python - 将值更改为数组会导致 NAN

Question

我开始学习 Python 并目睹了一个奇怪的问题，我无法在网上找到相关信息。

因为我不能发布实际代码，所以我将使用虚拟代码。

def model(x):
    A = list of length 4  
    B = list of length 4  
    C = list of length 4  
    D = list of length 4  

    # below methods of calculating X, one of which works
    # Method 1: WORKS
    X0 = A[0] + B[0] + C[0] + D[0]  
    X1 = A[1] + B[1] + C[1] + D[1]  
    X2 = A[2] + B[2] + C[2] + D[2]  
    X3 = A[3] + B[3] + C[3] + D[3]
    X = [X0, X1, X2, X3]

    # Method 2: DOES NOT WORK, 
    X = [A[i] + B[i] + C[i] + D[i] for i in range(4)]

    M = list of shape 4x4
    result = numpy.matmul(M, X)
    # result = [nan, nan, nan, nan] after some iterations

    return result

上面的模型返回一个dy/dt矩阵，我在下面的代码中使用它来整合

h   = 0.1;                              # timestep
sim_time = np.arange(0, 1500, h)        # timeseries
data = []
for i in sim_time:
    xdot    = model(x);                     
    x       = [a + h * b for a,b in zip(x, xdot)] 
    data.append(x)

我不明白我做错了什么。我确实发现这可能是由于浮点错误，但不确定是否是这种情况或如何克服这个问题。

任何帮助深表感谢。

这是在变为 nan 之前的一些迭代的结果

[-0.0383724   0.32659622 -0.02615451  0.01465148]
[-0.04153605  0.43814228 -0.02797397 -0.19174319]
[-0.04523129  0.07295898 -0.02615787  0.48396945]
[-0.0502085   0.88057674 -0.03373012 -0.90358412]
[-0.07082689 -0.20715555  0.00221235  1.75009264]
[-0.06635629  1.67610358 -0.04464798 -2.09508071]
[-0.0996444  -0.47602223  0.01246201  2.41644675]
[  -32.63095619   564.81941652   -14.72529351 -1025.05888268]
[-4.51778757e+09 -8.04785367e+10  2.09936804e+09  1.42336386e+11]
[-1.54489920e+31  2.75142984e+32 -7.17739692e+30 -4.86623834e+32]
[-3.08719207e+90 -5.49821772e+91  1.43426848e+90  9.72426681e+91]
C:\Users\Mod_18122018.py:323: RuntimeWarning: overflow encountered in double_scalars
  state_derivative =[X[0]*(U**2/L), X[1]*(U**2/L), X[2]*(U**2/L**2), X[3]*(U**2/L**2)]
C:\Users\Mod_18122018.py:205: RuntimeWarning: invalid value encountered in double_scalars
  u         =x[4]/U
C:\Users\Mod_18122018.py:206: RuntimeWarning: invalid value encountered in double_scalars
  v         =x[5]/U
C:\Users\Mod_18122018.py:207: RuntimeWarning: invalid value encountered in double_scalars
  p         =x[7]*L/U
C:\Users\Mod_18122018.py:208: RuntimeWarning: invalid value encountered in double_scalars
  r         =x[6]*L/U
C:\Users\Mod_18122018.py:233: RuntimeWarning: invalid value encountered in double_scalars
  u = (x[4] - uc)/U;
C:\Users\Mod_18122018.py:234: RuntimeWarning: invalid value encountered in double_scalars
  v = (x[5] - vc)/U;
C:\Users\Mod_18122018.py:265: RuntimeWarning: invalid value encountered in double_scalars
  n_c=n_c*U/L;
C:\Users\Mod_18122018.py:266: RuntimeWarning: invalid value encountered in double_scalars
  n=n*U/L;
[nan nan nan nan]
[nan nan nan nan]
[nan nan nan nan]
[nan nan nan nan]
[nan nan nan nan]
[nan nan nan nan]

标签： pythonlistintegration