Add airfoil transonic dataset (#46)

Author: annaivagnes

smithers/dataset/__init__.py

@@ -5,3 +5,4 @@
 from .datasets.graetz import GraetzDataset
 from .datasets.unsteady_heat import UnsteadyHeatDataset
 from .datasets.lid_cavity import LidCavity
+from .datasets.airfoil_transonic import AirfoilTransonicDataset

smithers/dataset/datasets/airfoil_transonic.py

@@ -0,0 +1,203 @@
+from ..abstract_dataset import AbstractDataset
+import os
+import numpy as np
+import matplotlib
+import matplotlib.pyplot as plt
+import matplotlib.tri as tri
+from mpl_toolkits.axes_grid1 import make_axes_locatable
+from scipy.interpolate import CubicSpline
+
+
+class AirfoilTransonicDataset(AbstractDataset):
+
+    parametric = True
+    time_dependent = False
+    description = "Transonic flow past a NACA 0012 airfoil"
+    data_directory = os.path.join(os.path.dirname(__file__),
+            'airfoil_transonic')
+
+    def __init__(self):
+        '''
+        Initialize the dataset and structure it in dictionaries.
+        '''
+        self.params = np.load(os.path.join(self.data_directory, 'params.npy'))
+        v_internal = np.load(os.path.join(self.data_directory,
+            'snapshots_v_internal.npy'))
+        p_internal = np.load(os.path.join(self.data_directory,
+            'snapshots_p_internal.npy'))
+        p_airfoil = np.load(os.path.join(self.data_directory,
+            'snapshots_p_airfoil.npy'))
+        wn_airfoil = np.load(os.path.join(self.data_directory,
+            'snapshots_wn_airfoil.npy'))
+
+        self.snapshots = {
+                'internal': {'mag(v)': v_internal, 'p': p_internal},
+                'airfoil': {'p': p_airfoil, 'wallshear_normal': wn_airfoil}
+        }
+        coords_internal = np.load(os.path.join(self.data_directory,
+            'coords_internal.npy'))
+        coords_airfoil = np.load(os.path.join(self.data_directory,
+            'coords_airfoil.npy'))
+        self.pts_coordinates = {
+                'internal': coords_internal, 'airfoil': coords_airfoil
+                }
+        self.auxiliary_triang = tri.Triangulation(coords_internal[0, :],
+                coords_internal[1, :])
+        self.auxiliary_triang.set_mask(self._mask_airfoil())
+
+    def _coords_airfoil(self, which='pos'):
+        '''
+        Get the positive (if which='pos') or negative (if which='neg')
+        coordinates of the airfoil.
+        '''
+        x_airfoil = self.pts_coordinates['airfoil'][0, :]
+        y_airfoil = self.pts_coordinates['airfoil'][1, :]
+        if which=='pos':
+            indices = y_airfoil >= 0
+        elif which=='neg':
+            indices = y_airfoil <= 0
+
+        x_airfoil_which = sorted(x_airfoil[indices])
+        y_airfoil_which = y_airfoil[indices][
+                np.argsort(x_airfoil[indices])]
+        return x_airfoil_which, y_airfoil_which
+
+    def _f_airfoil_neg(self, tol=1e-4):
+        '''
+        Return a function that represents the negative part of the airfoil.
+        '''
+        x_airfoil_neg, y_airfoil_neg = self._coords_airfoil(which='neg')
+        f_neg = CubicSpline(x_airfoil_neg, y_airfoil_neg)
+        array = []
+        for x in self.pts_coordinates['internal'][0, :]:
+            if x >= 0 and x <= 1:
+                array.append(f_neg(x) - tol)
+            else:
+                array.append(np.nan)
+        return np.array(array)
+
+    def _f_airfoil_pos(self, tol=1e-4):
+        '''
+        Return a function that represents the positive part of the airfoil.
+        '''
+        x_airfoil_pos, y_airfoil_pos = self._coords_airfoil(which='pos')
+        f_pos = CubicSpline(x_airfoil_pos, y_airfoil_pos)
+        array = []
+        for x in self.pts_coordinates['internal'][0, :]:
+            if x >= 0 and x <= 1:
+                array.append(f_pos(x) + tol)
+            else:
+                array.append(np.nan)
+        return np.array(array)
+
+    def _mask_airfoil(self):
+        '''
+        Mask the triangles that are inside the airfoil (useful for plots).
+        '''
+        coords = self.pts_coordinates['internal']
+        x_airfoil = self.pts_coordinates['airfoil'][0, :]
+        y_airfoil = self.pts_coordinates['airfoil'][1, :]
+        isbad = (np.greater(coords[0, :], x_airfoil.min()) &
