitable over time in the real imperfect world.

    But you’re smarter than you’re pretending here, and have been know to make effort not to be completely ill-informed. So knock it off!

    John Nachtigall Reply:

    So they lied in the article when it said High Speed?

    Why is Acela HSR when we want to talk about how cool it is but when other trains go that fast and crash they are not HSR?

    Max Wyss Reply:

    Unless this is a rhethoric question, it is simply USAn understanding of “High Speed”. In Europe, “High Speed” may mean something different. There are many kilometers of 200 km/h enabled lines (in Germany, in the UK, in France, in Italy, in Spain etc.) which are not specific “High Speed” Lines.

    In German terminology, the fast stretches in the North East Corridor would count as “Ausbaustrecken” (upgraded lines); conventional lines, conventional signalling, and so on.

    Nobody would call, for example, the trunk line betwen (Basel -) Mulhouse and Strasbourg, or the line between München and Augsburg, or between Hamburg and Hannover as “HSR”, but trains run routinely at 200 km/h, and have been doing so for decades.

    John Nachtigall Reply:

    And the train in Spain was HSR for sure, it was traveling at about 190 kph at the time of crash which is well within the most of the definitions supported in this blog

    Joey Reply:

    Most definitions in the real world (i.e. non USA special snowflake definitons) cut off at 200 km/h.

    Alon Levy Reply:

    It was overspeeding. It was HSR in the same sense that steam trains making up time by going 100+ mph on straight track in the 1930s were HSR.

    Peter Reply:

    Just because you’re going 65 mph doesn’t mean you’re on a freeway. You could just as easily be on a residential street with a speed limit of 25 mph.

    Useless Reply:

    It was not a TGV that crashed, but a regular express train.

  9. Emmanuel
    Jul 27th, 2013 at 22:24

    I don’t understand the point of this. You are saying, basically if we lived in a perfect world, this wouldn’t have happened. If we lived in a perfect world, there would be no traffic congestion, no plane crashes, no murders. If we lived in a world, there would be no humans…

    Neville Snark Reply:

    We don’t live in a world, then? ;)

    VBobier Reply:

    Nah, We live in Homer Simpson… ;)

    Emmanuel Reply:

    Oops. *perfect world.

  10. Keith Saggers
    Jul 28th, 2013 at 08:59
  11. Bill
    Jul 29th, 2013 at 08:27
  12. Reality Check
    Jul 29th, 2013 at 09:51

    Prospects dim for high-speed rail link for L.A., Vegas

    Even if the Buy America obstacle is somehow finessed away, the much more intractable financing issues remain. For starters, the price tag for XpressWest has doubled in the past five years to $6.9 billion at last count, and that was before even one spike had been purchased.

    Further, bullet trains have an international history of not generating enough revenue to pay for construction, as detailed in the Las Vegas Review-Journal in February 2012. A subsequent report by the Congressional Research Service came to the same conclusion, a not-so-trivial detail considering that XpressWest wants a loan, not a grant.


    Over the past year, XpressWest has tried some PR fixes, such as adopting that name to convey the impression that it was a key link in a Western rail system rather than a stand-alone tourist train. The company executives also obtained support from the Los Angeles Metropolitan Transit Authority to build a link from Victorville to a different line going to downtown Los Angeles, termed a “game changer” by cheerleaders masquerading as journalists.

    But nothing could detract from numbers that simply don’t work, never mind the questionable assumption that millions of people a year would get off I-15 mid-trip to board a train.

    In that regard, XpressWest joins Amtrak, the mag-lev, a different concept to run bullet trains south from Las Vegas to a desert juncture that would go to either Phoenix or Los Angeles and a company touting premium-service conventional trains as producing nothing that anyone can ride.

    The latter proposal even reached a $56 million agreement with Union Pacific to rework its freight tracks to accommodate passenger trains and eliminate the bottlenecks that plagued Amtrak on half the route. Unfortunately, Las Vegas Railway Express, the company pushing the idea, has never accomplished anything except to put out announcements that might move its penny stock and miss proposed start dates.

    In sum, the only viable solution to congestion that occasionally afflicts stretches of I-15 through the desert is to widen it.

