Eylem Seç
Advances in Uncertainty Quantification for Computational Electromagnetics
Başlık:
Advances in Uncertainty Quantification for Computational Electromagnetics
Yazar:
Ganta, Sathya Swaroop, author.
ISBN:
9780438005167
Yazar Ek Girişi:
Fiziksel Tanımlama:
1 electronic resource (82 pages)
Genel Not:
Source: Dissertation Abstracts International, Volume: 79-10(E), Section: B.
Includes supplementary digital materials.
Advisors: Susan C. Hagness; Barry Vanveen Committee members: John A. Gubner; Benjamin Peherstorfer.
Özet:
Uncertainties in the measured electromagnetic properties of materials and uncertainties in dimensions of fabricated circuits and devices give rise to statistical uncertainties in field values obtained in computational electromagnetics (CEM) simulations. It is desirable to efficiently quantify such uncertainties resulting in the computed fields.
We have developed a classification framework for existing methods of uncertainty quantification in CEM. The existing methods have an opaque terminology that hides the mathematical strategies employed by these methods for uncertainty quantification. The framework we have developed addresses this opaque terminology by classifying the existing methods into well-defined groups with transparent terminology. This classification elucidates the relationships between the existing methods, if any. The classification framework also reveals new directions for research in the field of uncertainty quantification for CEM.
We have analyzed the accuracy and the efficiency of using polynomial models in approximating statistical field variation in CEM simulations. Canonical scatterers with dielectric uncertainties were chosen as the performance evaluation testbeds. The accuracy of polynomial approximations was shown to be strongly dependent on the mean dielectric properties of the scatterer. Extremely high-order expansions are necessary to achieve accurate approximations in some instances. The computational cost of higher-order approximations limits the utility of polynomial basis function expansions in CEM uncertainty quantification. We have developed a new gradient estimation method using electric field integral equations to be used for sensitivity analysis in CEM. The cost of this new method does not scale with the number of uncertain parameters. Performance evaluation using the canonical dielectric cylinder and the 2D photonic crystal waveguide branch testbeds revealed that the new method is significantly better than the difference approximation technique in accuracy and that the new method is reliable even in realistic scenarios. The cost of the new method however scales with the number of observation points.
Finally, we have investigated the impact of sampling-based uncertainty quantification methods with regards to their computational efficiency. In sampling methods, computational efficiency depends on the sampling distribution and on whether the sampling is random or deterministic. Investigation was carried out using three distinct representative examples. Uniform deterministic sampling was observed to have a significant advantage over other forms of sampling in terms of computational efficiency for all the three examples.
Notlar:
School code: 0262
Tüzel Kişi Ek Girişi:
Mevcut:*
Yer Numarası | Demirbaş Numarası | Shelf Location | Lokasyon / Statüsü / İade Tarihi |
---|---|---|---|
XX(682409.1) | 682409-1001 | Proquest E-Tez Koleksiyonu | Arıyor... |
On Order
Liste seç
Bunu varsayılan liste yap.
Öğeler başarıyla eklendi
Öğeler eklenirken hata oldu. Lütfen tekrar deneyiniz.
:
Select An Item
Data usage warning: You will receive one text message for each title you selected.
Standard text messaging rates apply.