def pce_deg1(V, T, P, aerosol):
brackets = {'mu': ('uniform_n', 0.01, 0.25), 'sigma': ('uniform_n', 1.2, 3.0), 'V': ('weibull_n', 2.0, 1.0), 'kappa': ('uniform_n', 0.1, 1.2), 'N': ('uniform_n', 100.0, 10000.0)}
As = array([ 0.00022826, -0.00051625, -0.00020289, -0.00039829, -0.0001025 ,
-0.0001638 ])
def map_to_norm(x, param):
prior, args = brackets[param][0], brackets[param][1:]
if prior == "lognormal_n": args = map(np.log, args)
inv_func = inv_funcs[prior]
return inv_func(x, *args)
mu = map_to_norm(aerosol.mu, 'mu')
N = map_to_norm(aerosol.N, 'N')
V = map_to_norm(V, 'V')
sigma = map_to_norm(aerosol.sigma, 'sigma')
kappa = map_to_norm(aerosol.kappa, 'kappa')
V_1 = eval_hermite(1, V)
N_1 = eval_hermite(1, N)
kappa_1 = eval_hermite(1, kappa)
mu_1 = eval_hermite(1, mu)
sigma_1 = eval_hermite(1, sigma)
Smax = As.dot(np.array([1.0, mu_1, N_1, V_1, kappa_1, sigma_1, ]))
_, S_mode_crit = kohler_crit(T, aerosol.mu*1e-6, aerosol.kappa)
ui = 2.*np.log(S_mode_crit/Smax)/(3.*np.sqrt(2.)*np.log(aerosol.sigma))
N_act = 0.5*aerosol.N*erfc(ui)
act_frac = N_act/aerosol.N
return Smax, act_frac
def pce_deg3(V, T, P, aerosol):
brackets = {'mu': ('uniform_n', 0.01, 0.25), 'sigma': ('uniform_n', 1.2, 3.0), 'V': ('weibull_n', 2.0, 1.0), 'kappa': ('uniform_n', 0.1, 1.2), 'N': ('uniform_n', 100.0, 10000.0)}
As = array([ 9.14798604e-04, -9.22633928e-04, -8.06844177e-04,
-1.43940090e-03, -1.77269894e-04, -2.30463363e-04,
2.65464696e-04, 1.51795190e-04, 3.10878234e-04,
2.59723751e-05, 3.18675085e-05, -8.25970041e-05,
-6.73965549e-05, -2.32713884e-04, -3.38946258e-07,
9.64888514e-06, 1.70653952e-04, 4.44587400e-04,
7.76929961e-05, 1.90554293e-04, 4.03803792e-04,
6.04459159e-05, 9.44462548e-05, 1.07125979e-04,
1.13169278e-04, 5.70518250e-05, -9.24557555e-05,
-3.11843201e-04, -2.00457356e-05, -2.11958800e-05,
-5.27092388e-05, -4.79579441e-04, -1.51607324e-05,
-9.92390215e-06, -2.32074053e-04, -2.72153026e-04,
-1.84122185e-05, 1.02236473e-05, -2.28193188e-05,
-2.92921954e-05, -5.24197442e-05, -8.52954010e-07,
-6.09448717e-05, -4.54888486e-05, 7.46022148e-07,
-3.49603741e-06, -2.27098230e-04, -1.80186021e-05,
1.12536358e-05, -3.05864945e-05, 3.41873606e-05,
2.45392802e-05, -8.72265689e-05, -4.30851978e-05,
1.24256190e-06, 1.07910264e-04])
def map_to_norm(x, param):
prior, args = brackets[param][0], brackets[param][1:]
if prior == "lognormal_n": args = map(np.log, args)
inv_func = inv_funcs[prior]
return inv_func(x, *args)
mu = map_to_norm(aerosol.mu, 'mu')
N = map_to_norm(aerosol.N, 'N')
V = map_to_norm(V, 'V')
sigma = map_to_norm(aerosol.sigma, 'sigma')
kappa = map_to_norm(aerosol.kappa, 'kappa')
V_1 = eval_hermite(1, V)
kappa_3 = eval_hermite(3, kappa)
V_2 = eval_hermite(2, V)
N_1 = eval_hermite(1, N)
