Skip to main content
Warning: You are using the test version of PyPI. This is a pre-production deployment of Warehouse. Changes made here affect the production instance of TestPyPI (
Help us improve Python packaging - Donate today!

A vehicle simulator predicting CO2 emissions for NEDC using WLTP time-series

Project Description

CO2MPAS is backward-looking longitudinal-dynamics COsub(2) and fuel-consumption simulator for light-duty vehicles (cars and vans), specially crafted to back-translate consumption figures from WLTP cycles into NEDC ones.

It is an open-source project developed with Python-3.4, using Anaconda & WinPython under Windows 7, Anaconda under MacOS, and Linux’s standard python environment. It runs as a console command, with various graphical UIs on the making.


The European Commission is supporting the introduction of the WLTP cycle for Light-duty vehicles developed at the United Nations (UNECE) level, in the shortest possible time-frame. Its introduction requires the adaptation of COsub(2) certification and monitoring procedures set by European regulations. European Commission’s Joint Research Centre has been assigned the development of this vehicle simulator to facilitate this adaptation.

For recent activity, check the doc(changes).



About console-commands:

  • Console-commands beginning with $ symbol are for the bash shell (UNIX). You can install it on Windows with cygwin: along with these useful utilities:

    * git, git-completion
    * make, zip, unzip, bzip2, 7z, dos2unix
    * openssh, curl, wget
  • Console-commands beginning with > symbol are for Windows cmd.exe command-prompt. You can augment it with bash-like capabilities using Clink:

  • You can adapt commands between the two shells with minor modifications (i.e. ls <--> dir, rm -r <--> rmdir /s/q).

  • You may download and install the all-in-one archive which contains both shells configured in a console supporting decent copy-paste and resizing capabilities (see ref: all-in-one).

IF you have familiarity with v1 release AND IF you already have a full-blown python-3 environment (i.e. Linux or the all-in-one archive) you can immediately start working with the following bash commands; otherwise follow the detailed instructions under sections ref: install and ref: usage.

## Install co2mpas.
## NOTE: If behind proxy, specify additionally this option:
##    --proxy http://user:password@yourProxyUrl:yourProxyPort
$ pip install co2mpas

## Where to store input and output files.
## In *Windows* cmd-prompt use `md` command instead.
$ mkdir input output

## Create a template excel-file for inputs.
$ co2mpas template input/vehicle_1.xlsx

## Edit generated `./input/vehicle_1.xlsx` file. ##

## Run simulator.
$ co2mpas batch  input -O output

## Inspect generated results inside `./output/`. ##


The installation procedure has 2-stages:

  1. Install (or Upgrade) Python (2 choices under Windows).
  2. Install CO2MPAS:
    1. Install (or Upgrade) executable.
    2. (optional) Install documents.
    3. (optional) Install sources.

On Windows you may alternatively install the all-In-One archive instead of performing the above 2 steps separately.

All-In-One Installation under Windows

  • Download all-in-one archive from Ensure that you download the correct 32/64 architecture for your PC (the 64bit archive CANNOT run on 32bit PCs).
  • Use the original “7z” extraxtor, since “plain-zip” produces out-of-memory errors when expanding long directories. Prefer to extract it in a folder without any spaces in its path.
  • Run INSTALL.bat script contained in the root of the unzipped folder. It will install links for commons CO2MPAS tasks under your Windows Start-Menu.
  • Visit the guidelines for its usage: doc(allinone) (also contained within the archive).


If you have downloaded an all-in-one from previous version of CO2MPAS you may upgrade CO2MPAS contained within. Follow the instructions in the “Upgrade” section, below.

Python Installation

If you already have a suitable python-3 installation with all scientific packages updated to their latest versions, you may skip this 1st stage.


Installing Python under Windows:

The program requires CPython-3, and depends on numpy, scipy, pandas, sklearn and matplotlib packages, which depend on C-native backends and need a C-compiler to install from sources.

In Windows it is strongly suggested NOT to install the standard CPython distribution that comes up first(!) when you google for “python windows”, unless you are an experienced python-developer, and you know how to hunt down pre-compiled dependencies from the PyPi repository and/or from the Unofficial Windows Binaries for Python Extension Packages.

Therefore we suggest that you download one of the following two scientific-python distributions:

  1. WinPython python-3 (64 bit)
  2. Anaconda python-3 (64 bit)

Install WinPython

The WinPython distribution is just a collection of the standard pre-compiled binaries for Windows containing all the scientific packages, and much more. It is not update-able, and has a quasi-regular release-cycle of 3 months.

