International Journal of Numerical Modelling: Electronic Networks, Devices and Fields ( IF 1.6 ) Pub Date : 2020-09-25 , DOI: 10.1002/jnm.2813 Slawomir Koziel 1 , Anna Pietrenko‐Dabrowska 2
The design of modern‐day high‐frequency devices and circuits, including microwave/RF, antenna and photonic components, historically has relied on full‐wave electromagnetic (EM) simulation tools. Initially used for design verification, EM simulations are nowadays used in the design process itself, for example, for finding optimum values of geometry and/or material parameters of the structures of interest. In a growing number of cases, EM‐driven design closure is mandatory because alternative ways of evaluating the circuit performance (such as through equivalent network modeling) are grossly inaccurate and unable to account for cross‐coupling effects (eg, in densely arranged layouts of compact circuits or antenna arrays), or various environmental components that affect the circuit performance (eg, connectors or housing for antenna structures). Despite being imperative, simulation‐based design poses significant challenges, mostly due to the high computational cost of accurate, high‐fidelity analysis. Repetitive simulations entailed by conventional optimization routines and even more by uncertainty quantification procedures (eg, Monte Carlo analysis) or tolerance‐aware design tasks may generate the costs that are unmanageable or at least impractical. The availability of massive computational resources does not always translate into design speedup due to the need to account for interactions between devices and their surroundings as well as multiphysics (eg, EM‐thermal) effects.
Not surprisingly, traditional design procedures that directly utilize EM‐simulated responses often fail or are impractical. Alternatives to full‐wave simulation tools, therefore, are increasingly popular among EM designers. Among the many available options, fast surrogate models that accurately capture the electrical characteristics of the components of interest, recently have received significant attention. Replacing or supplementing EM analysis by the surrogates enables execution of simulation‐based design tasks at low computational cost. This is especially the case for data‐driven (approximation) models, which are by far the most popular ones due to their versatility and widespread availability. Some broadly used methods include polynomial regression, kriging, radial basis function, neural networks, and polynomial chaos expansion. The practical issue here is nonlinearity of high‐frequency component outputs, which along with the curse of dimensionality, hinders utilization of this class of techniques for multiparameter components. Physics‐based surrogates (eg, space mapping or various response correction methods) feature improved generalization capability at the expense of being problem specific: rendering the surrogate normally involves an underlying low‐fidelity model, for example, equivalent network or coarse‐mesh EM simulation. In addition to that, inverse modeling has been recently fostered as a practical alternative to forward models when solving certain types of design tasks, especially dimension scaling or high‐frequency structures. The new developments concerning the improvements and generalizations of the existing methods as well as addressing dimensionality and scalability issues are under way.
This special issue focuses on the current state of the art and future directions in forward and inverse surrogate modeling for high‐frequency design. The issue contains 13 articles covering various aspects of numerical modeling of microwave and antenna components as well as design applications. Some of these articles are review works that give an account for the recent developments of particular methodologies. The largest number of articles is devoted to forward surrogate modeling methods in the context of specific design problems. In particular, Feng et al1 overview the advancements of parametric modeling of microwave components using neural networks with the system output represented by the zeros and poles of its corresponding transfer function. Koziel and Pietrenko‐Dabrowska2 describe the developments of performance‐driven surrogate modeling methods, which is one of the approaches recently proposed to address the dimensionality and parameter range issues in high‐frequency modeling. Loukreziz et al3 propose a novel algorithm for sparse least squares‐based polynomial chaos expansion models involving sequential experimental designs, whereas Georg and Römer4 discuss the utilization of conformal maps to construct basis functions for generalized polynomial chaos (gPC) as a way of enhancing its convergence properties. The advantages of the method are demonstrated using optical components. De Ridder et al5 address statistical modeling of frequency responses using linear Bayesian vector fitting. Finally, Hassan et al6 present computationally efficient microwave design centering using space mapping and a trust‐region framework. A common theme of the aforementioned articles is to reduce the computational cost of EM‐simulation‐driven design processes, which is important from practical perspective, especially carrying out design closure in acceptable timeframes as well as design automation.
The second group of articles in this special issue is focused on inverse modeling techniques. The work by Jin et al7 summarizes the recent developments in neural network‐based inverse modeling for microwave design applications, whereas Koziel and Bekasiewicz8 discuss inverse surrogates for rapid dimension scaling of multiband antennas. Both works emphasize the fundamental advantages of inverse surrogates, that is, their capability of directly yielding near‐to‐optimum parameter sets corresponding to given performance requirements without the necessity of formal optimization.
The third group of articles addresses application of surrogate modeling methods for the design of microwave/RF and antenna components. Xhafa and Yelten9 apply neural network models for variability analysis of low‐noise amplifier, whereas Leifsson et al10 employ polynomial chaos‐based kriging for uncertainty quantification of multiband patch antennas. The special issue is concluded with the articles by Belen et al11 and Mahouti et al12 which discuss design of 3D printed ceramic reflectarrays using neural network surrogates.
As Guest Editors we would like to express our gratitude to Prof. Jianming Jin (Editor‐in‐Chief of Int J Numer Model) for the opportunity to publish this special issue. We would also like to thank all authors for their high‐quality contributions, as well as all reviewers who devoted their time and expertise to careful review of the submissions. We do believe that the journal readers will find the presented works useful and that the special issue will help raising the awareness and popularity of surrogate modeling techniques as viable methodologies to enable expedited EM‐driven design of high‐frequency structures.
