43 results about "Perfectly matched layer" patented technology

The invention provides a high-precision spatial and temporal arbitrary multiple variable grid finite difference forward modeling method, which comprises the steps of: step 1, establishing a forward modeling speed model of an underground medium; step 2, subjecting the forward modeling speed model to two-dimensional grid discretization, and subjecting an acoustic wave field in the forward modeling speed model to two-dimensional grid discretization, wherein the acoustic wave field is located on grid nodes; step 3, subjecting a perfectly matched layer boundary condition to grid discretization; and step 4, conducting time-domain finite difference forward modeling simulation through an acoustic wave equation, wherein time sampling step lengths are variable step lengths at grids of different sizes. The high-precision spatial and temporal arbitrary multiple variable grid finite difference forward modeling method can effectively suppress forward modeling frequency dispersion, significantly increases signal-to-noise ratio, and greatly improves precision and efficiency of forward modeling simulation.

The invention discloses a simulation method of ultra wideband electromagnetic waves. The invention proposes a new PML method on the basis of the PML method under a stretch coordinate system and a difference scheme of a higher-order time domain finite difference method (FDTD (2, 2M)). A piecewise linearity convolutional perfectly matched layer (PCPML, Piecewise linearity CPML) under self-adapting absorption parameters absorbs a boundary condition. PML absorption factors are set to be adjustable parameters by the method so as to absorb different ultra wideband electromagnetic waves from high frequency to low frequency, and the parameters are obtained by a mixed method of an optimization method of a real coding genetic algorithm and a linear iteration method in a self-adapting manner so as to adapt to the absorption of ultra wideband electromagnetic waves. The absorption boundary condition adopted by the method can absorb and process high frequency waves and low frequency waves, and also can reduce the reflection error of a matched layer caused by grid dispersion and edge reflection and improve the absorption efficiency and the accuracy of simulation result.

A frequency domain optimization mixture staggered grid finite difference forward modeling method comprises the following steps: 1) providing a time-domain two-dimension sound wave equation; 2) eliminating artificial boundary reflection and obtaining a time-domain two-dimension sound wave equation with a perfectly-matched layer boundary condition; 3) performing Fourier transform on time variables at the two sides of the equation and obtaining a frequency-domain sound wave equation; 4) carrying out finite difference discretization on the frequency-domain sound wave equation with the perfectly-matched layer boundary condition based on a conventional staggered grid and obtaining a finite difference discretization format; 5) carrying out finite difference discretization on the frequency-domain sound wave equation with the perfectly-matched layer boundary condition based on a rotation staggered grid and obtaining a finite difference discretization format; 6) performing optimization mixing on the conventional staggered grid and the rotation staggered grid, grid difference item being weighted average of the grid difference item in the two grid systems, and quality acceleration item being weighted average of a center point and eight points around the center point; and 7) under the criterion of the minimum phase velocity error, calculating an optimized coefficient. The weighting coefficient enables frequency dispersion error due to the finite difference discretization to be the smallest, and the precision of the frequency-domain forward modeling is greatly improved.

The invention discloses an implementation method for allowing two-dimension cylindrical coordinates to completely absorb boundary in a matching manner. The method comprises the specific steps of inputting a model file, initializing a magnetic field, an electric field and related auxiliary variables to be zero, calculating a one-dimension coefficient array in a PML (perfectly matched layer), updating and calculating the electric displacement vector in a whole calculation area, updating and calculating the electric field of the whole calculation area, updating and calculating a source of the electric field, updating and calculating the auxiliary variables of the electric field in the PML area, updating and calculating the magnetic induction in the whole calculation area, updating and calculating the magnetic field of the whole calculation area, updating and calculating the auxiliary variables of the magnetic field in the PML area, judging whether to continue to update according to the set simulation calculation time, judging to continue to carry out calculation and returning to the step 3 if the actual calculation time is less than the set value, and otherwise, outputting the calculated electric and magnetic fields to completion. The implementation method can be combined with CFS (complex frequency shift) parameters to more effectively absorb electromagnetic waves spreading outwards.