  1. Install the latest python-3.4+ 64 bit from Prefer an installation-folder without any spaces leading to it.

  2. Open the WinPython’s command-prompt console, by locating the folder where you just installed it and run (double-click) the following file:

    <winpython-folder>\"WinPython Command Prompt.exe"
  3. In the console-window check that you have the correct version of WinPython installed, and expect a similar response:

    > python -V
    Python 3.4.3
    REM Check your python is indeed where you installed it.
    > where python
  4. Use this console and follow ref: co2mpas-install instructions, below.

Install Anaconda

The Anaconda distribution is a non-standard Python environment that for Windows containing all the scientific packages we need, and much more. It is not update-able, and has a semi-regular release-cycle of 3 months.

  1. Install Anaconda python-3.4+ 64 bit from Prefer an installation-folder without any spaces leading to it.


    When asked by the installation wizard, ensure that Anaconda gets to be registered as the default python-environment for the user’s account.

  2. Open a Windows command-prompt console:

    "windows start button" --> `cmd.exe`
  3. In the console-window check that you have the correct version of Anaconda-python installed, by typing:

    > python -V
    Python 3.4.3 :: Anaconda 2.3.0 (64-bit)
    REM Check your python is indeed where you installed it.
    > where python
  4. Use this console and follow ref: co2mpas-install instructions, below.

CO2MPAS installation

  1. Install CO2MPAS executable internally into your python-environment with the following console-commands (there is no prob if the 1st uninstall command fails):

    > pip uninstall co2mpas
    > pip install co2mpas
    Collecting co2mpas
    Installing collected packages: co2mpas
    Successfully installed co2mpas-1.2.3


    Installation failures:

    The previous step require http-connectivity for pip command to Python’s “standard” repository ( and to co2mpas-site ( In case you are behind a corporate proxy, you may try one of the methods described in section Different ways of installation, below.

    If all methods to install CO2MPAS fail, re-run pip command adding extra verbose flags -vv, copy-paste the console-output, and report it to JRC.

  2. Check that when you run co2mpas, the version executed is indeed the one installed above (check both version-identifiers and paths):

    > co2mpas -vV
    co2mpas_version: 1.2.3
    co2mpas_rel_date: 2016-05-05 18:04:00
    co2mpas_path: d:\co2mpas_ALLINONE-64bit-v1.0.5.dev1\Apps\WinPython\python-3.4.3\lib\site-packages\co2mpas
    python_path: D:\co2mpas_ALLINONE-64bit-v1.0.5.dev1\WinPython\python-3.4.3
    python_version: 3.4.3 (v3.4.3:9b73f1c3e601, Feb 24 2015, 22:44:40) [MSC v.1600 XXX]
    PATH: D:\co2mpas_ALLINONE-64bit-v1.0.5.dev1\WinPython...


    The above procedure installs the latest CO2MPAS, which might be more up-to-date than the version described here!

    In that case you can either:

    1. Visit the documents for the newer version actually installed.
    2. “Pin” the exact version you wish to install with a pip command (see section below).

Install extras

Internally CO2MPAS uses an algorithmic scheduler to execute model functions. In order to visualize the design-time models and run-time workflows you need to install the Graphviz visualization library from:

If you skip this step, the modelgraph sub-command and the --plot-workflow option would both fail to run (see ref: debug).

Upgrade CO2MPAS

  1. Uninstall (see below) and re-install it.

Uninstall CO2MPAS

To uninstall CO2MPAS type the following command, and confirm it with y:

> pip uninstall co2mpas
Uninstalling co2mpas-<installed-version>
Proceed (y/n)?

Re-run the command again, to make sure that no dangling installations are left over; disregard any errors this time.

Different ways of installation

You may get multiple versions of CO2MPAS, from various places, but all require the use of pip command from a console to install:


In all cases below, remember to uninstall CO2MPAS if it’s already installed.

  • Latest STABLE: use the default pip described command above.

  • Latest PRE-RELEASE: append the --pre option in the pip command.

  • Specific version: modify the pip command like that, with optionally appending --pre:

    pip install co2mpas==1.0.1 ... # Other options, like above.
  • Specific branch from the GitHub-sources:

    pip install git+
  • Specific commit from the GitHub-sources:

    pip install git+
  • Speed-up download: append the --use-mirrors option in the pip command.

  • (for all of the above) When you are behind an http-proxy: append an appropriately adapted option --proxy http://user:password@yourProxyUrl:yourProxyPort.


    To avert any security deliberations for this http-proxy “tunnel”, JRC cryptographically signs all final releases with GPG key ID: 9CF277C40A8A1B08 so that you or your IT staff may validate their authenticity and detect man-in-the-middle attacks, however impossible.