    Reality Check Reply:

    The Flawed Federal Rule That Killed High-Speed Rail to Vegas

    The problem had nothing to do with the cost or promise of the Vegas HSR line — both of which have been questioned in the past — and everything to do with a rule requiring rail operators to buy their all material from U.S. manufacturers.


    The problem, of course, is that the United States doesn’t really have a domestic high-speed rail manufacturing industry at the present time.


    When the Obama administration made its high-speed rail push a few years back, it anticipated this exact problem with respect to Buy America principles. In a primer for loan applicants, the Federal Rail Administration explained how to satisfy the requirement when technology is “not readily available domestically.” In such instances the grantee is supposed to look for the bidder with the “highest domestic content” — and failing that, there’s always the possibility of applying for a Buy America waiver.


    Still, major transportation projects do get relief from Buy America now and then. The Systemic Failure blog astutely points out that, just this month, the federal government gave a billion-dollar highway project additional time to meet a new Buy America stipulation that requires “all” related contracts to meet the requirement — even, for instance, the relocation of utility lines. The fact that XpressWest wasn’t granted that same leniency raises the possibility that the Buy America provision was simply a pretense for some larger discontent with the plans.

    synonymouse Reply:

    “cheerleaders masquerading as journalists.”

    no comment required

    nslander Reply:

    “In sum, the only viable solution to congestion that occasionally afflicts stretches of I-15 through the desert is to widen it. ”

    Strawmen masquerading as a flying monkeys.

    synonymouse Reply:

    It is reasonable to anticipate that Reid will score some earmark highway money for I-15 as a palliative for the collapse of Deserted Xprss and related schemes.

    Derek Reply:

    Further, bullet trains have an international history of not generating enough revenue to pay for construction, as detailed in the Las Vegas Review-Journal in February 2012.

    That article quotes the World Bank report, High Speed Rail: The Fast Track to Economic Development? by Paul Amos, Dick Bullock, and Jitendra Sondhi, July 2010.

    Internationally (though not in China) many projects have taken very long periods to complete (over a decade is not unusual), creating a heavy capital and debt burden before any cash in-flows…Revenue from the initial Tokaido Shinkansen covered interest and depreciation by the third year. But other lines have been far less successful financially (unsurprising given the relative traffic volumes) and have been able to contribute very little, if anything, towards capital recovery…over 70 percent of the THSR line costs were associated with interest and depreciation and, despite a ridership of nearly 100,000 per day, the project has ended up having to be refinanced.

    The evidence is that it is very difficult for a stand-alone high-speed railway to recover much of its capital costs from the passenger revenue stream alone, except in the very densest corridors.

    Keith Saggers Reply:

    Who are all you HSR naysayers trying to convince? each other?

    John Burrows Reply:

    THSR (Taiwan High Speed Rail) did get off to a rocky start. Initial ridership was way under projections, and the project had to be bailed out by the Taiwanese government—But in the 3 years since the World Bank report that you quote above, the situation for high speed rail in Taiwan has improved. Starting in 2011, 5 years after operations began, and again in 2012 THSR made a profit on yearly operations. And ridership is steadily increasing (about 122,000 per day in 2012)—Way under initial wildly optimistic projections, but still, maybe not too bad for a small country with only about 60% of California’s population.

    If CAHSR were to start making a profit on its operations 5 years after the blended system was completed and if by that time daily ridership was 122,000 (44 million per year), that might not be too bad either.

    Derek Reply:

    It would be good to get the blended system done quickly in order to avoid “creating a heavy capital and debt burden.” People forget that tied up capital still costs money.

  13. Useless
    Jul 29th, 2013 at 12:20

    Damn, what’s happening in Europe? Another train crash in Europe, this time in Switzerland.

    Police say 44 injured in train collision, Swiss TV reports

    Swiss television, citing local police, reports that 44 people have been injured, four of them seriously, in a head-on train collision in the west of the country.

    Public TV station SRF quotes Vaud canton (state) police spokesman Pierre-Olivier Gaudard as saying one person has yet to be recovered from the wreckage.

    The crash happened late Monday on a regional line about 31 miles southwest of the capital, Bern.

    Pictures on the website of local daily 24 Heures showed the two regional trains locked together but still on the tracks.


    Richard Mlynarik Reply:

    Damn, what’s happening in Europe?


    Reedman Reply:

    A different still picture of the trains in this article.


