Seismic Inversion via Closed-Loop Fully Convolutional Residual Network and Transfer Learning

Geophysics ◽  
2021 ◽  
pp. 1-54
Author(s):  
Lingling Wang ◽  
Delin Meng ◽  
Bangyu Wu
Keyword(s):  
Deep Learning ◽  
Transfer Learning ◽  
Seismic Data ◽  
Closed Loop ◽  
Seismic Inversion ◽  
Open Loop ◽  
Residual Network ◽  
Velocity Inversion ◽  
Good Convergence ◽  
Loop Network

Because deep learning networks can 'learn' the complex mapping function between the labeled inputs and outputs, they have shown great potential in seismic inversion. Conventional deep learning algorithms require a large amount of labeled data for sufficient training. However, in practice, the number of well logs is limited. To address this problem, we propose a closed-loop fully convolutional residual network (FCRN) combined with transfer learning strategy for seismic inversion. This closed-loop FCRN consists of an inverse network and a forward network. The inverse network predicts the inversion target from seismic data, whereas the forward network calculates seismic data from the inversion target. The inverse network is initialized by pre-training on the Marmousi2 model and fine-tuned with the limited labeled data around the wells through transfer learning, to suit the target seismic data. The forward network is initialized by training with the limited labeled data around the wells. In this way, the closed-loop network is well initialized to ensure relatively good convergence. Then, the misfit of the limited labeled data and the error between the true and the forward seismic data are used to regularize the training of the initialized closed-loop network. The inverse network of the optimized closed-loop network is used to obtain the final inversion results. The proposed work flow can be used for velocity, density, and impedance inversion from post-stack seismic data. This paper takes velocity inversion as an example to illustrate the effectiveness of the method. The experimental results show that the closed-loop FCRN with transfer learning is superior than the open-loop FCRN with better lateral continuity and velocity details. The closed-loop FCRN can effectively predict the velocity with high accuracy on the synthetic data, has good anti-noise performance, and also can be effectively used for the field data with spatial heterogeneity.

Robust deep learning seismic inversion with a priori initial model constraint

2021 ◽  
Author(s):  
Jian Zhang ◽  
Jingye Li ◽  
Xiaohong Chen ◽  
Yuanqiang Li ◽  
Guangtan Huang ◽  
...  
Keyword(s):  
Deep Learning ◽  
Seismic Data ◽  
A Priori ◽  
Seismic Inversion ◽  
Training Dataset ◽  
Initial Model ◽  
Data Set ◽  
The Real ◽  
Model Constraint ◽  
Trained Network

Summary Seismic inversion is one of the most commonly used methods in the oil and gas industry for reservoir characterization from observed seismic data. Deep learning (DL) is emerging as a data-driven approach that can effectively solve the inverse problem. However, existing deep learning-based methods for seismic inversion utilize only seismic data as input, which often leads to poor stability of the inversion results. Besides, it has always been challenging to train a robust network since the real survey has limited labeled data pairs. To partially overcome these issues, we develop a neural network framework with a priori initial model constraint to perform seismic inversion. Our network uses two parts as one input for training. One is the seismic data, and the other is the subsurface background model. The labels for each input are the actual model. The proposed method is performed by log-to-log strategy. The training dataset is firstly generated based on forward modeling. The network is then pre-trained using the synthetic training dataset, which is further validated using synthetic data that has not been used in the training step. After obtaining the pre-trained network, we introduce the transfer learning strategy to fine-tune the pre-trained network using labeled data pairs from a real survey to acquire better inversion results in the real survey. The validity of the proposed framework is demonstrated using synthetic 2D data including both post-stack and pre-stack examples, as well as a real 3D post-stack seismic data set from the western Canadian sedimentary basin.

High resolution fixed-point seismic inversion

2021 ◽  
pp. 1-54
Author(s):  
Song Pei ◽  
Xingyao Yin ◽  
Zhaoyun Zong ◽  
Kun Li
Keyword(s):  
Fixed Point ◽  
Objective Function ◽  
Seismic Data ◽  
Target Function ◽  
Recursive Formula ◽  
Frequency Component ◽  
Seismic Inversion ◽  
Inversion Method ◽  
Good Convergence ◽  
Band Limited

Resolution improvement always presents the crucial task in geological inversion. Band-limited characteristics of seismic data and noise make seismic inversion complicated. Specifically, geological inversion suffers from the deficiency of both low- and high-frequency components. We propose the fixed-point seismic inversion method to alleviate these issues. The problem of solving objective function is transformed into the problem of finding the fixed-point of objective function. Concretely, a recursive formula between seismic signal and reflection coefficient is established, which is characterized by good convergence and verified by model examples. The error between the model value and the inverted value is reduced to around zero after few iterations. The model examples show that in either case, that is, the seismic traces are noise-free or with a little noise, the model value can almost be duplicated. Even if the seismic trace is accompanied by the moderate noise, the optimal inverted results can still be obtained with the proposed method. The initial model constraint is further introduced into the objective function to increase the low-frequency component of the inverted results by adding prior information into the target function. The singular value decomposition (SVD) method is applied to the inversion framework, thus making a high improvement of anti-noise ability. At last, the synthetic models and seismic data are investigated following the proposed method. The inverted results obtained from the fixed-point seismic inversion are compared with those obtained from the conventional seismic inversion, and it is found that the former has a higher resolution than the latter.

