Inversion of magnetic anomalies due to isolated thin dike-like sources using artificial neural networks

A new approach is proposed to interpret magnetic anomalies caused by isolated thin dike-like causative targets. The approach is essentially based on utilizing artificial neural network (ANN) inversion for estimating the problem parameters. Particularly, the modular neural network (MNN) is used for the inversion process in order to quantitatively interpret the magnetic anomalies. The MNN inversion has been first tested on a synthetic data with and without random white Gaussian noise. The effect of random noise has been clearly investigated where it showed that the approach provided satisfactory results. Furthermore, three field examples have been inverted in order to investigate the applicability of the proposed approach. The results showed good agreement with the techniques that have been stated in the literatures.     

Artificial neural network inversion of magnetic anomalies caused by 2D fault structures

A new approach is proposed to interpret magnetic anomalies caused by 2D fault structures. This approach is based on the artificial neural network inversion, utilizing particularly modular neural network algorithm. The inversion process is implemented to estimate the parameters of 2D fault structures where it has been verified first on synthetic models. The results of the inversion show that the parameters derived from the inversion agree well with the true ones. The analysis of noise has been studied in order to investigate the stability of the approach where it has been tested for contaminated anomalies with 5 and 10 % of white Gaussian noise. The results of the inversion provide satisfactory results even with contaminated signals.The validity of the approach has been demonstrated through real data taken from New South Wales, Australia. A comparable and satisfactory agreement is shown between the inversion results of the neural network and those from techniques published in literatures.     

Inversion of residual gravity anomalies using neural network

A new approach is presented in order to interpret residual gravity anomalies from simple geometrically shaped bodies such as horizontal cylinder, vertical cylinder, and sphere. This approach is mainly based on using modular neural network (MNN) inversion for estimating the shape factor, the depth, and the amplitude coefficient. The sigmoid function has been used as an activation function in the MNN inversion. The new approach has been tested first on synthetic data from different models using only one well-trained network. The results of this approach show that the parameter values estimated by the modular inversion are almost identical to the true parameters. Furthermore, the noise analysis has been examined where the results of the inversion produce satisfactory results up to 10% of white Gaussian noise. The reliability of this approach is demonstrated through two published real gravity field anomalies taken over a chromite deposit in Camaguey province, Cuba and over sulfide ore body, Nornada, Quebec, Canada. A comparable and acceptable agreement is obtained between the results derived by the MNN inversion method and those deduced by other interpretation methods. Furthermore, the depth obtained by the proposed technique is found to be very close to that obtained by drilling information.     

Optimization framework for calibration of constitutive models enhanced by neural networks

A two‐level procedure designed for the estimation of constitutive model parameters is presented in this paper. The neural network (NN) approach at the first level is applied to achieve the first approximation of parameters. This technique is used to avoid potential pitfalls related to the conventional gradient‐based optimization techniques, considered here as a corrector that improves predicted parameters. The feed‐forward NN (FFNN) and the modified Gauss–Newton algorithms are briefly presented. The proposed framework is verified for the elasto‐plastic modified Cam Clay model that can be calibrated based on standard triaxial laboratory tests, i.e. the isotropic consolidation test and the drained compression test. Two different formulations of the input data to the NN, enhanced by a dimensional reduction of experimental data using principal component analysis, are presented. The determination of model characteristics is demonstrated, first on numerical pseudo‐experiments and then on the experimental data. The efficiency of the proposed approach by means of accuracy and computational effort is also discussed. Copyright © 2008 John Wiley & Sons, Ltd.     

