Abstract
Modeling an electromagnetic (EM) structure with curved boundaries using a conformal finite-difference time-domain (CFDTD) method retains a second order accuracy while using a staircased FDTD one only gives a first order accuracy. Although the CFDTD algorithm has been demonstrated being very accurate for simulating some vacuum electronic devices which usually operated at fundamental modes, this is not necessarily true for higher order mode simulations such as modeling fusion gyrotrons. In this work, a benchmark model has been carefully designed for validating the CFDTD algorithm for accurate higher order mode simulations. The preliminary benchmark results show the calculation of a higher order mode exhibits a larger error and always lower frequency (red) shift. It is found that the error generated in higher order mode calculations using the CFDTD method is even larger than that of the staircased FDTD method. This indicates that the convergence of higher order mode simulations using the CFDTD method might not be as good as expected in lower order mode simulations compared to the staircased FDTD one. We attribute this contradiction to the dispersion error caused by a reduction of the time step in the Dey-Mittra algorithm required to retain most cut cells enclosing the full geometry. In order to achieve the required accuracy of the CFDTD method for higher order mode simulations, a compromised methodology has been carried out and validated using the delicate benchmark model. Detailed algorithm and benchmark results will be presented.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call