Evacuating along elevated platforms 24 May 2018

An experiment carried out in 2016 at the Skarpnäck subway station in Stockholm recorded human behaviour when evacuating the metro running tunnel along an elevated platform. A total of 111 people, approximately half men and half women, ranging from 18 to 77 years old and including three wheelchair users participated in the experiment.

Evacuation walkway created for the tests
Photo: Per Rohlén

As the running tunnels at the Skarpnäck subway station lack an elevated platform, a 93.6m-long screen wall was built along the edge of the station platform adjacent to the northern track. The width of the walkway for different tests was established by altering the distance between the wall and the edge of the platform. A train was parked next to the walkway along the first half of the elevated platform. The platform was level with the train floor, making the height difference between the walkway and the track 1.4m.

The experiment comprised five tests. In tests one, three and five, the entire test group moved from one end of the elevated platform to the other. In tests two and four, the test participants were divided into two groups; group A moved along the elevated platform as before, and group B began inside the train and joined the flow on the platform once group A passed the train doors. Reference tests were conducted in a 30m-long corridor built with two screen walls on the southern end of the platform. The platform widths tested were 1.2m, 1.05m and 0.9m. After the experiment, the participants were asked to complete a survey to further assist the findings.

The flow rate was measured at four points:

• the reference corridor,
• the first half of the elevated platform,
• the second half of the elevated platform and
• the doorway of the train.

Test results

The flow rate at the first part of the walkway, where the train was parked, was higher than the flow rate at the last part of the walkway where the platform was open to the track and higher also than the flow rate in the reference corridor. The gathered video material and survey clearly showed that the difference in height between the elevated platform and the track area affected the test participants, and so the difference in flow rate between the first and second parts of the walkway can be attributed to the drop from the platform to the track. Why the flow rate was lower in the reference corridor was not investigated further within the framework of the study, but one possible explanation is different lighting conditions.

To get an indication of the distribution between the flow on the walkway and the flow out of the train, the merging ratio was observed at the point where group B stepped out of the train and joined group A on the walkway. In general, the flow rate out of the train was larger than the flow rate along the walkway. This was true in both group test scenarios.

The participants did not walk close to the platform edge. No one used the 10cm closest to the track area. This was true for the full length of the walkway, independent of the width. The participants walked closer to the platform edge when there was a train parked next to the walkway than they did when it was open to the track area.

The various formations formed by test participants during the experiment

When the participants were asked about their ability to pass others, 41.8% answered that they never felt a need to overtake someone in the experiment. In a real evacuation situation, due to a fire for example, it can be assumed that the need to overtake others would increase. This is partly related to the fact that people who are related to one another try to stay together to a greater extent than the participants of the experiment did. Some of the participants who chose to keep to their own pace during the test, despite the fact that this separated them from friends and family, stated that they would not have done so in a real situation, where they would have adjusted their speed to stay together.

Survey feedback

Of the test participants, 32.7% stated that they did not overtake anyone during the experiment, but that they would have liked to do so on one or more occasions if they had had the opportunity. Together with those who answered that they overtook others, but with some difficulty, during the experiment (a total of 14.5%), 47.2% stated that they experienced problems related to overtaking others.

When there was enough room for participants to walk independently of one another, they generally chose to do so close to the screen wall. Those who knew one another prior to the experiment often chose to walk side by side, while the others generally preferred to walk in single file. When it became more crowded, the test participants adopted different formations depending on the width of the footpath. In tests with a 1.2m walkway, participants walked in two columns. In those with a 1.05m walkway, participants still walked in two columns but in a staggered fashion; in the test with a 0.9m walkway, they walked in single file.

Of the three wheelchair users, one felt uneasy due to the width of the walkway and/or the difference in height between the elevated platform and the track area. As a result, this person only participated in tests one and two in which the width of the footpath was 1.2m, and did not participate in the tests in which the walkway was narrower. This person stated that the width of the walkway was of great importance to a sense of safety. The fact that this person stopped and hesitated where the walkway opened up to the track area indicates that the height difference to the track also was of great importance.

In order to make evacuating along elevated walkways more effective, the possibilities for people to overtake others needs further review. Since nearly half of the participants experienced difficulties passing others, one suggestion for future studies is to investigate how the design of an elevated walkway can be improved to make overtaking easier.

The difference in height between the elevated platform and the track area below was of particular influence and affected the test participants in several ways. The flow rate along the walkway decreased and the width of the walkway was not used to the same extent where it was open to the track area as compared to the part of the walkway that the train was parked next to. The height difference also affected one of the wheelchair users to the extent that this person stopped in the middle of the walkway in two of the tests, and chose to not participate at all in the remaining ones.

Further studies should investigate how elevated platforms can be designed, and with what aids they can be equipped, in order for height differences to have less effect.