|
How do road
transport emissions work and how they are modelled
Emissions from
transport vehicles are modelled in a
number of different ways. Emissions are calculated as the sum of at least two
main components: ‘hot emissions’ and ‘cold start emissions’. If the modeller is
interested in VOC emissions, then also the ‘evaporative term’ has to be added.
Hot emissions are the emissions emitted when engine and abatement
devices have reached a sufficient (‘regime’) temperature. They are influenced
by a number of parameters: vehicle kinematics, gradient of the road, altitude,
maintenance level, vehicle age, vehicle loading,
electric loads addition (for lights and for air conditioning). Normally models
refer to hot emissions as a function of kinematics (e.g. average speed or
instantaneous speed and acceleration) and then multiply the ‘ideal’ hot
emission value for a number of corrective factors taking into account the other
mentioned parameters.
Cold Start emissions are the emissions emitted from the start up until the
vehicle reaches an almost steady state condition in terms of temperature of
engine and abatements system. The cold start emissions affect in practice the
first 3 – 4 kms of trips, and are particularly
relevant for catalyst vehicles: for CO and VOC the ‘cold start emission’ is
roughly 10 times higher than the hot emission. Since the EU fleets are getting
more and more ‘catalyzed’ this factor has a present and future relevance for
emissions modellers. Accuracy in modelling cold start emissions guarantees a
good global accuracy of the emission model.
Evaporative
emissions are the emissions of unburned
fuel from the ‘weak points’ of the vehicle: tank and canister. Current
modelling recognises three different contributions to evaporative emissions:
·
Running emissions,
emitted when vehicles are driven (emissions at tank level)
·
Hot Soak emissions,
emitted from the canister at trip conclusion
·
Diurnal emissions,
emitted at tank level by vehicles already parked (not included in usual traffic
modelling...)
Evaporative emissions
are a relevant (30 – 40 %) fraction of total transport related VOC emissions,
and so have a major role in the planning of measures for reducing VOC related
pollution (e.g. benzene pollution in southern European cities)
Several correction
factors can be added to these primary terms:
·
Gradient correction,
particularly important for heavy-duty vehicles;
·
Maintenance
correction, relevant for poorly maintained fleets;
·
Age correction,
important for older vehicles (‘aged fleets’);
·
Altitude correction,
relevant only in mountain areas crossed by vehicles ‘tuned’ at low altitude;
·
Load correction,
needed for correctly modelling light and heavy-duty vehicles (LDVs and HDVs);
·
Electric loading,
important especially for small and medium cars being equipped with Air
Conditioning systems.
The road transport
emission models available at international level (see reviews in COST319,
FP4 DGVII COMMUTE Project, COST346) include several
different ‘categories’ of models that join some common characteristics.
One of the most used
‘categorisations’ split these models into two main ‘families’: Aggregated
Models and Disaggregated Models (see COST 319 reports and web site).
Aggregated Models deal normally with an entire city or a whole Country and
model traffic as a global entity (total vehicles km driven in the area is the
traffic amount input). In some cases (e.g. COPERT tools) a whole country is
split into 3 entities corresponding to area types: cities, motorways and rural
roads. Vehicle kinematics is represented by average speed that is only
differentiated through vehicle classes and the three types of area.
Cold start is here
calculated from an average trip length (e.g. 12 km proposed for Italy) without any spatial differentiation. Some corrections
factors are applicable (age, maintenance) while for other factors the
representation is basically impossible (e.g. gradient effects).
Disaggregated Models consider the transport network and normally receive direct
input from transport models which produce input information such as link flows,
link speed etc. Emissions are calculated link by link. The vehicles kinematics
can have a different treatment (average speed on the link, or speed profile
along the link, or instantaneous speed and acceleration).
Basically all the
corrections factors can be applied. The important cold start fraction can here
be estimated link by link on the basis of information produced by the upstream
traffic model (e.g. average distance driven from the trip origin to the link
being considered).
Critical Issues at
Urban Scale
Vehicle kinematics plays a relevant role. Traffic models in general provide
the average speed as link attribute. Several tools nowadays provide more
detailed kinematics data. For the sake of emissions accuracy it is crucial to
take into account the kinematics. In fact the average speed is a poor local
indicator of the emission and fuel consumption level. There are infinite ways
in which a vehicle can experience an average speed level on a link, and these
conditions show very different consumption and emission levels (ranges – from
lowest to highest values – of around 300-400 % and more for some pollutants
have been observed!).
Cold start emissions have an essential role. The urban average trip length is
estimated from measurements to be around 3 to 6 km depending on the city type
and size. So the fraction of cold vehicles is in the average between 50 and 90
% approximately.
The distribution of
these cold vehicles is obviously uneven: higher in areas where trips start
(e.g. residential areas in the morning) and lower in areas where trips end
(e.g. Central Business Districts in business days at 9 a.m.). A correct representation of this variability is
essential for avoiding very large errors in the calculation of the most
important ‘emission term’ representing up to 90% of the emissions emitted in an
urban trip (case of catalyst cars for emissions of CO and VOC).
Evaporative emissions have a crucial role when cities cope with severe
challenges from high benzene and-or PAH concentrations. The accurate spatial
and temporal representation of this term involves the modelling of parking
processes. Information for this would include location of parked vehicles, and
characterisation of parking and inserting flows. This can give the basic input
for assigning the important emission contribution to the correct position in
the network and the right time evolution along the day. |