To evaluate all factors influencing the fuel consumption a simulation model (TRANSTEP) was developed, in order to estimate fuel consumption and greenhouse emissions for all the fleet in each line.
TRANSTEP main input data is schematically presented in Fig. 1 and is discriminated as follows:
• Fleet characterization: number of vehicles and vehicles characteristics such as:
• Engine maps (power, consumption and torque as a function of load and rpm);
• Transmission ratios and efficiency;
• Gear box management maps;
• Tires characteristics;
• Vehicle’s dimensions and weight;
• Network operational characterization: number of lines and trip frequency, occupancy rate, number of bus stops (urban and non-urban) and vehicle used in each line;
• Network topography characterizations: distances (urban and non-urban) and slopes of each line;
• Experimental measurements data: vehicles dynamics, topography data, registered events (e. g. time on a bus stop, stops because of traffic or lights);
• Instantaneous fuel consumptions for specific events (accelerations, decelerations, cruise velocity and idle) for different slopes and traffic conditions. This data can be obtained experimentally by installing a flux meter in the bus. An alternative solution is to use numerical simulation models that can estimate on a micro scale level the fuel consumption of a predefined bus running in a specific line. In the present study the authors used a numerical model developed at Institute Superior Tcnico named EcoGest 0.
Some of this data can be supplied by the bus operators or by the vehicles manufactures. However, experimental measurements in normal operational conditions are always needed leading to a complicated and cumbersome task.
Fig. 1. Scheme of interaction ofTRANSTEP and its components.
Since it takes too much time to evaluate every single line and bus for different periods, the methodology involves the selection of two or three characteristic lines for a detailed experimental evaluation. This experimental data is further on treated and fed into TRANSTEP.
The output results of TRANSTEP are effective fleet fuel consumptions disaggregated in:
• Fuel consumption for different events;
• Fuel consumptions for different lines;
• Annual fuel consumptions to the fleet/network;
• Elasticity of each parameter influencing fuel consumption.
With these results it becomes possible to identify measures in order to improve the company’s energy efficiency and, therefore, reduce the fuel consumption of the fleet.
These results are also used to estimate greenhouse emissions based on fuel conversion factors into C02 emissions (for diesel vehicles green house emissions are basically C02 emissions), see Table 1.
Table 1. Diesel conversion factor into CO2 emissions. 
kg C02 per unit
Sources: National Air Emissions Inventory, UK Greenhouse Gas Inventory, Digest of UK Energy Statistics DTI 1998, Greenhouse Gas Inventory Reference Manual IPCC 1996 (http://www. defra. gov. uk/environment/envrp/gas/05.htm).
normal operational conditions at different periods of the day and week, in order to allow the characterization of peak and off peak hours, and week days versus weekend days.
During experimental measurements data is collected regarding the topography of each analyzed line, vehicles dynamics (speed versus time and distance), significant events, occupancy rates, atmospheric conditions and road surface conditions.
For the case study presented, the equipment used for the topography and dynamic characterization was the following:
• Suunto’s Escape 203 altimeter to register altitude;
• Corrsys Datron Sensorsystems, М3, Doppler’s effect Microwave sensor, to register distances and instantaneous velocity and accelerations;
• Laptop, for data logging.
For this case study, it was not possible to install flux meters in the buses, therefore our in-house developed numerical model – EcoGest was used to estimate the instantaneous fuel consumption. These data was then calibrated with measured data regarding monthly and annual fuel consumption.
In this particular application, two measurements methods were adopted, one with the equipment installed directly in the bus and other with the equipment installed in a car that followed closely the bus along the route, thus emulating its dynamics.
In the first case the equipment was installed in the front of the bus and the electric feeding was provided by an auxiliary battery as presented in Fig. 1.
Fig. 2. Experimental measurement’s equipment installed in the front of the bus.
In the second case the equipment was installed on the car door as shown in Fig. 3
Fig. 3. Experimental measurement’s equipment installed in the lateral of a car.