Integrated full ERA5 netCDF-Interpolation, model-framework now considered to be complete

This commit is contained in:
Marcel Christian Frommelt 2021-03-08 09:16:10 +09:00
parent c87e08af99
commit 40583109b4
1 changed files with 387 additions and 0 deletions

387
main_netCDF_new.py Normal file
View File

@ -0,0 +1,387 @@
import numpy as np
import math
import xarray as xr
import matplotlib.pyplot as plt
from scipy.spatial import cKDTree
from scipy import interpolate
import cartopy.crs as ccrs
from datetime import datetime
start = datetime.now()
print(start)
begin_time = datetime.now()
from models.sun import sun_angles_analytical
from models.sun import sun_angles_astropy
from models.drag import drag, c_d
from input.user_input import *
from input.natural_constants import *
from astropy.time import Time
import astropy.units as unit
from models.thermal import AirMass
from netCDF4 import Dataset
import pandas as pd
data = pd.read_excel(r'C:\Users\marcel\PycharmProjects\MasterThesis\Mappe1.xls', sheet_name='Tabelle3')
comp_time = pd.DataFrame(data, columns= ['Time'])
comp_height = pd.DataFrame(data, columns= ['Height'])
def transform(lon, lat, t):
# WGS 84 reference coordinate system parameters
A = 6378137.0 # major axis [m]
E2 = 6.69437999014e-3 # eccentricity squared
t_s = 1000 * t
lon_rad = np.radians(lon)
lat_rad = np.radians(lat)
# convert to cartesian coordinates
r_n = A / (np.sqrt(1 - E2 * (np.sin(lat_rad) ** 2)))
x = r_n * np.cos(lat_rad) * np.cos(lon_rad)
y = r_n * np.cos(lat_rad) * np.sin(lon_rad)
z = r_n * (1 - E2) * np.sin(lat_rad)
return x, y, z, t_s
data = Dataset("test2021.nc", "r", format="NETCDF4")
ERAtime = data.variables['time'][:] # time
ERAlat = data.variables['latitude'][:] # latitude
ERAlon = data.variables['longitude'][:] # longitude
ERAz = data.variables['z'][:]/g # geopotential to geopotential height
ERApress = data.variables['level'][:] # pressure level
ERAtemp = data.variables['t'][:] # temperature in K
vw_x = data.variables['u'][:] # v_x
vw_y = data.variables['v'][:] # v_y
vw_z = data.variables['w'][:] # v_z
lon_era2d, lat_era2d, time_era = np.meshgrid(ERAlon, ERAlat, ERAtime)
xs, ys, zs, ts = transform(lon_era2d.flatten(), lat_era2d.flatten(), time_era.flatten())
tree = cKDTree(np.column_stack((xs, ys, zs, ts))) # !
lat = start_lat # deg
lon = start_lon # deg
t = 0 # simulation time in seconds
h = start_height
utc = Time(start_utc)
epoch_diff = (Time(start_utc).jd - Time('1900-01-01 00:00:00.0').jd) * 24.000000
t_epoch = epoch_diff
xt, yt, zt, tt = transform(lon, lat, t) # test coordinates
d, inds = tree.query(np.column_stack((xt, yt, zt, tt)), k=8) # longitude, latitude, time in h # !
