Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

How do fires affect the railway? - A Swedish study on what factors influence fire risk and how it affects railway traffic.

Gustafsson, Jonatan LU (2023) VTVL01 20231
Transport and Roads
Abstract
Forest fires can start in a variety of ways. "Fires started deliberately" and "fires started by sparks generated from a train braking" were two of the most common causes of fires between 2010 and 2020. Train braking accounted for 1.1% of all fires, whereas fires created by unknown reasons accounted for 43% of all fires.

The climate is expected to change in the future, and since different weather factors affect the risk of fire differently, it is interesting to observe how train traffic is currently affected by fires and where these fires occur, in order to prepare for and ease the consequences of future weather changes. In this thesis, the number of rescue missions reported was compared to two databases containing information concerning... (More)
Forest fires can start in a variety of ways. "Fires started deliberately" and "fires started by sparks generated from a train braking" were two of the most common causes of fires between 2010 and 2020. Train braking accounted for 1.1% of all fires, whereas fires created by unknown reasons accounted for 43% of all fires.

The climate is expected to change in the future, and since different weather factors affect the risk of fire differently, it is interesting to observe how train traffic is currently affected by fires and where these fires occur, in order to prepare for and ease the consequences of future weather changes. In this thesis, the number of rescue missions reported was compared to two databases containing information concerning fires near the railway. They were compared because rescue operations imply that the fire was so broad at the time that aid was required, which makes it interesting to examine. Between 2010 and 2020, these comparisons produced 1,337 matches. Where matches represent how many times a fire was reported on one of these databases on the same day and location that a rescue mission was established. When delays were calculated, just one database was compared to the number of rescue missions, because the other one did not possess information about delays. This comparison produced 621 matches with a total delay of 203,821 minutes.

The distribution of the counties where the matches and delays occurred was observed in order to determine which additional variables might have contributed. The population and land area were compared to the proportion of matches and delay minutes, and it was noticed that the counties of Stockholm, Västra-Götalands, and Skåne, which had the most people, were also in the top four for the most matches and the top eight for the most delay minutes. The municipalities with the largest values were also examined more closely, for both the 1,337 and 621 matches. For the 1,337 matches, the average temperature and average precipitation for the ten municipalities with the largest values in each category were compared to the average value for all 161 municipalities that were affected. For the 621 matches, the matches, the delay minutes and the number of affected trains for the ten municipalities with the largest values in each category were compared to the average value for all 130 municipalities that were affected.

Regressions were used to calculate the impact of various weather events on the number of matches and the number of delay minutes. In the regression for the number of matches, temperature and precipitation were used as the independent variables which were taken from the 1,337 matches, these gave a result that both variables explain a portion of the matches. The regression performed for the number of delay minutes used the same independent variables but instead took them from the 621 matches, this showed that the variables did not explain the outcome of the number of delay minutes.

Weather factors other than temperature and precipitation that have not been studied, such as relative humidity and wind speed, appear to have an impact on the number of matches, according to various sources. Where, for example, records on relative humidity from prior years correspond quite well with when a match occurs. Furthermore, the locations of matches and delays appear to be connected to the population, with more individuals implying a larger risk. (Less)
Abstract (Swedish)
Några utav de vanligaste orsakerna till att en brand uppkom mellan 2010 och 2020 var ”bränder som medvetet startades” samt ”bränder som startats av gnistor vid tågbromsning”. Av den totala mängden bränder som uppkom stod tågbromsning för cirka 1.1 % av dem och den anledning som stod för absolut störst andel bränder var ”bränder startade utav okänd anledning” vilket stod för 43%.

Klimatförändringar är beräknade att ske i framtiden vilket gör det intressant att observera hur dagens olika väderfenomen påverkar antalet bränder vid järnvägen samt var dessa bränder förekommer, detta för att förbereda och lindra konsekvenserna av de framtida väderförändringarna. Detta arbete genomfördes genom att två databaser innehållande information angående... (More)
Några utav de vanligaste orsakerna till att en brand uppkom mellan 2010 och 2020 var ”bränder som medvetet startades” samt ”bränder som startats av gnistor vid tågbromsning”. Av den totala mängden bränder som uppkom stod tågbromsning för cirka 1.1 % av dem och den anledning som stod för absolut störst andel bränder var ”bränder startade utav okänd anledning” vilket stod för 43%.

Klimatförändringar är beräknade att ske i framtiden vilket gör det intressant att observera hur dagens olika väderfenomen påverkar antalet bränder vid järnvägen samt var dessa bränder förekommer, detta för att förbereda och lindra konsekvenserna av de framtida väderförändringarna. Detta arbete genomfördes genom att två databaser innehållande information angående bränder runt järnvägen jämfördes med antalet rapporterade räddningsuppdrag. Jämförelsen gjordes då ett utfört räddningsuppdrag med största sannolikhet inneburit att branden vid tillfället varit utbredd till den mån att hjälp behövts, och därför indikerar att branden varit större vid detta tillfälle. Detta resulterade i 1,337 matchningar under de 11 åren mellan 2010 och 2020, där en match representerar att en brand rapporterats i minst en av dessa databaser under samma dag och plats som ett räddningsuppdrag har utförts. När beräkningar gjordes på förseningar jämfördes endast en utav databaserna med räddningsuppdragen då den andra databasen inte innehåller någon information angående förseningar. Jämförelsen mellan de två databaserna gav 621 matchningar som resulterade i totalt 203,821 förseningsminuter.

