Benchmarking and evaluation of railway operations performance

Transcription

1 Benchmarking and evaluation of railway operations performance Gemma Nicholson, David Kirkwood, Felix Schmid, Clive Roberts Birmingham Centre for Railway Research and Education, UK 25 March 2015

2 Outline Motivation Quality of Service (QoS) framework QoS quantitative evaluation Example of application ON-TIME project Conclusions and future work

3 Motivation Differing perspective depending on the stakeholder Evaluation of quantity and quality of operational behaviour Whole system approach Decomposition of system into properties Measure the effect of changing system properties, i.e., quantitative evaluation

4 Quality of Service framework Key measures Quantitative output KPIs Railway system System properties Railway system decomposition Influencing factors

5 QoS framework applications The QoS framework broadly has 4 applications: benchmarking of simulation tools comparison of tables or operational control systems visualisation of delay propagation linking of real operational data and microscopic simulation

6 QoSQE methodology

7 The full Quality of Service framework

8 Key measures The full framework KPIs Railway system System properties Influencing factors QoS framework outline

9 QoSQE performance indicators deviation area crew freight tonne km to recover delays at destination rolling stock passenger km journey no transfer interchange crowding maximum deviation station arrival delays energy consumption track Transport volume Journey Passenger Connectivity Resilience Punctuality Energy comfort Resource Railway system

10 TV key measures freight tonne km passenger km Transport volume journey no transfer Journey Transport volume interchange crowding deviation area to recover maximum deviation delays at destination passenger seat km = number of km travelled number of seats available on the service station arrival delays freight tonne km = number of km travelled freight train cargo capacity in tonnes energy consumption Passenger Connectivity total values Resilience Punctuality Energy comfort selected O-D pairs during T crew rolling stock track Resource Railway system

11 JT key measures freight tonne km Journey deviation area to recover delays at destination crew rolling stock passenger km journey no transfer interchange maximum deviation station arrival delays energy consumption The average journey crowding [seconds] of all journeys that make scheduled stops at O and D, in that order track Transport volume Journey Passenger Connectivity Resilience Punctuality Energy comfort Resource Railway system

12 CN key measures freight tonne km Connectivity deviation area to recover delays at destination crew rolling stock passenger km journey no transfer The average interchange interchange of maximum all interchanges station at I for crowding journeys that both depart O and deviation arrive at D arrival during delays the simulation T energy consumption track Transport volume Journey Passenger Connectivity Resilience Punctuality Energy comfort Resource Railway system

13 RS key measures freight tonne km passenger km Transport volume journey no transfer Journey Resilience deviation area to recover delays at destination Based on the system deviation measurement: interchange maximum station maximum deviation crowding during period T [seconds] deviation arrival delays to recover [seconds] deviation area [seconds 2 ] energy consumption Passenger Connectivity Resilience Punctuality Energy comfort crew rolling stock track Resource Railway system

14 PT key measures freight tonne km Punctuality deviation area to recover delays at destination crew rolling stock passenger km journey no transfer interchange maximum deviation station arrival delays energy consumption At station S, during crowding period T, the sum of arrival delays to all services departing S during T track Transport volume Journey Passenger Connectivity Resilience Punctuality Energy comfort at intermediate stations at terminal stations Resource Railway system

15 EG key measures freight tonne km Energy deviation area to recover delays at destination crew rolling stock passenger km journey no transfer For a interchange given O-D pair, the average maximum energy consumed station per crowding service for all services that both deviation depart from arrival O and delays arrive at D during period T energy consumption track Transport volume Journey Passenger Connectivity Resilience Punctuality Energy comfort Resource Railway system

16 RU key measures freight tonne km passenger km Transport volume journey no transfer Resource interchange crowding deviation area to recover maximum deviation delays at destination RU1: track : the average number of trains passing a point per hour during period T station arrival delays energy consumption RU2: rolling stock: the total number of rolling stock units in use Journey during period T Passenger Connectivity Resilience Punctuality Energy comfort crew rolling stock track Resource Railway system

17 Example: ONTIME project Project aim: an improvement in capacity by reducing delays and improving traffic fluidity Evaluation and comparison of traffic management approaches For minor perturbations For major perturbations

18 QoSQE applied in ON-TIME Benchmarking Network, table, period Simulation Evaluation Simulation Timetable delay delay delay (pre described scenarios) Simulation + real- perturbation management method delay introduced delay introduced Comparison between simulation and schedule at service level Key measure values Outputs Visualisation of KPIs