  • (for all of the above) Without internet connectivity or when the above proxy cmd fails:

    1. With with a “regular” browser and when connected to the Internet, pre-download locally all files present in the packages folder located in the desired CO2MPAS version in the CO2MPAS site (e.g.

    2. Install co2mpas, referencing the above folder. Assuming that you downloaded the packages in the folder path/to/co2mpas_packages, use a console-command like this:

      pip install co2mpas  --no-index  -f path/to/co2mpas_packages

Install Multiple versions in parallel

In order to run and compare results from different CO2MPAS versions, you may use virtualenv command.

The virtualenv command creates isolated python-environments (“children-venvs”) where in each one you can install a different versions of CO2MPAS.


The virtualenv command does NOT run under the “conda” python-environment. Use the conda command in similar manner to create child conda-environments instead.

  1. Ensure virtualenv command installed in your “parent” python-environment, i.e the “WinPython” you use:

    > pip install virtualenv


    The pip command above has to run only once for each parent python-env. If virtualenv is already installed, pip will exit gracefully.

  2. Ensure co2mpas uninstalled in your parent-env:

    > pip uninstall co2mpas


    It is important for the “parent” python-env NOT to have CO2MPAS installed! The reasone is that you must set “children venvs” to inherit all packages installed on their “parent” (i.e. numpy and pandas), and you cannot update any inherited package from within a child-env.

  3. Move to the folder where you want your “venvs” to reside and create the “venv” with this command:

    > virtualenv --system-site-packages co2mpas_v1.0.1.venv.venv

    The --system-site-packages option instructs the child-venv to inherit all “parent” packages (numpy, pandas).

    Select a venv’s name to signify the version it will contains, e.g. co2mpas_v1.0.1.venv. The .venv at the end is not required, it is just for tagging the venv folders.

  4. Workaround a virtualenv bug with TCL/TK on Windows!

    This is technically the most “difficult” step, and it is required so that CO2MPAS can open GUI dialog-boxes, such as those for selecting the input/output dialogs.

    1. Open with an editor the co2mpas_v1.0.1.venv.venv\Scripts\activate.bat script,

    2. locate the set PATH=… line towards the bottom of the file, and append the following 2 lines:

      set "TCL_LIBRARY=d:\WinPython-64bit-3.Y.Y.Y\python-3.Y.Y.amd64\tcl\tcl8.6"
      set "TK_LIBRARY=d:\WinPython-64bit-3.Y.Y.Y\python-3.Y.Y.amd64\tcl\tk8.6"


    If you don’t modify the activation-script, you will receive the following message while running CO2MPAS:

    This probably means that Tcl wasn't installed properly.

    Of course you have to adapt the paths above to match the TCL & TK folder in your parent python-env. For instance, in ALLINONE the lines above would become:

    set "TCL_LIBRARY=%WINPYTHON%\tcl\tcl8.6"
    set "TK_LIBRARY=%WINPYTHON%\tcl\tk8.6"


    The ALLINONE archives already include this workaround ;-)

  5. “Activate” the new “venv” by running the following command (notice the dot(.) at the begining of the command):

    > .\co2mpas_v1.0.1.venv.venv\Scripts\activate.bat

    Or type this in bash:

    $ source co2mpas_v1.0.1.venv.venv\Scripts\activate.bat

    You must now see that your prompt has been prefixed with the venv’s name.

  6. Install the co2mpas version you want inside the activated venv. See the ref: co2mpas-install section, above.

    Don’t forget to check that what you get when running co2mpas is what you installed.

  7. To “deactivate” the active venv, type:

    > deactivate

    The prompt-prefix with the venv-name should now dissappear. And if you try to invoke co2mpas, it should fail.


  • Repeat steps 2–>5 to create venvs for different versions of co2mpas.
  • Use steps (6: Activate) and (9: Deactivate) to switch between different venvs.


In order to press [Tab] and get completions, do the following in your environment (ALLINONE is pre-configured with them):

  • For the Clink environment, on cmd.exe, add the following lua script inside clink’s profile folder: clink/profile/co2mpas_autocompletion.lua

    --[[ clink-autocompletion for CO2MPAS
    local handle = io.popen('co2mpas-autocompletions')
    words_str = handle:read("*a")
    function words_generator(prefix, first, last)
        local cmd = 'co2mpas'
        local prefix_len = #prefix
        --print('P:'..prefix..', F:'..first..', L:'..last..', l:'..rl_state.line_buffer)
        if prefix_len == 0 or rl_state.line_buffer:sub(1, cmd:len()) ~= cmd then
            return false
        for w in string.gmatch(words_str, "%S+") do
            -- Add matching app-words.
            if w:sub(1, prefix_len) == prefix then
            -- Add matching files & dirs.
            full_path = true
            nf = clink.match_files(prefix..'*', full_path)
            if nf > 0 then
        return clink.match_count() > 0
    sort_id = 100
  • For the bash shell just add this command in your file(~/.bashrc) (or type it every time you open a new console):

    complete -fdev -W "`co2mpas-autocompletions`" co2mpas



The following commands are for the bash console, specifically tailored for the all-in-one archive. In cmd.exe the commands are rougly similar, but remember to substitute the slashes (/) in paths with backslashes(\).

