 We are told, automation is inevitable. I mean, it will be eventually, will be it many, many years away. The two most mentioned places for automation will come the quickest, at least here in the UK, are that of cars and trains. Personally, I think cars are a long way off as how can an automated system deal with our pretty narrow roads here in the UK? I mean, just look at some of this video. Well, trains have to be easier to make automatic, right? Well, yes and no. Like everything, it comes down to cost. Just look at the money pit the four lines upgrade has become on the underground. It's much easier to build an automation right from the start. And that was the case for today's subject, the Washington DC Metro. However, automation would be the cause of today's disaster in which a train operating automatically would speed right into the rear of another train requiring the system to be operated manually in the aftermath. Today, we're going to have a look at the Washington Metro crash of 2009. And my name is John. Welcome to Plainly Difficult. Forward. So I've got a real soft spot for 1960s and 70s Metro projects in the US. When they were built, many tried to incorporate the cutting edge of technology. This was, and still kind of is today, automatic train control. But microprocessors and computers weren't really powerful enough at the time. So they exist in a weird realm of modern yet old. The networks used fixed block signaling principles and were made up of individual subsystems. Unlike, say, the modern day cell tracks moving block system where movement authority, supervision and control are all in one neat little package. Raging hard ons for networks like the barter side, this period of rail technology for me is at least a very fascinating one. And one such interesting network, which I hadn't really considered until working on this script, is of the Washington Metro. It really sits in that golden era of US Metro building. So as always, the video will be making use of a few different sources. Most notably the NTSB report, as well as a few other things along the way. And as always, the linkies will be in the description below. Background. So our story begins in the 1960s. The network's beginnings were, as said, on the WM-ATA website. The Washington Metropolitan Area Transit Authority, Metro, was created by an interstate compact in 1967 to plan, develop, build, finance and operate a balanced regional transport system in the national capital area. You see, the capital city needed a better public transport system. And an all new and shiny rail network was planned to provide this, as well as integrating buses, connections with the new metro system. Think like transport for London. Building began in 1969 for the Metro rail system, opening its first section of five stations of the red line in 1976. And over the next few decades, the network would expand with a new station as recently as May, 2023. But we aren't here for a full blown history of the network. We are here for the geeky stuff. Well, I am at least. And this is in the system in which trains are signalled and controlled. Right, so from the start, the network was envisioned as modern as possible for the time. And that was an automatic train operation system. That is, train control was taken out of the hands of the operator. Although manned with the driver's cab, normal stopping, starting and driving was done by the automatic train control system. The system in use on the Washington Metro gets rid of the majority of associated track size signals, opting for an in-cab system. So the whole ATC system comprises of three subsystems, each taking care of an important aspect of signalling and controlling the trains, ATP, ATS and ATO. So the first and most essential is the automatic train protection subsystem. This essentially is the signals that the trains run to and the detection along the line. This is done by track circuits, which detect the presence of a train. The ATC system takes this information and gives the train a maximum safe speed. The track circuits used in the network are like this. Each block has an impedance bond on each end. One transmits a coded modulated sine wave audio signal from a local train control room, along the rails and to the end of the block, where the other impedance bond receives the signal. If the signal reaches the receiver and matches the transmitted signal after it's been filtered and amplified, a relay is energized. Then this will tell the ATC system that the block is unoccupied and a speed code can be transmitted to the next train. Now if a train is in the section, the train's wheels will interrupt the audio signal and the relay will de-energize, thus telling the ATC system that there is a train in the section. Now this is a really basic explanation. It's pretty much as far as I can understand it anyway. So with all this happening, the ATC system will give the maximum speed a train can run to, which if the next block is occupied, will be of the speed of zero miles an hour. The ATC system also takes into consideration any diverging or converging sets of points, and if the route is not set for the train, then the value of zero miles an hour will be generated. The speed signal is generated and transmitted via the rail to the train from the train control room, of which there is usually one at every station. This system is meant to be fail safe in that if a train loses its speed code, it will default to zero and basically stop the train. The next part of the system is the automatic train supervision subsystem or ATS. This system is not fail safe and is tasked with regulating the service, that is to set a maximum speed in which the train will travel to enforce the schedule. I should say that in the cab, there are three speeds displayed to the operator. The first is the maximum safe speed, which is set up by the ATP. Then there is the maximum allowed speed, which is set by the ATS. And then there's the actual speed of the train. The ATS speed can't be higher than the ATP speed. So at the beginning of each trip, the train's journey is transmitted to and stored on the train. At each station, if an adjustment is needed, the speed profile, acceleration profile and dwell time is transmitted to the train from the local train control room. And finally, the final subsystem is ATO or automatic train operation. This does the actual driving of the train. When a train approaches a station, line side equipment tells the ATO system to initiate a station stop. Stops are controlled by tuned passive calls on the track, which tells the train a distance remaining before a station stop. And that's a very simple explanation of the ATC system on the network. Now it's time to get out your bingo card. I'm already going to mark off complex system as my head hurts after reading about the Washington Metro's ATC system. The disaster. So it is the afternoon of the 22nd of June, 2009. And trains around the Fort Totten area on the red line are backing up. This had been caused by an earlier fault on the train. And due to the congestion, it caused several trains to be lining up, waiting to move along the line. Strangely, as trains had gone over a section of track, they were receiving