    Richard Mlynarik Reply:

    If only there were some way of not having to post train crash pictures one by one. (And Fox News? Oh dear.)

    And for rail enthusiasts, a link complete with run numbers, unit numbers, and link to the relevant graphic timetable

    Max Wyss Reply:

    And we get to the discussion about signalling again… The Bahnonline article (for the rail enthusiasts), states that that station has “Gruppenausfahrsignale”. That means that one single signal is used for all tracks for trains leaving the station. This is a solution which is to be phased out because it does contain risks.

    Other sources (bahnforum.ch, where railfans but also a good number of professionals participate) state that there was no ZUB, which could have stopped the train leaving the station). The same sources guesstimates an impact speed of 80 to 100 km/h. The vehicles involved were built in the late 70s to 80s, and the motor unit got a rebuild a few years ago, whereas the driving trailer was not rebuilt.

    The person to be recovered is one of the drivers, and according to the messages, his state is still unknown.

    John Nachtigall Reply:

    Is it really too much to ask to have a signaling system that ovens trains from running into each other? I mean really, what is the system for if it does not prevent that.

    Max Wyss Reply:

    A few more things coming out during the day:

    The train which was stopping at the station left prematurely. Apparently the driver noticed the error, slammed the brakes and could flee into the baggage compartment behind the cab. The driver on the incoming train had less chance. He could only be found dead.

    In some ways, it is a bit of a side effect of fixed interval scheduling; That station has metings only four times a day (additional morning and evening peak trains), and it seems that the drivers running the regular services are rather frequently on that line, meaning that it gets a rather repetitive job.

    The signalling equipment (I guess in US terminology, it would be the “tower”) dates from 1958, and has been updated in 1975. This is adequate for a secondary line, even if the line now has hourly trains throughout the day, plus some freight. Upgrading to ETCS L1 LS is planned for 2017.

  14. Reality Check
    Jul 29th, 2013 at 12:34

    Don’t Believe What You Read About China

    The tone of western reporting about Chinese high speed rail has been fairly negative. Stories of safety concerns, lack of ridership, excessively high ticket prices that the Chinese cannot afford, inability to compete with the discount airlines, “white elephants,” etc.

  15. Keith Saggers
    Jul 29th, 2013 at 16:14

    100,000 people are killed every year in car accidents
    It will take 435 years for the operators to pay back the tax payers for their investment in sustainable high speed rail for our grandchilden

    Keith Saggers Reply:


    Derek Reply:

    Phase 1 Blended is expected to cost $68 billion (YOE), and is expected to make an operating profit of $1.830 billion per year (YOE) in 2034, according to the business plan. So your payoff estimate is off by about 400 years.

    Paul Dyson Reply:

    Apples and Oranges Derek. I believe brother Saggers is talking about investment, not operating margin. Make a gift of plant and machinery to an enterprise and they’d be in an extremely sorry state if they could not operate at a profit. We’ll have to forget nasty things like interest, taxes, depreciation and amortization, of course. There will be an awful lot of the i,d and a accumulated by 2034 and no t paid either. Such a deal for taxpayers.
    Mind you, not sure how competitive the “blended” plant and machinery will be in 2034 by the time it is operating. Maybe even the operating profit is optimistic.

    John Nachtigall Reply:

    Make that 38k killed


    Only off by a factor of 3

    Keith Saggers Reply:

    in the world

    Alon Levy Reply:

    Globally it’s 2 million a year.

    Eric Reply:

    In some countries, more than one person has been killed by car for every 10 cars.

  16. Clem
    Jul 29th, 2013 at 20:12

    Some interesting news is starting to trickle out of Spain (in El Pais). The driver apparently stated that he briefly lost situational awareness of what part of the line (which consists of a repetitive series of tunnels and bridges) he was located on. By the time he clued in to where he really was, it was too late for him to slow down. Combined with the surprising fact that there was no automatic system to catch his error (none!!), the root cause of this accident is probably going to reach far beyond one mistake made by one man staring into the rhythmic onrush of rails and catenary masts for 20 minutes straight with hardly anything to do but wait for the fateful curve.

    The parallels with airline crashes are hard to avoid, with the same issues of human/machine interaction, crew resource management, automation, workload, situational awareness, etc.

