CAREER: EleXIR: A Top-Down Approach to Research and Learning in Efficient Hierarchical Electromagnetic Simulation Methods for Complex Structures and Systems

计算电磁学（CEM）是一个多样化的领域，每年都有几个新的应用。由于计算算法和硬件的进步，分析越来越复杂的电磁问题已经成为可能。事实上，未来几代计算机的发展本身将严重依赖于高效、集成的CEM芯片和互连工具的可用性。随着全球对空前复杂的高速系统的日益重视，很明显，电磁效应正变得越来越占主导地位。无线网络在美国和世界上一些地方正以惊人的速度扩张。高速片上系统现在是无数商业、国防和研究领域的焦点。包括下一代互联网、战场通信和情报以及大规模分布式计算在内的几个具有前所未有潜力的关键当前和未来领域都依赖于复杂的无线网络和片上系统。这些领域将对日益信息化的经济和美国的国防，以及未来的全球高科技经济产生巨大影响。这些复杂系统和结构的设计和开发需求导致了对准确、分层的电磁分析的关键要求。为了开发新的设计和范例，并节省迭代物理原型和测试的高昂成本和时间要求，这些模拟是必要的。模拟的复杂性是如此之高，以至于直到最近才有可能在合理的计算资源和时间内获得甚至中等精度的结果。虽然电磁模拟的进化进步无处不在，但本提案中的电磁模拟：教学和研究（EleXIR）方法提出了范式转换，基于物理的革命性电磁模拟方法，这些方法显示出在未来几年内实现大规模结构和系统的准确，高效和集成的电磁模拟的潜力。这些方法将导致先进工具的发展，这些工具能够执行严格和自动化的三维全波电磁仿真，并且具有当今电路级建模的易用性和透明度。由此产生的进展将具有强大的潜力，影响当前和未来EM应用的几个仿真方面，包括基于无线的网络和片上系统问题，例如：高速射频电路和片上系统的信号完整性分析复杂目标雷达横截面的计算新型天线阵列的设计和分析现代机电、光学和量子器件的分析这些应用虽然不是详尽的，但却是广泛而多样的。为了解决这些问题，提出了一种统一的面向类的方法，该方法依赖于基于物理的冗余提取方法来对EM问题进行分层积分方程建模。高效的积分方程公式、针对特定类别问题的快速算法、分层和降阶建模方案的组合，是未来几年为几个片上系统和无线相关的EM应用开发革命性的新CEM工具的范例。随着这些技术的进步，为这些高科技领域的可持续研究和发展创造必要的人才储备的必要性和挑战，将通过新的综合课程以自上而下的方式在感兴趣的领域（如经典EM，高速电路，无线通信和设备）中纳入新的CEM方法来解决。这些对美国和全球经济具有战略重要性的研究、教学和技术转让方面的关键要求构成了本文提出的EleXIR方法的重点。EleXIR的总体目标包括开发，教学和技术转移无缝自上而下的方法，以在高科技领域实现现代CEM，在这些领域需要或将在不久的将来需要，包括无线和高速电路应用

0093102JandhyalaComputational electromagnetics (CEM) is a diverse field that sees several new applications each year. Owing to advances in computational algorithms and hardware, it has become feasible to analyze progressively more complex electromagnetic (EM) problems. Indeed, the development of future generations of computers will itself be heavily dependent on the availability of efficient, integrated CEM tools for chips and interconnects.With the enhanced global emphasis on high-speed systems of unparalleled complexity, it is evident that EM effects are becoming increasingly dominant. Wireless networks are expanding at a dramatic rate in the United States and in several parts of the world. High-speed systems-on-chip are now the focal points in myriad commercial, defense, and research areas. Several critical current and future domains with unprecedented potential including the next-generation internet, battlefield communications and intelligence, and large-scale distributed computing do or will rely on complex wireless networks and systems-on-chip. These areas will have a massive impact on the increasingly information-based economy and defense of the United States, and on the global high-technology economy of the future.The design and development needs of these complex systems and structures have led to a critical requirement for accurate, hierarchical EM analyses. These simulations are necessitated in order to develop new designs and paradigms, and to save on the exorbitant costs and time requirements of iterated physical prototyping and testing. The sheer complexity of the simulations is such that the possibility of achieving even moderately accurate results with reasonable computing resources and time did not exist until recently.While evolutionary advances in CEM are omnipresent, the Electromagnetic-Simulation: Instruction and Research (EleXIR) approach in this proposal presents paradigm-shifting, physics-based, revolutionary CEM methodologies that exhibit the potential to render feasible, over the next few years, the accurate, efficient, and integrated EM simulation of large-scale structures and systems. These approaches will lead to the development of advanced tools with the ability to perform rigorous and automated three-dimensional full-wave EM simulation with the ease and transparency of present-day circuit-level modeling. The resulting advances will have a strong potential to affect several simulation aspects of present and future EM applications, including wireless-based network and system-on-chip issues, such as: Propagation through random media and scattering from rough surfaces Signal integrity analysis for high-speed radio-frequency circuits and systems-on-chip Computation of radar cross sections of complex targets Design and analysis of novel antenna arrays Analysis of modem electromechanical, optical, and quantum devicesThese applications, while not being exhaustive, are extensive and varied. To address these, a unified class-oriented approach is proposed that relies on physics-based, redundancy-extracting approaches to hierarchical integral-equation modeling of EM problems. A combination of efficient integral equation formulations, fast algorithms tuned to the physics of specific classes of problems, and hierarchical and reduced-order modeling schemes, is the paradigm presented here for the development of revolutionary new CEM tools for several system-on-chip and wireless-related EM applications over the next few years. Along with the advances in such techniques comes the necessity and challenge of creating a pool of talent with the critical mass necessary for sustainable research and development in these high-technology areas, which will be addressed through new integrated curricula incorporating new CEM methods in a top-down manner in areas of interest i.e. classical EM, high-speed circuits, wireless communications, and devices. These crucial requirements in research, instruction, and technology-transfer that are of strategic importance to the United States and to global economics form the focus of the EleXIR approach proposed here.The general goals of EleXIR include the development, teaching, and technology transfer of seamless topdown approaches to modem CEM in high-technology areas where it is necessitated, or will be necessitated in the near future, including wireless and high-speed circuit applications.***