Distributionen för både matchningarna för de tre databaserna samt matchningarna för de två databaserna undersöktes, detta för att utvärdera hur matchningarna var uppdelade i de svenska länen. Genom att erhålla denna information kunde fler variablers påverkan studeras. Population samt landarean jämfördes med andelen matchningar och andelen förseningsminuter vilket visade att Stockholms, Västra-Götalands samt Skånes län vilka har störst befolkningsmängd också var bland de fyra länen med flest matchningar. Dessa var också bland de åtta länen med mest förseningsminuter. Närmare studerades också kommunerna med störst medelvärden för de olika variablerna för både de 1,337 matchningarna och de 621 matchningarna. För de 1,337 matchningarna jämfördes matchningar, medeltemperatur samt medelnederbörd för de tio kommunerna med störst värden i varje enskild kategori med medelvärdet för alla 161 kommunerna som blivit påverkade. För de 621 matchningarna jämfördes istället matchningarna, förseningsminuterna och antalet påverkade tåg för de tio kommunerna med störst värden i varje enskild kategori med medelvärdet för alla 130 kommunerna som blivit påverkade.

För att bedöma väderfenomenens påverkan i antalet matchningar och antalet förseningsminuter gjordes två separata regressioner. Regressionen för antalet matchningar undersökte temperaturens samt nederbördens påverkan vilka togs från de 1,337 matchningarna. Detta visade att båda variabler hade en påverkan över en del av matchningarna. Den andra regressionen som undersökte hur antalet förseningsminuter blivit påverkad använde samma oberoende variabler fast från de 621 matchningarna istället och visade att dessa variabler inte hade någon påverkan över antalet förseningsminuter.



Väderfaktorer som inte undersökts då information inte fanns att tillgå, däribland relativ fuktighet och vindhastighet, verkade ha en inverkan på antalet matchningar enligt olika källor. Exempelvis har relativ fuktighet från tidigare år överensstämt väldigt väl med när matchningarna inträffat. Utöver detta verkade lokaliseringen av matchningar och förseningar förekomma där befolkningsmängden varit större vilket kan tyda på en större risk för bränder vid större populationer. (Less)
Please use this url to cite or link to this publication:
author
Gustafsson, Jonatan LU
supervisor
organization
course
VTVL01 20231
year
type
M3 - Professional qualifications ( - 4 Years)
subject
keywords
Train, railway, fire, forest fires, delays, weather, Sweden.
language
English
id
9125218
date added to LUP
2023-06-15 11:01:22
date last changed
2023-06-15 11:01:22
@misc{9125218,
  abstract     = {{Forest fires can start in a variety of ways. "Fires started deliberately" and "fires started by sparks generated from a train braking" were two of the most common causes of fires between 2010 and 2020. Train braking accounted for 1.1% of all fires, whereas fires created by unknown reasons accounted for 43% of all fires.

The climate is expected to change in the future, and since different weather factors affect the risk of fire differently, it is interesting to observe how train traffic is currently affected by fires and where these fires occur, in order to prepare for and ease the consequences of future weather changes. In this thesis, the number of rescue missions reported was compared to two databases containing information concerning fires near the railway. They were compared because rescue operations imply that the fire was so broad at the time that aid was required, which makes it interesting to examine. Between 2010 and 2020, these comparisons produced 1,337 matches. Where matches represent how many times a fire was reported on one of these databases on the same day and location that a rescue mission was established. When delays were calculated, just one database was compared to the number of rescue missions, because the other one did not possess information about delays. This comparison produced 621 matches with a total delay of 203,821 minutes.

The distribution of the counties where the matches and delays occurred was observed in order to determine which additional variables might have contributed. The population and land area were compared to the proportion of matches and delay minutes, and it was noticed that the counties of Stockholm, Västra-Götalands, and Skåne, which had the most people, were also in the top four for the most matches and the top eight for the most delay minutes. The municipalities with the largest values were also examined more closely, for both the 1,337 and 621 matches. For the 1,337 matches, the average temperature and average precipitation for the ten municipalities with the largest values in each category were compared to the average value for all 161 municipalities that were affected. For the 621 matches, the matches, the delay minutes and the number of affected trains for the ten municipalities with the largest values in each category were compared to the average value for all 130 municipalities that were affected.

Regressions were used to calculate the impact of various weather events on the number of matches and the number of delay minutes. In the regression for the number of matches, temperature and precipitation were used as the independent variables which were taken from the 1,337 matches, these gave a result that both variables explain a portion of the matches. The regression performed for the number of delay minutes used the same independent variables but instead took them from the 621 matches, this showed that the variables did not explain the outcome of the number of delay minutes.

Weather factors other than temperature and precipitation that have not been studied, such as relative humidity and wind speed, appear to have an impact on the number of matches, according to various sources. Where, for example, records on relative humidity from prior years correspond quite well with when a match occurs. Furthermore, the locations of matches and delays appear to be connected to the population, with more individuals implying a larger risk.}},
  author       = {{Gustafsson, Jonatan}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{How do fires affect the railway? - A Swedish study on what factors influence fire risk and how it affects railway traffic.}},
  year         = {{2023}},
}