Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Designing a road traffic model for the cross-sectoral analysis of future national infrastructure

Milan Lovrić, Simon Blainey & John Preston, University of Southampton

11–13 September 2017

International Symposia for Next Generation Infrastructure (ISNGI) meeting in 2017 brought together a global community of infrastructure academics, policy and industry professionals. The event aimed to share progress, knowledge and new thinking, and a number of ITRC-MISTRAL’s researchers presented their work. 

  • Login to see the comments

  • Be the first to like this

Designing a road traffic model for the cross-sectoral analysis of future national infrastructure

  1. 1. Designing a Road Traffic Model for the Cross-sectoral Analysis of Future National Infrastructure Milan Lovrić, Simon Blainey, John Preston University of Southampton ISNGI London, 11-13 September 2017
  2. 2. ITRC Consortium • ITRC (Infrastructure Transitions Research Consortium): • Interdependent infrastructure systems (transport, energy, water, waste, digital communications). • NISMOD v1 (2011 – 2015) – the first family of models to analyse the long-term performance, plans, and risks and vulnerabilities of the national infrastructure under future uncertainty. • Used by the UK’s NIC to inform its National Infrastructure Assessment.
  3. 3. Aspirations for MISTRAL (NISMOD v2) • MISTRAL (Multi-Scale InfraSTRucture AnaLytics) project (2016 – 2020). • Integration of capacity, demand and risk modelling frameworks. • System model with packages of policy interventions: – New road development. – New bus/rail services (e.g. HS2). – New technology/modes (e.g. autonomous vehicles). – Electrification of vehicles. – Congestion charging. • Global connectivity: integration with international demand/supply nodes at model boundaries. • Risk and resilience: identification of most vulnerable points on networks.
  4. 4. Fast-track Case Study (Highway Demand Model) freightpassenger• Transport model predicts highway demand (OD matrix): – For passenger and freight vehicles. – Elasticity-based simulation. – Network assignment to major road network. – Implemented in Java (GeoTools). • Fast-track case study: – Four local authority districts (LADs). – Three interventions: • Road expansion • Road development • Vehicle electrification – Cross-sectoral interdependencies: • Input: electricity price (per kWh). • Output: total electricity consumption.
  5. 5. Passenger Vehicle Demand Model • Passenger demand (passenger vehicle flows) are predicted using the following formula: Where: - 𝐹"#$ is the flow between origin zone i and destination zone j in year y. - 𝑃"$ is the population in zone i in year y. - 𝐼"$ is the GVA per head in zone i in year y. - 𝑇"#$ is average travel time between zone i and zone j. - 𝐶"#$ is average fuel cost between zone i and zone j. - elasticity parameters are taken from previous studies. 𝐹"#$ = 𝐹"#$+, 𝑃"$ 𝑃"$+, -. 𝐼"$ 𝐼"$+, -/ 𝑃#$ 𝑃#$+, -. 𝐼#$ 𝐼#$+, -/ 𝑇"#$ 𝑇"#$+, -0 𝐶"#$ 𝐶"#$+, -1 𝜂3 = 1.0 𝜂4 = 0.63 𝜂5 = -0.41 𝜂6 = -0.215
  6. 6. Freight Vehicle Demand Model • Freight demand (freight vehicle flows) are predicted using the following formula: • Freight model uses different elasticity values and different travel time/cost matrices. • Three types or freight vehicles: artics, rigids and vans. • Freight zones can be: LADs, major distribution centres, airports and seaports. • Adopted from the DfT’s Base-Year Freight Matrices study (2006). 𝐹"#$ = 𝐹"#$+, 𝑃"$ 𝑃"$+, -. 𝐼"$ 𝐼"$+, -/ 𝑃#$ 𝑃#$+, -. 𝐼#$ 𝐼#$+, -/ 𝑇"#$ 𝑇"#$+, -0 𝐶"#$ 𝐶"#$+, -1 𝜂3 = 1.0 𝜂4 = 0.7 𝜂5 = -0.41 𝜂6 = -0.1
  7. 7. Network Assignment (Node Choice) • Origin and destination zones (LADs) are relatively large compared to the road network. • Finer census output areas with their population size are used for the node choice.
