A1/1998R Train accident at Jyväskylä on March 6, 1998

On Friday 6 March, 1998, at 13.49 hours, a passenger train travelling from Turku towards Pieksämäki derailed when arriving at the Jyväskylä railway yard. The train consisted of one locomotive and eleven coaches carrying a total of about 300 passengers.

The train approached Jyväskylä at a speed of 143 km/h. The engine driver operated the brakes three times. The first time was at 840 m before the derailment turnout. The second time, i.e. a more efficient braking was operated after the main signal at 320 m before the derailment turnout. The emergency braking took place only two seconds after the second braking, but because of a braking delay of a few seconds, the emergency braking did not became effective until at about 100 m before the derailment turnout. The train arrived at the turnout at a speed of 110 km/h. The maximum speed allowed at the point was 35 km/h.

At the turnout the locomotive turned on its side, skidding across two lanes of a highway and turning over on its other side. The locomotive crashed against a bridge pillar roof first at about 200 m from the point where it turned over on its side. On derailing the locomotive pulled along its first two passenger coaches which having become detached from the locomotive twisted in the direction of travel and then turned over.

There were ten fatalities in the accident, the engine driver of the train and nine passengers in the two coaches having turned over. A total of ninety-four persons were injured, of which eight were seriously injured.

Of the fatalities, two passengers were travelling in the first coach and seven in the second coach of the train. These persons died primarily as a result of injuries sustained when falling through the windows of a coach turning over and then being crushed under the coach.

The accident was generated by five central factors:

The engine driver was tired. During the five-hour break, he had failed to have a good rest because of the poor sound insulation and the echo effect in the rest premises.

The engine driver had failed to react to the distant signal, because he was concentrating on making coffee. For this same reason the look-out driver had failed to react to the fact that the train had not been started to brake.

The engine driver had most probably thought that he was entering a different arrival track than he had actually been. A year earlier, all passenger trains arriving at Jyväskylä from the south used a different track through another turnout located further away a track which permitted a speed of 80 km/h. The following facts seem to imply that the intention of engine driver had been to use
track 1 instead of track 3:

The engine driver had driven or served as look-out driver in passenger trains running through a turnout (in which the maximum speed is 80 km/h) located further away from the derailment turn-out more often than he had driven or served as look-out driver in passenger trains running through the derailment turnout.

The speed adjusting device had been set at 60 km/h. The engine crews were used to entering a turnout where maximum speed is 80 km/h with this particular speed setting.

Prior to the emergency braking, the engine driver had started braking in a way which would have been normal for running through the point at which the maximum speed is 80 km/h.

In test runs operated by the Accident Investigation Commission, it was discovered that the first and second brakings of the accident train would have permitted the running of the train to track 1.

Due to the above-mentioned facts, the engine driver had also failed to react to the main signal indicating 35 km/h as the maximum speed on the approaching turnout. In both the situations mentioned the signal aspects (yellow and green light) are identical. When heading towards a turnout in which the maximum speed is 80 km/h, the additional sign 8 would have appeared above the main signal, which in this case did not happen.

The look-out driver had paid no attention to the high speed of the train when approaching the main signal; he was concentrating on making coffee. The look-out driver had perceived a restriction indication on the main signal and had immediately expressed his observation aloud. Due to the time of reaction of the driver of the train and the braking delay, the emergency braking did not start to have an effect until at a distance of about 100 m from the derailment turnout. There was not enough time for the emergency braking to become fully effective, and the train travelled at 110 km/h to the turnout where the maximum speed allowed was 35 km/h.

Certain factors linked with the operational culture of the organization feature as background causes of the accident. In this operational culture there were problems both on the operational and the management levels. On the operational level the problems manifested themselves in a careless attitude toward some safety regulations and instructions, which then compromised train safety. For example had the brake adjuster been set according to the instructions and had the approaching braking been performed adequately, the speed of the Jyväskylä accident train would probably have slowed down to a point influencing the consequences of the accident. On the management level, the problems involved the norms, standards, regulations and rules, and instructions has not having been updated.

The subsystem may have had an effect on the generation of the accident. When running to turnouts with a maximum speed allowed either 35 or 80 km/h, the signal aspects (a yellow and green light) are identical. The only difference is the additional number display below the distant signal and above the main signal. The display is dark when running to an Sn35 turnout while it shows the number 8 when running to an Sn80 turnout. The possibility of mixing up the signals becomes real in very rapidly evolving situations, in situations where the driver concentrates on things other than operating the train, or in case a certain future situation is so strongly anticipated by the driver that the differences in the aspects are not big enough to ensure a sufficient stimulus to the perception.

From the point of view of rescue operations, the accident took place at a central location with good communications. For example Jyväskylä Central Hospital and the Jyväskylä Fire Station were at a two minutes distance. The notice of the accident reached the Alarm Centre in about one minute from the accident. The first rescue units were at the scene of the accident in about seven minutes. In about thirty-seven minutes from the notice of the accident, the last injured person was ready for transportation to hospital. The rescue operation worked out quite well with the exception of the alarm and supervision methods which in certain aspects could be improved.

The Accident Investigation Commission visited nine families in order to inform them about the progress of the accident investigation and leaving them contact data should they have any further questions. The families had felt confused about receiving news and information on the accident only from the media.

The final report by the Accident Investigation Commission includes recommendations for improved train safety. One central recommendation focuses on speed recommendations for distant signals and main signals, optional seat belts to be introduced in the day coaches of new and renovated long-distance trains, certain aspects linked with railway yard planning, updating of instructions for locomotive operation, improvements in the rest premises in the depots, and finally some development measures involving rescue operations.

The total economic loss generated by the accident was about USD 4 million, excluding social costs.

Preface, Summary, Contents, pdf-file partly in English (pdf, 0.5 Mt)

A1/1998R Report (pdf, 0.75 Mt)

Appendices (pdf, 0.26 Mt)

Appendix photos (pdf, 0.48 Mt)

• Recommendation S18
• Recommendation S78
• Recommendation S79
• Recommendation S80
• Recommendation S81
• Recommendation S82
• Recommendation S83
• Recommendation S100
• Recommendation S101
• Recommendation S102