kappa_2 = eval_hermite(2, kappa)
N_2 = eval_hermite(2, N)
sigma_2 = eval_hermite(2, sigma)
mu_2 = eval_hermite(2, mu)
N_3 = eval_hermite(3, N)
kappa_1 = eval_hermite(1, kappa)
mu_3 = eval_hermite(3, mu)
mu_1 = eval_hermite(1, mu)
sigma_3 = eval_hermite(3, sigma)
V_3 = eval_hermite(3, V)
sigma_1 = eval_hermite(1, sigma)
Smax = As.dot(np.array([1.0, mu_1, N_1, V_1, kappa_1, sigma_1, mu_2, N_2, V_2, kappa_2, sigma_2, mu_3, N_3, V_3, kappa_3, sigma_3, mu_1*N_1, mu_1*V_1, mu_1*kappa_1, mu_1*sigma_1, N_1*V_1, N_1*kappa_1, N_1*sigma_1, V_1*kappa_1, V_1*sigma_1, kappa_1*sigma_1, mu_1*N_2, mu_1*V_2, mu_1*kappa_2, mu_1*sigma_2, N_1*mu_2, N_1*V_2, N_1*kappa_2, N_1*sigma_2, V_1*mu_2, V_1*N_2, V_1*kappa_2, V_1*sigma_2, kappa_1*mu_2, kappa_1*N_2, kappa_1*V_2, kappa_1*sigma_2, sigma_1*mu_2, sigma_1*N_2, sigma_1*V_2, sigma_1*kappa_2, mu_1*N_1*V_1, mu_1*N_1*kappa_1, mu_1*N_1*sigma_1, mu_1*V_1*kappa_1, mu_1*V_1*sigma_1, mu_1*kappa_1*sigma_1, N_1*V_1*kappa_1, N_1*V_1*sigma_1, N_1*kappa_1*sigma_1, V_1*kappa_1*sigma_1]))
_, S_mode_crit = kohler_crit(T, aerosol.mu*1e-6, aerosol.kappa)
ui = 2.*np.log(S_mode_crit/Smax)/(3.*np.sqrt(2.)*np.log(aerosol.sigma))
N_act = 0.5*aerosol.N*erfc(ui)
act_frac = N_act/aerosol.N
return Smax, act_frac
def pce_deg2(V, T, P, aerosol):
brackets = {'mu': ('uniform_n', 0.01, 0.25), 'sigma': ('uniform_n', 1.2, 3.0), 'V': ('weibull_n', 2.0, 1.0), 'kappa': ('uniform_n', 0.1, 1.2), 'N': ('uniform_n', 100.0, 10000.0)}
As = array([ 1.14326819e-03, -7.17119470e-04, -2.87207700e-04,
-6.49925186e-04, -1.32394919e-04, -2.03550833e-04,
4.19256128e-04, 1.95758520e-04, 4.28430253e-04,
5.43726035e-05, 4.01075063e-05, 1.78061247e-04,
6.89565065e-04, 1.06794441e-04, 2.34925303e-04,
5.96974401e-04, 9.09500909e-05, 7.94986009e-05,
1.49255164e-04, 1.22341600e-04, 5.20229943e-05])
def map_to_norm(x, param):
prior, args = brackets[param][0], brackets[param][1:]
if prior == "lognormal_n": args = map(np.log, args)
inv_func = inv_funcs[prior]
return inv_func(x, *args)
mu = map_to_norm(aerosol.mu, 'mu')
N = map_to_norm(aerosol.N, 'N')
V = map_to_norm(V, 'V')
sigma = map_to_norm(aerosol.sigma, 'sigma')
kappa = map_to_norm(aerosol.kappa, 'kappa')
V_1 = eval_hermite(1, V)
V_2 = eval_hermite(2, V)
N_1 = eval_hermite(1, N)
kappa_2 = eval_hermite(2, kappa)
N_2 = eval_hermite(2, N)
sigma_2 = eval_hermite(2, sigma)
mu_2 = eval_hermite(2, mu)
kappa_1 = eval_hermite(1, kappa)
mu_1 = eval_hermite(1, mu)
sigma_1 = eval_hermite(1, sigma)
Smax = As.dot(np.array([1.0, mu_1, N_1, V_1, kappa_1, sigma_1, mu_2, N_2, V_2, kappa_2, sigma_2, mu_1*N_1, mu_1*V_1, mu_1*kappa_1, mu_1*sigma_1, N_1*V_1, N_1*kappa_1, N_1*sigma_1, V_1*kappa_1, V_1*sigma_1, kappa_1*sigma_1]))