Analysis of subsurface magnetic features of the southeast of Al-Muwayh quadrangle,Saudi Arabia,using aeromagnetic data

Aeromagnetic data have been utilized to investigate the subsurface features of the southeast of Al-Muwayh quadrangle. Several techniques have been comprehensively used in an integrative way to reach the goals. Local phase and normalized standard deviation filters are used in this study as edge detectors, showing the possible occurrences of structural lineaments/faults in the quadrangle. Magnitude magnetic transform filters are used to produce anomalies that are closer to the true horizontal position of magnetic sources to enhance the interpretation. Among these transforms, a transform which has been used as edge detectors and the other two transforms are used to show the shallow and the shallowest magnetic sources within the study area. Tilt angle is mainly used to delineate the main magnetic contacts (faults), their locations, and their expected depths. The integration between these different filters show clearly the possible occurrences of edges (contacts/faults), the direction of these lineaments, the source locations of magnetic anomalies, the shallow and the shallowest causative targets, and the location and the depths of the main faults deduced from the tilt angle approach.     

A direct inversion scheme for deep resistivity sounding data using artificial neural networks

Initialization of model parameters is crucial in the conventional 1D inversion of DC electrical data, since a poor guess may result in undesired parameter estimations. In the present work, we investigate the performance of neural networks in the direct inversion of DC sounding data, without the need ofa priori information. We introduce a two-step network approach where the first network identifies the curve type, followed by the model parameter estimation using the second network. This approach provides the flexibility to accommodate all the characteristic sounding curve types with a wide range of resistivity and thickness. Here we realize a three layer feed-forward neural network with fast back propagation learning algorithms performing well. The basic data sets for training and testing were simulated on the basis of available deep resistivity sounding (DRS) data from the crystalline terrains of south India. The optimum network parameters and performance were decided as a function of the testing error convergence with respect to the network training error. On adequate training, the final weights simulate faithfully to recover resistivity and thickness on new data. The small discrepancies noticed, however, are well within the resolvability of resistivity sounding curve interpretations.     

基于卷积神经网络的磁异常反演

针对传统反演方法存在的初始模型依赖、计算时间较长等问题,提出一种基于卷积神经网络的磁异常反演方法。该方法首先设计大量磁异常体模型,进行正演模拟产生样本数据集;接着借鉴经典的卷积神经网络VGG-13设计了一种全新的VGG磁异常反演网络(VGGINV);然后使用样本数据集训练该网络,并优化网络参数;最后对理论模型和实测数据进行反演实验。实验结果表明,该方法可以准确地反演出磁异常体的位置和磁化强度,具有较强的学习能力和一定的泛化能力,能有效解决磁异常数据反演问题。     

Geoacoustic inversion using adaptive neuro-fuzzy inference system

The geoacoustic parameters form significant input for underwater acoustic propagation studies and geoacoustic modeling. Conventional inversion techniques commonly used as indirect approach for extraction of geoacoustic parameters from acoustic or seismic data are computationally intensive and time-consuming. In the present study, we have tried to exploit the advantage of soft computing techniques like, reasoning ability of fuzzy logic and learning abilities of neural networks, in inversion studies. The network model based on the combined approach called adaptive neuro-fuzzy inference system (ANFIS), is found to be very promising in inversion of the acoustic data. The network model once built is capable of invert a few thousand data sets instantaneously, to a reasonably good accuracy. In the case of conventional approaches, repetition of the entire inversion process with each new data set is required. A limited number of sensor’s data are sufficient for simulation of the network model and provides an advantage to use short hydrophone array data. Inversion results of a few hundred test data sets, representing different geoacoustic environments, show the prediction error is much less than 0.01 g/cc, 10 m/s, 10 m and 0.1 against first layer’s density, compressional sound speed, thickness and attenuation respectively for a three-layer geoacoustic model. However, the error is relatively large for the second- and third-layer parameters, which need to be improved. The model is efficient, robust and inexpensive.     