19 ONTIME case studies Iron Ore line, Sweden and Norway East Coast main line, UK Section of Dutch network, The Netherlands Minor perturbations e.g. temporary speed restriction, signal failure Major perturbations e.g. complete line blockage

20 ONTIME case studies Iron Ore line, Sweden and Norway East Coast main line, UK Section of Dutch network, The Netherlands Minor perturbations e.g. temporary speed restriction, signal failure Major perturbations e.g. complete line blockage

21 deviation area crew freight tonne km to recover delays at destination rolling stock passenger km journey no transfer interchange crowding maximum deviation station arrival delays energy consumption track Transport volume Journey Passenger Connectivity comfort Resilience Punctuality Energy Resource delay Railway system Capability Dependability Rolling stock Infrastructure Timetable Operational rules Traffic management Operational management Human factors System maintenance Environmental factors dynamic performance stations pattern train operational rules priority resource allocation planners condition monitoring technical protection facilities static characteristics track sections allowance infrastructure operational rules conflict detection incident management dispatchers maintenance plan environmental incident handling signalling buffer traffic operational rules conflict resolution drivers power network recovery crew operational rules delay management passengers communication network third parties passenger information system

22 IOL simulator benchmarking Comparison between tabled and simulated event s for IOL simulation

23 IOL simulator benchmarking stations NoNk NoSms NoRom NoKat NoBjf Rgn Kja Vj Lak Ka Bln Akt Ak Soa Kpe Sbk Tnk Bfs Rsn Rut Kv Pea Kia Rsi Apt Svv Kmb Kra tabled but not logged in simulation tabled train service indices table matches simulation Difference between tabled and simulated departure [seconds]

24 IOL quantitative evaluation Quantification of key measures 1. Demonstration of resilience KPI using in-built simulator logic only 2. Key measure results using perturbation management algorithms

25 NoNk NoSms NoRom NoKat 9922B NoBjf Rgn Kja Vj Lak 9921B stations Ka Bln Akt Ak Soa Kpe Sbk Tnk Bfs Rsn 9919B 9175B Rut Kv Pea 9926B [hours]

26 NoNk NoSms NoRom NoKat 9922B NoBjf Rgn Kja Vj Lak 9921B stations Ka Bln Akt Ak Soa Kpe Sbk Tnk Bfs Rsn 9919B 9175B Rut Kv Pea 9926B [hours]

27 deviation (min) (hours)

28 NoNk stations NoSms NoRom NoKat NoBjf Rgn Kja Vj Lak Ka Bln Akt Ak Soa Kpe B 9175B 9922B Sbk Tnk 9919B Bfs Rsn Rut Kv Pea 9926B [hours]

29 deviation (min) (hours)

30 deviation (min) (hours)

31 240 Resilience KPI 180 deviation (min) maximum deviation deviation area 0 to recover (hours)

32 Key measure results with perturbation management Punctuality

33 deviation (s) FCFS (hours) Resilience deviation (s) Algorithm Algorithm 2 deviation (s) (hours) (hours)

34 Summary of the other measures JT O-D Reference simulation FCFS Algorithm 1 Algorithm 2 Mean journey [min] Narvik - Bergfors (100%) (126%) (101%) (102%) RS FCFS Algorithm 1 Algorithm 2 Deviation area [s 2 ] 1.519e e e+08 (100%) Maximum delay [s] (94%) (83%) (100%) (98%) (66%) Time to recover [s] EG O-D Reference simulation Total energy consumption [kwh] Narvik - Bergfors (100%) FCFS Algorithm 1 Algorithm (101%) (100%) (101%) RU2 # trains 00:00:00-07:00:00 Reference simulation FCFS Algorithm 1 Algorithm 2 Passenger Freight Passenger Freight Passenger Freight Passenger Freight

35 Conclusions The QoSQE showed that the perturbation management approaches applied in ON-TIME resulted in improvements in particular to the resilience and punctuality KPIs less significant, but still positive outcome for the other KPIs Since the built-in simulator dispatching logic is used in the benchmark simulations, no quantitative conclusions can yet be drawn about the effects of perturbation management systems in real railway networks General implications Extension to platform independence Applicable to the assessment of operational data

36 Further work Formalisation of the key measures Application in other situations future projects using real operational data e.g. evaluation of traffic on the Birmingham Cross City line

37

38 Thank you Benchmarking and evaluation of railway operations performance Gemma Nicholson, David Kirkwood, Felix Schmid, Clive Roberts Birmingham Centre for Railway Research and Education, UK