The invention provides an implementation algorithm for truncating a one dimensional Debye medium Crank-Nicolson perfectly matched layer, belongs to the technical field of numerical simulation, and aims to truncate a Debye dispersive medium by using the perfectly matched layer and simulate a limited memory space of a computer into an infinite space to simulate the propagation characteristic of electromagnetic wave in the Debye dispersive medium. The algorithm is characterized by comprising the following steps: a plurality of stretch coordinate variables are converted from a frequency domain to a z domain by using a bilinear transformation method, then a maxwell equation is dispersed in the time domain by using a Crank-Nicolson finite difference time domain method, an explicit iterative equation of an electric field is induced, and finally a value of an electromagnetic field component is solved. The algorithm has the advantages of unconditional stability and capabilities of improving the electromagnetic field calculation speed and saving memory.

The invention provides a PML boundary condition absorbing method based on a second-order partial differential wave equation. The method includes the first step of loading a longitudinal wave field of an earthquake source at a certain sampling moment, the second step of calculating a second-order partial derivative of the longitudinal wave field in a certain direction of a three-dimensional space at the sampling moment, and carrying out attenuation by absorption of a time first-order derivative wave field of the longitudinal wave field in a perfectly matched layer area in the direction according to the second-order partial derivative in the direction, the third step of repeating the second step, and carrying out attenuation by absorption of time first-order derivative wave fields of the longitudinal wave field in perfectly matched layer areas in the other two directions of the three-dimensional space at the sampling moment respectively, and the fourth step of repeating the preceding steps until processing of the longitudinal wave fields at all sampling moments is finished. By the adoption of the method, achieving complexity of boundary condition absorbing on a perfectly matched layer is low, and the calculation storage amount is small.

The present invention provides a PML (Perfectly Matched Layer) boundary three-dimensional seismic wave propagation simulation method utilizing CUDA (Compute Unified Device Architecture), comprising the steps of: reading in an initial parameter and a Ricker wavelet; decomposing a stress in a first acoustic wave fluctuation equation of a three-dimensional seismic wave into stress components in three directions which are mutually perpendicular, and adding a PML boundary condition to form a second acoustic wave fluctuation equation; using a finite difference method to discretize the second acoustic wave fluctuation equation, so as to form a finite difference acoustic wave fluctuation equation; bordering a first three-dimensional speed model to form a second three-dimensional speed model; calculating attenuation coefficients of edges, angles and surfaces of the second three-dimensional speed model in a CPU (Central Processing Unit) according to the finite difference acoustic wave fluctuation equation; calculating stresses of net points of the second three-dimensional speed model in a GPU (Graphics Processing Unit) according to the attenuation coefficients; and outputting a three-dimensional seismic wave simulation record at the last moment according to the stresses. The PML boundary three-dimensional seismic wave propagation simulation method of the present invention can achieve high speed-up ratio and shorten simulation time.

The invention discloses a D type photonic crystal fiber refractive index sensor using double-loss-peak detection. The sensor comprises a base material, cladding air holes, a metal film, to-be-detectedanalyte liquid and a perfectly matching layer. The base material is made of a quartz material. The cladding air holes include 25 air holes arranged vertically in parallel in five layers and one largeair hole arranged below the center of the fiber. According to the invention, a gold film is deposited on a polished surface of the D type fiber and is used as an inducing material for surface plasmonresonance effect generation. A liquid analyte is arranged at the outermost layer. When the sensor is used, the fiber is required to be immersed into the to-be-detected analyte liquid completely. According to the specific embodiment, the sensor not only can use two loss peaks to detect the to-be-detected analyte but also can select different types of spectrum detectors flexibly based on specific experimental conditions and application requirements or select one single peak randomly from the two loss peaks for detection.

