Hello,
The Natural Modes result uses the linear hydrostatic stiffness matrix at the cone's equilibrium position to find the natural periods of the system. On the other hand, in your time domain analysis the Aqwa solver calculates the nonlinear hydrostatics over the instantaneous wetted surface of the cone at each time step.
In the time domain Analysis Settings, try setting the 'Use Linear Stiffness Matrix to Calculate Hydrostatic' option to 'Yes'. This should give you a better agreement between the Natural Modes result and your estimated natural period in the time domain.
You can also try setting the 'Convolution' option to 'No', which causes the Aqwa solver to use the acceleration RAOs to calculate the radiation forces/moments (rather than a convolution integral method). This may give you a better match with your own eigenvalue calculation.
I should point out that the natural period from your original time domain analysis (with nonlinear hydrostatics and convolution method for radiation terms) should be the most accurate estimate of the real system behaviour.
I hope this helps!
Mike