Design-Oriented Two-Stage Surrogate Modeling of Miniaturized Microstrip Circuits With Dimensionality Reduction (original) (raw)
Related papers
Surrogate Modeling of RF Circuit Blocks
Mathematics in Industry, 2010
Surrogate models are a cost-effective replacement for expensive computer simulations in design space exploration. Literature has already demonstrated the feasibility of accurate surrogate models for single radio frequency (RF) and microwave devices. Within the European Marie Curie project O-MOORE-NICE! (Operational Model Order Reduction for Nanoscale IC Electronics) we aim to investigate the feasibility of the surrogate modeling approach for entire RF circuit blocks. This paper presents an overview about the surrogate model type selection problem for low noise amplifier modeling.
Kriging, co-kriging and space mapping for microwave circuit modeling
2011
Space mapping (SM) is a popular technique that allows creating computationally cheap and reasonably accurate surrogates of EM-simulated microwave structures (so-called fine models) using underlying coarse models, typically equivalent circuits. Here, we consider various ways of enhancing SM surrrogates by exploiting additional training data as well as two function approximation methodologies, kriging and co-kriging. To our knowledge, it is the first application of co-kriging for microwave circuit modeling. Based on the three examples of microstrip filters, we present a comprehensive numerical study in which we compare the accuracy of the basic SM models as well as SM enhanced by kriging and co-kriging. Direct kriging interpolation of fine model data is used as a reference.
Rf circuit block modeling via kriging surrogates
2008
The use of replacement metamodels (global surrogate models) has become commonplace as a cost effective alternative for performing complex high fidelity computer simulations. Due to their compact formulation and negligible evaluation time, global surrogate models are very useful tools for exploring the design space, what-if analysis, optimization, and sensitivity analysis. In addition, multiple surrogate models can be chained together to easily model large scale systems where a direct, fullwave simulation would be too cumbersome. Two crucial aspects of global surrogate modeling are data collection (sequential design) and hyperparameter optimization. Kriging models have become very popular in many domains but have seen little application in electronics and Electro-Magnetism. In this paper we study the impact of different hyperparameter optimization algorithms and sampling strategies on the accuracy of Kriging models as applied to a Low Noise Amplifier (LNA) RF circuit block.
2012
Simulation-based optimization has become an important design tool in microwave engineering. Yet, employing electromagnetic (EM) solvers in the design process is a challenging task, primarily due to a high-computational cost of an accurate EM simulation. This paper is focused on efficient EM-driven design optimization techniques that utilize physically-based low-fidelity models, normally based on coarse-discretization EM simulations. The presented methods attempt to exploit as much of the knowledge about the system or device of interest embedded in the low-fidelity model as possible, so as to reduce the computational cost of the design process. Unlike many other surrogate-based approaches, the techniques discussed here are non-parametric ones, i.e., they are not based on analytical formulas. The paper presents several specific methods, including those based on correcting the low-fidelity model response (adaptive response correction and shape-preserving response prediction), as well as on suitable modification of the design specifications. Formulations, application examples and the discussion of advantages and disadvantages of these techniques are also included.
IET Microwaves, Antennas & Propagation, 2012
In this study, the authors propose a robust and computationally efficient algorithm for simulation-driven design optimisation of microwave structures. Our technique exploits variable-fidelity electromagnetic models of the structure under consideration. The low-fidelity model is optimised using its local response surface approximation surrogates. The high-fidelity model is refined by space mapping with polynomial interpolation of the low-fidelity model data used as an underlying coarse model. Our algorithm is carefully developed to minimise the number of evaluations of both the low-and high-fidelity model in the optimisation process. The operation and efficiency of the approach is demonstrated through design of a microstrip filter, a double-ring antenna and an ultra-wideband monopole antenna. A comparison with other design approaches, including the direct high-fidelity model optimisation, is also presented.
