University of Thessaly (UTH)
Order reduction of large thermal and electrical models with system-theoretic techniques and matrix equation algorithms
Abstract
dc:descriptionComputer simulation of modern IC subsystems, like power grids, interconnect structures and substrate regions, possesses a fundamental role in the EDA industry. The electric models of the aforementioned systems are enormous, and their functional simulation requires solving systems of equations with dimension that can reach several millions or even billions. Model Order Reduction (MOR) techniques provide attractive ways to reduce these highly complex models, replacing them by models with much smaller internal dimensions whose behavior at the input/output ports approximates that of the original models. Balanced Truncation (BT) type MOR methods of internal states that are at the same time insufficiently controllable at the inputs and insufficiently observable at the outputs have the advantage of very reliable estimates for the accuracy of the reduced model. However, they have significant computational and storage costs for producing the reduced-order models, as they require the solution of Lyapunov matrix equations which is an intensive computational procedure, and also involve the storage of dense matrices, making them applicable only to models of few thousand states. To eliminate computational costs and storage needs, low-rank factorization methods have been developed that allow the use of these methods in real-world applications. The emphasis in the present work is given in circuit and thermal models derived from integrated circuits and the calculation of the reduced-order models on specific frequency or time windows, as well as on the efficient implementation of Balanced Truncation type methods. More specifically, in most applications the circuit is intended to operate only in specific frequency windows, which means that the reduced-order model can become unnecessarily large to achieve approximation over all frequencies. Correspondingly, for the thermal models there is usually a final time in which all thermal effects can be assumed to have reached steady state. The focus is on model order reduction methods in specific frequency or time windows, combined with the efficient implementation of sparse methods for calculating Lyapunov-type matrix equations in order to manipulate large-scale input models.
Degree
thesis:*- Grantor dc:publisher
- University of Thessaly (UTH)
- Year dc:date
- 2019
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Floros, Georgios
- Φλώρος, Γεώργιος
Subjects
dc:subject × 16- Μείωση τάξης μοντέλου
- Προσομοίωση κυκλωμάτων
- Θερμική προσομοίωση
- Μέθοδοι εξισορρόπησης και αποκοπής
- Εξισώσεις πινάκων
- Ψηφιακά ολοκληρωμένα κυκλώματα
- Model order reduction
- Circuit simulation
- Thermal simulation
- Balanced truncation methods
- Matrix equations
- Integrated circuits
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
dc:rights- Statement dc:rights
-
- BY
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/46633