_, S_mode_crit = kohler_crit(T, aerosol.mu*1e-6, aerosol.kappa)
ui = 2.*np.log(S_mode_crit/Smax)/(3.*np.sqrt(2.)*np.log(aerosol.sigma))
N_act = 0.5*aerosol.N*erfc(ui)
act_frac = N_act/aerosol.N
return Smax, act_frac
def pce_deg4(V, T, P, aerosol):
brackets = {'mu': ('uniform_n', 0.01, 0.25), 'sigma': ('uniform_n', 1.2, 3.0), 'V': ('weibull_n', 2.0, 1.0), 'kappa': ('uniform_n', 0.1, 1.2), 'N': ('uniform_n', 100.0, 10000.0)}
As = array([ 7.56052516e-04, -9.34563322e-04, -5.60707772e-04,
-2.81045850e-04, -8.60005463e-05, -1.82949921e-04,
3.64882107e-04, 3.49237541e-04, -3.86741535e-04,
9.02462861e-05, 8.08886457e-05, -1.33596686e-05,
-3.66592784e-05, 1.84829967e-04, 6.31816281e-06,
1.43484844e-05, -2.21552918e-05, 7.80306040e-06,
-1.63991275e-04, -3.31969381e-06, -2.77832570e-06,
1.31325527e-04, 3.36410592e-04, 1.29343179e-04,
2.53607557e-04, 4.72821983e-04, 1.10686337e-04,
-7.76721989e-06, 5.59677016e-05, -2.03761928e-04,
1.49402457e-04, -5.81947331e-05, -2.87190526e-04,
-3.32166236e-05, -3.11984126e-05, -3.12737856e-05,
-2.90655230e-04, -3.87123623e-06, 6.32799902e-06,
-1.99773540e-04, -2.14317445e-04, 1.88693178e-06,
4.01362982e-05, -2.82253395e-05, -9.73456375e-06,
5.36536549e-05, 2.32915895e-06, -5.70239230e-05,
-1.01407528e-05, 3.66706996e-05, 1.34141867e-06,
-2.51711061e-04, -1.77305074e-05, 1.77902144e-05,
-1.92713812e-04, -1.36248503e-04, -6.39443575e-05,
-7.64443869e-05, -7.75791871e-05, 8.91567781e-06,
7.53736769e-07, 6.30223806e-06, -1.34681343e-04,
-4.42559165e-06, -1.00314540e-05, -1.41222799e-05,
-1.02167268e-06, -1.73354939e-06, -7.14665553e-06,
-9.84241564e-06, 6.97345884e-05, 6.02591893e-06,
-1.40684526e-05, -6.74866315e-06, 3.57188698e-06,
-6.56244944e-05, 1.71048935e-06, -9.80075449e-06,
4.05502943e-06, -1.47650326e-04, 4.27525023e-06,
1.48744812e-05, 8.65829131e-05, 1.58888194e-05,
1.50391139e-05, 2.04711595e-04, 8.58208991e-06,
2.56625369e-06, 3.83611977e-05, 4.11932517e-05,
8.16629817e-06, 7.99776514e-06, -3.15884420e-06,
-1.86066573e-05, 3.63951273e-05, 1.75620847e-05,
-6.68152115e-06, -1.32085279e-05, 2.22445829e-05,
1.13519416e-05, -2.76824182e-05, 6.90459595e-05,
1.70119847e-05, -3.14717813e-05, 2.62866496e-05,
-3.92082853e-05, -2.89646068e-05, 8.58123514e-05,
-4.97675282e-06, -4.48064665e-05, -1.34016155e-05,
4.09449969e-06, 3.68871564e-05, 7.32857832e-05,
3.70926853e-05, -3.92935660e-06, 2.27675971e-05,
4.16571690e-05, 5.83095830e-06, 3.52573949e-06,
1.00259494e-04, -4.64968600e-05, -6.63668001e-05,
-3.33467254e-05, -4.04622167e-05, 1.43066345e-05])
def map_to_norm(x, param):
prior, args = brackets[param][0], brackets[param][1:]
if prior == "lognormal_n": args = map(np.log, args)
inv_func = inv_funcs[prior]
return inv_func(x, *args)
mu = map_to_norm(aerosol.mu, 'mu')
