Help on Monte Carlo Simulation of temperature relaxation of He-Ar mixture

[jorvz]

Hey, so i`ve been working on this code for a while for my research but it just doesn't seem to work well, i think it is misscalculating in some step, probaly on the relative velocities, because it keeps exploding the value for some really high values, but i am not shure. Can someone help me finding the source of the problem here?

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312

import numpy as np import matplotlib.pyplot as plt import pandas as pd import random as rd from matplotlib import animation from matplotlib.animation import PillowWriter   # Constants kB = 1.38e-23  # Boltzmamn constant amu = 1.66e-27  # Atomic mass unity m_He = 4 * amu  # He mass m_Ar = 38.9 * amu  # Ar mass   # Initial conditions n = 100000  # Total number os particles L = 10.  # Length of the box T_He = 300 # initial temperature of He T_Ar = 300 #T_mistura = 500 #Temperature of the mixture conc_He = 0.90  # molar fraction of He in the mixture #T_Ar = (T_mistura - conc_He*T_He)/(1.-conc_He) #initial temperature of Ar T_mistura = conc_He*T_He + (1. - conc_He)*T_Ar   n_He = int(n * conc_He)  # Number of He particles n_Ar = n - n_He  # Number of Ar particles sigma_He = np.sqrt((kB * T_He) / m_He)  # standard deviation for He speeds sigma_Ar = np.sqrt((kB * T_Ar) / m_Ar)  #standard deviation for Ar speeds   T = 100 #Total time for the simulations dt = 0.00002 # time interval   va = np.sqrt(2.*kB*T_mistura/(m_He + m_Ar)) #termal speed of the mixture def generate():     # Generating the initial speed and position of the particles     x = L * np.random.rand(n, 3)  # random positions     v = np.zeros((n, 3),dtype='float64')  # velocities array     v[:n_He,:] = np.random.normal(0, sigma_He, size=(n_He, 3))  # He velocities       v[n_He:,:] = np.random.normal(0, sigma_Ar, size=(n_Ar, 3))  # Ar velocities           return v/va, x #va to let it dimentionless   def plot(v):     fig, ax = plt.subplots(1,2, figsize=(25,10))           a = ax[0]     v_mag_He = np.linalg.norm(v[:n_He,:], axis=1)  # Magnitude das velocidades do hélio     a.hist(v_mag_He, bins=50, density=True, alpha=0.5, label='He')       # Plot the histogram for the initial velocities of Argon     v_mag_Ar = np.linalg.norm(v[n_He:,:], axis=1)  # Magnitude das velocidades do argônio     a.hist(v_mag_Ar, bins=50, density=True, alpha=0.5, label='Ar')       # Configuring the graph     a.set_title('Distribuição de velocidades')     a.set_xlabel('Velocidade (m/s)')     a.set_ylabel('Densidade de probabilidade')     a.legend(fontsize=20)           a.grid()           a = ax[1]     v_mag_He = v[:n_He,0]  # Magnitude das velocidades do hélio     a.hist(v_mag_He, bins=50, density=True, alpha=0.5, label='He')       # Plot the histogram for the initial velocities of Argon     v_mag_Ar = v[n_He:,0]  # Magnitude das velocidades do argônio     a.hist(v_mag_Ar, bins=50, density=True, alpha=0.5, label='Ar')       # Configuring the graph     a.set_title('Distribuição de velocidades')     a.set_xlabel('Velocidade (m/s)')     a.set_ylabel('Densidade de probabilidade')     a.legend(fontsize=20)           a.grid()                 plt.show() #Importing the matrix for viscosity muHA = pd.read_csv('tables/muHA.dat', sep="\s+",index_col=False).set_index(['0'])   #Importing the matrix for cosx and the cross sections xiHe4He4 = pd.read_csv('tables/xiHe4He4.dat', sep="\s+", skiprows=4, header=None) xiHe4Ar = pd.read_csv('tables/xiHe4Ar.dat', sep="\s+", skiprows=4, header=None) xiArAr = pd.read_csv('tables/xiArAr.dat', sep="\s+", skiprows=4, header=None)   def mover_particulas(v, x):     # Move the particles in every Δt of the simulation     #x = x + v*dt     x,v = bordas(x,v)     return x   def temp_cinetica(vel):           u = (1./n)*np.sum(v,axis=0)           T_He = (m_He*va**2)/(3. * n_He * kB)* np.sum(v[:n_He]**2)     T_Ar = (m_Ar*va**2)/(3. * n_Ar * kB)* np.sum(v[n_He:]**2)     T = conc_He*T_He + (1.