Quickstart

This is a quick example that shows how to fit a Foster network model to thermal impedance data from a spreadsheet.

First, we need to import the necessary packages and load the data. The thermal-network package does not include openpyxl as a dependency. To run this example, you will need to install it: pip install openpyxl.

1. Load data from the spreadsheet

   The excel file should be in the same folder as the script

   import numpy as np
   from openpyxl import load_workbook, workbook
   from thermal_network.fitting import fit_optimal_foster_network, trim_steady_state
   from thermal_network.conversions import foster_to_cauer
   
   # Load data from the spreadsheet
   workbook = load_workbook('excel_example.xlsx', read_only=True, data_only=True)
   worksheet = workbook["Thermal Data"]
   
   # Load time data
   cells = worksheet["C6:C103"]
   time_data = np.array([row[0].value for row in cells])
   
   # Load temperature data
   cells = worksheet["E6:E103"]
   impedance_data = np.array([row[0].value for row in cells])

2. Trim steady-state data points

   The raw data may contain many data points at the end of the curve that are already in steady-state. These points can be removed to avoid overweighting steady-state in the loss function of the optimization algorithm (steady-state is enforced exactly by reparametrization).

   time_data, impedance_data = trim_steady_state(
       time_data=time_data,
       impedance_data=impedance_data
   )

3. Fit a model to the data

   We specify that the optimal model can have at most 10 layers

   model = fit_optimal_foster_network(
       time_data=time_data,
       impedance_data=impedance_data,
       max_layers=10,
       tau_floor= 1e-3
   )

4. Convert the resulting Foster network to the Cauer topology

   best_model = model.best_model
   foster_network = best_model.network
   cauer_network = foster_to_cauer(foster_network)
   
   print("Fit completed!")
   print(f"Convergence was successful: {best_model.convergence_info['converged']}")
   print(f"Final Loss (MSE): {best_model.final_loss:.6f}")
   print("\nFitted Foster Network Parameters:")
   print(foster_network)
   print("\nEquivalent Cauer Network Parameters:")
   print(cauer_network)

The fit_optimal_foster_network function returns the best Foster network model found. We can then convert it to a Cauer network using the foster_to_cauer function.