コード例 #1
0
def diag_weight(fig, Results, time_point, ch, DistWaveFile=None, ax=None):
    # Currently only RELAX/LUKE distributions supported
    # Extension for GENE trivial though
    if (ax is None):
        ax = fig.add_subplot(111)
    harmonic_n = 2
    itime = np.argmin(np.abs(time_point -
                             Results.Scenario.plasma_dict["time"]))
    time_cor = Results.Scenario.plasma_dict["time"][itime]
    EQObj = EQDataExt(Results.Scenario.shot)
    EQObj.set_slices_from_ext(Results.Scenario.plasma_dict["time"],
                              Results.Scenario.plasma_dict["eq_data"])
    B_ax = EQObj.get_B_on_axis(time_cor)
    if (DistWaveFile is not None):
        dist_obj = load_f_from_mat(DistWaveFile, True)
        f_inter = make_f_inter(Results.Config["Physics"]["dstf"],
                               dist_obj=dist_obj,
                               EQObj=EQObj,
                               time=time_cor)[0]
    else:
        dist_obj = None
        f_inter = make_f_inter("Th", EQObj=EQObj, time=time_cor)[0]
    m = 40
    n = 80
    if (dist_obj is None):
        # Estimate grid for thermal distribution assuming X-mode
        rhop_Te_signed = np.concatenate([- Results.Scenario.plasma_dict[Results.Scenario.plasma_dict["prof_reference"]][itime][::-1][:-1], \
                                        Results.Scenario.plasma_dict[Results.Scenario.plasma_dict["prof_reference"]][itime]])
        Te_signed_rhop = np.concatenate([
            Results.Scenario.plasma_dict["Te"][itime][::-1][:-1],
            Results.Scenario.plasma_dict["Te"][itime]
        ])
        Te_spl = InterpolatedUnivariateSpline(rhop_Te_signed, Te_signed_rhop)
        Te_weighted_spl = InterpolatedUnivariateSpline(Results.BPD["rhopX"][itime][ch - 1], Results.BPD["BPDX"][itime][ch - 1] * \
                                                       Te_spl(Results.BPD["rhopX"][itime][ch - 1]))
        Te_av = Te_weighted_spl.integral(-1.0, 1.0)
        beta_th = rel_thermal_beta(cnst.c**2 * cnst.electron_mass / cnst.e /
                                   Te_av)
        u_th = beta_th / np.sqrt(1.0 - beta_th**2)
        u_perp_grid = np.linspace(0.0, 5.0 * u_th, m)
        u_par_grid = np.linspace(-5.0 * u_th, 5.0 * u_th,
                                 n)  # five sigma should be good enough!
    else:
        u_perp_grid = np.linspace(0.0, np.max(dist_obj.u), m)
        u_par_grid = np.linspace(-np.max(dist_obj.u), np.max(f_inter.x), n)
    diag_weight_f = np.zeros((m, n))
    diag_weight_rel = np.zeros((m, n))
    for ir in range(Results.Config["Physics"]["N_ray"]):
        cur_BDOP = make_3DBDOP_for_ray(Results,
                                       time_cor,
                                       ch,
                                       ir,
                                       harmonic_n,
                                       B_ax,
                                       f_inter=f_inter)
        for irhop, in range(len(cur_BDOP.rho)):
            print(irhop + 1, " / ", len(cur_BDOP.rho))
            intercep_points = find_cell_interceps(u_par_grid, u_perp_grid,
                                                  cur_BDOP, irhop)
            for i_intercep, intercep_point in enumerate(intercep_points[:-1]):
                i = np.argmin(np.abs(intercep_point[0] - u_par_grid))
                j = np.argmin(np.abs(intercep_point[1] - u_perp_grid))
                if (u_par_grid[i] > intercep_point[0]):
                    i -= 1
                if (u_perp_grid[j] > intercep_point[1]):
                    j -= 1
                if (i < 0 or j < 0):
                    continue  # only happens at the lower bounds, where u_perp is very small and, therefore, also j is very small
                # Compute arclength
                t = np.zeros(cur_BDOP.u_par[irhop].shape)
                for i_res_line in range(1, len(cur_BDOP.u_par[irhop])):
                    t[i_res_line] = t[i_res_line - 1] + np.sqrt((cur_BDOP.u_par[irhop][i_res_line] - cur_BDOP.u_par[irhop][i_res_line - 1])**2 + \
                                                                (cur_BDOP.u_perp[irhop][i_res_line] - cur_BDOP.u_perp[irhop][i_res_line - 1])**2)
                t /= np.max(t)  # Normalize this
                # Sort
                t_spl = InterpolatedUnivariateSpline(cur_BDOP.u_par[irhop], t)
                try:
                    BPD_val_spl = InterpolatedUnivariateSpline(
                        t, cur_BDOP.val[irhop])
                except Exception as e:
                    print(e)
                BPD_val_rel_spl = InterpolatedUnivariateSpline(
                    t, np.abs(cur_BDOP.val[irhop] - cur_BDOP.val_back[irhop]))
                t1 = t_spl(intercep_point[0])
                t2 = t_spl(intercep_points[i_intercep + 1][0])
                diag_weight_f[j,i] += Results.weights["ray"][itime][ch][ir] * \
                                    BPD_val_spl.integral(t1, t2)
                diag_weight_rel[j,i] += Results.weights["ray"][itime][ch][ir] * \
                                    BPD_val_rel_spl.integral(t1, t2)
    ax.contour(u_perp_grid, u_par_grid, diag_weight_f.T / np.max(diag_weight_f.flatten()), \
