[skbake94]

Hi, I'm trying to build a neural network for fluid simulations. I'm using Ansys fluent for my ground truth simulation. So, I have some data from ansys fluent and I'm trying to handle the data using python. In this process, I'd like to be able to calculate gradients of velocity and pressure using python. They don't have to exactly match the gradients calculated by fluent but I want the general shape of fields generated by the spatial derivative variables match. However, importing the fluent data as csv and using np or torch gradient gives me values that are completely off from the ansys fluent data. Has anyone tried the same thing as me and experienced this situation? I need some suggestions on how to make the calculation done in python to at least follow the general trend of actual spatial gradient values from ansys flluent.

Thanks