Metodi Avanzati per L'Analisi Transitoria di Sistemi Elettrici Lineari

The purpose of this doctoral work is to explore alternative TD simulation techniques in the context of the PEEC method. The key motivation arises from the necessity of stable full-wave models, able to include propagation delays in the PEEC formulation. Quasi-Static (QS) PEEC models have been widely employed in the past decades as efficient and reliable tools for the analysis and verification of the EM behavior of common structures typically employed in the electronic industry. Such models have demonstrated to be robust and, above all, always stable. In the last three decades, the technological evolution in electronics and the growing necessity of small and compact devices has led to an increase of the geometrical complexity of signal interconnecting structures. Moreover, for efficiency and functional purposes, an impressive increase of the working frequencies has been observed throughout the years, especially in the signal electronics and microwave areas. Such context has brought to the necessity of analyzing complex electrically long structures in the simulation stage. This has required the development of full-wave solvers, able to catch correctly the propagation delays and to represent more accurately all the non-ideal loss mechanisms affecting such devices. In the PEEC method, such delays are buried in the partial elements, complicating the circuit representation and giving rise to computationally demanding models, inevitably lengthening the simulation times. In order to keep the computational burden within an acceptable level, several approximations have been introduced for the representation of delayed partial elements. Unfortunately, those approximations, can lead to unpredictable instabilities showing up in the TD final results. For the reasons already explained, two different approaches for obtaining stable TD PEEC models have been investigated in this work. A possibility is to resort to Inverse Laplace Transform (ILT) techniques, naturally grounded in the frequency domain, collecting and combining the results obtained through the evaluation of the Laplace-domain PEEC model over specified points of the complex plane. Such a technique is relatively simple in its implementation, being necessary few changes to standard FD PEEC solvers. Another investigated option is the rigorous TD representation of partial elements through the Cagniard deHoop (CdH) technique, giving rise to direct TD convolution-based PEEC models.