The doc(allinone) contains additionally batch-files (e.g. file(RUN_CO2MPAS.bat), file(NEW_TEMPLATE.bat), etc) that offer roughly the same capabillities described below. When you double-click them, the output from these commands gets to be written in the file(ALLINONE/CO2MPAS/co2mpas.log) file.

First ensure that the latest version of CO2MPAS is properly installed, and that its version match the version declared on this file.

The main entry for the simulator is the co2mpas console-command, which is not visible, but it is installed in your PATH. To get the syntax of the co2mpas console-command, open a console where you have installed CO2MPAS (see ref: install above) and type:

Predict NEDC CO2 emissions from WLTP.

:Copyright:    2015-2016 European Commission (JRC-IET <>
:License:       EUPL 1.1+ <>

Use the `batch` sub-command to simulate a vehicle contained in an excel-file.

  co2mpas batch       [-v | --logconf=<conf-file>] [--gui]
                      [--out-template=<xlsx-file> | --charts]
                      [--plot-workflow]  [-O=<output-folder>]
                      [--only-summary]  [--soft-validation]
  co2mpas demo        [-v | --logconf=<conf-file>] [--gui]
                      [-f] [<output-folder>]
  co2mpas template    [-v | --logconf=<conf-file>] [--gui] [-f] [<excel-file-path> ...]
  co2mpas ipynb       [-v | --logconf=<conf-file>] [--gui] [-f] [<output-folder>]
  co2mpas modelgraph  [-v | --logconf=<conf-file>] [-O=<output-folder>]
                      (--list | [--graph-depth=<levels>] [<models> ...])
  co2mpas sa          [-v | --logconf=<conf-file>] [-f] [-O=<output-folder>]
                      [--soft-validation] [--only-summary] [--overwrite-cache]
                      [--out-template=<xlsx-file> | --charts]
                      [<input-path>] [<input-params>] [<defaults>]...
  co2mpas             [--verbose | -v]  (--version | -V)
  co2mpas             --help

Syntax tip:
  The brackets `[ ]`, parens `( )`, pipes `|` and ellipsis `...` signify
  "optional", "required", "mutually exclusive", and "repeating elements";
  for more syntax-help see:

  <input-path>                Input xlsx-file or folder. Assumes current-dir if missing.
  -O=<output-folder>          Output folder or file [default: .].
  --gui                       Launches GUI dialog-boxes to choose Input, Output
                              and Options. [default: False].
  --only-summary              Do not save vehicle outputs, just the summary file.
  --overwrite-cache           Overwrite the cached file.
  --charts                    Add basic charts to output file.
  --soft-validation           Validate only partially input-data (no schema).
  --out-template=<xlsx-file>  Clone the given excel-file and appends model-results into it.
                              By default, results are appended into an empty excel-file.
                              Use `--out-template=-` to use input excel-files as templates.
  --plot-workflow             Open workflow-plot in browser, after run finished.
  -l, --list                  List available models.
  --graph-depth=<levels>      An integer to Limit the levels of sub-models plotted
                              (no limit by default).
  -f, --force                 Overwrite template/demo excel-file(s).

  -h, --help                  Show this help message and exit.
  -V, --version               Print version of the program, with --verbose
                              list release-date and installation details.
  -v, --verbose               Print more verbosely messages - overridden by --logconf.
  --logconf=<conf-file>       Path to a logging-configuration file, according to:
                              If the file-extension is '.yaml' or '.yml', it reads a dict-schema from YAML:

    batch                   Simulate vehicle for all <input-path> excel-files & folder.
                            If no <input-path> given, reads all excel-files from current-dir.
                            Read this for explanations of the param names:
    demo                    Generate demo input-files for the `batch` cmd inside <output-folder>.
    template                Generate "empty" input-file for the `batch` cmd as <excel-file-path>.
    ipynb                   Generate IPython notebooks inside <output-folder>; view them with cmd:
                              jupyter --notebook-dir=<output-folder>
    modelgraph              List or plot available models. If no model(s) specified, all assumed.
    sa                      (undocumented - subject to change)


