Hi --

We've had a nasty train accident over here in Oslo, Norway, today - 16 empty flatcars rolled away at the freight yard at Alnabru, and rolled about five miles downhill before careening off the track at the docks, hitting and throwing several tractor/trailers into the harbor and making a office/warehouse building collapse. News reports say three dead, one missing and four or five injured down at the docks :-(

My brother works at the yards up there, went in to work this afternoon, and sent me a text message this afternoon saying that place is full of police, railroad and traffic safety board inspectors, who are interviewing those of his coworkers that was working when this happened.

Only "good" thing is that someone at traffic control seemingly made a fast (and almost certainly correct) decision that when attempts to detour the runaway cars into a single ended siding further up the valley came too late, it was better to route them towards the harbor docks than towards Oslo Central passenger station. I shudder at the thought of what fifteen freight cars moving at 80+ kilometers and hour would have done with an EMU filled with commuters. Too bad the police did not have time to warn people down at the docks - those poor blokes never knew what hit them :-(

Lord knows what set the cars rolling. Hopefully the investigation will find the answer to that question.

This is what the accident scene looks like from the air:


Stein

Thanks for posting the map and the pictures. I read it in the German online news but there was no useful explanation.
My condolence and best wished to the relatives of the poor victims.

I agree. My sympathies to all those who lost a loved one. And prayers for those injured.

I heard some news of this story on Public Radio but you've provided more details. Its a terrible event but I'm glad some quick action apparently saved many lives.
Ralph

Sorry to hear the bad news about the loss of life.

A little more information has emerged in the media during the night, it seems.

Several newspapers report that according to a named manager in the transportation safety board, the roll-away apparently started a short while after a switcher had added a 16th car to a string of 15 cars on a track at the yard and uncoupled from the cars.

So (this is my inference, not part of the statement from the transportation safety board person), it seems like it might be a case of not enough wheel brakes being turned on to hold the string of cars with a safety margin, before leaving the cars on a incline. Maybe combined with one or more wheel brakes failing.

Our couplers are of a type that cannot be coupled while the cars are rolling, so the switcher crew did not have a chance to chase down the rolling cars and coupling to them at speed.

It has also emerged that the cars were not routed away from the passenger station. They were on the freight track down towards the docks all the time after they started rolling. They could have been routed towards the passenger station or out on the mainline to the south to avoid them hitting the docks, but that was never a realistic option.

It also seems that what was attempted was not to route the cars into a single ended siding on their way down the four mile valley from the yard to the curve leading into the dock area, as was initially reported.

What was attempted was activating what is described in the media as a "track lock" at an intermediate small yard at Loenga about 3 miles down the valley, but by the time the string of cars got there, they were moving with such speed (possibly as high as 100 kph) and force that they blew straight through the "track lock" thing. I assume this may be a remote controlled derail of some kind.

It also has been reported that we might have been fortunate (relatively speaking - loss of life is of course never fortunate!) about the accident happened at 1:15 pm instead of an hour later.

About 10 of the 16 cars careened off the track at the curve shown by the first star in the picture above. At the tank terminal tracks. At a little after 2 pm the daily jet fuel shuttle train between Oslo harbor and Oslo airport would likely have been there pulling a string of loaded tank cars from the tank terminal, and the loss of life could have been far worse.

The last six or so cars went on for about a quarter or half a mile, knocking over several trucks on the way, before going off the track and hitting and collapsing a small freight terminal building, killing three people and wounding four, one of them critically.

Three cars stopped in the wreckage of the building, three went through the building and over the dockside, with one of them coming to a rest across the stern of a small harbor tug or a similar boat.

During the night they have searched the building for more victims and demolished the remains of the building to make it safer to recheck the rubble, and started preparations for retrieving cars from the water to examine their brakes etc.

The dead were three men, the seriously wounded are two men (age 40 and 65) and one woman (age 35). The woman is in critical condition, the men are listed as seriously, but not life threateningly injured. A fourth injured person was treated for smaller injuries at a local clinic and released.

One of the dead and two of the seriously injured were working for the same small dock terminal company, which only has 10 employees.