w = 1.0 / d[0] ** 2
lat_sel = np.unravel_index(inds[0], lon_era2d.shape)[0]
lon_sel = np.unravel_index(inds[0], lon_era2d.shape)[1]
time_sel = np.unravel_index(inds[0], lon_era2d.shape)[2]
interp4d_height = np.ma.dot(w, ERAz[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
interp4d_temp = np.ma.dot(w, ERAtemp[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
interp4d_vw_x = np.ma.dot(w, vw_x[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
interp4d_vw_y = np.ma.dot(w, vw_y[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
interp4d_vw_z = np.ma.dot(w, vw_z[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
pressure_hPa = np.array([1, 2, 3, 5, 7, 10, 20, 30, 50, 70, 100, 125, 150, 175, 200,225, 250, 300, 350, 400,
450, 500, 550, 600, 650, 700, 750, 775, 800, 825, 850, 875, 900, 925, 950, 975, 1000])
pressure = 100 * pressure_hPa
print(pressure)
# height_interp1d = interpolate.interp1d(interp4d_height, pressure)
temp_interp1d = interpolate.interp1d(interp4d_height, interp4d_temp)
press_interp1d = interpolate.interp1d(interp4d_height, pressure)
vw_x_interp1d = interpolate.interp1d(interp4d_height, interp4d_vw_x)
vw_y_interp1d = interpolate.interp1d(interp4d_height, interp4d_vw_y)
vw_z_interp1d = interpolate.interp1d(interp4d_height, interp4d_vw_z)
#u = 0
#v = 0
#w = 0
T_air = temp_interp1d(h)
p_air = press_interp1d(h)
rho_air = p_air/(R_air * T_air)
u = vw_x_interp1d(h)
v = vw_y_interp1d(h)
w = -1 / g * vw_z_interp1d(h) * R_air * T_air / p_air
v_x = 0
v_y = 0
v_z = 0
v_rel = 0
T_gas = T_air
T_film = T_gas
t_list = []
h_list = []
v_list = []
lat_list = []
lon_list = []
p_list = []
Temp_list = []
rho_list = []
while t <= t_end and h >= 0:
t_list.append(t)
h_list.append(h)
v_list.append(v_rel)
lat_list.append(lat)
lon_list.append(lon)
p_list.append(p_air)
Temp_list.append(T_air)
rho_list.append(rho_air)
t_epoch = epoch_diff + t/3600
xt, yt, zt, tt = transform(lon, lat, t_epoch) # current balloon coordinates in cartesian coordinates: x [m], y [m], z [m], time [h since 1900-01-01 00:00:00.0]
d, inds = tree.query(np.column_stack((xt, yt, zt, tt)), k=8) # longitude, latitude, time in h # !
w = 1.0 / d[0] ** 2
lat_sel = np.unravel_index(inds[0], lon_era2d.shape)[0]
lon_sel = np.unravel_index(inds[0], lon_era2d.shape)[1]
time_sel = np.unravel_index(inds[0], lon_era2d.shape)[2]
interp4d_height = np.ma.dot(w, ERAz[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
interp4d_temp = np.ma.dot(w, ERAtemp[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
interp4d_vw_x = np.ma.dot(w, vw_x[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
interp4d_vw_y = np.ma.dot(w, vw_y[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
interp4d_vw_z = np.ma.dot(w, vw_z[time_sel, :, lat_sel, lon_sel]) / np.sum(w)
press_interp1d = interpolate.interp1d(interp4d_height, pressure)
temp_interp1d = interpolate.interp1d(interp4d_height, interp4d_temp)
vw_x_interp1d = interpolate.interp1d(interp4d_height, interp4d_vw_x)
