⚙️ Configuration & Execution

Q: How to run a simulation?

TRICYS offers several ways to run simulations:

1. Command-Line Interface (CLI):

# Use the default configuration file config.json
tricys

# Specify a configuration file
tricys -c my_config.json

# Run an analysis task
tricys analysis -c analysis_config.json

2. Graphical User Interface (GUI):

tricys gui

3. Interactive Examples:

# Run an interactive menu for all examples
tricys example

# Run basic simulation examples
tricys basic example

# Run analysis examples
tricys analysis example

Q: How to understand the output files?

After the simulation is complete, the results are saved in a timestamped subdirectory under results/:

Filename	Description
simulation_result.csv	Detailed results for a single parameter combination, containing all variables over time.
sweep_results.csv	Aggregated results for multiple parameter combinations (parameter sweep).
sensitivity_analysis_summary.csv	[Analysis Tasks Only] Summary metrics for sensitivity analysis, where each row is the KPI for one run.
requierd_tbr_summary.csv	[Analysis Tasks Only] Optimization results generated when performing goal-seeking tasks like "Required TBR".
simulation_*.log	Detailed runtime log, including debugging information.
config.json	A backup of the full configuration used for this run.
*_report.md	[Analysis Tasks Only] Auto-generated AI analysis report.
*.png / *.csv	Various charts and data exports.

Result CSV File Structure

Whether running one or multiple parameter combinations, the final CSV files follow similar column naming conventions.

• Base Case (No Scanned Parameters): If simulation_parameters is empty, the column names are simply the variable names defined in variableFilter.

  time,sds.I[1],blanket.I[1],...
  0.0,10.5,2.3,...
  

• Parameter Sweep Case: When simulation_parameters is not empty, the column names will have parameter information appended.

  time,sds.I[1]&blanket.TBR=1.05,sds.I[1]&blanket.TBR=1.1,...
  

  • Column Name Format: <variable_name>&<param1>=<value1>&<param2>=<value2>...
  • The time column remains unchanged.
  • Each variable under each parameter combination becomes a separate column. The column name is formed by joining the variable name and parameter-value pairs, separated by the & symbol.

Q: How to define complex parameter scans?

TRICYS supports several parameter scan formats:

Feature	Format	Example	Description
Discrete List	[v1, v2, ...]	[6, 12, 18]	A set of discrete values
Arithmetic Series	"start:stop:step"	"1.05:1.15:0.05"	Start, stop, and step size
Linear Spacing	"linspace:start:stop:num"	"linspace:10:20:5"	Generate num evenly spaced points
Logarithmic Spacing	"log:start:stop:num"	"log:1:1000:4"	Generate num logarithmically scaled points
Read from File	"file:path:column"	"file:data.csv:voltage"	Read from a specified column in a CSV file

Example Configuration:

{
    "simulation_parameters": {
        "blanket.TBR": [1.05, 1.1, 1.15, 1.2],
        "plasma.fb": "linspace:0.01:0.1:10",
        "tep_fep.to_SDS_Fraction[1]": "log:0.1:1.0:5"
    }
}

Q: How to filter output variables?

Use the variableFilter parameter to select the variables to be saved. This parameter supports regular expressions, but be mindful of the syntax to match Modelica's variable naming rules.

Configuration Example:

{
    "simulation": {
        "variableFilter": "time|sds.I[1]|blanket.I[1-5]|div.I[1-5]"
    }
}

Common Patterns: * time: The time variable (must be included). * sds.I[1]: Exact match for a single variable. * sds.I[1-5]: Matches array variables from sds.I[1] to sds.I[5]. * blanket.I[1-5]|div.I[1-5]: Matches multiple specific array variables.

Q: The simulation is very slow, how to speed it up?

You can take the following optimization measures:

1. Enable concurrent execution:

{
    "simulation": {
        "concurrent": true,
        "max_workers": 4
    }
}

2. Reduce the number of output variables:

{
    "simulation": {
        "variableFilter": "time|sds.I[1]"  # Only save key variables
    }
}

3. Increase the time step (trade-off with accuracy):

{
    "simulation": {
        "step_size": 1.0  # Increase from 0.5 to 1.0
    }
}

4. Reduce the number of scan points:

{
    "simulation_parameters": {
        "blanket.TBR": "linspace:1.05:1.15:5"  # Reduce from 20 to 5
    }
}