    # Don't enter lines starting with `#`.

    # Create work folders and then fill `input` with sample-vehicles:
    md input output
    co2mpas  demo  input

    # Launch GUI dialog-boxes on the sample-vehicles just created:
    co2mpas  batch  --gui  input

    # or specify them with output-charts and workflow plots:
    co2mpas  batch  input  -O output  --charts  --plot-workflow

    # Create an empty vehicle-file inside `input` folder:
    co2mpas  template  input/vehicle_1.xlsx

    # View a specific submodel on your browser:
    co2mpas  modelgraph  co2mpas.model.physical.wheels.wheels

    # View full version specs:
    co2mpas -vV

The default sub-command (batch) accepts either a single input-excel-file or a folder with multiple input-files for each vehicle, and generates a summary-excel-file aggregating the major result-values from these vehicles, and (optionally) multiple output-excel-files for each vehicle run.

Demo files

The simulator contains input-files for demo-vehicles that are a nice starting point to try out:

id manual precon cal WLTP-H cal WLTP-L theoretical S/S BERS correct_f0
0 X X X X       X
1 X   X X   X X  
2   X X X        
3 X   X X   X   X
4   X   X     X  
5 X   X X     X X
6     X X X X    
7     X     X X  
8 X   X X       X
9     X X        
10 X   X X   X X X

To run them, do the following:

  1. Choose a folder where you will store the input and output files:

    ## Skip this if ``tutorial`` folder already exists.
    $ mkdir tutorial
    $ cd tutorial
    ## Skip also this if folders exist.
    $ mkdir input output


The input & output folders do not have to reside in the same parent, neither to have these names. It is only for demonstration purposes that we decided to group them both under a hypothetical some-folder.

  1. Create the demo vehicles inside the input-folder with the demo sub-command:

    $ co2mpas demo input
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-0.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-1.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-10.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-2.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-3.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-4.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-5.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-6.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-7.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-8.xlsx'...
    INFO:co2mpas.__main__:Creating INPUT-DEMO file 't\co2mpas_demo-9.xlsx'...
    INFO:co2mpas.__main__:You may run DEMOS with:
        co2mpas batch input
  2. Run the simulator on all demo-files (note, it might take considerable time):

    $ co2mpas batch input -O output
    Processing ['input'] --> 'output'...
    Processing: co2mpas_demo-0
    Done! [499.579 sec]
  3. Inspect the results (explained in the next section):

    $ start output/*summary.xlsx       ## More summaries might exist in the folder from previous runs.
    $ start output                     ## View the folder with all files generated.

Output files

The output-files produced on each run are the following:

  • One file per vehicle, named as <timestamp>-<inp-fname>.xls: This file contains all the inputs and calculation results for each vehicle contained in the batch-run: scalar-parameters and time series for target, calibration and prediction phases, for all cycles. In addition, the file contains all the specific submodel-functions that generated the results, a comparison summary, and information on the python libraries installed on the system (for investigating reproducibility issues).
  • A Summary-file named as <timestamp>-summary.xls: Major CO2 emissions values, optimized CO2 parameters values and success/fail flags of CO2MPAS submodels for all vehicles in the batch-run.


Additionally, a sample output file is provide here:

Entering new vehicles

You may modify the samples vehicles and run again the model. But to be sure that your vehicle does not contain by accident any of the sample-data, use the template sub-command to make an empty input excel-file:

  1. Decide the input/output folders. Assuming we are still in the tutorial folder and we wish to re-use the input/output folders from the example above, we may clear all their contents with this:

    $ rm -r ./input/* ./output/*      ## Replace `rm` with `del` in *Windows* (`cmd.exe`)
  2. Create an empty vehicle template-file (eg. vehicle_1.xlsx) inside the input-folder with the template sub-command:

    $ co2mpas template input/vehicle_1.xlsx  ## Note that here we specify the filename, not the folder!
    Creating TEMPLATE INPUT file 'input/vehicle_1.xlsx'...
  3. Open the template excel-file to fill-in your vehicle data (and save it afterwards):

    $ start input/vehicle_1.xlsx      ## Opens the excel-file. Use `start` in *cmd.exe*.

    The generated file contains help descriptions to help you populate it with vehicle data. For items where an array of values is required (i.e. gear-box ratios) you may reference different parts of the spreadsheet following the syntax of the “xlref” mini-language.


    You may also read the “annotated” input excel-file to get an understanding of each scalar paramet and series required, but DO NOT USE THIS “fatty” xl-file (~10Mb) when running the model.

    For an explanation of the naming of the fields, read below the ref: excel-model section

    You may repeat these last 2 steps if you want to add more vehicles in the batch-run.