The invention discloses a method for realizing a perfectly matched absorbing boundary of expanding coordinates in a plasma. The method comprises the following steps of inputting a model file; initializing parameters and setting a PML (perfectly matched layer) coefficient and an absorbing boundary parameter; updating and calculating the magnetic field component coefficient H<q>(z) of the whole calculation area, adding a field source to an electric field component coefficient E<q>(y), updating and calculating the electric field component coefficient E<q>(y) of the whole calculation area, updating and calculating the electric field component coefficient E<q>(x) of the whole calculation area, updating and calculating the electronic average speeds U<q>(ex), U<q>(ey) of the whole calculation area, and updating and calculating the auxiliary variables of the electric and magnetic field component coefficients of the whole calculation area; calculating electric and magnetic field components at an observing point; assigning q+1 to q, judging whether q reaches a preset value or not, and if q does not reach the preset value, returning to step 3; if q reaches the preset value, ending. According to the method for realizing the perfectly matched absorbing boundary of the expanding coordinates in the plasma, due to the fact that the absorbing boundary can be very conveniently combined with parameters in CFS factors, the method has better absorbing effect on low frequencies and evanescent waves.

The invention relates to a finite difference calculation method based on the quasi-spatial domain acoustic wave equation and belongs to the field of seismic exploration. The main idea of the method isto transform a conventional acoustic wave equation into a quasi-spatial domain acoustic wave equation so that the equally spaced 'distance' grid length steps in the spatial domain is transformed intonon-equally spaced 'propagation time' length steps. Thus, a user can accurately calculate the 'propagation time' on both sides of a velocity interface based on a strictly defined velocity model. Then, the equation of a 2N-order (N is a positive integer) precision finite difference expression for its perfectly matched layer boundary condition is given based on this. Accordingly, the finite difference wavefield extension of seismic waves in the process of reverse time migration can be realized. The method can well solve the problem of the shape distortion of the velocity interface in the reverse-time migration profile of a conventional acoustic wave equation. In addition, the wave field extension based on the quasi-spatial domain acoustic wave equation can significantly weaken the interfacepseudo-scattering and the inter-layer reflection wave so as to further improves the quality of offset imaging.

The invention discloses a reverse time migration method of a carbonatite reservoir. By aiming at a phenomenon that a fracture-cavity pore construction of carbonatite is always companied with fracture and big-angle form distortion formed due to construction movement, the reverse time migration method is used for precisely depicting the carbonatite. A finite difference method is used for solving a wave equation to simulate a seismic wave field; a processing method for the wave field to transmit to a boundary adopts a PML (Perfectly Matched Layer) absorbing boundary condition; a seismic origin wave field recovery part adopts a reverse PML wave field recovery technology; and a calculation part adopts a GPU (Graphics Processing Unit) parallel acceleration technology. The PML absorbing boundary condition is adopted, so that the invention can more effectively eliminate boundary reflection interference than a random boundary condition; since the reverse PML wave field recovery technology is adopted, compared with a traditional wave field storage strategy, the invention can greatly reduce requirements on a storage amount on the premise that a calculated amount is not increased, a wave field recovery effect is good, amplitude fading is avoided, and operation efficiency is improved by adopting the GPU parallel acceleration technology.

The invention relates to a piecewise linear cyclic convolution-based one-dimensional left-handed material Crank-Nicolson perfectly matched layer realizing algorithm, belongs to the technical field of numerical simulation, and aims at shortening the left-handed material FDTD computational domain and simulating the finite memory spaces of computers into infinite spaces. The algorithm is technically characterized by comprising the following steps: in the process of transforming plural stretching coordinate variables from a frequency domain to a time domain, eliminating the second-order differential in the stretching variables by utilizing a piecewise linear cyclic convolution method so as to remarkably decrease the number of imported auxiliary variables and optimize the memory; dispersing a time domain Maxwell equation by utilizing a Crank-Nicolson time domain finite difference method so as to derive an explicit iterative equation of an electric field; and finally solving the value of an electromagnetic field component. The algorithm has the advantages of providing unconditional stability, improving the electromagnetic field calculation speed and saving the memory.