a spurious zero mile an hour code. However, these dropouts were fairly short as trains entered the next section they received the continuation of their 55 mile an hour code. And this brings us onto train 214. It's been operated in mode two. Basically means it's being driven manually but under the protection and supervision of the ATC system. It's closely following the train ahead. And as such, the operator is having to regulate the train speed. Soon enough, train 214 gets that owner a zero mile an hour speed code. Train 214 runs into block 304. And due to the slower speed from the operator working the train manually, doesn't get to the end of the block to then receive the needed resumption of speed code. The operator of 214 was now stopped in block 304 and was awaiting a step up in code. But what no one that day knows is that block 304 has a fault that is not detecting trains. Meanwhile, the next train behind is being operated in mode one, i.e. automatic mode. This is train 112. It's following as train 214 was ahead. And before entering the failed block, 112 also got a zero mile an hour code. 112 came to a stand. But 39 seconds later was given a code of 55 miles an hour. This was because 214 was now fully in block 304. And was pretty much invisible to the signaling system. Train 112 speeds up to just shy of 50 miles an hour. As the train goes around the corner, the operator of 112 sees the tail lights of 214. She hits the emergency stop button, but it's too late. 112 slams into the rear of 214, riding the leading carriage over the rear of 214. The following carriage then telescope. The resulting crash pushed 214 forward. This was at around two minutes past five in the evening. Soon after, the first 911 calls started coming in. The controller noticed traction current going out in the area. He attempted to contact train 112, but no reply. At around the same time, the operator of 214 was making his way to the rear of his train to investigate the impact. He raided into control and informed them of the collision. The controller got a track worker to go to the accident site to confirm the report. And sadly, he saw two mangled trains. The controller then held all trains in stations on the red line, which would create crazy amounts of congestion for passengers trying to get home in the evening rush. Fire engines and ambulances were dispatched to the scene. Access was difficult, requiring a block on the nearby CSX lines and for the boundary fences to be cut. By the time the first responders had reached the site, some passengers had self-evacuated out onto the tracks. In order to get to some passengers, workers had to cut away at the train's bodywork. By an hour and a half post-event, all living victims were recovered and sent off for medical treatment around the capital city. A crane would be employed to help with the search and rescue, which would lead up to the next day. And finally, around 9 a.m. on the 23rd, everyone was recovered, including the dead. Nine would be confirmed dead, including the operator of train 112. So, with such a public disaster in the nation's capital, of course, the cause had to be found out and this would enter the NTSB. Investigation. In the immediate aftermath, the crash site was photographed and inspected by the FBI and Washington Metro's engineering team. This allowed the NTSB to investigate whilst cleanup work could be conducted. The NTSB would need to find out how an automatic train could rear-end another. This would lead investigators to look into the system's past as a tragically similar near-miss incident had happened just a few years before. In 2005, a track circuit between Foggy Bottom and Roslyn failed to detect the presence of a stopped train. The near-collision was avoided by the actions of the train operators, but it showed that the train could just disappear off the signalling system. After looking into this further, the NTSB found that the track circuits could be susceptible to a thing called parasitic oscillation. So remember how the train detection is done via an audio signal sign wave running through the rails and if this signal is interrupted, then the signal system knows a train is occupying the section? Well, if there is a fault with the ouster module then a spurious signal could enter the circuit which mimics what the receiver is looking for, basically a sign wave and thus report the block as clear when it actually wasn't. What's bizarre is that in 1999, the Metro undertook a program of track circuit replacement with the aforementioned Alstom system. Interestingly, after 2005, incidents Washington Metro and the module manufacturer Alstom didn't bring in enhanced tests to check the network's track circuits. The NTSB would even say this in the investigation. Contributing to the accident were, one, WMATA's lack of safety culture, two, WMATA's failure to effectively maintain and monitor the performance of this automatic train control system, three, GRS Alstom sign inks, failure to provide a maintenance plan to detect spurious signals that could cause its track circuit modules to malfunction. Four, ineffective safety oversight by the WMATA Board of Directors. Five, a tri-state oversight committee's ineffective oversight and lack of safety oversight authority. And finally six, the Federal Transit Administration's lack of statutory authority to provide federal safety oversight. The NTSB set out several recommendations, such as regular inspection of track circuits and a replacement of all 1,000 series Metro cars. This is because of the telescoping issue the Unix experienced in the crash and the very unsatisfactory crash-worthiness that had been seen. The Washington Metro would note this in an update. Retirement of the 1,000 series cars, the oldest in the fleet is Metro's number one safety priority. The NTSB recommended that Metro's 1,000 series rail cars, the oldest of the fleet be replaced with more modern vehicles that will be equipped with advanced crash-worthiness systems technology. The disaster ended up costing millions of dollars for a Metro as well as nine lives. In its wake, the disaster undermined the public and operators' confidence in the ATC system as such, all trains would be operated manually, albeit with the protection of ATP. Ironically, if the following train was operating manually, it was still very likely the disaster would have still have occurred as the failure was with the ATP system and train detection and not the automatic train operation. So what did you get on your bingo card? I got this and I'm going to give it a four on my disaster scale. Do you agree? Let me know down below. This is a plainly full production. All videos on the channel created commons that we should share like licensed. Plainly full videos produced by me, John, the currently wet and windy corner of Southern London, UK. I have Instagram and Twitter as well as the second YouTube channel to check out all the links in the description below. And I'd like to say a very warm thank you to my Patreon and YouTube members for your financial support as well as the rest of you for tuning in every week to watch my videos. And all that's left to say is Mr. Music, play us out please and thank you for watching.