  4. Run the simulator. Specify the single excel-file as input:

    $ co2mpas batch ./input/vehicle_1.xlsx -O output
    Processing './input/vehicle_1.xlsx' --> 'output'...
    Processing: vehicle_1
    Done! [12.938986 sec]
  5. Assuming you do receive any error, you may now inspect the results:

    $ start output/*summary.xlsx      ## More summaries might open from previous runs.
    $ start output                    ## View all files generated (see below).
  6. In the case of errors, or if the results are not satisfactory, repeat the above procedure from step 3 to modify the vehicle and re-run the model. See also ref: debug, below.

Synchronizing time-series

The model might fail in case your time-series signals are time-shifted and/or with different sampling rates. Even if the run succeeds, the results will not be accurate enough.

As an aid tool, you may use the datasync command-line tool to “synchronize” your data-tables. This command reads one or more tables from excel-files and synchronizes their columns. The syntax of this utility command is given by typing datasync --help in the command line (listing below just the main fields):

Shift and resample excel-tables; see

  datasync  [(-v | --verbose) | --logconf <conf-file>]
            [--force | -f] [--no-clone] [--prefix-cols] [-O <output>]
            <x-label> <y-label> <ref-table> [<sync-table> ...]
  datasync  [--verbose | -v]  (--version | -V)
  datasync  --help

  <x-label>              Column-name of the common x-axis (e.g. 'times') to be resampled if needed.
  <y-label>              Column-name of y-axis cross-correlated between all <sync-table>
                         and <ref-table>.
  <ref-table>            The reference table, in *xl-ref* notation (usually given as  `file#sheet!`);
                         synced columns will be appended into this table.
                         The captured table must contain <x_label> & <y_label> as column labels.
                         If hash(`#`) symbol missing, assumed as file-path and
                         the table is read from its 1st sheet .
  <sync-table>           Sheets to be synced in relation to <ref-table>, also in *xl-ref* notation.
                         All tables must contain <x_label> & <y_label> as column labels.
                         Each xlref may omit file or sheet-name parts; in that case,
                         those from the previous xlref(s) are reused.
                         If hash(`#`) symbol missing, assumed as sheet-name.
                         If none given, all non-empty sheets of <ref-table> are synced
                         against the 1st one.

All input tables must share 2 common columns: <x-label> and <y-label>, as if those tables describe 2D cartesian data, with a common X-axis and multiple data-series on the Y-Axis.


The <x-label> usually refers to the “time” dimension.

The 1st table given (<ref-table>) is considered to contain the “reference” X/Y values; the data-columns to shift-and-resample are contained in one or more tables (<sync-table>) specified subsequently in the command line, that are possibly read from different excel work-books.

  • Shifting is based on the cross-correlation of <y-label> columns;
  • resampling is based on the values of <x-label> columns among the different tables.

All tables are read from excel-sheets using the xl-ref syntax, which is best explained with some examples.


  • Read the full contents from all wbook.xlsx sheets as tables and sync their columns using the table from the 1st sheet as reference:

    datasync times  velocity  folder/Book.xlsx
  • Sync Sheet1 using Sheet3 as reference:

    datasync times  velocity  wbook.xlsx#Sheet3!  Sheet1!
  • The same as above but with integeres used to index excel-sheets:

    datasync times  velocity  wbook.xlsx#2!  0


    Sheet-indices are zero based!

  • A more complex xlr-ref example which reads the synce-table from sheet2 of wbook-2 starting at D5 cell, or more Down ‘n Right if that was empty, till the first empty cell Down n Right, and synchronizes that based on 1st sheet of wbook-1:

    datasync times  velocity wbook-1.xlsx  wbook-2.xlsx#0!D5(DR):..(DR)
  • Typical usage for CO2MPAS velocity time-series from Dyno and OBD:

    datasync -O ../output times  velocities  ../input/book.xlsx#WLTP-H  WLTP-H_OBD

Using custom output xl-files as templates

You may have defined customized xl-files for summarizing time-series and scalar parameters. To have CO2MPAS fill those “output-template” files with its results, execute it with the --out-template option.

To create/modify one output-template yourself, do the following:

  1. Open a typical CO2MPAS output-file for some vehicle.

  2. Add one or more sheets and specify/referring CO2MPAS result-data using named-ranges.


    Do not use simple/absolute excel references (e.g. “=B2”). Use excel functions (indirect, lookup, offset, etc.) and array-functions together with string references to the named ranges (e.g. “=indirect(“nedc_predictions_time_series!_fuel_consumptions”)”).

  3. (Optional) Delete the old sheets and save your file.

