-
Notifications
You must be signed in to change notification settings - Fork 13
4. Exploring the output of the model.
CO2MPAS produces by default 1 output file per each input file plus an additional batch summary file that contains a set of results from each input file. CO2MPAS produces within the output file a tab named output_report containing the main results and a tab named dice_report that is issued for sampling. Note that the dice_report is hidden in the CO2MPAS output file. CO2MPAS is built in such a way that the nomenclature of the output and summary files is as follows:
- output file: YYYYMMDD_HHMMSS-inputfilename.xlsx
- batch summary file: YYYYMMDD_HHMMSS-summary.xlsx
The output file contains the following set of Excel tabs:
This Excel tab contains an individual code for the simulated vehicle Vehicle family ID, the CO2MPAS version used, the date and time when CO2MPAS was used, and whether CO2MPAS was used in type approval or engineering mode.
The second table presents the declared and the simulated average specific CO2 emissions ki-factor corrected and it shows the calculated deviation as (Prediction - Target)/Target*100. For each WLTP interpolation family, the manufacturer shall declare to the approval authority the NEDC CO2 combined value for vehicle H, and, where applicable, vehicle L. That value should be Ki factor corrected. The NEDC CO2 reference values for vehicle H and, where applicable, vehicle L simulated by the CO2MPAS and Ki factor corrected can be found in second raw. The CO2MPAS deviation (%) is shown in the third raw.
The final NEDC CO2 value for test vehicle H or L to be used in the type approval certificate shall be the manufacturer declared value, if CO2MPAS deviation does not exceed 4 per cent (4.0000001 is not accepted as valid result). CO2MPAS deviation may be negative without any limitation. If the CO2MPAS deviation exceeds the manufacturer declared value by more than 4 per cent, the NEDC CO2 simulated by the CO2MPAS may be used in the type approval certificate for test vehicle H or L, or the manufacturer may request that a physical measurement is performed.
The third table presents the information about CO2 emissions simulated by CO2MPAS for UDC and EUDC phases of NEDC. These values are not Ki factor corrected and should be used for the purpose of calculations outlined in section 3.3. of Annexes to the correlation regulation.
The fourth table presents the submodel scores for the simulations. The CO2MPAS model is comprised of 8 submodels each one of them is calibrated based on the input WLTP data provided. The calibrated model is then used to predict NEDC. Essentially, a model fails the calibration when CO2MPAS is unable to accurately (according to the empirical scores set by JRC) reproduce the behaviour over WLTP. A warning is announced to the user but CO2MPAS proceeds to the prediction phase. The metrics and scores for each submodel are described in the site. A color code is assigned to the score in order to help the user to qualitatively assess the performance of each submodel (dark green meaning Good performance and graduately changing to orange which indicates a problem in the calibration). Additionaly, the output file summarizes basic inputs of the simulated vehicle (RLs, masses, engine capacity, etc.). This summary may contain confidential data but it is not in any case subject to online circulation.
This tab contains non-confidential information, including the CO2MPAS deviation, to be communicated to the time-stamp server via a timestamped email. After this communication the DICE will be rolled selecting randomly 10% of the cases for a pysical test. The type approval authority or the designated technical service, as well as the supervisory body (DG CLIMA), will receive the dice stamp stating the dice decision (either SAMPLE or OK), that is to say whether to run a physical NEDC test (taking into account the precisions specified in section 3.3. of Annexes to the correlation regulation) or not. Note that the sampling is implemented from the "Stamp" release and it is to be used by means of the command console.
It contains a comprehensive overview of the main model inputs, outputs and scores regarding the functioning of the CO2MPAS sumodels.
These tabs contain a graphical overview of the calibration, and prediction of the CO2MPAS model for the aforementioned cycles. They aim at understanding the general behaviour of a simulated vehicle for the main temporal series (electric currents, engine power, engine speed, velocity, temperature, fuel consumption).
The following tabs contain all the data that was input the CO2MPAS model as well as all the data that was produced during the simulation. This information is explicit for transparency purposes. Note that a CO2MPAS output file can be run as input since it contains all the necessary input data.
The proc_info Excel tab contains a list of the model version (together with the Python libraries version), as well as the runtime of the execution and the flags that were used. This information is essential to replicate the run if needed in the future to get the same outputs.
The summary file contains the CO2 target and prediction values for every phase as well as a list of the model CO2 parameters, and the information of success/fail flags of CO2 submodels for all the vehicles that were launched together in the same simulation. It is a light file that can be easily stored for future reference/comparisons.