For the same reason why we still use wind tunnels and a lot of other types of physical simulations when it comes to fluid dynamics and many other fields of “cla

For the same reason why we still use wind tunnels and a lot of other types of physical simulations when it comes to fluid dynamics and many other fields of “classical” applied physics. The reality is that all the equations and formulas we use are approximation, even simple things like the actual direction of flow in a pipe aren’t as simple as you think take any pipe and flow fluid through it in one direction, now introduce a dye in the middle and without a recirculating system the dye will eventually creep up against the flow. None of the equations you can nominally apply to fluid flows would enable you to predict nor explain this. There are now diffusion calculations that you can apply to calculate how much and how fast things could diffuse upstream, but these are essentially statistical models that we derived from experiments. So back to the Tesla valve question, the simulation likely work on ideal flows and geometries which are nearly impossible to actually produce through engineering so the slightest of variations in the surface, fluid density etc. can have drastic impacts on the outcome.