scholarly journals Simulasi Modifikasi Konfigurasi Jaringan Spindel Menjadi Loop Tertutup Untuk Meningkatkan Keandalan Sistem Distribusi Menggunakan Software ETAP

2021 ◽  
Vol 13 (1) ◽  
pp. 33-39
Author(s):  
Eko Nio Rizki
Keyword(s):  
Closed Loop ◽  
Network Reliability ◽  
Short Circuit ◽  
Circuit Breaker ◽  
Open Loop ◽  
Network Configuration ◽  
Ground Fault ◽  
Overcurrent Relay ◽  
Level 3 ◽  
Loop Network

Distribution reliability network rely on to any factor such as material quality, maintenance, operational pattern, protection device and also network configuration. In the spindle network the level of network reliability is level 3 (SPLN  52-3, 1983: 5). In order to leveling up the network reliability from level 3 to level 5 (zero down time)[2][3] we need to modify the protection system from overcurrent relay and ground fault relay to line differential relay in each distribution substation. Beside that Load Break Switch in each customer cubicle substation and in the connection substation should replaced by circuit breaker. Spindle network which operated open loop in the connection substation switch to normally close operated, so it can be called as closed loop network. This modification purpose is ther is no down time in case off ground fault or phase to phase sort circuit on the network cable. Before this kind of modification and the setting applied into real network, we make a simulation using an application called ETAP and no missmatch trip from 7 time experiment  consist of ground fault and phase to phase short circuit in 7  cable

scholarly journals Fault Diagnosis under Variable Working Conditions Based on STFT and Transfer Deep Residual Network

2020 ◽  
Vol 2020 ◽  
pp. 1-18 ◽  
Author(s):  
Yan Du ◽  
Aiming Wang ◽  
Shuai Wang ◽  
Baomei He ◽  
Guoying Meng
Keyword(s):  
Deep Learning ◽  
Fault Diagnosis ◽  
Transfer Learning ◽  
Working Conditions ◽  
Classification Accuracy ◽  
Working Condition ◽  
Learning Ability ◽  
Residual Network ◽  
Time Frequency ◽  
Residual Learning

Fault diagnosis plays a very important role in ensuring the safe and reliable operations of machines. Currently, the deep learning-based fault diagnosis is attracting increasing attention. However, fault diagnosis under variable working conditions has been a significant challenge due to the domain discrepancy problem. This problem is also unavoidable in deep learning-based fault diagnosis methods. This paper contributes to the ongoing investigation by proposing a new approach for the fault diagnosis under variable working conditions based on STFT and transfer deep residual network (TDRN). The STFT was employed to convert vibration signal to time-frequency image as the input of the TDRN. To address the domain discrepancy problem, the TDRN was developed in this paper. Unlike traditional deep convolutional neural network (DCNN) methods, by combining with transfer learning, the TDRN can make a bridge between two different working conditions, thereby using the knowledge learned from a working condition to achieve a high classification accuracy in another working condition. Moreover, since the residual learning is introducing, the TDRN can overcome the problems of training difficulty and performance degradation existing in traditional DCNN methods, thus further improving the classification accuracy. Experiments are conducted on the popular CWRU bearing dataset to validate the effectiveness and superiority of the proposed approach. The results show that the developed TDRN outperforms those methods without transfer learning and/or residual learning in terms of the accuracy and feature learning ability for the fault diagnosis under variable working conditions.

Image Dedusting with Deep Learning based Closed-Loop Network

2021 ◽  
Author(s):  
Xibin Song ◽  
Dingfu Zhou ◽  
Jin Fang ◽  
Liangjun Zhang
Keyword(s):  
Deep Learning ◽  
Closed Loop ◽  
Loop Network

Physics-guided deep learning for seismic inversion with hybrid training and uncertainty analysis

Geophysics ◽  
2021 ◽  
pp. 1-64
Author(s):  
Jian Sun ◽  
Kristopher A. Innanen ◽  
Chao Huang
Keyword(s):  
Deep Learning ◽  
Elastic Property ◽  
Seismic Data ◽  
Degrees Of Freedom ◽  
Seismic Inversion ◽  
Training Data ◽  
Data Driven ◽  
Model Data ◽  
Data Set ◽  
Propagation Network