vw_y_interp1d = interpolate.interp1d(interp4d_height, interp4d_vw_y)
vw_z_interp1d = interpolate.interp1d(interp4d_height, interp4d_vw_z)
try:
p_air = press_interp1d(h)
T_air = temp_interp1d(h)
except:
if (h > interp4d_height[0]):
h = interp4d_height[0]
else:
h = interp4d_height[-1]
p_air = press_interp1d(h)
T_air = temp_interp1d(h)
#print("height")
#print(h)
#print("temperature")
#print(temp_interp1d(h))
#print("pressure")
#print(press_interp1d(h))
#print("vw_x")
#print(vw_x_interp1d(h))
p_gas = p_air
rho_air = p_air / (R_air * T_air)
u = vw_x_interp1d(h)
v = vw_y_interp1d(h)
w = -1/g * vw_z_interp1d(h) * R_air * T_air / p_air
v_relx = u - v_x
v_rely = v - v_y
v_relz = w - v_z
v_rel = (v_relx ** 2 + v_rely ** 2 + v_relz ** 2) ** (1/2)
rho_gas = p_gas/(R_gas * T_gas) # calculate gas density through ideal(!) gas equation
V_b = m_gas/rho_gas # calculate balloon volume from current gas mass and gas density
if V_b > V_design:
V_b = V_design
#m_gas = V_design * rho_gas
else:
pass
m_gross = m_pl + m_film
m_tot = m_pl + m_film + m_gas
m_virt = m_tot + c_virt * rho_air * V_b
d_b = 1.383 * V_b ** (1/3) # calculate diameter of balloon from its volume
L_goreB = 1.914 * V_b ** (1/3)
h_b = 0.748 * d_b
A_surf = 4.94 * V_b ** (2/3)
A_surf1 = 4.94 * V_design ** (2/3) * (1 - np.cos(np.pi * L_goreB/L_goreDesign))
A_eff = 0.65 * A_surf + 0.35 * A_surf1
A_top = np.pi/4 * d_b ** 2
D = drag(c_d, rho_air, d_b, v_rel) # calculate drag force
if v_rel == 0:
Drag_x = 0
Drag_y = 0
Drag_z = 0
else:
Drag_x = D * v_relx/v_rel
Drag_y = D * v_rely/v_rel
Drag_z = D * v_relz/v_rel
I = g * V_b * (rho_air - rho_gas) # calculate gross inflation
W = g * m_gross # calculate weight (force)
F = I - W + Drag_z
AZ, ELV = sun_angles_analytical(lat, lon, utc)
A_proj = A_top * (0.9125 + 0.0875 * np.cos(np.pi - 2 * np.deg2rad(ELV)))
# CALCULATIONS FOR THERMAL MODEL
if ELV >= -(180 / np.pi * np.arccos(R_E / (R_E + h))):
tau_atm = 0.5 * (np.exp(-0.65 * AirMass(p_air, p_0, ELV, h)) + np.exp(-0.095 * AirMass(p_air, p_0, ELV, h)))
tau_atmIR = 1.716 - 0.5 * (np.exp(-0.65 * p_air / p_0) + np.exp(-0.095 * p_air / p_0))
else:
tau_atm = 0
tau_atmIR = 0
doy = int(utc.doy)
MA = (357.52911 + 0.98560028 * (utc.jd - 2451545)) % 360 # in degree, reference: see folder "literature"
TA = MA + 2 * e * np.sin(np.deg2rad(MA)) + 5 / 4 * e ** 2 * np.sin(np.deg2rad(2 * MA))
I_Sun = 1367.5 * ((1 + e * np.cos(np.deg2rad(TA))) / (1 - e ** 2)) ** 2
I_SunZ = I_Sun * tau_atm
q_sun = I_SunZ
q_IRground = epsilon_ground * sigma * T_ground ** 4
q_IREarth = q_IRground * tau_atmIR
if ELV <= 0:
q_Albedo = 0
else:
q_Albedo = Albedo * I_Sun * np.sin(np.deg2rad(ELV))
my_air = (1.458 * 10 ** -6 * T_air ** 1.5) / (T_air + 110.4)
my_gas = 1.895 * 10 ** -5 * (T_gas / 273.15) ** 0.647
k_air = 0.0241 * (T_air / 273.15) ** 0.9
k_gas = 0.144 * (T_gas / 273.15) ** 0.7
Pr_air = 0.804 - 3.25 * 10 ** (-4) * T_air
Pr_gas = 0.729 - 1.6 * 10 ** (-4) * T_gas
Gr_air = (rho_air ** 2 * g * np.abs(T_film - T_air) * d_b ** 3) / (T_air * my_air ** 2)