中文翻译:
有关高频设计的正向和反向代理建模进展的特刊的社论
过去,包括微波/射频,天线和光子组件在内的现代高频设备和电路的设计一直依靠全波电磁(EM)仿真工具。EM仿真最初最初用于设计验证,如今已用于设计过程本身,例如,用于查找目标结构的几何形状和/或材料参数的最佳值。在越来越多的情况下,强制执行由EM驱动的设计,因为评估电路性能的替代方法(例如,通过等效的网络建模)非常不准确,并且无法考虑交叉耦合效应(例如,在密集布置的布线中)。紧凑型电路或天线阵列),或影响电路性能的各种环境组件(例如,天线结构的连接器或外壳)。尽管势在必行,但基于仿真的设计还是面临着严峻的挑战,这主要是由于准确,高保真度分析的计算成本较高。传统优化例程所进行的重复仿真,甚至不确定性量化程序(例如,蒙特卡洛分析)或具有公差意识的设计任务所带来的重复仿真,都可能产生难以控制或至少不切实际的成本。由于需要考虑设备与其周围环境之间的相互作用以及多物理场效应(例如,EM热效应),因此大量计算资源的可用性并不能总是转化为设计加速。传统优化例程所进行的重复仿真,甚至不确定性量化程序(例如,蒙特卡洛分析)或具有公差意识的设计任务所带来的重复仿真,都可能产生难以控制或至少不切实际的成本。由于需要考虑设备与其周围环境之间的相互作用以及多物理场效应(例如,EM热效应),因此大量计算资源的可用性并不能总是转化为设计加速。传统优化例程所进行的重复仿真,甚至不确定性量化程序(例如,蒙特卡洛分析)或具有公差意识的设计任务所带来的重复仿真,都可能产生难以控制或至少不切实际的成本。由于需要考虑设备与其周围环境之间的相互作用以及多物理场效应(例如,EM热效应),因此大量计算资源的可用性并不能总是转化为设计加速。
毫不奇怪,直接利用EM模拟响应的传统设计程序通常会失败或不切实际。因此,全波仿真工具的替代方案在EM设计人员中越来越受欢迎。在许多可用选项中,快速替代模型可以准确捕获目标组件的电气特性,最近受到了极大的关注。用代理替代或补充EM分析可以以较低的计算成本执行基于仿真的设计任务。数据驱动(近似)模型尤其如此,由于其多功能性和广泛的可用性,它们是迄今为止最受欢迎的模型。一些广泛使用的方法包括多项式回归,克里金法,径向基函数,神经网络和多项式混沌展开。这里的实际问题是高频分量输出的非线性,再加上维数的诅咒,阻碍了此类技术用于多参数分量的使用。基于物理的替代(例如空间映射或各种响应校正方法)具有改进的泛化能力,但要以特定于问题为代价:渲染替代通常涉及底层的低保真度模型,例如等效网络或粗糙网格EM仿真。除此之外,在解决某些类型的设计任务(尤其是尺寸缩放或高频结构)时,最近已经提出将逆建模作为正向模型的一种实用替代方案。
本期专刊着眼于高频设计的正向和反向代理建模的最新技术和未来方向。本期包含13篇文章,涵盖微波和天线组件的数值建模以及设计应用的各个方面。这些文章中有一些是评论性著作,它们解释了特定方法的最新发展。最多的文章致力于在特定设计问题的背景下提出替代建模方法。特别是,Feng等人1概述了使用神经网络对微波组件进行参数化建模的进展,其系统输出由其相应传递函数的零点和极点表示。Koziel和Pietrenko-Dabrowska 2描述性能驱动的替代建模方法的发展,这是最近提出的解决高频建模中尺寸和参数范围问题的方法之一。Loukreziz等3提出了一种算法用于涉及连续的实验设计,而乔治和Römer稀疏最小二乘基于多项式混沌扩张模型4讨论共形的利用映射来构造广义多项式混乱(GPC)基函数作为加强的方式其收敛性。使用光学组件证明了该方法的优势。De Ridder等人5利用线性贝叶斯向量拟合解决了频率响应的统计建模问题。最后,哈桑等6目前使用空间映射和信任区域框架进行计算有效的微波设计居中。前述文章的一个共同主题是降低EM仿真驱动的设计过程的计算成本,这从实践的角度来看很重要,尤其是在可接受的时间范围内进行设计封闭以及设计自动化。
本期特刊的第二组文章集中于逆建模技术。Jin等人7的工作总结了微波设计应用中基于神经网络的逆建模的最新进展,而Koziel和Bekasiewicz 8讨论了多波段天线快速尺寸缩放的逆替代。两项工作都强调了反向代理的基本优势,即它们可以直接产生与给定性能要求相对应的接近最佳的参数集,而无需进行形式上的优化。
第三组文章介绍了替代建模方法在微波/射频和天线组件设计中的应用。Xhafa和Yelten 9将神经网络模型应用于低噪声放大器的变异性分析,而Leifsson等人10将基于多项式混沌的Kriging用于多频带贴片天线的不确定性量化。Belen等人11和Mahouti等人12的文章总结了这一特殊问题,这些文章讨论了使用神经网络替代技术设计3D打印陶瓷反射阵列的方法。
作为客座编辑,我们要感谢金建明教授(Int J Numer Model的主编)有机会出版此期特刊。我们还要感谢所有作者的高质量贡献,以及所有花费时间和专业知识仔细审查提交内容的审稿人。我们确实相信,期刊读者会发现所介绍的作品很有用,并且本期特刊将有助于提高替代建模技术作为可行方法的认识和普及度,从而能够加快由EM驱动的高频结构的设计。