Interferometric electromagnetic Green's functions representations using propagation invariants
Interferometric electromagnetic Green's functions representations using propagation invariants
复制标题
DOI：
10.1111/j.1365-246x.2006.03296.x
复制
发表时间：
2007-04
影响因子：
2.8
通讯作者：
E. Slob;D. Draganov;K. Wapenaar
中科院分区：
文献类型：
--
作者：
E. Slob;D. Draganov;K. Wapenaar
摘要 根据在不同位置测量的响应的互相关创建新的响应被称为干涉测量。每个新创建的响应代表在一个接收器位置测量的场，就好像另一个位置有源一样。在这里，我们在开放配置中制定了电磁干涉格林函数表示。域内或域外介质的异质性和各向异性原则上没有限制。时间相关型公式依赖于总波能守恒，并且它们不能用于以直接方式显示某种形式的弛豫的介质。时间卷积型传播不变量与介质弛豫机制无关，并且可以通过将测量的响应与另一个响应的时间反演互相关来将它们用于干涉测量。这种类型的干涉测量只能在域外具有一个接收器的配置中制定。对于时间卷积干涉测量来说，对介质异质性、各向异性或弛豫机制没有限制。为了使这些干涉公式具有实际用途，主要的简化是对源坐标中的法向导数进行高频近似。这些精确结果的近似值会导致两种不同类型的错误。我们讨论这些错误的原因和后果，并用数值例子进行说明。
SUMMARY Creating new responses from cross-correlations of responses measured at different locations is known as interferometry. Each newly created response represents the field measured at one of the receiver locations as if there were a source at the other. Here, we formulate electromagnetic interferometric Green's functions representations in open configurations. There are in principle no restrictions on the heterogeneity and anisotropy of the medium inside or outside the domain. Time-correlation type formulations rely on conservation of total wave energy and they cannot be used for media showing relaxation of some form in a straightforward way. Time-convolution type propagation invariants are independent of the medium relaxation mechanisms and they can be used for interferometry by cross-correlating a measured response with the time-reverse of another response. This type of interferometry can only be formulated in the configuration with one receiver outside the domain. For time-convolution interferometry no restrictions on the medium heterogeneity, anisotropy or relaxation mechanisms are made. For these interferometric formulations to be of practical use, the main simplification is to make a high-frequency approximation for the normal derivative in the source coordinate. These approximations of the exact result lead to two different types of errors. We discuss the causes and consequences of these errors and illustrate them with numerical examples.