New coder here. I have 2 rows of inputs and 3 of outputs, trying to interpolate between the inputs. The outputs are experimentally determined coefficients. The input data does form a grid.
Inputs are temperature, %concentration. Outputs are coefficients A, B, C.
T %
[[316, 3],
[316, 6],
[322, 3],
[322, 6],
[333, 3],
[333, 6]]
A, B, C
[[0.2925*10**-11, 7517.9, -0.0027],
[0.1275*10**-14, 9826.53, -0.0471],
[0.2506*10**-13, 8923.77, -0.0010],
[0.2506*10**-15, 10669.89, -0.2284],
[0.7319*10**-10, 6770.42, -0.0467],
[0.1800*10**-13, 9259.93, -0.0564]]
I've tried multidimensional linear interpolation using numpy and scipy.linalg as taught in class but I think it only works for data with known functions. The columns of the square matrix are [1], T, %, T*%, T/%, %/T. I figured it only works with a square matrix but I'm still not getting good values. This one just tries to solve for the coefficient A for the moment.
import numpy
import scipy.linalg
N = 6
coorA = numpy.zeros((N,N))
coorA = [[1, 316, 3, 948, 105, 0.0095],
[1, 316, 6, 1896, 53, 0.0190],
[1, 322, 3, 966, 107, 0.0093],
[1, 322, 6, 1932, 54, 0.0186],
[1, 333, 3, 999, 111, 0.0090],
[1, 333, 6, 1998, 56, 0.0180]]
solnA = numpy.zeros(N)
solnA = [0.2925*10**-11, 0.1275*10**-14, 0.2506*10**-13, 0.2691*10**-15, 0.7319*10**-10, 0.18*10**-13]
cA = numpy.zeros(N)
cA = scipy.linalg.solve(coorA, solnA)
x = 316
y = 6
A = cA[0] + cA[1]*x + cA[2]*y + cA[3]*x*y + cA[4]*x/y + cA[5]/x*y
print (coorA)
print (solnA)
print (cA)
print (A)
# A output was -9.966296504331275e-11
# A output should be 0.1275*10**-14 as it's the 2nd line of inputs
Any idea on how to move forward? I feel like there's an easy answer but this class is a crash course into coding.
