2.20. Custom Class for coupled PDEs

This example shows how to solve a set of coupled PDEs, the spatially coupled FitzHugh–Nagumo model, which is a simple model for the excitable dynamics of coupled Neurons:

\[\begin{split}\partial_t u &= \nabla^2 u + u (u - \alpha) (1 - u) + w \\ \partial_t w &= \epsilon u\end{split}\]

Here, \(\alpha\) denotes the external stimulus and \(\epsilon\) defines the recovery time scale. We implement this as a custom PDE class below.

  0%|          | 0/100.0 [00:00<?, ?it/s]
Initializing:   0%|          | 0/100.0 [00:00<?, ?it/s]
  0%|          | 0/100.0 [00:00<?, ?it/s]
  0%|          | 0.24/100.0 [00:00<01:26,  1.15it/s]
  1%|          | 0.75/100.0 [00:00<00:32,  3.02it/s]
  3%|3         | 3.32/100.0 [00:00<00:12,  7.49it/s]
  9%|9         | 9.15/100.0 [00:00<00:08, 10.40it/s]
 18%|#7        | 17.97/100.0 [00:01<00:07, 11.65it/s]
 29%|##8       | 28.81/100.0 [00:02<00:05, 12.23it/s]
 41%|####      | 40.83/100.0 [00:03<00:04, 12.57it/s]
 54%|#####3    | 53.54/100.0 [00:04<00:03, 12.74it/s]
 67%|######6   | 66.57/100.0 [00:05<00:02, 12.87it/s]
 80%|#######9  | 79.78/100.0 [00:06<00:01, 12.95it/s]
 93%|#########3| 93.07/100.0 [00:07<00:00, 13.01it/s]
 93%|#########3| 93.07/100.0 [00:07<00:00, 12.13it/s]
100%|##########| 100.0/100.0 [00:07<00:00, 13.03it/s]
100%|##########| 100.0/100.0 [00:07<00:00, 13.03it/s]

from pde import FieldCollection, PDEBase, UnitGrid


class FitzhughNagumoPDE(PDEBase):
    """FitzHugh–Nagumo model with diffusive coupling"""

    def __init__(self, stimulus=0.5, τ=10, a=0, b=0, bc="auto_periodic_neumann"):
        super().__init__()
        self.bc = bc
        self.stimulus = stimulus
        self.τ = τ
        self.a = a
        self.b = b

    def evolution_rate(self, state, t=0):
        v, w = state  # membrane potential and recovery variable

        v_t = v.laplace(bc=self.bc) + v - v**3 / 3 - w + self.stimulus
        w_t = (v + self.a - self.b * w) / self.τ

        return FieldCollection([v_t, w_t])


grid = UnitGrid([32, 32])
state = FieldCollection.scalar_random_uniform(2, grid)

eq = FitzhughNagumoPDE()
result = eq.solve(state, t_range=100, dt=0.01)
result.plot()