The Inversion Method of Surface-Wave Frequency Dispersion Curve Based on Neural Network

2009 ◽  
Author(s):  
Zhang Jin ◽  
Liu Huaishan ◽  
Meng Lin ◽  
He Yi
Keyword(s):  
Neural Network ◽  
Surface Wave ◽  
Dispersion Curve ◽  
Frequency Dispersion ◽  
Wave Frequency ◽  
Inversion Method

scholarly journals Imaging the subsurface using induced seismicity and ambient noise: 3-D tomographic Monte Carlo joint inversion of earthquake body wave traveltimes and surface wave dispersion

2020 ◽  
Vol 222 (3) ◽  
pp. 1639-1655
Author(s):  
Xin Zhang ◽  
Corinna Roy ◽  
Andrew Curtis ◽  
Andy Nowacki ◽  
Brian Baptie
Keyword(s):  
Surface Wave ◽  
Ambient Noise ◽  
Velocity Structure ◽  
Joint Inversion ◽  
Body Wave ◽  
Source Parameters ◽  
Wave Dispersion ◽  
Inversion Method ◽  
Surface Wave Dispersion ◽  
Wave Data

SUMMARY Seismic body wave traveltime tomography and surface wave dispersion tomography have been used widely to characterize earthquakes and to study the subsurface structure of the Earth. Since these types of problem are often significantly non-linear and have non-unique solutions, Markov chain Monte Carlo methods have been used to find probabilistic solutions. Body and surface wave data are usually inverted separately to produce independent velocity models. However, body wave tomography is generally sensitive to structure around the subvolume in which earthquakes occur and produces limited resolution in the shallower Earth, whereas surface wave tomography is often sensitive to shallower structure. To better estimate subsurface properties, we therefore jointly invert for the seismic velocity structure and earthquake locations using body and surface wave data simultaneously. We apply the new joint inversion method to a mining site in the United Kingdom at which induced seismicity occurred and was recorded on a small local network of stations, and where ambient noise recordings are available from the same stations. The ambient noise is processed to obtain inter-receiver surface wave dispersion measurements which are inverted jointly with body wave arrival times from local earthquakes. The results show that by using both types of data, the earthquake source parameters and the velocity structure can be better constrained than in independent inversions. To further understand and interpret the results, we conduct synthetic tests to compare the results from body wave inversion and joint inversion. The results show that trade-offs between source parameters and velocities appear to bias results if only body wave data are used, but this issue is largely resolved by using the joint inversion method. Thus the use of ambient seismic noise and our fully non-linear inversion provides a valuable, improved method to image the subsurface velocity and seismicity.

A neural network-based inversion method of a feedback linearization controller applied to a hydraulic actuator

2021 ◽  
Vol 43 (5) ◽  
Author(s):  
Fábio Augusto Pires Borges ◽  
Eduardo André Perondi ◽  
Mauro André Barbosa Cunha ◽  
Mario Roland Sobczyk
Keyword(s):  
Neural Network ◽  
Feedback Linearization ◽  
Inversion Method ◽  
Hydraulic Actuator

Surface Wave Dispersion Imaging Using Improved τ-p Transform Approach

2013 ◽  
Vol 353-356 ◽  
pp. 1196-1202 ◽  
Author(s):  
Jian Qi Lu ◽  
Shan You Li ◽  
Wei Li
Keyword(s):  
Surface Wave ◽  
Phase Shift ◽  
Dispersion Curve ◽  
Wave Dispersion ◽  
S Wave ◽  
Surface Wave Dispersion ◽  
Narrow Frequency Range ◽  
Channel Analysis ◽  
Rayleigh Wave Dispersion

Surface wave dispersion imaging approach is crucial for multi-channel analysis of surface wave (MASW). Because the resolution of inversed S-wave velocity and thickness of a layer are directly subjected to the resolution of imaged dispersion curve. The τ-p transform approach is an efficient and commonly used approach for Rayleigh wave dispersion curve imaging. However, the conventional τ-p transform approach was severely affected by waves amplitude. So, the energy peaks of f-v spectrum were mainly gathered in a narrow frequency range. In order to remedy this shortage, an improved τ-p transform approach was proposed by this paper. Comparison has been made between phase shift and improved τ-p transform approaches using both synthetic and in situ tested data. Result shows that the dispersion image transformed from proposed approach is superior to that either from conventionally τ-p transform or from phase shift approaches.

