University of Illinois at Urbana-Champaign
Efficient numerical modeling of miniature RF devices based on acoustic wave technologies
Abstract
dc:descriptionIn this dissertation, efficient numerical approaches are developed for modeling miniature RF devices based on acoustic wave technologies. The problems under consideration fall into two parts. One is the fast electromagnetic (EM) analysis of electrical layouts using eigenmode decomposition and lumped equivalent circuit models. The other is efficient electro-mechanical modeling of acoustic wave (AW) resonators based on the finite element tearing and interconnect (FETI) algorithm. In the first part, two different fast frequency sweep approaches are developed for analyzing electrical layouts of acoustic wave devices, i.e., the analytic extension of eigenvalues (AEE) and network characteristic mode analysis (NCMA). In those approaches, network parameters such as impedance (Z-) or scattering (S-) parameters are first computed from full-wave simulation at sampling frequencies and then decomposed into eigen modes. The obtained eigenvalues and eigenvectors are used to extract modal-based lumped equivalent circuit parameters which can generate device responses at all other frequencies with very little cost. The proposed methods not only significantly reduce computational time but also create equivalent circuits that provide physical insights to designers, facilitating diagnosis in circuit designs and optimizations. The accuracy of these approaches is evaluated by comparing results with full-wave solutions. For RF circuits whose electrical sizes are small and whose frequency range of interest is relatively small, the proposed fast frequency sweep approaches are found to be sufficiently accurate and highly practical for engineering applications. In the second part, solutions to electro-acoustic problems are formulated using the finite element method for accurate modeling of AW resonators. To accelerate the simulation of large-scale devices, the FETI method is developed by decomposing the entire domain into many smaller non-overlapping subdomains whose FE subsystems can be factorized with a direct sparse solver at a low cost. In the FETI algorithm, subdomain degrees-of-freedom (DOFs) are reduced to local interface DOFs, and transmission conditions are enforced to interconnect neighboring subdomains to form a global interface system that is solved iteratively. A dual-primal version of FETI (FETI-DP) is also formulated to extend the simulation from 3D periodic structures to general 3D finite devices by introducing a global corner system that contains all primal DOFs corresponding to corner nodes. To accelerate the iterations, transmission conditions are carefully designed for piezoelectric materials to make the interfaces as transparent as possible for waves across the boundaries. A forward-backward preconditioner (FBP) and a diagonal FBP (diagFBP) are proposed to further speed up the convergence of iterative solvers in 3D periodic and 3D finite simulations, respectively. The constructed preconditioners are found to be very efficient and effective and can dramatically reduce the number of iterations without adding extra computational cost. Parallelization is also implemented to enhance computational efficiency by distributing subdomain operations to different cores and solving local problems in parallel. Once the admittances of acoustic resonators are calculated, combining them with the frequency responses of electrical layouts yields the electrical response of an entire AW device. The accuracy, efficiency, and capability of the proposed methods are demonstrated through many examples including highly complicated RF devices.
Degree
thesis:*- Name thesis:degree_name
- Ph.D.
- Level thesis:degree_level
- Dissertation
- Discipline thesis:degree_discipline
- Electrical & Computer Engr
- Grantor
- University of Illinois at Urbana-Champaign
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Li, Hongliang
- Contributors dc:contributor
-
- Jin, Jianming
- Feng, Milton
- Gong, Songbin
- Schutt-Aine, Jose E
Subjects
dc:subject × 7Rights
dc:rights- Statement dc:rights
-
- Copyright 2022 Hongliang Li
- Language dc:language
- en, eng
Identifiers
dc:identifier.*- Handle dc:identifier
- https://hdl.handle.net/2142/117775