The invention discloses a combined absorbing boundary condition applied to sound wave finite difference numerical simulation, and belongs to the field of seismic exploration numerical simulation. The combined absorbing boundary condition applied to the sound wave finite difference numerical simulation specifically comprises the following steps of: when sound wave equation numerical simulation is performed based on a 2N-order (N>0) accuracy staggered-mesh finite difference scheme, at first, setting L (L>N) perfectly matched layers (PML) at an artificial truncated boundary, and absorbing boundary reflected waves from a central wave field by using a PML boundary condition; and then, for N layers of boundaries outside the PML, by using an Higdon three-order absorbing boundary condition, absorbing outer boundary reflexes of the PML. According to the method provided by the present invention, by fully utilizing the advantages of both the PML boundary condition and the Higdon three-order absorbing boundary condition are fully utilized, the boundary reflexes of inner layers and outer layers of artificial boundaries can be effectively absorbed, thereby achieving high-accuracy finite difference numerical simulation.

The invention discloses an omnidirectional perfectly matched transparent material conforming to a uniaxial perfectly matched layer model. The transparent material comprises at least one triangular sub-wavelength unit. Each sub-wavelength unit comprises a first dielectric plate, a triangular metal copper ring and a second dielectric plate which are superposed in turn, and a cylinder is embedded into the two dielectric plates. By adjusting the physical dimensions of the transparent material, the equivalent relative dielectric constant and permeability of the transparent material meet the diag (a, a, 1/a) form. The equivalent wave impedance of the transparent material is perfectly matched with the free space, and the incident TE polarized electromagnetic wave can pass through the material completely without reflection. The transparent material is simple, thin and light, is perfectly matched with the free space, and can be widely used in electromagnetic compatibility, anti-electromagnetic-interference and other fields.

The invention relates to a one-dimension left-handed material Crank-Nicolson perfectly matched layer realizing algorithm based on an auxiliary differential equation, and belongs to the technical field of numerical simulation. The method aims at reducing the left-handed material FDTD (Finite-Different Time-Domain) computational domain, and simulating a computer finite memory space into an infinite space. The one-dimension left-handed material Crank-Nicolson perfectly matched layer realizing algorithm has the technical characteristics that in a process of converting a plurality of stretched coordinate variables from the frequency domain to the time domain, the second-order differential in the stretched variables is eliminated by an improved auxiliary differential equation method, so that the number of the introduced auxiliary variables is obviously lowered, and a memory is optimized; then, a time domain Maxwell equation is subjected to discretization by a Crank-Nicolson time domain finite differential method; an explicit iteration equation of an electric field is deduced out; and finally, a value of an electromagnetic field component is obtained. The one-dimension left-handed material Crank-Nicolson perfectly matched layer realizing algorithm has the advantages that the unconditional stability is realized; the electromagnetic calculation is accelerated; and the memory is saved.

The invention belongs to the technical field of computational electromagnetics, and particularly relates to a transparent excitation source implementation method applied to the finite-difference time-domain method. A Yee network is used, two electric field excitation sources are added to an excitation source grid, and a feedback electric field is reduced, so that the excitation sources are equivalent to hard sources and transparent. Perfectly matched layer absorbing boundary truncated waveguide structures are added in an input port and an output port to be suitable for first-dimensional, second-dimensional and third-dimensional structures. According to the method, extra computation needs to be conducted only on electromagnetic field components on the surfaces of the excitation sources, compared with other methods, the required auxiliary calculation amount is very small, and influences on calculation efficiency can be ignored. The method is transparent, time for auxiliary calculation is avoided for waveguide type calculation simulation problems, an incident field and a scattering field are efficiently obtained, and the method has a good effect on improving the simulation calculation efficiency.

