Premium
Generalization of circuit simulator by automatic differentiation
Author(s) -
Kato Toshiji
Publication year - 2005
Publication title -
electrical engineering in japan
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.136
H-Index - 28
eISSN - 1520-6416
pISSN - 0424-7760
DOI - 10.1002/eej.20075
Subject(s) - generalization , jacobian matrix and determinant , simple (philosophy) , computer science , automatic differentiation , class (philosophy) , power (physics) , electronic circuit , simulation , algorithm , electrical engineering , engineering , mathematics , artificial intelligence , mathematical analysis , philosophy , physics , epistemology , quantum mechanics , computation
Circuit simulators have often been used for power electronic system design. However, most of them are limited in their application class to one with fixed elements because they can only handle programmed or fixed element characteristics. To overcome this disadvantage, a simulator should be flexible enough to handle any combination of characteristics of available mathematical functions and operations. It is a problem how to compute Jacobian elements or partial derivatives with respect to circuit variables. This paper proposes a way to compute the derivatives by using the automatic differentiation method which is simple and accurate. The principle and application examples are described. © 2005 Wiley Periodicals, Inc. Electr Eng Jpn, 152(3): 47–54, 2005; Published online in Wiley InterScience ( www.interscience.wiley.com ). DOI 10.1002/eej.20075