Closed Jinbo0201 closed 1 year ago
By default fuel output is in mg/s (to obtain fuel/s, set option --emissions.volumetric-fuel)
Parts of the documentation are outdated and still refer to ml/s unfortunately.
Nevertheless, there may still be problems with the default HBEFA3 model. The following plot shows consumption for a car accelerating at 1m/s^2 (in mg/s) the values match your data but are much higher than for the newer HBEFA4 model.
@behrisch should we maybe change the default model to HBEFA4?
The below plot is for ml/s and the lower data point at the right is for driving at steady speed. This multiplies out to about 10l/100km while driving at 144km/h for HBEFA3 and an unbelievable 2l/100km for HEBA4.
despite this, the error metric for this class stands at 4.32% (https://sumo.dlr.de/docs/Models/Emissions/HBEFA4-based.html#passenger_cars).
@piwagner can you take a look at the original HBEFA data?
It turns out that the HBEFA4 coefficient added in release 1.14.0 were fitted against km/h instead of m/s which results in the observed divergence.
Just a follow-up question: For HBEFA4 in the current version of SUMO, what is the speed and acceleration unit? -- km/h and m/s^2 -- m/s and m/s^2 -- km/h and km/h^2
everything is now working with m/s and m/s^2
I obtain the edge information by traci.edge.getFuelConsumption(edgeID) and traci.edge.getCO2Emission(edgeID). The length of edge is 2 km. The data shows that, in step#0, there are 2 running vehicles and 10536ml/s Fuel Consumption and 33401mg/s CO2 emission. The data is unbelievable.