The invention provides a method for obtaining a stable nearly perfectly matched layer absorption boundary condition. The method comprises the following steps of (1) converting an elastic wave first-order speed-stress equation of a time domain to a frequency domain; (2) performing coordinate conversion on the elastic wave first-order speed-stress equation of the frequency domain; (3) setting a parameter of a complex frequency shift stretching function in a CFS-NPML; (4) transferring the stretching function into a derived wave field function; (5) performing conversion on the complex frequency shift stretching function, performing variable substitution, and substituting the function into the stretched elastic wave first-order speed-stress equation of the frequency domain; (6) introducing an auxiliary wave field variable for performing variable substitution on the stretched elastic wave first-order speed-stress equation of the frequency domain, and converting the auxiliary variable to be in the form of a first-order ordinary differential equation; and (7) inversely converting the stretched elastic wave first-order speed-stress equation of the frequency domain and the auxiliary equationto the time domain, thereby obtaining an elastic wave first-order speed-stress equation and an auxiliary ordinary differential equation of the CFS-NPML of the time domain.

The invention discloses a Laguerre-polynomials-based perfectly matched layer implementing method with fewer splitting errors, comprising the specific steps of inputting a model file; initializing and setting parameters; adding a field source to electric field component coefficient in direction y, and using factor splitting WLP-FDTD (Weighted Laguerre Polynomials and finite difference time domain) method to calculate the electric field component coefficient denoted as initial field value; update-calculating electric field component coefficient in direction y in a whole calculation area; update-calculating electric field component coefficient in direction x in the whole calculation area; judging whether the number of iterations k reaches a preset value or not; update-calculating electromagnetic field component of an observation point; judging whether order q of Laguerre polynomials reaches a preset value or not. The Laguerre-polynomials-based perfectly matched layer implementing method with fewer splitting error has high calculating speed and high precision, and provides more effective absorption of low frequency and evanescent wave.

Described here are systems and methods for efficiently simulating wave propagation in one or more viscoelastic media. The systems and methods described here implement an elastodynamic model that includes memory variables for one or more viscoelastic properties. Boundary conditions can also be generated based on a perfectly matched layer (PML) condition. A hybrid implicit/explicit method is used to generate wave simulation data. In this method, equations related to the memory variables and PML are solved using an implicit integrator, and other equations in the elastodynamic model are solved using an explicit method.

The invention relates to an algorithm for realizing a perfectly matched layer by truncating plasmas based on a CNDG algorithm, and belongs to the technical field of numerical simulation. The algorithm is technically characterized in that a complex frequency offset stretch coordinate variable is substituted into a two-dimensional corrected Maxwell equation set, and an auxiliary differential equation method is used. In a process of deriving a relationship between dielectric flux density D and an electric field E, a piecewise linear recursive convolution method is applied. By applying a CNDG method, the electric field is solved by solving a three-diagonal equation, so that the calculation efficiency is improved and the time is saved. The algorithm has the advantages of unconditional stability and increased electromagnetic field calculation speed.

The invention relates to a two-dimensional Debye medium and Lorentz medium truncation Crank-Nicolson perfectly matched layer implementation algorithm, and belongs to the technical field of numerical simulation.The method aims at reducing a two-dimensional dispersive medium FDTD calculation domain and simulating limited memory space of a computer into infinite space.The method is technically characterized in that in the process of transforming a two-dimensional corrected Maxwell equation with multiple tension coordinate variables into finite difference time domain from a frequency domain, an auxiliary differential equation method and a Crank-Nicolson-Approximate-Decoupling (CNAD) algorithm are utilized, two tridiagonal matrix equations related to an electric field are obtained, and thus the calculation efficiency is obviously improved.The method has the advantages that unconditional stability is achieved, and the electromagnetic field calculation speed is increased.