Neuronal architecture for waveguide inductive iris bandpass filter optimization

We present a simple and very accurate neuronal architecture approach matching, in a wide range of iris aperture, thickness and frequency, the numerical results obtained by using a precise high frequency electromagnetic simulator for symmetrical inductive irises in rectangular waveguide. For this purpose, a smoothed piecewise linear model has been chosen because this approach permits smooth transitions between linear regions through the use of logarithm of hyperbolic cosine functions, well suited for the frequency behavior of these inductive irises and circuit optimization. The model has been easily implemented into MMICAD(R), by using their MDL capability. Comparisons for high order high frequency waveguide filters for satellite applications has been made, showing an excellent agreement with full 3D electromagnetic HP-HFSS(R) simulations