Study Of The Extended Finite-Difference Time-Domain Algorithm

With the increasingly higher frequencies and higher density of the microwave integrated circuits, the use of the simulators able to combine the full-wave numerical method with the models of the active electronic devices is essential due to the fact that EMI and radiation effects in active circuits become significant. The finite-difference time-domain (FDTD) method is one of the best full-wave simulations. In this thesis, some approaches of modeling the microwave active circuits based on the FDTD method are discussed.The dissertation firstly introduces the history of FDTD, and then introduces the fundamental knowledge and some key techniques of FDTD. An emphasis is placed on some classical modeling approaches of the extended FDTD algorithm, including equivalent current source approach, equivalent voltage source approach, SPICE sub-circuit approach and finite bandwidth device modeling approach. We illustrate the disadvantages of the available approaches, based on which some improved techniques are presented.The thesis presents a novel FDTD model of matched load terminating a microstrip line and discusses the modeling process in detail. Numerical results demonstrate this proposed model can effectively decrease the reflection and improve the matching effect. With this proposed model, we can effectively extract the S-parameters of microwave circuits using the matched load extracting approach.To increase the computational efficiency, then the extended ADI-FDTD approach is discussed in this thesis, and its numerical stability and numerical dispersion characteristic are theoretically studied for the first time. Three common lumped models are investigated:resistor, capacitor and inductor, and three difference schemes are discussed. It provides a theoretical proof to the numerical stability and dispersion of the extended ADI-FDTD approach.To incorporate an arbitrary linear circuit network into FDTD, the piecewise linear recursive convolution (PLRC) technique is combined to the FDTD codes, and the FDTD iterative formulas of one-port linear network and two-port linear network are generally derived. Numerical results show that the proposed approach is of the same computational speed and accuracy compared with the Z-transform technique.On the other hand, three novel techniques are proposed to analyze microwave active circuits including lumped devices characterized by finite bandwidth parameters by combing the vector fitting technique into FDTD algorithm. Compared with the earlier approaches, there are two attracting advantages to be gained. One is that all the proposed schemes do not need the tedious time-domain convolution products. The other is that it avoids the time-domain non-causality brought by the inverse Fourier transform technique. It is useful for designer to analyze microwave active circuits in global-domain.At last, how to apply FDTD to full-wave analyze the effect of high-power pulse on the microwave active circuits is fundamentally discussed, and global-domain modeling process is introduced. Numerical experiments show that the FDTD method is an efficient analyzing tool for this problem. It will be of practical value to enhance the survival ability of electronics equipments under the complicated electromagnetism environment.