-conc_He)*T_Ar     return T_He, T_Ar, T     def bordas(pos, vel):     """     Mantém as partículas dentro de uma caixa de tamanho L,     fazendo com que elas refletam nas paredes da caixa quando     colidem com elas.           Args:         pos (numpy array): Matriz Nx3 com as posições das N partículas.         vel (numpy array): Matriz Nx3 com as velocidades das N partículas.         L (float): Tamanho da caixa quadrada.               Returns:         numpy array: Matriz Nx3 com as novas posições das N partículas.         numpy array: Matriz Nx3 com as novas velocidades das N partículas.     """     # Verifica se alguma partícula está fora da caixa.     new_pos = pos.copy()     new_vel = vel.copy()       # Verifica se as partículas estão dentro da caixa e corrige as posições e velocidades caso contrário     for i in range(pos.shape[0]):         for j in range(pos.shape[1]):             if new_pos[i,j] < 0:                 new_pos[i,j] = abs(new_pos[i,j])                 new_vel[i,j] = -new_vel[i,j]             elif new_pos[i,j] > L:                 new_pos[i,j] = 2*L - new_pos[i,j]                 new_vel[i,j] = -new_vel[i,j]       return new_pos, new_vel     def colisao(v, tipo,sg):     # type 1 = He-He     # type 2 = Ar-Ar     # type 3 = He-Ar     p1 = 0     p2 = 0           while p1==p2:         if tipo==1:             #Selecting particles in the interval [0, n_He]             p1 = int(np.floor((n_He)*rd.random()))             p2 = int(np.floor((n_He)*rd.random()))                           #Particles masses             m1 = m_He             m2 = m_He                           #G constant for the type of collision             G = 400.                           #Cross section matrix             σ = xiHe4He4[100].to_numpy()                           #matrix for the values of cosX             ξ = xiHe4He4.to_numpy()                   elif tipo == 2:             #Selecting particles in the interval (n_He, n_Ar]             p1 = int(n_He + np.floor(rd.random()*n_Ar))             p2 = int(n_He + np.floor(rd.random()*n_Ar))                           #Particles masses             m1 = m_Ar             m2 = m_Ar                           #G constant for the type of collision             G = 100.                           #Cross section matrix             σ = xiArAr[100].to_numpy()             ξ = xiArAr.to_numpy()                   else:             #Selecting particles in the interval [0, n_He] and (n_He, n_Ar]             p1 = int(np.floor((n_He)*rd.random()))             p2 = int(n_He + np.floor(rd.random()*n_Ar))                           #Particle masses             m1 = m_He             m2 = m_Ar                           #G constant for the type of collision             G = 300.                           #Cross section matrix             σ = xiHe4Ar[100].to_numpy()                           #matrix for the values of cosX             ξ = xiHe4Ar.to_numpy()                           v1 = v[p1,:] #speed of the particle 1     v2 = v[p2,:] #speed of the particle 2           #Calculating relative speed     g = v1 - v2     gr = np.linalg.norm(g) #norm of g           #Post-collisional relative speed     gl = np.zeros(3,dtype='float64')                 #Center of mass velocity     G_cm = (m1*v1 + m2*v2)/(m1 + m2)           #Calcuting the index of cross section array     k = int(np.floor((np.log(1. + gr*va/G)/np.log(1.005)) + 0.5))          if k>899:k=899               vr = np.sqrt(g[1]*g[1] + g[2]*g[2])                  if ((σ[k]*gr)/(sg) > rd.random()): #Accept-rejection condition                   print(σ[k], gr)          n = int(np.floor(rd.random()*100))                   #Monte Carlo parameters fot the post collisional velocities         e =  2. * rd.random() * np.pi         cosx = ξ[k][n]         