Transient Simulation

For time-dependent analyses, pycanha provides the TSCNRLDS solver — a Crank-Nicolson scheme with radiative linearization and direct sparse factorization.

Setting up a transient run

from pycanha.solvers import TSCNRLDS

solver = TSCNRLDS(tmm)
solver.set_simulation_time(
    start_time=0.0,        # [s]
    end_time=100_000.0,    # [s]
    dtime=100.0,           # integration timestep [s]
    output_stride=1000.0,  # store results every 1000 s
)
solver.initialize()
solver.solve()
solver.deinitialize()

The output_stride controls how often results are written to the output model. A larger stride saves memory when the timestep is small.

Retrieving results

Transient results are stored in the model’s ThermalData container as named output models:

output_model = tmm.thermal_data.models.get_model(solver.output_model_name)

The temperature series is exposed as a dense time series:

• output_model.T.times — time samples [s]

• output_model.T.values — temperature matrix with shape (n_steps, n_nodes)

Each temperature column is ordered by internal index.

Use get_idx_from_node_num() to map user node numbers to column indices:

import numpy as np
import matplotlib.pyplot as plt

idx1 = tmm.nodes.get_idx_from_node_num(1)

times = output_model.T.times
T1    = output_model.T.values[:, idx1]

plt.plot(times / 3600, T1)
plt.xlabel("Time [h]")
plt.ylabel("Temperature [K]")
plt.show()

Initial conditions

The transient solver uses the current node temperatures as initial conditions. Set them before calling initialize():

for node_num in range(1, 11):
    tmm.nodes.set_T(node_num, 273.15)   # uniform 0 °C start