Nu_air = 2 + 0.45 * (Gr_air * Pr_air) ** 0.25
HC_free = Nu_air * k_air / d_b
Re = np.abs(v_relz) * d_b * rho_air / my_air
HC_forced = k_air / d_b * (2 + 0.41 * Re ** 0.55)
HC_internal = 0.13 * k_gas * ((rho_gas ** 2 * g * np.abs(T_film - T_gas) * Pr_gas) / (T_gas * my_air ** 2)) ** (
1 / 3)
HC_external = np.maximum(HC_free, HC_forced)
HalfConeAngle = np.arcsin(R_E / (R_E + h))
ViewFactor = (1 - np.cos(HalfConeAngle)) / 2
Q_Sun = alpha_VIS * A_proj * q_sun * (1 + tau_VIS / (1 - r_VIS))
Q_Albedo = alpha_VIS * A_surf * q_Albedo * ViewFactor * (1 + tau_VIS / (1 - r_VIS))
Q_IREarth = alpha_IR * A_surf * q_IREarth * ViewFactor * (1 + tau_IR / (1 - r_IR))
Q_IRfilm = sigma * epsilon * alpha_IR * A_surf * T_film ** 4 * 1 / (1 - r_IR)
Q_IRout = sigma * epsilon * A_surf * T_film ** 4 * (1 * tau_IR / (1 - r_IR))
Q_ConvExt = HC_external * A_eff * (T_air - T_film)
Q_ConvInt = HC_internal * A_eff * (T_film - T_gas)
# RoC = -v_z
# dT_gas = (Q_ConvInt / (gamma * m_gas * c_v) - (gamma - 1) / gamma * rho_air(h) * g / (rho_gas * R_gas) * RoC) * dt
# dT_film = ((Q_Sun + Q_Albedo + Q_IREarth + Q_IRfilm + Q_ConvExt - Q_ConvInt - Q_IRout) / (c_f * m_film)) * dt
# v_w = w
# v = v_z
s = h
a_x = Drag_x/m_virt
a_y = Drag_y/m_virt
a_z = F/m_virt
# DIFFERENTIAL EQUATIONS
dx = dt * v_x + 0.5 * a_x * dt ** 2 # lon
dy = dt * v_y + 0.5 * a_y * dt ** 2 # lat
lon = lon + np.rad2deg(np.arctan(dx/((6371229.0 + h) * np.cos(np.deg2rad(lat)))))
lat = lat + np.rad2deg(np.arctan(dy/(6371229.0 + h)))
v_xn = v_x + dt * a_x
v_yn = v_y + dt * a_y
s_n = s + dt * v_z + 0.5 * a_z * dt ** 2
dh = s_n - s
v_n = v_z + dt * a_z
T_gn = T_gas + dt * (Q_ConvInt / (gamma * c_v * m_gas) - g * R_gas * T_gas * dh / (c_v * gamma * T_air * R_air))
T_en = T_film + (Q_Sun + Q_Albedo + Q_IREarth + Q_IRfilm + Q_ConvExt - Q_ConvInt - Q_IRout) / (c_f * m_film) * dt
T_g = T_gn
T_e = T_en
s = s_n
v_x = v_xn
v_y = v_yn
v_z = v_n
h = s
T_gas = T_g
T_film = T_e
# v_z = v
t += dt # time increment
utc = Time(start_utc) + t * unit.second
#print(len(lon_list), len(lat_list))
#"""
#plt.plot(start_lon, start_lat, 'rx')
#plt.plot(ax=ax, lon_list, lat_list, transform=ccrs.PlateCarree())
#plt.show()
#"""
### WORKS!
###
###dset = xr.open_dataset("test2021.nc")
###
###print(dset['t'][1][0]) # [time] [level]
###
#fig = plt.figure() #figsize=[120,50])
###
#ax = fig.add_subplot(111, projection=ccrs.PlateCarree())
###
###dset['t'][1][0].plot(ax=ax, cmap='jet', transform=ccrs.PlateCarree())
###ax.coastlines()
###plt.show()
#print(len(ERAlat))
#print(len(ERAlon))
#print(len(ERAz))
#"""
plt.plot(comp_time, comp_height, 'r--', label='PoGo Flight Test')
#plt.plot(t_list, v_list, 'k--', label='Ballon v_rel')
plt.plot(t_list, h_list, 'k-', label='Balloon Altitude')
#plt.plot(t_list, p_list, 'r-', label='Air Pressure [Pa]')
#plt.plot(t_list, Temp_list, 'b-', label='Air Temperature [K]')
#plt.plot(t_list, rho_list, 'g-', label='Air Density in [kg/m$^3$]')
plt.xlabel('time in s')
plt.ylabel('Balloon Altitude in m')
plt.legend()
plt.show()
#"""