+                np.less(coords[0, :], x_airfoil.max()) &
+                np.less(coords[1, :], self._f_airfoil_pos()) &
+                np.greater(coords[1, :], self._f_airfoil_neg()))
+        triangles = self.auxiliary_triang.triangles
+        mask = np.all(np.where(isbad[triangles], True, False), axis=1)
+        return mask
+
+    def plot_internal(self, snaps_array, idx=0, title='Snapshot', namefig=None,
+            lim_x=(-0.5, 2), lim_y=(-0.5, 1), figsize=(8, 4), logscale=False):
+        '''
+        Plot a snapshot of the internal flow.
+
+        Parameters
+        ----------
+        snaps_array : numpy.ndarray
+            Snapshots of the internal flow (also external, not only
+            self.snapshots['internal']).
+            The shape should be (nparams, npoints).
+        idx : int, optional
+            Index of the snapshot to plot. The default is 0.
+        title : str, optional
+            Title of the plot. The default is 'Snapshot'.
+        namefig : str, optional
+            Name of the file where the plot is saved. The default is None.
+        lim_x : tuple, optional
+            Limits of the x-axis. The default is (-0.5, 2).
+        lim_y : tuple, optional
+            Limits of the y-axis. The default is (-0.5, 1).
+        figsize: tuple, optional
+            Size of the output figure. The default is (8, 4).
+        '''
+        fig, ax = plt.subplots(1, 1, figsize=figsize)
+        snapshot = snaps_array[idx, :]
+        if logscale:
+            lognorm = matplotlib.colors.LogNorm(vmin=snapshot.min()+1e-12,
+                vmax=snapshot.max())
+            c = ax.tripcolor(self.auxiliary_triang, snapshot, cmap='rainbow',
+                shading='gouraud', norm=lognorm)
+        else:
+            c = ax.tripcolor(self.auxiliary_triang, snapshot, cmap='rainbow',
+                shading='gouraud')
+        ax.plot(self._coords_airfoil()[0], self._coords_airfoil()[1],
+                color='black', lw=0.5)
+        ax.plot(self._coords_airfoil(which='neg')[0],
+                self._coords_airfoil(which='neg')[1],
+                color='black', lw=0.5)
+        ax.set_aspect('equal')
+        if lim_x is not None:
+            ax.set_xlim(lim_x)
+        if lim_y is not None:
+            ax.set_ylim(lim_y)
+        if title is not None:
+            ax.set_title(title)
+        divider = make_axes_locatable(ax)
+        cax = divider.append_axes("right", size= "5%", pad=0.1)
+        plt.colorbar(c, cax=cax)
+        if namefig is not None:
+            fig.savefig(namefig, dpi=300)
+        plt.show()
+
+    def plot_airfoil(self, snaps_array, idx=0, title='Snapshot', namefig=None,
+            lim_x=(-0.1, 1.1), lim_y=(-0.2, 0.2), figsize=(8, 4)):
+        '''
+        Plot a snapshot on the airfoil.
+
+        Parameters
+        ----------
+        snaps_array : numpy.ndarray
+            Snapshot on the airfoil (also external, not only
+            self.snapshots['airfoil']).
+            The shape should be (nparams, npoints).
+        idx : int, optional
+            Index of the snapshot to plot. The default is 0.
+        title : str, optional
+            Title of the plot. The default is 'Snapshot'.
+        namefig : str, optional
+            Name of the file where the plot is saved. The default is None.
+        lim_x : tuple, optional
+            Limits of the x-axis. The default is (-0.1, 1.1).
+        lim_y : tuple, optional
+            Limits of the y-axis. The default is (-0.2, 0.2).
+        figsize: tuple, optional
+            Size of the output figure. The default is (8, 4).
+        '''
+        fig, ax = plt.subplots(1, 1, figsize=figsize)
+        snapshot = snaps_array[idx, :]
+        coords = self.pts_coordinates['airfoil']
+        ax.grid()
+        c = ax.scatter(coords[0, :], coords[1, :], c=snapshot, s=10)
+        ax.set_aspect('equal')
+        if lim_x is not None:
+            ax.set_xlim(lim_x)
+        if lim_y is not None:
+            ax.set_ylim(lim_y)
+        if title is not None:
+            ax.set_title(title)
+        divider = make_axes_locatable(ax)
+        cax = divider.append_axes("right", size= "5%", pad=0.1)
+        plt.colorbar(c, cax=cax)
+        if namefig is not None:
+            fig.savefig(namefig, dpi=300)
+        plt.show()
+