  4. Use that file together with the --out-template argument.

Launch CO2MPAS from Jupyter(aka IPython)

You may enter the data for a single vehicle and run its simulation, plot its results and experiment in your browser using IPython.

The usage pattern is similar to “demos” but requires to have ipython installed:

  1. Ensure ipython with notebook “extra” is installed:


    This step requires too many libraries to provide as standalone files, so unless you have it already installed, you will need a proper http-connectivity to the standard python-repo.

    $ pip install ipython[notebook]
    Installing collected packages: ipython[notebook]
    Successfully installed ipython-x.x.x notebook-x.x.x
  2. Then create the demo ipython-notebook(s) into some folder (i.e. assuming the same setup from above, tutorial/input):

    $ pwd                     ## Check our current folder (``cd`` alone for Windows).
    $ co2mpas ipynb ./input
  3. Start-up the server and open a browser page to run the vehicle-simulation:

    $ ipython notebook ./input
  4. A new window should open to your default browser (AVOID IEXPLORER) listing the simVehicle.ipynb notebook (and all the demo xls-files). Click on the *.ippynb file to “load” the notebook in a new tab.

    The results are of a simulation run already pre-generated for this notebook but you may run it yourself again, by clicking the menu:

    "menu" --> `Cell` --> `Run All`

    And watch it as it re-calculates cell by cell.

  5. You may edit the python code on the cells by selecting them and clicking Enter (the frame should become green), and then re-run them, with Ctrl + Enter.

    Navigate your self around by taking the tutorial at:

    "menu" --> `Help` --> `User Interface Tour`

    And study the example code and diagrams.

  6. When you have finished, return to the console and issue twice Ctrl + C to shutdown the ipython-server.

Debugging and investigating results

  • Make sure that you have installed graphviz, and when running the simulation, append also the --plot-workflow option.

  • Use the modelgraph sub-command to plot the offending model (or just out of curiosity). For instance:

    $ co2mpas modelgraph co2mpas.model.physical.wheels.wheels
  • Inspect the functions mentioned in the workflow and models and search them in CO2MPAS documentation ensuring you are visiting the documents for the actual version you are using.


Execution Model

The execution of CO2MPAS model for a single vehicle is a stepwise procedure of 3 stages: precondition, calibration, and prediction. These are invoked repeatedly, and subsequently combined, for the various cycles, as shown in the “active” flow-diagram of the execution, below:


The models in the diagram are nested; explore by clicking on them.

  1. Precondition: identifies the initial state of the vehicle by running a preconditioning WLTP cycle, before running the WLTP-H and WLTP-L cycles. The inputs are defined by the input.precondition.wltp_p node, while the outputs are stored in output.precondition.wltp_p.
  2. Calibration: the scope of the stage is to identify, calibrate and select (see next sections) the best physical models from the WLTP-H and WLTP-L inputs (input.calibration.wltp_x). If some of the inputs needed to calibrate the physical models are not provided (e.g. initial_state_of_charge), the model will select the missing ones from precondition-stage’s outputs (output.precondition.wltp_p). Note that all data provided in input.calibration.wltp_x overwrite those in output.precondition.wltp_p.
  3. Prediction: executed for the NEDC and as well as for the WLTP-H and WLTP-L cycles. All predictions use the calibrated_models. The inputs to predict the cycles are defined by the user in nodes. If some or all inputs for the prediction of WLTP-H and WLTP-L cycles are not provided, the model will select from `output.calibration.wltp_x nodes a minimum set required to predict CO2 emissions.

Excel input: data naming conventions

This section describes the data naming convention used in the CO2MPAS template (.xlsx file). In it, the names used as sheet-names, parameter-names and column-names are “sensitive”, in the sense that they construct a data-values tree which is then fed into into the simulation model as input. These names are splitted in “parts”, as explained below with examples:

  • sheet-names parts:

                  └─┬─┘ └────┬─────┘ └─┬──┘
    usage───────────┘        │         │
    stage────────────────────┘         │

    All 3 parts above are optional, but at least one of them must be present on a sheet-name; those parts are then used as defaults for all parameter-names contained in that sheet.

  • parameter-names/columns-names parts:

                  └─┬─┘ └────┬────┘ └──────────┬──────────┘ └──┬─┘
    usage(optional)─┘        │                 │               │
    stage(optional)──────────┘                 │               │
    parameter──────────────────────────────────┘               │

    OR with the last 2 parts reversed:

                                    └──┬─┘ └──────────┬──────────┘
    cycle(optional)────────────────────┘              │


  • The dot(.) may be replaced by space.
  • The usage and stage parts may end with an s, denoting plural, and are case-insensitive, e.g. Inputs.