sinx = np.sqrt(1. - cosx*cosx)                   if vr>1e-15:         #Components of the post collisional relative speed             gl[0] = g[0]*cosx + vr*sinx*np.sin(e)             gl[1] = g[1]*cosx + (gr*g[2]*np.cos(e) - g[0]*g[1]*np.sin(e))*sinx / vr             gl[2] = g[2]*cosx - (gr*g[1]*np.cos(e) + g[0]*g[2]*np.sin(e))*sinx / vr         else:             gl[0] = g[0]*cosx             gl[1] = g[0]*sinx*np.cos(e)             gl[2]= g[0]*sinx*np.sin(e)                                    #calculating the post collision velocities         v1 = G_cm + 0.5*gl         v2 = G_cm - 0.5*gl                   #setting the new velocities         v[p1,:] = v1         v[p2,:] = v2         return v, σ[k]*gr     else:         return v, sg             t=0. μ = muHA[f"{conc_He}"][np.floor(T_mistura)] σg11 = 15. σg12 = 15. σg22 = 15. v, x = generate() v_t = np.zeros((n, 3, T)) Temp = np.zeros((T,3)) for t in range(0,T,1):           #plot(v)     #x = mover_particulas(v,x)                 np.append(Temp[t,:],temp_cinetica(v))     N_col_HH = int((1./n)*n_He*(n_He - 1.)*μ*dt*σg11)     N_col_AA = int((1./n)*n_Ar*(n_Ar - 1.)*μ*dt*σg22)     N_col_HA = int((2./n)*n_He*n_Ar*μ*dt*σg12)           print(f'N11 = {N_col_HH}, N12 = {N_col_HA}, N22 ={N_col_AA}')     print(f'σ11 = {σg11}, σ12 = {σg12}, σ22 ={σg22}')       for i in range(0,N_col_HH):         v, sg11 = colisao(v, 1,σg11)         if (sg11 > σg11): σg11 = sg11     for i in range(0,N_col_AA):         v, sg22 = colisao(v, 2,σg22)         if (sg22 > σg22): σg22 = sg22     for i in range(0,N_col_HA):         v, sg12 = colisao(v, 3,σg12)         if (sg12 > σg12): σg12 = sg12           print(t) plt.plot(np.arange(0,T,1), Temp[:,0],label='He') plt.plot(np.arange(0,T,1), Temp[:,1],label='Ar') plt.plot(np.arange(0,T,1), Temp[:,2], label='T') plt.legend() fig, ax = plt.subplots(1,1, figsize=(10,10))   def animate(i):     ax.clear()     v_mag_He = np.linalg.norm(v_t[:n_He,:,i], axis=1)  # Magnitude das velocidades do hélio     ax.hist(v_mag_He, bins=50, density=True, alpha=0.5, label='He')       # Plot the histogram for the initial velocities of Argon     v_mag_Ar = np.linalg.norm(v_t[n_He:,:,i], axis=1)  # Magnitude das velocidades do argônio     ax.hist(v_mag_Ar, bins=50, density=True, alpha=0.5,label='Ar')       # Configuring the graph     #ax.title('Distribuição de velocidades')     ax.set_xlabel('Velocidade (m/s)',fontsize=20)     ax.set_ylabel('Densidade de probabilidade',fontsize=20)           #ax.set_xlim(0,10)     #ax.set_ylim(0,2)     ax.grid()             ani = animation.FuncAnimation(fig, animate,frames=T,interval=10) ani.save('ani.gif',writer='pillow')

[farshid]

there are some steps that you can take to debug the code and narrow down the potential causes of the issue:

Check the inputs and outputs of the functions to make sure that they are consistent and make sense.

Print out the intermediate results of the calculations to see where things might be going wrong.

Use a debugger to step through the code and see where the program might be diverging from what you expect.

Use unit tests to validate the output of the code against known inputs and outputs.

Based on the code, it looks like you are simulating the motion of particles in a box, and calculating their velocities and positions. Some potential sources of the problem could be the calculation of relative velocities or the handling of the boundary conditions.

One thing you can do is to print out the values of the velocities and positions at each step of the simulation to see if there are any extreme or unexpected values. You can also plot the distributions of the velocities to see if they look reasonable.

Another thing you can try is to use smaller time steps and see if the problem persists. Sometimes, if the time step is too large, the simulation can become unstable and produce unexpected results.