IEEE Transactions on Microwave Theory and Techniques, 1997
Efficient utilization of commercial electromagnetic (EM) simulators for design and optimization of microwave (MW) and millimeter-wave (MMW) circuits is achieved by classifying design problems into three categories-characterization of circuit elements, optimization of circuit elements, and creation of circuit element libraries such as scalable libraries. Practical aspects of the methods are illustrated by several examples. An equivalent circuit extraction technique suitable for n n n-port coupled structures is provided. The derived equivalent circuit is useful for extrapolating data, optimization, and deriving scalable models.
Simplified space-mapping approach to enhancement of microwave device models
International Journal of RF and Microwave Computer-Aided Engineering, 2006
In this article, we present advances in microwave and RF device modeling exploiting the space mapping (SM) technology. New SM-based modeling techniques are proposed that are easy to implement entirely in the Agilent ADS framework. A simplified SM-based model description is discussed. Using a two-section transformer example, we show how the modeling accuracy is affected by the model flexibility. Tables, diagrams, and flowcharts are developed to help in understanding the concepts. This makes the SM modeling concept available to engineers through widely used commercial software. Our approach permits the creation of library models that can be used for model enhancement of microwave elements. Frequency-interpolation techniques are discussed and implemented. A set of four different SM-based models is presented along with corresponding implementations in the ADS schematic for a microstrip right-angle bend and a microstrip shaped T-junction. We use a three-section transformer to illustrate the implementation procedure in full details. We apply the technique to a more complicated HTS filter modeling problem. Fine-model data is obtained from Sonnet's em. We discuss the relation between the model complexity and accuracy as well as further improvement of the model.
Physics-based Surrogates for Low-cost Modeling of Microwave Structures
Procedia Computer Science, 2013
High-fidelity electromagnetic (EM) simulation is a very accurate but computationally expensive way of evaluating the performance of microwave structures. In many situations, it has to be done multiple times when conducting various design tasks, such as parametric optimization or statistical analysis. Fast and accurate models, so-called surrogates, are therefore indispensable in contemporary microwave engineering. The most popular way of creating such models is by approximation of sampled EM-simulation data using, for example, low-order polynomials, support vector regression or neural networks. Unfortunately, initial cost of creating such models may be extremely high because of a large number of samples necessary to ensure reasonable accuracy. An alternative approach is to use physics-based models, where the surrogate is created by correcting an auxiliary low-fidelity model, e.g., equivalent circuit. In this paper, we review several modeling techniques exploiting this idea, including some variations of space mapping as well as shape-preserving response prediction. Our considerations are illustrated using examples of typical microwave components such as filters and antennas.
Rapid EM-Driven Design of Compact RF Circuits By Means of Nested Space Mapping
IEEE Microwave and Wireless Components Letters, 2000
A methodology for rapid design of RF circuits constituted by compact microstrip resonant-cells (CMRCs) is presented. Our approach exploits nested space mapping (NSM) technology, where the inner SM layer is used to correct the equivalent circuit model at the CMRC level, whereas the outer layer enhances the coarse model of the entire structure under design. We demonstrate that NSM dramatically improves performance of surrogate-based optimization of composite CMRC-based structures. It is validated using four examples of UWB microstrip matching transformers (MTs) and compared to previous attempts to surrogate-based compact structure design.
IEEE Transactions on Microwave Theory and Techniques, 2005
Accurate and efficient empirical model generation techniques of microwave devices, for a large range of geometric and material parameters opportunely chosen, are presented. The empirical models are based on multiple linear regression approach, which compensates the error between an initial inaccurate empirical model and an electromagnetic (EM) full-wave solver (or measurement data). The aim of these techniques is to generate accurate empirical models, which are computationally very efficient with respect to any EM technique. These simple models could be integrated in a toolbox of any commercially available computed-aided design tools for RF/microwave circuits. Comparisons with artificial neural networks and linear-regression-based models are listed and discussed for the dispersion of a microstrip transmission line propagating the quasi-TEM mode and a microwave tunable phase shifter propagating the even mode.