Techniques for stochastic simulation of complex electromagnetic and circuit systems with uncertainties

Chen, Xu

Permalink

https://hdl.handle.net/2142/101666 Copy

Description

Title

Techniques for stochastic simulation of complex electromagnetic and circuit systems with uncertainties

Author(s)

Chen, Xu

Issue Date

2018-07-03

Director of Research (if dissertation) or Advisor (if thesis)

Cangellaris, Andreas C.

Doctoral Committee Chair(s)

Cangellaris, Andreas C.

Committee Member(s)

• Schutt-Ainé, José E.
• Başar, Tamer
• Ravaioli, Umberto

Department of Study

Electrical & Computer Eng

Discipline

Electrical & Computer Engr

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Date of Ingest

2018-09-27T16:30:15Z

Keyword(s)

stochastic, modeling, simulation, latency insertion method, circuit simulation, packaging, uncertainty quantification, stochastic galerkin method, stochastic collocation, monte carlo

Abstract

This thesis presents a set of tools and methodologies that perform fast stochastic characterization and simulation of uncertainties in electromagnetic and circuit systems. Background information on polynomial chaos and fast stochastic numerical techniques is reviewed, and discussion is offered on comparison of different approaches to stochastic simulations. The formulation for Stochastic LIM, a Stochastic Galerkin Method-based time-domain circuit solver, is presented, and some simulation results are shown comparing the new solver to Monte Carlo techniques using a commercial circuit solver. The simulator is then used to simulate several transmission line problems, including single- and multi-conductor, crosstalk, and coupled-line on a printed circuit board substrate with fiber-weave effect. Stochastic Collocation technique is discussed as a method to characterize multi-level electromagnetic-circuit simulations. A method to use Monte Carlo integration to evaluate interpolation residual is presented. The effectiveness of the proposed method is demonstrated with a high-order problem of electromagnetic waves causing interference on a printed circuit board inside a vehicle with apertures. The effectiveness of the multi-level analysis methodology is demonstrated using eye diagram opening as cost function. Additionally, a wavelet-based Stochastic Collocation technique is introduced to solve circuit problems with resonant behavior. Finally, we discuss the overall work presented in this thesis and discuss several future research directions to extend the results presented here.

Graduation Semester

2018-08

Type of Resource

text

Permalink

http://hdl.handle.net/2142/101666

Copyright and License Information

Copyright 2018 Xu Chen

Owning Collections