🔬 Advanced Features

Q: What is co-simulation and how to use it?

Co-simulation allows TRICYS to interact with external software (like Aspen Plus):

Workflow: 1. Run preliminary simulation 2. Call external processor (Handler) 3. External software calculates new data 4. Inject data back into the Modelica model 5. Run the final full simulation

Configuration Example:

{
    "co_simulation": [
        {
            "mode": "interceptor",
            "submodel_name": "example_model.I_ISS",
            "instance_name": "i_iss",
            "handler_module": "tricys.handlers.i_iss_handler",
            "handler_function": "run_aspen_simulation",
            "params": {
                "bkp_path": "path/to/aspen/file.bkp"
            }
        }
    ]
}

* mode: interceptor (default) or replacement. * handler_module: The module where the handler is located. * handler_script_path: Alternatively, provide the direct path to the handler script.

For details, see: Co-simulation Module

Q: How to create a custom post-processing module?

Post-processing modules are Python functions that receive simulation results and perform analysis:

1. Create the handler function:

# my_postprocess.py
def analyze_results(df, output_filename="my_report.txt"):
    """
    Custom post-processing function

    Args:
        df: Pandas DataFrame containing simulation results
        output_filename: Output file name
    """
    # Perform analysis
    max_inventory = df['sds.I[1]'].max()

    # Save results
    with open(output_filename, 'w') as f:
        f.write(f"Maximum tritium inventory: {max_inventory} g\n")

Note: Please place your custom module (e.g., my_postprocess.py) in the root directory of the tricys project, or ensure it is in a path that Python can import.

2. Reference it in the configuration file:

{
    "post_processing": [
        {
            "module": "my_postprocess",
            "function": "analyze_results",
            "params": {
                "output_filename": "custom_report.txt"
            }
        }
    ]
}

Q: How to perform sensitivity analysis?

TRICYS offers several sensitivity analysis methods:

1. Single-parameter sensitivity analysis:

tricys analysis -c single_param_analysis.json

Studies the impact of a single parameter on KPIs.

2. Multi-parameter sensitivity analysis:

tricys analysis -c multi_param_analysis.json

Studies the coupling effects between parameters.

3. SOBOL global sensitivity analysis:

tricys analysis -c sobol_analysis.json

Quantifies the contribution of parameters and their interactions.

4. Latin Hypercube uncertainty quantification:

tricys analysis -c latin_analysis.json

Assesses the impact of input uncertainty on the output.

For details, see: Sensitivity Analysis Tutorial

Q: How to define custom performance metrics?

Performance metrics (KPIs) are defined in sensitivity_analysis.metrics_definition:

Using built-in metrics:

{
    "metrics_definition": {
        "Max_Inventory": {
            "source_column": "sds.I[1]",
            "method": "max_value"
        }
    }
}

Built-in metric methods: * get_final_value * max_value, min_value, mean_value * time_of_max, time_of_min * time_of_turning_point * calculate_startup_inventory * calculate_doubling_time * calculate_required_tbr (bisection search)

For a detailed physical explanation of the built-in metrics, see Core Performance Metrics.

Creating custom metrics:

# my_metrics.py
def calculate_peak_to_peak(series):
    """Calculate peak-to-peak value"""
    return series.max() - series.min()

Register your function in tricys/analysis/metric.py or reference it directly in the configuration.

---