Back to overview.
In order to allow running and comparing various simulations swiftly and make SIMSON useable to non-programmers, the Simulation Module was added.
In this page the module will be described covering the following aspects:
For every simulation, a set of configurations or parameters can be set. Hence, the results of the simulation can be compared to show the effect that the change of a certain parameter has on the steel cycle predictions.
The set of changeable configurations that we call simulation parameters here are described in the table below. They are comprised of static factors like the simulation name, various estimates of factors in the steel cycle like the maximum scrap rate in production as well as assumptions on the development in the five shared socio-economic pathway (SSP) scenarios.
| Parameter Name | Default value | Possible range | Notes |
|---|---|---|---|
| Simulation Name | Default | Any text without | |
| Max. scrap share in production | 0.6 | 0-1 | For base model, e.g. 0.6 = 60 % |
| Scrap share in BOF | 0.22 | 0-1 | Percentage of secondary steel in basic oxygen furnace route |
| Forming yield | ca. 0.93 | 0-1 | Yield of steel during the steel forming process, taken from Cullen (2012) |
| Model economy | False | True / False | Use the economy module or not |
| Model economy - base year | 2008 | 2008-2099 | the starting year of the economic module |
| Elasticity of steel, scrap recovery rate, scrap share in production | -0.2,-1,-0.8 | Negative, 0 | Price elasticities of demand for the economic module |
| Initial scrap recovery rate and share in production | 0, 0 | 0-1 | Recov. rate at free scrap cost, Scrap share at free dissassembly cost |
| Steel price change (by SSP scenario) | 0 | Real numbers | e.g. -0.2 would result in a linear decrease of steel price from the basee year to -20 % of that in 2100 |
| Inflow change (by SSP scenario and/or product) | 0 | -1 to ∞ | Like steel price change, steel demand can be manually adapted |
| Reuse rate (by SSP scenario and/or product) | 0 | 0 to ∞ | Like steel price change, steel reuse rates can be manually adapted, e.g. assuming 0.2 (20 %) of construction materials being reused by 2100 in SSP1 |
The simplest way to run several simulations is to use the Excel
SIMSON simulation interface. It is located in simulation/interface/excel
and called simson_simulation_interface.xlsx (see picture above).
Users can simply fill in the simulation parameters (rows) they wish to test
in the columns from column D on. Every column represents one simulation.
Some parameters are grouped together, like configurations for each SSP scenario.
If Users intend to parametrise their simulations to that degree of detail,
they can use the + and - buttons to expand these rows.
Once the Excel sheet is saved, the simulations can be run
via the run_simulations.py Pythons script as described below.
Similarly to the Excel Interface, simulations can also be run using YAML.
YAML is commonly used to store configurations in .yml or .yaml files.
In SIMSON, such a file represents the configurations for one simulation. When
they are stored in simulation/interface/yaml and the run_simulations.py is run,
the simulations are run.
An example is shown in the figure above, depicting the default.yml file stored
in the same folder that can be used to create custom YAML files for simulations.
Using YAML files like this might allow Users to let other models communicate with SIMSON.
After installing SIMSON as described in section 2, the simulations
parametrised in the Excel interface or via YAML files can be run
via the run_simulations.py in the root directory.
This can be down without any programming environment like Pycharm or VSCode but simply via a command line/Terminal. For this, Python and all required libraries need to be installed and the data needs to have been loaded via the git submodules. You simply need to enter change the working directory to the root directory of SIMSON and then run the following command:
python run_simulations.py
Depending on how you installed Python, the first term might need to
be changed to python2 or python3.
After running the simulations, the output is stored in simulation/output.
Each simulation is stored in a seperate folder labelled by the simulation name and time.
The folders each contain three aspects: the SIMSON model as an instance of the ODYM MFASystem class
stored in a pickle (.p) file, a data directory containing .csv files for all flows and stocks
across all dimensions as well as a figure directory containing some select aspects of the
predicted steel cycle.
Go to section 4: Project Structure