Joint inversion of two-dimensional magnetotelluric and surface wave dispersion data with cross-gradient constraints

2020 ◽  
Vol 221 (2) ◽  
pp. 938-950
Author(s):  
Pingping Wu ◽  
Handong Tan ◽  
Changhong Lin ◽  
Miao Peng ◽  
Huan Ma ◽  
...  
Keyword(s):  
Surface Wave ◽  
Velocity Structure ◽  
Joint Inversion ◽  
Wave Dispersion ◽  
Uniqueness Problem ◽  
Inversion Method ◽  
Inversion Algorithm ◽  
Surface Wave Dispersion ◽  
Gradient Constraints ◽  
Dispersion Data

SUMMARY Multiphysics imaging for data inversion is of growing importance in many branches of science and engineering. Cross-gradient constraint has been considered as a feasible way to reduce the non-uniqueness problem inherent in inversion process by finding geometrically consistent images from multigeophysical data. Based on OCCAM inversion algorithm, a direct inversion method of 2-D profile velocity structure with surface wave dispersion data is proposed. Then we jointly invert the profiles of magnetotelluric and surface wave dispersion data with cross-gradient constraints. Three synthetic models, including block homogeneous or heterogeneous models with consistent or inconsistent discontinuities in velocity and resistivity, are presented to gauge the performance of the joint inversion scheme. We find that owning to the complementary advantages of the two geophysical data sets, the models recovered with structure coupling constraints exhibit higher resolution in the classification of complex geologic units and settle some imaging problems caused by the separate inversion methods. Finally, a realistic velocity model from the NE Tibetan Plateau and its corresponding resistivity model calculated by empirical law are used to test the effectiveness of the joint inversion scheme in the real geological environment.

WAVEFORM INVERSION AND MULTI-LAYER NEURAL NETWORK

1995 ◽  
Vol 03 (03) ◽  
pp. 175-202 ◽  
Author(s):  
DONGYU FEI ◽  
JOHN T. KUO ◽  
YU-CHIUNG TENG
Keyword(s):  
Neural Network ◽  
Continuous Function ◽  
Mathematical Problem ◽  
Initial Guess ◽  
Acoustic Velocity ◽  
Inversion Method ◽  
True Parameter ◽  
Seismic Waveform ◽  
Neural Network Inversion ◽  
The Neural Network

This paper presents the concept of neural network inversion. The basic mathematical problem is to establish the mapping between two multi-dimensional spaces, addressing the continuous function mapping between two close intervals. It introduces the 3-layer neural network existence theorem and proves that the multi-layer neural network can approximate any continuous function in the sense of supernorm or mean-squares-norm, provided that the activation function is locally Riemann integrable and nonpolynomial. With an initial guess of the target parameter, which, in the present case, is acoustic velocity, in a prescribed sphere, which contains the true parameter, the neural network inversion method ensures the search reaching the global minima. The principle of the neural network inversion on the basis of the least-squares minimization (L2 norm) is developed. As its application, this method is employed to perform seismic waveform inversions — Model 1, for a homogeneous isotropic earth with a 2-D rectangle embedded body, and Model 2, for a layered earth with an elliptically elongated inclusion. A fast computation algorithm of the finite element method is adopted to generate a series of synthetic shot records for training the 3-layer neural network. The trained neural network possesses the capability to find the acoustic velocity of the embedded body in both Model 1 and 2 with a real-time solution within a sufficient accuracy.

Ü-Net: Deep-Learning Schemes for Ground Penetrating Radar Data Inversion

2020 ◽  
Vol 25 (2) ◽  
pp. 287-292
Author(s):  
Longhao Xie ◽  
Qing Zhao ◽  
Chunguang Ma ◽  
Binbin Liao ◽  
Jianjian Huo
Keyword(s):  
Neural Network ◽  
Dielectric Constant ◽  
Data Compression ◽  
Ground Penetrating Radar ◽  
Quantitative Imaging ◽  
Inversion Method ◽  
Output Unit ◽  
Data Inversion ◽  
Ground Penetrating ◽  
Constant Distribution

Electromagnetic (EM) inversion is a quantitative imaging technique that can describe the dielectric constant distribution of a target based on the EM signals scattered from it. In this paper, a novel deep neural network (DNN) based methodology for ground penetrating radar (GPR) data inversion, known as the Ü-net is introduced. The proposed Ü-net consists of three parts: a data compression unit, U-net, and an output unit. The novel inversion approach, based on supervised learning, uses a neural network to generate the dielectric constant distribution from GPR data. The GPR data can be compressed and reshaped the size using data compression unit. The U-net maps the object features to the dielectric constant distribution. The output unit meshes the dielectric constant distribution more finely. A novel feature of the proposed methodology is the application of instance normalization (IN) to the DNN EM inversion method and a comparison of its performance to batch normalization (BN). The validity of this technique is confirmed by numerical simulations. The Mean-Square Error of the test data sets is 0.087. These simulations prove that the instance normalization is suitable for GPR data inversion. The proposed approach is promising for achieving quality dielectric constant images in real-time.

Sign in / Sign up

Export Citation Format

Share Document