Characterisation of an urban bus network for environmental purposes

Since pollutant emissions are closely related to the operating conditions of vehicles, their evaluation usually involves studying these operating conditions (through bus instrumentation and monitoring under actual operation), the design of representative driving or engine test cycles and the measurement of pollutant emissions. A preliminary characterisation of the routes on a bus network should make it possible to identify typical routes, the driving conditions and pollutant emissions of which are then studied. Two approaches are envisaged and applied to the Paris area, for which a wealth of information is available, which should be transferable to other bus networks. Both approaches are based on factorial analysis and automatic clustering, to allow optimum description and the identification of a pertinent typology of the bus routes in several classes. The first attempt at characterisation is based on statistics relating to bus operations: route characteristics (length, dedicated bus lanes, number of stops, location of stops: schools, tourist sites, hospitals, railways or underground stations), travel time, commercial speed, annual statistics (number of passengers, number of vehicles per hour, total kilometres), the irregularity of travel (variation of travel times, injuries, congestion.), as well as information on the problems encountered (congestion, distribution of the passenger load, junctions, bends). A second approach is based on the analysis of the "urban context" in which buses are driven. Population, employment, housing, road network, traffic and places that generate or disturb traffic (schools, railway stations, shopping areas, etc.) are calculated for the Ile de France region, by cells of 100 x 100 m, and collected in a geographical information system (GIS). Statistical analyses enable a typology of these urban cells to be established, the main parameters being density, type of housing, road types and traffic levels. The bus routes are then analysed according to their itineraries across these typical areas (distances travelled in each type of area) using a similar approach. A comparison of the typologies obtained from operational data and from urban data highlights the advantages and disadvantages of the two approaches. The first result from these typologies is the selection of routes which are representative of the different classes, in order to instrument buses and record driving patterns. This method should also make it possible to link driving conditions and urban characteristics, and then to allocate pollutant emission factors to given geographical situations, in particular, in the context of emission inventories or impact studies.