My deepest sympathy for the people who lost family members, friends and colleagues yesterday. As the Ww2 poet Nordahl Grieg wrote it in his poem "May 17th 1940" : "we are so few in this country - every fallen is a friend or brother".

Also, my deepest sympathy for those on the docks who just dodged the bullet, and last night probably had the shakes - it is a sometimes sobering and frightening experience to be reminded of how quickly and unexpectedly your life can end. In particular, I expect that the two people who were on the harbor tug that had a freight car land on the stern of their boat were rather shaken.

And of course, my sympathy for those railroad employees at the yard and at the traffic control center, who last night, today, and probably for a long time to come will be wondering "should I have done something else?".

They had to make some quick decisions there and then, and probably did their best. But usually, some of the decisions made under pressure by stressed people will turn out to be less optimal than what someone with the benefit of days or weeks to calmly ponder the situation and 20/20 hindsight would have chosen to do.

That is all that is known at this time. Some of this information (especially my inference about what might have happened when the cars started rolling) may be inaccurate, and/or based on misunderstandings or misinterpretations by either the journalists or by me.

An official report with recommendations on what should or could be done to prevent a similar of this tragic accident will probably take weeks, months or even years - more likely months than weeks or years.

Stein

Last update on this thread for a while - I won't keep bombarding you with these posts.

Today the media reports quotes the same transport safety board official as yesterday, now speculating that the retarders on the hump yard may have failed, giving a car too much speed and sending other cars flying down the bowl track towards the yard throat at the south end of the yard.

It also is reported that cars should not have been able to roll out of the yard unless someone deliberately sets the path at the south end of the yard from the default path (which leads into a single ended track than ends in a big gravel pile) to one of the two paths out of the south end of the freight yard - the harbor track to the right in the picture below or the freight bypass around the north east side of town to the left in the picture below.



A test of the remote controlled turnouts from the control tower shows that all turnouts in the path the cars took seems to be working correctly and signal indications seems to be correct, plus that the retarders at the hump in the middle of the yard (about 700 yards north of the area shown in the picture above) seems to work as designed.

Looks like the cause of the accident might be a combination of factors. At the moment, it is not looking very good for the poor guy who was at the control panel in the tower - the exit switches should not have been lined for the harbor track.

But I guess it is time to stop speculating and wait for the official report. And hope that it was some kind of bizarre technical malfunction, rather than what it is starting to look like now - human error.

Stein

Thanks for the update.

Please...don't be concerned about "bombarding" us with more info!!!! I, for one, would like to hear of what is being said, and what may be speculated. Please, keep us informed. Thanks.

Another update. The picture of what happened (or at least - the picture you get through the media reports) is now fairly clear. Bear in mind that the reporters may have misunderstood some things - very few reporters have any clue whatsoever about how a railroad works.

The cut of 16 cars that ran away was initially located above the hump of the yard, at the north end of the yard, in arrival/departure track A5 (se overview picture below):



Here is a closeup of the area around the hump:



And here is a second closeup that shows the situation on the north side of the tower more clearly - observe that the orientation of the last picture is such that north is left, and south is right:




A switcher engine, working from the north end of the yard had just added a 16th car to a cut of 15 cars on A/D track 5, just above the hump. The cars were held in position not by having their own brakes applied, but by track brakes - having their wheels pressed against the tracks by hydraulic operated brakes along the side of the track, controlled from the tower.

As far as anyone can tell, the last car was added to the cut in the proper way (given our type of couplers), with one guy on the engine and one guy going between the cars and hooking the car to the other 15 cars, before the engine moved away.

Some time after the 16th car was added, the track brakes holding the string of 16 cars, was released, and the cars started drifting down the track past the control tower.

Would not have been a big problem, except for a series of unfortunate choices having been made previously:

- the tower operator could not slow down or stop the cars using the hump retarders, because the switch closest to track A5 had been lined for the engine bypass track along the side of the yard, rather than for the hump. By the time it was realized that the cars were not being pushed by an engine, but rather was rolling away on their own, they were past the switch and on the engine bypass track, that runs the length of the yard.