                 levels = np.linspace(0.01,1,10), cmap = plt.get_cmap("plasma"))
    m = cm.ScalarMappable(cmap=plt.cm.get_cmap("plasma"))
    m.set_array(np.linspace(0.01, 1.0, 10))
    cb_diag = fig.colorbar(m, pad=0.15, ticks=[0.0, 0.5, 1.0])
    cb_diag.set_label(r"$D_\omega [\si{{a.u.}}]$")
    ax.set_ylabel(r"$u_\parallel$")
    ax.set_xlabel(r"$u_\perp$")
    ax.set_aspect("equal")
    return fig
コード例 #2
0
def ECRH_weight(fig,
                Result_file,
                time_point,
                ibeam,
                DistWaveFile,
                beam_freq=105.e9,
                ax=None):
    # Currently only RELAX/LUKE distributions supported
    # Extension for GENE trivial though
    if (ax is None):
        ax = fig.add_subplot(111)
    harmonic_n = 2
    Results = ECRadResults()
    Results.from_mat_file(Result_file)
    itime = np.argmin(np.abs(time_point -
                             Results.Scenario.plasma_dict["time"]))
    time_cor = Results.Scenario.plasma_dict["time"][itime]
    EQObj = EQDataExt(Results.Scenario.shot)
    EQObj.set_slices_from_ext(Results.Scenario.plasma_dict["time"],
                              Results.Scenario.plasma_dict["eq_data"])
    B_ax = EQObj.get_B_on_axis(time_cor)
    EqSlice = EQObj.GetSlice(time_point)
    dist_wave_mat = loadmat(DistWaveFile)
    dist_obj = load_f_from_mat(DistWaveFile, True)
    f_inter = make_f_inter(Results.Config["Physics"]["dstf"],
                           dist_obj=dist_obj,
                           EQObj=EQObj,
                           time=time_cor)[0]
    linear_beam = read_waves_mat_to_beam(dist_wave_mat,
                                         EqSlice,
                                         use_wave_prefix=None)
    itme = np.argmin(np.abs(Results.Scenario.plasma_dict["time"] - time_point))
    Te_spl = InterpolatedUnivariateSpline(Results.Scenario.plasma_dict[Results.Scenario.plasma_dict["prof_reference"]][itime], \
                                          np.log(Results.Scenario.plasma_dict["Te"][itme]))
    ne_spl = InterpolatedUnivariateSpline(Results.Scenario.plasma_dict[Results.Scenario.plasma_dict["prof_reference"]][itime], \
                                          np.log(Results.Scenario.plasma_dict["ne"][itme]))
    m = 40
    n = 80
    u_perp_grid = np.linspace(0.0, np.max(dist_obj.u), m)
    u_par_grid = np.linspace(-np.max(dist_obj.u), np.max(f_inter.x), n)
    diag_weight_f = np.zeros((m, n))
    for ray in linear_beam.rays[ibeam]:
        tot_pw_ray, cur_PDP = make_PowerDepo_3D_for_ray(ray, beam_freq, "Re", harmonic_n, \
                                                        B_ax, EqSlice, Te_spl, ne_spl, f_inter, \
                                                        N_pnts=100, fast= True)
        for irhop in range(len(cur_PDP.rho)):
            print(irhop + 1, " / ", len(cur_PDP.rho))
            intercep_points = find_cell_interceps(u_par_grid, u_perp_grid,
                                                  cur_PDP, irhop)
            for i_intercep, intercep_point in enumerate(intercep_points[:-1]):
                i = np.argmin(np.abs(intercep_point[0] - u_par_grid))
                j = np.argmin(np.abs(intercep_point[1] - u_perp_grid))
                if (u_par_grid[i] > intercep_point[0]):
                    i -= 1
                if (u_perp_grid[j] > intercep_point[1]):
                    j -= 1
                if (i < 0 or j < 0):
                    continue  # only happens at the lower bounds, where u_perp is very small and, therefore, also j is very small
                # Compute arclength
                t = np.zeros(cur_PDP.u_par[irhop].shape)
                for i_res_line in range(1, len(cur_PDP.u_par[irhop])):
                    t[i_res_line] = t[i_res_line - 1] + np.sqrt((cur_PDP.u_par[irhop][i_res_line] - cur_PDP.u_par[irhop][i_res_line - 1])**2 + \
                                                                (cur_PDP.u_perp[irhop][i_res_line] - cur_PDP.u_perp[irhop][i_res_line - 1])**2)
                t /= np.max(t)  # Normalize this
                # Sort
                t_spl = InterpolatedUnivariateSpline(cur_PDP.u_par[irhop], t)
                try:
                    PDP_val_spl = InterpolatedUnivariateSpline(
                        t, cur_PDP.val[irhop])
                except Exception as e:
                    print(e)
                t1 = t_spl(intercep_point[0])
                t2 = t_spl(intercep_points[i_intercep + 1][0])
                diag_weight_f[j,i] += tot_pw_ray * \
                                        PDP_val_spl.integral(t1, t2)
    ax.contourf(u_perp_grid, u_par_grid, diag_weight_f.T / np.max(diag_weight_f.flatten()), \
                 levels = np.linspace(0.01,1,10), cmap = plt.get_cmap("Greens"))
    m = cm.ScalarMappable(cmap=plt.cm.get_cmap("Greens"))
    m.set_array(np.linspace(0.01, 1.0, 10))
    cb_diag = fig.colorbar(m, pad=0.15, ticks=[0.0, 0.5, 1.0])
    cb_diag.set_label(r"$\mathrm{d}P/\mathrm{d}s [\si{{a.u.}}]$")
    ax.set_ylabel(r"$u_\parallel$")
    ax.set_xlabel(r"$u_\perp$")
    ax.set_aspect("equal")
    return fig