The TRIMODE Integrated Model for Europe

The TRIMODE Integrated Model for Europe


Davide Fiorello, TRT Trasporti E Territorio, Angelo Martino, TRT Trasporti E Territorio, Klaus Nökel, PTV Group


This paper presents the TRIMODE integrated model for Europe that combines the simulation of transport, economy and energy systems for the assessment of major transport infrastructure projects and policies


This paper presents the TRIMODE integrated model for Europe that combines the simulation of transport, economy and energy systems for the assessment of major transport infrastructure projects and policies. Within a single software platform, the TRIMODE model components include a full four stage transport model of passenger and freight movements across Europe, an energy model with dynamic vehicle fleets for all transport modes and an economy model representing the complete macroeconomic system of European countries.
Covering the whole European Union and it neighbouring countries, the model has a very detailed spatial resolution of the transport system, whose multimodal network and zoning systems are based on the NUTS III level and below, coupled with a very detailed disaggregation of transport demand by socio-economic groups and trip purposes for passengers and by commodity groups and modes of appearance for freight. From the energy point of view the model incorporates complete vehicle fleet models for cars, trucks, trains, vessels and planes, including all types of energy sources and technologies. The economic impacts of transport are dealt with by a two layers General Equilibrium model that works at regional basis.
The main focus of the paper will be on the passenger modelling in the specific context of a EU wide scale model, considering both generation, distribution and mode sequence choice within long distance intermodal passenger trips. The passenger demand model consists of the main stages embedded within the PTV VISUM software plus several appended modules developed as Python scripts to manage in a consistent and transparent form some tasks that are not part of standard software functions. Main features of the passenger transport modelling in TRIMODE are:
• Demand generation managed by means of trip rates covering the whole mobility (i.e. including local pedestrian trips) for several trip purposes and population groups based on employment status, car availability and income. This approach allows us to estimate the number of trips and to explain differences across regions without the need for estimating region-specific parameters (the differences are the result of different socioeconomic composition of regions’ population) as well as to forecast future changes without the need for updating parameters (the changes are the result of modifications of the socioeconomic characteristics of population, e.g. income growth, motorisation, etc.).
• Modelling of intra-zone trips based on a classification of zones in different types and on distribution of trips among different distance bands. Zone types – from rural areas to metropolitan areas – reflect the heterogeneity of transport supply and level of congestion. Distance bands allow to describe the local component of mobility with some level of detail but in manageable form also in terms of calculation burden.
• Estimation of the split of demand for mode “car” among various usage segments (LDV, private car, taxi, car sharing, car-pooling) in order to be able to incorporate in the model the separation between car use and car ownership
• The explicit modelling of multi-modal chains in the assignment algorithm, in order to cope with the fact that long-distance passenger journeys will use a wide range of mode combinations, including even within individual journeys modes from both private and public transport. The TRIMODE assignment is therefore formulated as a bi-level problem:
• At the upper level (the choice is between mode sequences and mode transfer points (called hubs for short). Each alternative consists of a mode sequence and a choice of hubs between the mode legs (e.g. Chester Car Manchester Air Frankfurt Rail Karlsruhe).
• At the lower level for each mode the O-D demand from the upper level is assigned in a separate assignment for that mode.
The model construction is split in two phases: Phase I version of the model will be ready by summer 2017 and will include six European Countries: UK, NL, DE, AT, IT, SI; Phase II will cover the whole Europe and will be completed in 2019. This paper will include Phase I results.


Association for European Transport