I tested your program and as I had difficulties to get the numpy.net version to run I changed the "Test" function to use the commented Accord implementation.
The result looks ok, but the amplitude of the IMFs is too small by a factor of 2, e.g. the maximum of the first IMF is not 1 but 0.5
I wasn't able to find the source of this problem.
lariamato
commented
2 years ago 
I tested your program and as I had difficulties to get the numpy.net version to run I changed the "Test" function to use the commented Accord implementation. The result looks ok, but the amplitude of the IMFs is too small by a factor of 2, e.g. the maximum of the first IMF is not 1 but 0.5 I wasn't able to find the source of this problem.