    Joey Reply:

    Is there no cab signaling?

  17. Reality Check
    Jul 29th, 2013 at 20:30

    Valadao’s staff on message, but off point?


    While no answers were forthcoming Tuesday, Valadao’s chief of staff was kind enough to email a new photo of the congressman.

    It wasn’t until Thursday that Valadao spokeswoman Anna Vetter responded — sort of — to the latest questions.

    “Congressman Valadao has submitted all of the proper documentation of his economic interest, including property, both during his time in the California Assembly and United States Congress,” her statement read. “Congressman Valadao will not be bullied into silence by outlandish accusations by extreme liberal organizations for their own political gain. He will continue to fight this wasteful project in order to best represent the values and opinions of those who elected him.”

    The economic disclosure filed with the House clerk, however, doesn’t detail any specific parcel or property ownership, but the statement represents a continuation of a consistent message over the past two weeks. As for being “bullied into silence,” so far the only thing Valadao’s office is being silent about is actually answering the central questions:

    Is Valadao’s position that there is no conflict of interest by advocating for legislation that could ultimately spare his own property (along with the property of many others) from losing value because of the rail line? Did he ever directly tell his fellow committee members that his family owns property along the route? And did he have an obligation to tell his colleagues about the property interests?

  18. Reality Check
    Jul 29th, 2013 at 20:30

    High Speed Rail Authority ramping up IT projects

    The California High Speed Rail Authority (HSR) will be partnering with suppliers of cloud services for a construction package that will start up in about a month, according to CIO Jeff Vargas.

    In an interview with Techwire’s Christina Gagnier, Vargas said the HSR has doubled its workforce within the past six months, particularly in the engineering and IT fields, and plans to double again within the next six months due to project CP1, or Construction Package 1. HSR is currently hiring for two senior supervisor positions in IT, with 15 additional staff positions to be filled in the next few months.

    In a follow-up interview, Vargas explained that CP1 would be for an initial operating segment for high speed rail in the Central Valley, which will require support systems for public reports, notifications and workflow processes.

  19. Reality Check
    Jul 29th, 2013 at 20:50

    Thomas D. Elias cranks out another idiotic column on CA HSR:
    Spanish train wreck may have a California lesson: Thomas Elias


    Extrapolate this to California, and with four intermediate stations, trains will run well below maximum speed for at least an hour. Combine this with relatively low speeds in the San Fernando Valley portion of Los Angeles and on the Peninsula. Then do the math, and it’s clear trains will not easily complete the run in 2:40 or less.

    Which could mean that drivers will be rushed, as the Spanish engineer may have been.

    Still, insists [CA HSRA spokeswoman Lisa Marie] Alley, there will be no risk of a Spanish-style accident. “We will have a fully unified control system covering the entire route where things talk to each other, not a piecemeal one,” she said. “Where we’ll make full speed is yet to be determined, but we will build absolutely the safest mode of transport for Californians that’s ever been seen.”

    But no control system is fail-safe. So why not remove factors that might someday encourage an engineer worried about job ratings and on-time performance to do something risky?

    Stops at intermediate stations (only lightly used during high-speed rides this columnist has taken in Europe) now appear to be the most obvious items that might create a rushed feeling.

    So one California lesson from the Spanish crash may be to take a second look at those stations and figure a way to eliminate one or two before spending many millions to build them.

    Jonathan Reply:

    I call bullshit. Ms. Alley doesn’t know what she’s talking about. The quoted statement is simply untrue. CHSRA’s own relevant Technical Memorandum lays out requirements for native HSR signalllng. That TM notes that ETCS Level 2 is the only technical solution meeting those criteria. ETCS is not “compatible” with the signalling systems which will be in use at either end of a “blended” system.

    Running HSR trains with ETCS Level 2 on the HSR portion, and (not-yet-implemented!) “legacy” CBOSS on the Peninsula “blended” corridor is exactly, I repeat exactly the scenario encountered in Spain. As several of us have noted, here and on Clem’s blog, changing over from one signalling system to another is a high-risk point. Same goes for transitioning between HSR signalling and V-ETMS, or whatever is in use in SoCal.

    Peter Reply:

    Wasn’t there a plan for an overlay system, with Caltrain using CBOSS and HSR using ERTMS on the same tracks?

