The Urda Collision Followed an LZB Failure and Degraded Operation | The Calibre Brief Ep. 67
About this episode
This episode examines the 17 minutes between the failure of the LZB train-control system and the collision near Urda, Toledo. LZB initially stopped several trains automatically, including an Iryo train and a Renfe AVE train behind it. The Iryo train remained stopped near signal E242 while the Renfe train was later authorised to move.
The verified sequence shows that the danger emerged during the shift to degraded operation, when fallback signalling, communications, operating rules and ASFA replaced LZB’s continuous supervision. Reporting describes a possible restricted-speed instruction, but the applicable rule and the meaning of the authorisation remain under investigation.
It is not yet established why the Renfe train was allowed to proceed, what information the control centre had, whether train-position or signalling data were inaccurate, or whether one failure or several interacting failures caused the collision. A proposed AI monitor might have flagged contradictions, but it could not replace certified train protection or make unreliable data reliable.
0:00 A safety system stopped two trains on its own 0:34 ASFA did not stop Renfe 0:53 The sequence is clear, but the cause is not 1:18 But an alert cannot fix bad data 1:37 The lesson is not simply a system that stopped 1:51 This is The Calibre Brief from Calibre Code USA
Stock footage: tiburi via Pexels (https://pixabay.com/videos/id-9005/) Stock footage: MEISTERvideo via Pexels (https://pixabay.com/videos/id-32000/)
Key topics
Key moments
Transcript
Show transcript
Welcome to The Calibre Brief, a Calibre Code USA production.
A safety system stopped two trains on its own. Seventeen minutes later, one was moving toward the other. What changed after the first warning?
At signal E242, an Iryo train remained stopped on the Madrid-Seville high-speed line near Urda, Toledo. Behind it, a Renfe AVE train had also stopped when the LZB protection system failed across roughly 63 kilometres. The first safety step appeared to work. The harder question is what happened after both trains stopped.
LZB normally watches movement, speed, and distance from danger. When it failed, trains had to use backup rules: signals, messages, ways to keep trains apart, and possibly “march at sight,” meaning slow movement. ASFA did not stop Renfe. Reports say it was allowed to move while Iryo remained ahead. That was the key change.
The sequence is clear, but the cause is not. Investigators know LZB failed, the Iryo train stayed ahead, Renfe movement was allowed, and a collision followed. They do not yet know what the control centre saw, what words were used, or what speed applied. Location data may have been wrong or incomplete. “Inexplicable” names the puzzle, not proof of negligence or blame for one person.
Spain has proposed an AI monitor that would compare train positions, signals, and messages between the control room and train. It might flag a mismatch: LZB is down, one train is stopped ahead, and another is told to move. But an alert cannot fix bad data. It cannot replace certified protection that enforces movement and braking.
The lesson is not simply a system that stopped working on its own. Danger grew when train locations and movement orders no longer fit. AI may expose that gap, but cannot replace backup rules.
This is The Calibre Brief from Calibre Code USA. Should AI only alert operators when protection fails, or block movements itself? Comment below, like and subscribe on YouTube, and follow the podcast.