No track brakes or retarders on the engine bypass track. Safety mechanism on that track is cutting the power to the overhead catenary wire, which doesn't do squat for stuff that isn't electric engines. So the cars were rolling down hump hill and along the side of the yard.

At the south end of the yard, they should have been caught by the trap - switches should normally be lined for a single ended track that goes straight ahead and ends in a couple of concrete blocks or a pile of gravel. Switches should only be lined for one of the freight tracks out of the yard if there is a train about to depart or arrive.

Those switches were lined for the freight track that goes downhill to the left, not for the trap or for the freight track that goes straight or slightly uphill to the right at the south end of the yard.

There now was a runaway cut of cars on the freight track. The accident had already happened - only thing left was the crash.

The people at central traffic control made two attempts to stop the cars that were picking up speed heading down the 2.5% grade down towards downtown and the harbor.

They first lined a switch about 2 1/2 miles down the track for a single ended spur that ended up in a cliff, but by the time the cars got there they were moving too fast, and just ran straight ahead through the switch instead of taking the deviating path into the siding.

About a mile further down a track block was activated, pulling a steel bar across the rails. The runaway cars hit that block at a speed probably exceeding 100 kph, and shattered it - pieces of the steel bar mechanism was found 90 yards away from the point of impact.

The cars then continued through the small yard at Loenga.

A decision was made to not divert the cars into Oslo Central station or into the southbound main - the risk of hitting passenger trains was deemed to big.

In the meantime, other people at the central traffic control had frantically called the police emergency number, which called the harbor office to warn them.

The people at the harbor office was still on the phone with the police by the time the cars careened into the harbor area, moving at speeds exceeding 100 kph. There were no routine for sounding a "take cover" alarm all over the harbor.

So the people less than a mile from the impact at that point had no chance to get warned in the 30 seconds or so that was left until the crash. Three people were killed - one in a car that was crossing the tracks when it was impacted by the moving cut of cars, two in a container transload facility were the office was smashed by the runaway cars.

Well, now we know roughly what happened.

What we still don't know is why it happened - why the switches was lined as they were, bypassing the normal safety mechanisms, when the track brakes holding the cars were released.

No word yet on whether the operator on duty will face charges of manslaughter.

In these cases it is always hard to know exactly where you need to draw the line on personal responsibility versus designing the system so that the consequences of operator error is minimized. Because operator error will happen - nobody has yet designed a human being that never makes mistakes.

No matter whether that operator will be charged or not, he or she will have to live with the knowledge that whatever was done with the switches and with the track brakes in the end killed three people, left one critically wounded and two seriously wounded.

I guess we will just have to leave that question of whether to charge the operator to the proper authorities.

The important thing is to try to prevent it from happening again. No doubt there will be changes made in operating practices at the yard immediately.

Obvious technical changes would be maybe making it necessary for an outbound train to trip a signal at the south end of the yard, and then stop and wait for a timer circuit before the exit switch at the south end of the yard can be thrown to allow the train out of the yard or some such thing.

There perhaps should be found a good way of automatically returning the exit switches to the trap position after the last outbound car has passed over the switch, and not returning the exit switches to the trap position right in front of an inbound train coming up the 2.5% hill from the harbor.

Or at least sounding an audible alarm (which can not easily be turned off by the operator) if the switches remain lined for the freight line instead of the trap for longer than the time it normally would take a train to enter or leave the yard at the south end.

Maybe they also need to make changes to the track configuration where the futile attempts of stopping the runaway cars was made - making the single ended spur go straight ahead and the main track take the deviating path through the turnout. If there is room to make those changes down there - it is a pretty cramped area.

But the most important thing would be to prevent cars from rolling out of the yard when out of control - i.e. to prevent the exit switches at the south end of the yard from being set for the harbor track when there is no train just arriving or departing.

EDIT: I was looking at the <!-- m --><a class="postlink" href="http://www.bing.com/maps">http://www.bing.com/maps</a><!-- m --> again, and realized that I probably have shown the wrong path for the runaway cars at the top of the hump.

Looking again at the south end of the yard, I now realize that the runaway cars could not have rolled down the bypass along the right side of the yard (closest to the yard tower) - there is no way that they could have rolled out on the harbor track that way. The right hand bypass goes either into the trap or on the freight track to Grorud.

