| The Insulated Gate Bipolar Transistor (IGBT) combines the advantages of the high current density bipolar operation that results in low conduction losses with the advantages of the fast switching and low drive power of MOS gated devices. Since their introduction in 1983, IGBTs have become widely accepted as the preferred switching devices in a variety of power converters and motor drive applications. Several models have been proposed in the literature to describe the static and dynamic behavior of IGBTs. These models can be broadly classified into two main categories: physics based models and behavioral models also known as compact models. In this thesis, we present a new behavioral model valid for both punch through and non punch through IGBT. The do part of the model is based on an empirical formula for the IGBT and needs three data points on the I-V curves to generate the entire DC characteristics. The dynamic part of the model is based on a Hammerstein-like current source. An algorithm for the extraction of the dynamic model parameters is outlined. A new template written in the MAST language is created in Saber based on this new model. The theoretical predictions of the model are in excellent agreement with the experimental data available for IGBTs fabricated by various manufacturers. Furthermore, the new model is consistently faster than its Hefner counterpart in simulating various circuit topologies. A new technique for calculating the switching losses Eon and Eoff in IGBT is introduced. The technique is compatible with the graphical capability of the simulator Saber and yields faster results than the conventional method. |
