The rail network is divided up into sections, each controlled by a signal that acts as a traffic light. If a section of track has a train on it the signal will show red to the next train. This process stops two trains being in the same section of track and keeps trains a safe distance apart.
Each train acts like a switch when it enters a section of track. This means that the two rails the train runs on are electrically connected. The signalling system then knows the track is occupied and stops another train entering that same section.
Sometimes the joints between the sections of track can become worn which would activate a safety system. This will make the signal, in that section, red. Debris likes trees and shopping trolleys can also connect the two rails which could act like a train. Flooding affects this process and could cause a temporary connection between the rails also.
Less often signals are unable to identify a train because there has been a build up of materials on the track or on the train wheels. This is most common in the Autumn when leaves on the line stop the electrical connection between the train and the track.
A train cannot progress past a red signal without special permission from a Network Rail signaller. When authorised to do so it must travel very slowly.
The process of allowing a train to pass a red signal involves the driver calling the signaller and asking for permission to continue safely onwards. The driver is then warned that the signaller is unable to confirm the track is clear, and the train must travel slowly. Both the signaller and driver have to complete a short form and record a reference number. This process adds approximately 10 minutes to each train. When there are lots of trains this process can take longer as a queue of trains may have formed.
Matthew Broad asked the BBC to help him find out what causes signal failures, you can see what Matthew discovered in this BBC news video.