So my labels for the A/D tracks must be wrong - A5 must be the track furthest from the yard tower, not the A/D track closest to the tower, and the runaway cars must have rolled down the bypass along the left side of the yard, on the opposite side of the bowl tracks from the tower.

Doesn't change much in how the accident happened - the cars should still have been caught by the trap at the south end, if the switches had been lined into the trap instead of towards the freight track down to the harbor.

Stein

The one thing that sticks in my mind is if a switch engine was working a cut of cars, why did the switch crew uncouple without first setting the brakes on the cars? If they were using the hump to classify cars, why have the switch lined for engine bypass? Is it normal procedure to hold a cut of cars in position using just the track mounted retarders? The accident sounds a lot like an accident on the U.P. here in Los Angeles that occurred about 5 or 6 years ago. A switch crew was working a cut of cars in Montclaire. They took a short cut of uncoupling the cars and then planned to set the hand brakes to hold the cars in position, instead of setting the brakes before uncoupling. From Montclaire there is a gradual grade all the way into the U.P. yard on Washington Blvd in Los Angeles probably 20 miles away. When the crew uncoupled the cars they started rolling toward Los Angeles and they got away from the crew. The cars ran the 20 or 25 miles downhill toward Los Angeles at speeds of 100 miles per hour. Grade crossings were tripped by the automatic sensors to turn on the lights and drop the arms to stop auto traffic. Eventually the cars were switched onto a dead end track where they crashed and dumped 5 or 6 loads of lumber. Fortunately in that case no one was injured, but the U.P. issued new orders on how to handle cuts of cars in yards.

Russ Bellinis Wrote:The one thing that sticks in my mind is if a switch engine was working a cut of cars, why did the switch crew uncouple without first setting the brakes on the cars? If they were using the hump to classify cars, why have the switch lined for engine bypass? Is it normal procedure to hold a cut of cars in position using just the track mounted retarders?

All good questions.

I suspect that they were not humping cars - that they were just doing some switching in the A/D yard above the hump. It sounds from media reports like they had just added a 16th empty car to a cut of 15 empty cars on one of the A/D tracks. If they had been using the hump, I would perhaps have expected them to drop the 16th car down into the bowl, onto the same track as the other 15 cars, rather than the cut being assembled on top first.

I don't know what the plan was for those cars - whether they were planning to eventually send them over the hump and down into a bowl track, or whether they were planning to have an engine pick up the cars from the A/D yard and take it somewhere - either north or out one of the freight tracks at the south end of the yard.

I also don't know why they seemingly were using just the track brakes at the south end of the A/D tracks to hold the cars.

The formal investigation will probably let us know whether just holding the cars with the track brakes was an approved procedure or a shortcut that the switching crews had developed on their own over time.

Stein

what kind of brakes do these cars have??

I worked for the SRA in NSW (government railways) and uncoupling a cars airlines automatically put the brakes on full bore

we dont have these `hump yards' either- for a carriage to move it has to be coupled to a diesel shunter

seems to me that any system that gives a carriage a shove and lets it roll till it bangs into something is bound by its nature to go horribly wrong sooner or later.....

:-(

(all shunting done now has the loco with a horn blast that its moving the carriage and the loco itself has one of those annoying `beep beep' style beepers whenever its in motion)

A fascinating, if alarming thread. I initially saw the story on the BBC News website which aroused lots of questions in my head. A lot of those have now been answered by the excellent posts by Stein in this topic. Many thanks for them.

A pity it always seems to take a tragedy to highlight safety shortcomings, but surely many operating practises will be changed now to prevent this happening again, here and maybe other similar places.

Hump yards have been around almost as long as there have been trains. I'm sure some railroads may still use them. Yes, they are an accident waiting to happen. But then, what isn't? We've come a long way in railroad safety. And we have yet a long way to go. I'm sure the local (US) economy has a lot to do with the reduction of crew size, and the subsequent increase in the possibilitiy of disasters. I don't know how the economy is in Europe, but it may be very similar to ours. Too few people to do too much work.