CDI-Type II: Collaborative Research: Computing Supermassive Black Hole Mergers in Astrophysics

超大质量黑洞合并被认为发生的频率很高，但通过引力波观测直接探测仍然只是未来的希望。寻找它们的另一种途径是通过光子观察它们对附近气体的影响，但对于可能产生哪种光，我们知之甚少。为了解决这个问题，本项目对这些事件的电磁输出进行了系统的理论研究，从双黑洞的前体阶段，到事件本身，包括合并后的弛豫阶段。因为辐射光的数量和特征在根本上取决于有多少质量被带到了离黑洞给定的距离，所以系统的初始演化（当二元分离很大且牛顿动力学适用时）和合并本身（当广义相对论控制系统时）之间的协调是计划的重要组成部分。通过对这些系统进行传统的望远镜观测，其目的是创建一个全新的子领域——黑洞合并天文学。这些科学目标只能通过构建高度复杂的计算系统来实现。高性能计算是必不可少的，因为跟踪这些事件的动力学涉及同时求解广义相对论时空中的辐射输运和磁流体动力学方程，以及高度非线性的爱因斯坦场方程。数据通信需求也极具挑战性，因为即使定义黑洞合并的动态时空也可能需要几个pb。由于需要多种专业知识：数值相对论、MHD模拟、数据压缩、3-d可视化和观测天文学，因此这项工作具有很强的多学科性。该项目将产生重大的额外影响，从在一个全新的跨学科领域（动态时空中的流体运动）培养研究生和博士后到各种公共宣传工作。初级研究人员将需要学习先进的计算技术，包括并行计算、使用非常大的数据集和科学可视化。公共教育工作将包括传统的演示，在多媒体教育计划中使用模拟动画，并将模拟动画整合到音乐和舞蹈中。必要的计算发展，如计划中的积极数据压缩程序，将对探索更大动力系统的其他领域的科学家有很大的价值。

Super-massive black hole mergers are believed to happen with significant frequency, but direct detection through gravitational wave observations remains only a hope for the future. An alternate path toward finding them lies through photon observations of their effects on nearby gas, but little is known for certain about what sorts of light might be produced. To solve this problem, this project is a systematic theoretical study of the electromagnetic output from these events, from the precursor stage of binary black holes, through the event itself, and including the post-merger relaxation phase. Because the quantity and character of the light radiated depends in a fundamental way on how much mass is brought to a given distance from the black holes, coordination between the initial evolution of the system, when the binary separation is large and Newtonian dynamics applies, and the merger itself, when general relativity controls the system, is an essential part of the plan. By enabling conventional telescopic observation of these systems, the aim is to create an entirely new subfield - the astronomy of black hole mergers.These science goals can be reached only by the construction of a highly-sophisticated computational system. High-performance computation is essential because tracking the dynamics of these events involves simultaneously solving the equations of radiation transport and magneto-hydrodynamics (MHD) in a general relativistic spacetime, as well as the highly nonlinear Einstein Field Equations. Data communication requirements are also extremely challenging because even defining the dynamical spacetime of a black hole merger can require as much as several petabytes. This effort is therefore strongly multidisciplinary because of the multiple kinds of expertise needed: numerical relativity, MHD simulation, data compression, 3-d visualization, and observational astronomy.This project will have significant additional impact, ranging from training graduate students and post-doctoral fellows in a brand-new interdisciplinary field (fluid motion in dynamical spacetime) to a variety of public outreach efforts. The junior researchers will need to learn advanced computational techniques, including parallel computing, the use of very large data sets, and scientific visualization. Public education efforts will include conventional presentations, the use of simulation animations in multi-media educational programs, and integration of simulation animations into music and dance. The necessary computational developments, such as the planned aggressive data compression procedure, will be of great value to scientists in other fields exploring ever larger dynamical systems.