N = map_to_norm(aerosol.N, 'N')
V = map_to_norm(V, 'V')
sigma = map_to_norm(aerosol.sigma, 'sigma')
kappa = map_to_norm(aerosol.kappa, 'kappa')
V_1 = eval_hermite(1, V)
mu_4 = eval_hermite(4, mu)
kappa_3 = eval_hermite(3, kappa)
V_2 = eval_hermite(2, V)
N_1 = eval_hermite(1, N)
kappa_2 = eval_hermite(2, kappa)
N_2 = eval_hermite(2, N)
sigma_2 = eval_hermite(2, sigma)
V_4 = eval_hermite(4, V)
mu_2 = eval_hermite(2, mu)
N_3 = eval_hermite(3, N)
kappa_1 = eval_hermite(1, kappa)
mu_3 = eval_hermite(3, mu)
N_4 = eval_hermite(4, N)
mu_1 = eval_hermite(1, mu)
sigma_3 = eval_hermite(3, sigma)
V_3 = eval_hermite(3, V)
sigma_1 = eval_hermite(1, sigma)
sigma_4 = eval_hermite(4, sigma)
kappa_4 = eval_hermite(4, kappa)
Smax = As.dot(np.array([1.0, mu_1, N_1, V_1, kappa_1, sigma_1, mu_2, N_2, V_2, kappa_2, sigma_2, mu_3, N_3, V_3, kappa_3, sigma_3, mu_4, N_4, V_4, kappa_4, sigma_4, mu_1*N_1, mu_1*V_1, mu_1*kappa_1, mu_1*sigma_1, N_1*V_1, N_1*kappa_1, N_1*sigma_1, V_1*kappa_1, V_1*sigma_1, kappa_1*sigma_1, mu_1*N_2, mu_1*V_2, mu_1*kappa_2, mu_1*sigma_2, N_1*mu_2, N_1*V_2, N_1*kappa_2, N_1*sigma_2, V_1*mu_2, V_1*N_2, V_1*kappa_2, V_1*sigma_2, kappa_1*mu_2, kappa_1*N_2, kappa_1*V_2, kappa_1*sigma_2, sigma_1*mu_2, sigma_1*N_2, sigma_1*V_2, sigma_1*kappa_2, mu_1*N_1*V_1, mu_1*N_1*kappa_1, mu_1*N_1*sigma_1, mu_1*V_1*kappa_1, mu_1*V_1*sigma_1, mu_1*kappa_1*sigma_1, N_1*V_1*kappa_1, N_1*V_1*sigma_1, N_1*kappa_1*sigma_1, V_1*kappa_1*sigma_1, mu_1*N_3, mu_1*V_3, mu_1*kappa_3, mu_1*sigma_3, N_1*mu_3, N_1*V_3, N_1*kappa_3, N_1*sigma_3, V_1*mu_3, V_1*N_3, V_1*kappa_3, V_1*sigma_3, kappa_1*mu_3, kappa_1*N_3, kappa_1*V_3, kappa_1*sigma_3, sigma_1*mu_3, sigma_1*N_3, sigma_1*V_3, sigma_1*kappa_3, mu_2*N_2, mu_2*V_2, mu_2*kappa_2, mu_2*sigma_2, N_2*V_2, N_2*kappa_2, N_2*sigma_2, V_2*kappa_2, V_2*sigma_2, kappa_2*sigma_2, mu_1*N_1*V_2, mu_1*N_1*kappa_2, mu_1*N_1*sigma_2, mu_1*V_1*N_2, mu_1*V_1*kappa_2, mu_1*V_1*sigma_2, mu_1*kappa_1*N_2, mu_1*kappa_1*V_2, mu_1*kappa_1*sigma_2, mu_1*sigma_1*N_2, mu_1*sigma_1*V_2, mu_1*sigma_1*kappa_2, N_1*V_1*mu_2, N_1*V_1*kappa_2, N_1*V_1*sigma_2, N_1*kappa_1*mu_2, N_1*kappa_1*V_2, N_1*kappa_1*sigma_2, N_1*sigma_1*mu_2, N_1*sigma_1*V_2, N_1*sigma_1*kappa_2, V_1*kappa_1*mu_2, V_1*kappa_1*N_2, V_1*kappa_1*sigma_2, V_1*sigma_1*mu_2, V_1*sigma_1*N_2, V_1*sigma_1*kappa_2, kappa_1*sigma_1*mu_2, kappa_1*sigma_1*N_2, kappa_1*sigma_1*V_2, mu_1*N_1*V_1*kappa_1, mu_1*N_1*V_1*sigma_1, mu_1*N_1*kappa_1*sigma_1, mu_1*V_1*kappa_1*sigma_1, N_1*V_1*kappa_1*sigma_1]))
_, S_mode_crit = kohler_crit(T, aerosol.mu*1e-6, aerosol.kappa)
ui = 2.*np.log(S_mode_crit/Smax)/(3.*np.sqrt(2.)*np.log(aerosol.sigma))
N_act = 0.5*aerosol.N*erfc(ui)
act_frac = N_act/aerosol.N
return Smax, act_frac