  8. 8. Network Assignment (Node Choice) • Population weighted centroids are assigned to the nearest neighbour nodes. • Nodes are then ranked based on the gravitating population.
  9. 9. Network Assignment (Routing)
  10. 10. Network Assignment (Routing) • AADF UK major road network (A roads and motorways). • OD flow is assigned to the least-cost path between origin and destination node. • Fastest path (based on congested link travel times, using a heuristic search algorithm). • Alternative implementation based on a route-choice model. • Off-line route generation will be performed on a computer cluster.
  11. 11. Link Travel Time Update • Link travel times are updated as (BPR): 𝑇7 = 𝑇8 1 + 𝛼 < 6 = , – Tc is a congested travel time on a link, – To is a free-flow travel time on a link, – V is hourly volume [PCU/lane/hour], – C is max. road capacity [PCU/lane/hour], – α, β are parameters. • Alternative specification using fundamental diagrams of traffic flow (FORGE, DfT).
  12. 12. Skim Matrices Update • Contain inter- and intra-zonal travel times and travel costs. • Calculated after network assignment as average travel time/cost across all chosen paths. • Feeds back into the elasticity-based simulation: 𝑇"#$ 𝑇"#$+, -0 𝐶"#$ 𝐶"#$+, -1
  13. 13. Capacity Utilisation • After the network assignment of passenger and freight vehicle flows, the capacity utilisation of the road network can be assessed. • Capacity “pinch points” can be identified – candidates for policy interventions.
  14. 14. Interventions (Road Expansion and Development) • Road expansion = building new lanes. • Road development = building new links. • Expected impact: – Lower capacity utilisation and decreased travel times. – Somewhat increased demand due to lower travel times (see the elasticity-based model). Intervention: road expansion Intervention: road development
  15. 15. Interventions (Road Expansion and Development) (a) No intervention (b) Road expansion (c) Road development • Predicted road capacity utilisation after policy interventions:
  16. 16. Interventions (Vehicle Electrification) 5% 45% 35% 10% 5% 15% 40% 30% 10% 5% ELECTRICITY PETROL DIESEL LPG HYDROGEN ELECTRICITY PETROL DIESEL LPG HYDROGEN 2015 (base year) 2020 (no intervention) 2020 (electrification) (a) Fuel type market shares (b) Predicted car fuel consumptions • Increased total electricity consumption → energy demand model. • Reduced environmental impact.
  17. 17. Cross-sectoral Interdependencies • TR – transport • E – energy • DC – digital communications • SW – solid waste • W – water TR E DC WSW Energy demand Energy supply Bandwidth demand Bandwidth supply Transport demand Power outage Service disruption Flood • Interdependencies between transport and the energy sector: – Energy supply → electricity unit price (kWh) → Transport – Transport → total electricity consumption → Energy demand
  18. 18. Full-scale Highway Model • Major road network for Great Britain (A roads and motorways). • Adding ferry lines. • OD matrix estimation (TEMPRO trip end data, trip length distr., AADF count data). • Code optimization. • Offline route generation. • Calibration with traffic counts.
  19. 19. Other Major Tasks • National rail model. • Airport and seaport model. • Global interconnectivity. • Cross-sectoral interdependencies. (T + E + DC + SW + WS) • Integration with risk & resilience models. • Environmental impacts. • Validation and calibration.
  20. 20. Acknowledgments The authors acknowledge funding of the work described here by the EPSRC (Engineering and Physical Sciences Research Council of the UK) under Program Grants EP/I01344X/1 and EP/N017064/1 as part of the Infrastructure Transitions Research Consortium (ITRC, and MISTRAL projects. We also thank all ITRC colleagues for their continuing help in developing and adapting the modelling approach presented here. This presentation contains Ordnance Survey data ©Crown copyright and database right (2017). Further InformationFurther Information