The determination of subsurface elastic property models is crucial in quantitative seismic data processing and interpretation. This problem is commonly solved by deterministic physical methods, such as tomography or full-waveform inversion. However, these methods are entirely local and require accurate initial models. Deep learning represents a plausible class of methods for seismic inversion, which may avoid some of the issues of purely descent-based approaches. However, any generic deep learning network capable of relating each elastic property cell value to each sample in a seismic data set would require a very large number of degrees of freedom. Two approaches might be taken to train such a network: first, by invoking a massive and exhaustive training data set and, second, by working to reduce the degrees of freedom by enforcing physical constraints on the model-data relationship. The second approach is referred to as “physics-guiding.” Based on recent progress in wave theory-designed (i.e., physics-based) networks, we have developed a hybrid network design, involving deterministic, physics-based modeling and data-driven deep learning components. From an optimization standpoint, a data-driven model misfit (i.e., standard deep learning) and now a physics-guided data residual (i.e., a wave propagation network) are simultaneously minimized during the training of the network. An experiment is carried out to analyze the trade-off between two types of losses. Synthetic velocity building is used to examine the potential of hybrid training. Comparisons demonstrate that, given the same training data set, the hybrid-trained network outperforms the traditional fully data-driven network. In addition, we performed a comprehensive error analysis to quantitatively compare the fully data-driven and hybrid physics-guided approaches. The network is applied to the SEG salt model data, and the uncertainty is analyzed, to further examine the benefits of hybrid training.

Perfect Multivariable Feedback Theory

1997 ◽  
Vol 07 (02) ◽  
pp. 129-151
Author(s):  
M.-D. Tong ◽  
W.-K. Chen
Keyword(s):  
Closed Loop ◽  
Continuous System ◽  
Open Loop ◽  
Feedback Network ◽  
Loop Network ◽  
The Stability ◽  
Return Difference ◽  
And Control ◽  
Z Transformation ◽  
Multivariable Feedback

The paper nicely combines the state-space description with the input-output description and elegantly formulates the multivariable feedback theory as well as obtains a number of useful results for modern network and control theory. In particular, it reveals various kinds of duality between a multivariable feedback network and its associated inverse network, such as structure duality, transfer function matrix (determinant) duality and duality on controllability (observability). It also thoroughly studies four pairs of the (null) return difference matrices of a multivariable feedback network and its associated inverse network, and presents not only the dual properties about these (null) return diferrence matrices and their respective determinants, but also treats the relationships among these determinants, the characteristic polynomials of a closed-loop network (the multivariable feedback network), its associated closed-loop inverse network, and their respective corresponding open-loop networks. Finally, the stability criteria, the testing criteria for a minimum-phase matrix and the sensitivity matrices are discussed. Although all of these results are obtained for a continuous system, they are also suitable for a discrete system provided that we use the z-transformation instead of the Laplace transformation.

scholarly journals Seismic inversion with deep learning

2021 ◽  
Author(s):  
Silvia L. Pintea ◽  
Siddharth Sharma ◽  
Femke C. Vossepoel ◽  
Jan C. van Gemert ◽  
Marco Loog ◽  
...  
Keyword(s):  
Deep Learning ◽  
Radon Transform ◽  
Network Optimization ◽  
Seismic Data ◽  
Seismic Inversion ◽  
Single Step ◽  
Background Model ◽  
Learning Methods ◽  
Type Classes ◽  
Type Classification

AbstractThis article investigates bypassing the inversion steps involved in a standard litho-type classification pipeline and performing the litho-type classification directly from imaged seismic data. We consider a set of deep learning methods that map the seismic data directly into litho-type classes, trained on two variants of synthetic seismic data: (i) one in which we image the seismic data using a local Radon transform to obtain angle gathers, (ii) and another in which we start from the subsurface-offset gathers, based on correlations over the seismic data. Our results indicate that this single-step approach provides a faster alternative to the established pipeline while being convincingly accurate. We observe that adding the background model as input to the deep network optimization is essential in correctly categorizing litho-types. Also, starting from the angle gathers obtained by imaging in the Radon domain is more informative than using the subsurface offset gathers as input.

On closed-loop equilibrium strategies for mean-field stochastic linear quadratic problems

2020 ◽  
Vol 26 ◽  
pp. 41
Author(s):  
Tianxiao Wang
Keyword(s):  
Optimal Control ◽  
Closed Loop ◽  
Optimal Control Problems ◽  
Mean Field ◽  
Open Loop ◽  
Time Inconsistency ◽  
Control Problems ◽  
Linear Quadratic ◽  
Linear Quadratic Optimal Control ◽  
Equilibrium Strategies

This article is concerned with linear quadratic optimal control problems of mean-field stochastic differential equations (MF-SDE) with deterministic coefficients. To treat the time inconsistency of the optimal control problems, linear closed-loop equilibrium strategies are introduced and characterized by variational approach. Our developed methodology drops the delicate convergence procedures in Yong [Trans. Amer. Math. Soc. 369 (2017) 5467–5523]. When the MF-SDE reduces to SDE, our Riccati system coincides with the analogue in Yong [Trans. Amer. Math. Soc. 369 (2017) 5467–5523]. However, these two systems are in general different from each other due to the conditional mean-field terms in the MF-SDE. Eventually, the comparisons with pre-committed optimal strategies, open-loop equilibrium strategies are given in details.

Sign in / Sign up

Export Citation Format

Share Document