Description of the name-parts

  1. usage:
    • input [default]: values provided by the user as input to CO2MPAS.
    • data: values selected (see previous section) to calibrate the models and to predict the CO2 emission.
    • output: CO2MPAS precondition, calibration, and prediction results.
    • target: reference-values (NOT USED IN CALIBRATION OR PREDICTION) to be compared with the CO2MPAS results. This comparison is performed in the report sub-model by compare_outputs_vs_targets() function.
  2. stage:
    • precondition [imposed when: wltp-p is specified as cycle]: data related to the precondition stage.
    • calibration [default]: data related to the calibration stage.
    • prediction [imposed when: nedc is specified as cycle]: data related to the prediction stage.
  3. cycle:
    • nedc data related to the NEDC cycle.
    • wltp-h data related to the WLTP High cycle.
    • wltp-l data related to the WLTP Low cycle.
    • wltp-precon: data related to the preconditioning WLTP cycle.
    • wltp-p: is a shortcut of wltp-precon.
    • wltp: is a shortcut to set values for both wltp-h and wltp-l cycles.
    • all [default]: is a shortcut to set values for nedc, wltp, and wltp-p cycles.
  4. param: any data node name (e.g. vehicle_mass) used in the physical model.

Sheet types

There are two sheet types, which are parsed according to their contained data:

  • parameters [parsed range is #B2:C_]: scalar or not time-depended values (e.g. r_dynamic, gear_box_ratios, full_load_speeds).
  • time-series [parsed range is #A2:__]: time-depended values (e.g. times, velocities, gears). Columns without values are skipped. COLUMNS MUST HAVE THE SAME LENGTH!

When cycle is missing in the sheet-name, the sheet is parsed as parameters, otherwise it is parsed as time-series.

Calibrated Physical Models

There are potentially eight models calibrated from input scalar-values and time-series (see doc(reference)):

  1. AT_model,
  2. electric_model,
  3. clutch_torque_converter_model,
  4. co2_params,
  5. engine_cold_start_speed_model,
  6. engine_coolant_temperature_model,
  7. engine_speed_model, and
  8. start_stop_model.

Each model is calibrated separately over WLTP_H and WLTP_L. A model can contain one or several functions predicting different quantities. For example, the electric_model contains the following functions/data:

  • alternator_current_model,
  • alternator_status_model,
  • electric_load,
  • max_battery_charging_current,
  • start_demand.

These functions/data are calibrated/estimated based on the provided input (in the particular case: alternator current, battery current, and initial SOC) over both cycles, assuming that data for both WLTP_H and WLTP_L are provided.


The co2_params model has a third possible calibration configuration (so called ALL) using data from both WLTP_H and WLTP_L combined (when both are present).

Model selection

To select which is the best calibration (from WLTP_H or WLTP_L or ALL) to be used in the prediction phase, the results of each stage are compared against the provided input data (used in the calibration). The calibrated models are THEN used to recalculate (predict) the inputs of the WLTP_H and WLTP_L cycles. A score (weighted average of all computed metrics) is attributed to each calibration of each model as a result of this comparison.


The overall score attributed to a specific calibration of a model is the average score achieved when compared against each one of the input cycles (WLTP_H and WLTP_L).

For example, the score of electric_model calibrated based on WLTP_H when predicting WLTP_H is 20, and when predicting WLTP_L is 14. In this case the overall score of the the electric_model calibrated based on WLTP_H is 17. Assuming that the calibration of the same model over WLTP_L was 18 and 12 respectively, this would give an overall score of 15.

In this case the second calibration (WLTP_L) would be chosen for predicting the NEDC.

In addition to the above, a success flag is defined according to upper or lower limits of scores which have been defined empirically by the JRC. If a model fails these limits, priority is then given to a model that succeeds, even if it has achieved a worse score.

The following table describes the scores, targets, and metrics for each model:

Release History

Release History

This version
History Node


Download Files

Download Files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

File Name & Checksum SHA256 Checksum Help Version File Type Upload Date
co2mpas-1.2.2-py2.py3-none-any.whl (9.0 MB) Copy SHA256 Checksum SHA256 py2.py3 Wheel May 6, 2016

Supported By

WebFaction WebFaction Technical Writing Elastic Elastic Search Pingdom Pingdom Monitoring Dyn Dyn DNS Sentry Sentry Error Logging CloudAMQP CloudAMQP RabbitMQ Heroku Heroku PaaS Kabu Creative Kabu Creative UX & Design Fastly Fastly CDN DigiCert DigiCert EV Certificate Rackspace Rackspace Cloud Servers DreamHost DreamHost Log Hosting