Learnings from MoCho-TIMES - Modal choice within bottom-up - PowerPoint PPT Presentation

Learnings from MoCho-TIMES - Modal choice within bottom-up optimization energy system models ETSAP Meeting College Park, 10 th -11 th July 2017 Jacopo Tattini PhD Student Energy System Analysis group

Motivation MoCho-TIMES model Discussion Motivation • Bottom-up (BU) energy system models describe in detail the technical, economic and environmental dimensions of an energy system • They are weak in representing consumer behaviour : only one central- decision maker is considered • The behavioural dimension is fundamental in decision making in the transportation sector  It shall not be neglected • Essential to represent real households ’ preferences For more info: Venturini et al., Improvements in the representation of behaviour in integrated energy and transport models, 2017 (Under revision) 2 19 July 2017

Motivation MoCho-TIMES model Discussion MoCho-TIMES model • MoCho-TIMES ( Mo dal Cho ice in TIMES) is an approach to incorporate modal choice directly in BU optimization energy system models • The methodology consists in two main steps: 1. Divide transport users into heterogeneous consumer groups 2. Incorporate intangible costs • Other constraints: -Monetary budget -Availability of transport infrastructures -Travel Time Budget (TTB) -Travel patterns -Maximum shift potential -Maximum rate of shift For more info refer to working paper: Tattini et al., Improving the representation of modal choice into bottom-up optimization energy system models – The MoCho-TIMES model, 2017 3 19 July 2017

Motivation MoCho-TIMES model Discussion Demand side heterogeneity • Heterogeneity differentiates modal perception among subgroups of transport users DENMARK Modes have different levels of service Region Region DENMARK DENMARK 2 1 Region EAST WEST URBAN SUBURBAN RURAL Urbanization URBAN SUBURBAN RURAL area VERY LOW INCOME VERY LOW INCOME VERY LOW INCOME VERY LOW INCOME VERY LOW INCOME VERY LOW INCOME MEDIUM INCOME MEDIUM INCOME MEDIUM INCOME MEDIUM INCOME MEDIUM INCOME MEDIUM INCOME HIGH INCOME HIGH INCOME HIGH INCOME HIGH INCOME HIGH INCOME HIGH INCOME LOW INCOME LOW INCOME LOW INCOME LOW INCOME LOW INCOME Different LOW INCOME Income perceptions of level levels of service 4 19 July 2017

Motivation MoCho-TIMES model Discussion Intangible costs Intangible costs are introduced for two reasons: 1. To capture other non-economic factors into the expression of the generalized cost, accounting modal perception 2. To differentiate modal perceptions across consumer groups through monetization. Varies across types Varies across of urbanisation income classes 5 19 July 2017

Motivation MoCho-TIMES model Discussion Overall structure of MoCho-TIMES Perceived Infrastructure Demands Travel cost Fuel Consumer Group 1 Consumer Group 2 Consumer Group 3 Consumer Group 24 time ... Intangible cost CG1 MONETARY BUDGET NON Intangible MOTORIZED cost CG2 … Intangible cost CG24 TRAVEL TIME Intangible BUDGET cost CG1 PUBLIC Intangible TRANSPORT cost CG2 … EXISTING Intangible INFRA- cost CG24 STRUCTURE Intangible cost CG1 NEW INFRA- STRUCTURE Intangible PRIVATE CAR cost CG2 … Intangible cost CG24 6 19 July 2017

Motivation MoCho-TIMES model Discussion Data requirement • Many new data are required: – Spatial distribution of the population (region, type of urbanization) – Income distribution across the population – Mileage distribution across the population – LoS attributes: free travel time, congestion travel time, waiting time, walking time, access/egress time, etc – Value of time (VoT) – Infrastructure data: investment and O&M costs, capacity utilization level – Travel pattern: share of km in the urban/suburban/rural areas – Public transport fares – Car parking cost – ….. • Need a rich and reliable data-source, consistent with the energy system model that will incorporate modal choice 7 19 July 2017

Motivation MoCho-TIMES model Discussion Support model • The development of MoCho-TIMES requires a support model: -Transport model able to simulate modal choice -Consistent with the geographical scope of the energy system model • The support model is used to draw data and parameters for MoCho-TIMES • The transport model might have a different time horizon than the energy system model  Assumptions required • In case support model is not available, a travel survey (travel diary) could be used Transport Model 8 19 July 2017

Motivation MoCho-TIMES model Discussion Reflections • Modal choice is determined at aggregated level, for macro clusters of consumers, but is able to capture variability acorss population • Dimensions for heterogeneity is crucial • Finer resolution is achievable, but trade-off trade-off between model size and representation of the population shall be pursued • Additional variability to modal perception achieved through the ” clones ” • Vague spatial resolution  Focus is not trip, but entire energy system • Heterogeneity overcomes the “mean -decision maker” perspective • Perfect-information, perfect-foresight and perfect-rationality 9 19 July 2017

Motivation MoCho-TIMES model Discussion Shall MoCho-TIMES be incorporated into an integrated energy system model? • Modal shift as an option to decarbonize energy system,within a unique model framework. • Effect of energy system dynamics on modal shares and vice versa • Transport sector is expected to become increasingly integrated into the energy system • New policy and scenario analyses: effect of variations of LoS and consumers ’ perception of modes on rest of energy system and viceversa • Intangible costs act as a barrier to decarbonisation of the transport sector  Required consistency across sectors Compare MoCho-TIMES and soft-linking of TIMES with external transport model (ABM+system dynamic model) 10 19 July 2017

…questions, suggestions?!?! Jacopo Tattini jactat@dtu.dk 11 DTU Management Engineering, Technical University of Denmark

Soft link of TIMES-DK and LTM Modal choice in LTM and technology choice in TIMES-DK Outputs from TIMES-DK: Fuels prices • Inputs to ABM+SD model: Socioeconomic description: • Iterations gender, income class, car ownership, age, nr. of children, marital status, Interface TIMES-DK GDP , employment Infrastructure: existing and • planned Average mode travel cost • … • Outputs from LTM (2010-2030): Passenger travel demand per mode , • location, purpose (pkm) Freight travel demand per mode, location, • ABM+System purpose (tkm) …….. • Dynamic 13 DTU Management Engineering, Technical University of Denmark

Motivation MoCho-TIMES model Discussion MoCho-TIMES vs Soft-link with external model Soft link with transport model MoCho-TIMES Advantages: Advantages: • Transport • Wider scope of analysis, including models have suitable structure and mathematical the energy system • Enables expression (MNL) for computing assessing cross-sectoral modal shares influences • Spatial disaggregated • Flexible for scenario analysis • Household/Individual resolution • Catch some variability of preferences Disadvantages: Disadvantages: • Macro-clusters of consumers • Long computational time of transport • Aggregated spatial resolution model • Low sensitivity to price changes • Iterations required? 14 19 July 2017

Disaggregated modal shares 15 19 July 2017

Disaggregated modal shares 16 19 July 2017

Disaggregated modal shares 17 19 July 2017

Disaggregated modal shares 18 19 July 2017