    Jonathan Reply:

    You are joking, right??

    adirondacker12800 Reply:

    Why not? From what little I’ve been able to find about the interoperability of whatever the Class Is are calling their PTC this week and ACSES, the back office will send an appropriately formatted message to each train. Why you would do that when you don’t have to is a different question.

    Jonathan Reply:

    You are assuming that the two distinct protocols are commensurate: that the “back end” can decide what the trains should be doing, and then map that decision onto two (or more) different message encoding. Each of the encodings has an accompanying interpretation –or semantics — of what the messages (the bits in each field in the message layout) actually mean; what to do, when you get a specific message with specific vakues. This combination of message-layout, and semantics (interpretation, or what-to-do) is what is technically known as a “protocol”).

    It’s not at all clear, _a priori_, that a back-end system designed specifically for just _one_ such protocol, can have its internal (back-end) state transitions mapped onto a very different protiocol.
    I suggest braking curves — and how hard to brake, when (if the driver hasn’t) as a specific example.

    Richard M. could contribute very constructively, here. If he happens to know details of ITCS message formats.

    That said, the major feature which differentiates CBOSS from ITCS, is that CBOSS is supposed to let level-crossing gates immediately in front of station platforms _open_ — if the only train which would otherwise lower the gates, is in fact stopped at the immediately-adjacent station There is absolutely no way that you are going to get ETCS Level-2 equipped HSR trainsets to do _that_.

    Not unless one equips all the HSR train-sets with both ETCS Level 2 _and_ CBOSS signalling; and go through the error-prone changeover from one to the other when entering or leaving the Peninsula corridor. Which is the whole point at issue, is it not?

    thatbruce Reply:

    To play Devil’s Advocate, the circuit controlling the crossing gates don’t need to know how the line’s safety system is communicating with the trains. The line’s safety system is intended to know both the position and identity of each train on the line, and by extension, whether a given train is scheduled to stop at a given station, and by further extension, whether or not to override the crossing circuit’s desire to start dingdingdinging and waving its boom about.

    All that being said, the very idea of crossing gates not coming down when a train is approaching is anathema to the basic idea of safely keeping cars and trains separated.

    Jonathan Reply:

    No, Wrong. Broken. I have personally travelled on (late-night, weekend) trains where the train crew walked through the train, asking, by station name, if anyone was going to get off at certain very low-patronage stations. The train crew then slowed, enough to stop if intending passengers showed themselves — but did not actually stop, at said stations.

    You do not, repeat _not_ bake the scheduled intention-to-stop, into the decision to raise the crossing gates immediately in advance of t the end of the platform.
    Go study Peter Cook and Elizabeth Hurley some more ;)

    That said: I entirely agree with your last sentence. Now I raise you, to the tune of ignoramuses (albeit well-intended) saying “Oh, the back-end will transliterate between CBOSS and ETCS”.
    I already have to skip eating, after seeing that bandied about: too sickening.

    thatbruce Reply:


    Good, good. I can feel the force of your violent agreement. Strike my post down with all of its meanings and agree with it in righteous anger and your journey to the dark side will be complete.

    ( I think you’re somewhat confused as to what the term ‘Devil’s Advocate‘ involves. )

    adirondacker12800 Reply:

    that the “back end” can decide what the trains should be doing, and then map that decision onto two (or more) different message encoding.

    I realize that Californians think that their local designers are idiots but I suspect that someone somewhere is taking into account that there will be trains on the tracks that don’t use CBOSS. Why they want CBOSS is a different question.

  20. Reedman
    Jul 30th, 2013 at 11:17

    The Spanish train driver received a call from Renfe just before the crash.
    He was on the phone with them, possibly looking for paperwork, when it happened.


    Keith Saggers Reply:

    By by Brazilian contract

    thatbruce Reply:


    Minutes before the derailment, Garzon received a call on his work phone, apparently receiving instructions on the way to Ferrol from a Renfe staff member, the court said Tuesday. Background noise suggested he was looking at or shuffling papers, the court said.

    Unlike Chatsworth where the driver was distracted by personal business on his cellphone, Santiago de Compostela had the driver distracted by an official call at the critical time? That makes the accident and the factors leading to it so much worse.

Comments are closed.