Simulation-based average -value modeling of power electronic converters and subsystems
Abstract
The ability to rapidly and accurately establish the frequency-domain characteristics of power electronic converters and subsystems is essential for the stability analysis of modern power systems. Consequently, various average-value and sampled-data modeling methodologies have been developed as a means to establish these characteristics. However, the derivation of an average-value model typically requires intimate knowledge of the operation of the converter and requires significant effort if the circuit topology is complex and/or the power converter exhibits multiple modes of operation. In this thesis, a simulation-based averaging methodology is introduced applicable to power converters of arbitrary complexity including those containing nonlinear and/or time-varying components. In addition to eliminating the need for the analytical derivation of average-value models, this methodology readily permits the inclusion of secondary effects such as conduction and switching losses, and magnetic nonlinearities, to name a few. This methodology has been successfully applied to numerous examples including five distinct power converter topologies. In each of the cases considered, reasonable agreement was observed between the frequency-domain characteristics predicted using the newly developed methodology and those established using detailed simulations and/or hardware measurements. However, in two of the examples, small discrepancies were observed at higher frequencies, which were subsequently analyzed and quantified. Criteria are set forth in which it is shown that the proposed method yields convergent results for disturbances whose frequencies are sufficiently less than the switching frequency. In addition to providing a theoretical foundation, an easy-to-apply correction term is introduced that improves the high-frequency accuracy of the new approach.
Degree
Ph.D.
Advisors
Wasynczuk, Purdue University.
Subject Area
Electrical engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.