A study of the performance and utilization of the Swedish railway network Anders Lindfeldt Royal Institute of Technology 2011-02-03
Introduction The load on the railway network increases steadily, and with the up-coming deregulation of railway traffic, it will increase even more. Need to get a picture how the railway network is utilized and how it is performing. Analyzing an entire network manually is time consuming Use available data in automatic procedures to calculate useful measurements of performance.
Swedish railway 8 100 km of single track, 1 00 km of double track. High capacity utilization on the main lines. Great speed differences between trains on the same track (high speed trains and freight trains).
Method Extract relevant data from several different databases regarding: - Infrastructure t - Timetable (planned) - Train properties (operational data) - Delays (operational data) Divide the network into smaller parts according to traffic patterns and type of infrastructure.
Method Define parameters describing the properties of the lines. Compute the parameters for each part of the network. Use maps to present the data to make it easy to get an overview of the status of the network. Perform an analysis to find out if any of the parameters can be used to explain the delays.
Parameters Infrastructure - Station distance, station track lengths - Share of stations ti with more than three tracks - Share of stations with parallel movement facility Timetable - Number of trains running during different parts of the day
Parameters Train properties - Mean speeds - Speed heterogeneity - Freight train mass, length, nr of axles - Gross weight / day Delays - Share of trains that have received an additional delay - Median delay/km of the delayed trains
Data sources Infrastruktur Tidtabell Trafik Förseningar Enkelspår: Antal tåg per dag Godståg Persontåg/Godståg Avstånd mellan mötestationer (km) Totalt/Persontåg/Godståg min Antal tåg per dag Vikt (ton) Andel merförsenade tåg max min medel Antal tåg per timme: max Median merförsening normerat med standardavvikelse Totalt/Persontåg/Godståg medel sträckans längd [min/100 km] Under maxtimmen Maxtimmen standardavvikelse Andel 3-spårs stationer På morgonen 06-0 Standardavvikelse merförsening normerat På eftermiddagen 15-18 Längd (m) med sträckans längd [min/100 km] Andel stationer med samtidig infart På eftermiddagen 16-17 min Under dagtid -15 och 18-20 max Dubbellspår: Under natten 20-06 medel Avstånd mellan förbigångsstationer (km) standardavvikelse min Hastighet (km/h) max Persontåg/Godståg Antal axlar medel max min standardavvikelse min max medel medel Alla banor och stationer (m): median standardavvikelse Hinderfri längd min Hastighetskillnader Axellast medel standardavvikelse d min max standardavvikelse/medel max 5 percentilen/10 percentilen medel standardavvikelse Bruttoton/dag g( (ton) Persontåg Andel med 12 Andel med > 12 Databas: Databas: Databas: Databas: BIS T08.3 BANSTAT TFÖR Mätperiod: Mätperiod: Mätperiod: Mätperiod: 2008-12-1 2008-10-0 2008-10 2008-0 och 2008-10
Number of trains / day Black: Total number of trains. Green: Freight trains. Range: 0-267 trains/day and direction. Freight trains dominate in the north and passenger trains around the bigger cities. ff h d h Mix of freight and passenger trains on the double track lines in the south.
50 7 Average speed (all trains) Green: Low speed. Red: High speed. Range: 23-142 km/h. High speed on double tracks, lower speed on single tracks due to frequent stops at crossing stations. 70 73 46 54 75 77 53 52 60 61 65 63 61 77 86 78 73 58 4 83 76 34 78 75 6 63 73 122 65 3 87 127 81 68 105 87 67 65 123 67 136 116 7 61 56 1 50 108 6 76 62 76 63 61 64 88 8 71 78 60 4 66 73 72 66 83 112 12 7 86 105 8 70 84 66 107 73 104 8 81 7 75 100 64 82
1,11 Speed mix (all trains) 5 percentile divided by the 10 percentile. Green: Homogeneous speed. Red: Heterogeneous speed. Range: 1-26 2.6 [-] [] Lines with the most heterogeneous traffic are the double track lines with extensive freight and passenger traffic. 1 2 2 1,1 1,8 1,2 1,6 1,1 1, 1,3 1,5 1,1 1 1,4 1,3 2 1,3 1,8 1,6 2 1, 1,5 1, 1, 1,5 2,3 1,2 2,1 1,7 1, 1 1,6 1,5 2 1,7 2,2 1 2 1,3 1,8 1,6 1,4 1,2 1,2 2 1,7 1,7 1,1 1,7 1,7 1,3 1,1 1,3 1,4 2,3 1, 8 1,22,3 1,2 1,11 2,2 1,7 1,5 1,81, 1,3 1,8 2,4 1 2,6 1,6 1 1,1 1,4 1,1 1 1,1
32 37 27 Share of delayed trains (passenger trains) 31 34 Share of the passenger trains that have received an additional delay [%]. 45 26 36 30 61 18 20 Green: A small share of the trains are delayed. 57 45 37 58 41 48 35 31 50 Red: A big share of the trains are delayed. 44 54 68 23 46 27 45 30 44 50 21 48 45 40 38 26 26 60 44 3 28 55 42 27 18 36 26 25 325 36 45 8 34 82 52 22 56 36 42 1 26 22 64 15 Range: 14-8 %. The highest values are due to temporary speed restrictions. 3 41 24
27 28 Station distance (mean) Red: Long station distance. Green: Short station distance. Range: 5.5 5 8 km. Usually longer distances between siding stations (double track) than between crossing stations (single track). 18 15 50 15 10 6 12 8 15 6 40 16 17 8 12 22 44 11 10 12 13 6 25 10 11 12 20 10 10 47 16 8 12 7 33 17 18 17 16 16 14 11 16 12 10 26 8 10 13 20 15 17 6 23 14 15 16 26 18 14 6 16 10 7 11 6
Station track length vs. freight train length Share of the freight trains that are longer than the mean track length of the stations on the line. Range: 0 61 %. (Green-Red) The effective station distance increase when crossings/sidings cannot be performed on all stations.
Delay correlations Investigate if the calculated parameters correlates to the delays on the lines. Stepwise multilinear regression used. Correlations found significant at the 0.05 level: Train type Delay type Parameter Slope Passenger Share Total nr of trains/day 0.44 Passenger Share Mean speed (all trains) 0.42 Passenger Share Speed mix (all trains) 0.2 Freight Share Total nr of trains/day 0.57 Freight Share Share of stations with at least 3 tracks -0.46
General conclusions By defining suitable parameters based on data about infrastructure, timetable, train properties and delays, it is possible to perform a time efficient analysis of a nationwide railway network. It has been shown that the following properties are correlated to the delays: - The total nr of trains/day - Mean speed and speed mix of the trains - Share of stations ti with at least 3 tracks Data quality issues.