基于ANFIS的单桩竖向极限承载力预测模型

人工神经网络已应用在岩土工程的各个方面。针对常用的BP网络的不足之处,建立了基于自适应神经模糊推理系统（ANFIS）的单桩竖向极限承载力预测模型。利用文献中桩的载荷试验数据来训练ANFIS网络,确定了网络参数。研究结果表明,同常用的BP网络相比,ANFIS预测模型具有学习速度快,拟合能力较好,训练结果唯一等优点,该方法是一种有效地预测单桩极限承载力的方法。     

A comparative performance of least-square method and very fast simulated annealing global optimization method for interpretation of self-potential anomaly over 2-D inclined sheet type structure

Inversion of self-potential anomaly for 2-D inclined sheets of infinite horizontal extent has been studied. Least-square inversion and very fast simulated annealing global optimization has been used to model the five parameters of self potential anomaly. The method of least square and very fast simulated annealing global optimization method is compared and analyzed. Very fast simulated annealing can model the noisy and field data of self potential anomaly very precisely than linear inversion technique. However, time taken by very fast simulated annealing inversion is larger than linearized inversion. The comparative analysis has been done on synthetic data (noise free and noisy) and two field data from Bavarian woods anomaly, Germany and Surda anomaly, India to show the efficacy of both the methods. The estimated parameters were compared with those from previous studies using various global optimization algorithms, mainly neural network, genetic algorithm and particle swarm optimization on the same field data sets. It can be concluded that the global optimization algorithms considered in this study were able to yield compatible solutions with those from least-square methods. The present global optimization method is in good agreement with the other global optimization methods in terms of results and computation time.     

A new approach to interpret self-potential anomaly over a two-dimensional inclined sheet using complex gradient analysis

A new approach is developed to determine the model parameters of a two-dimensional inclined sheet from self-potential anomaly. In this method, the numerical horizontal self-potential gradient obtained from self-potential anomaly is convolved using Hilbert transform to obtain the vertical self-potential gradient. The complex gradient is the sum of horizontal and vertical gradient anomalies. The horizontal and vertical gradients are plotted in one graph to form the complex gradient graph. By defining few characteristic points and distances along the complex gradient profile, procedures are then formulated using the analytical functions of the complex gradients to obtain the model parameters of sheet-like structures. The validity of the new proposed method has been tested on synthetic data with and without random noise. The obtained parameters are in congruence with the model parameters when using noise-free synthetic data. After adding 10% random error in the synthetic data, the maximum error in model parameters is 11.8%. Moreover, the method have been applied to analyze and interpret the self-potential anomaly measured on a graphite ore body at southern Bavarian woods, Germany to prove its efficiency where an acceptable agreement has been noticed between the obtained results and the other published results.     

A new method of controlled-source frequency induction soundings

We suggest a new inversion method for frequency induction data implying the use of a new parameter, which has a simple analytical form in the case of a uniform subsurface. The new parameter is found from induction numbers measured in the field of a vertical magnetic dipole or a vertical magnetic dipole combined either with a horizontal electrical dipole or with a horizontal magnetic dipole. Compared with the classical methods, the new technique provides better resolved resistivity curves and faithful images of the subsurface at higher frequencies and smaller transmitter-receiver separations. The inversion algorithm is applied to amplitude and amplitude-phase data and provides reliable depth assignment of the detected resistivity layers in the latter case.     

Tilt-depth方法适用性研究及其应用

Tilt-depth方法是一种可以快速反演磁源上顶深度的新兴方法。二维、三维模型试验表明,tiltdepth法反演误差与地质体的上顶埋深、厚度及水平尺度均有关,同时叠加异常也会对反演结果产生影响。实例中,将tilt-depth法应用于广东省下庄矿田航磁数据反演中,tilt梯度图上展示出了两条明显的近东西走向条带异常,推断为浅部基性辉绿岩脉在深部汇聚形成,并反演了两条高磁异常带上11个位置的磁源上顶埋深。反演结果揭示了下庄矿区内北侧高磁性体深度较浅,且具有南浅北深的构造特征,而南侧高磁性体规模较大,且埋深较厚,为该区深部矿产勘查提供了有利依据。     

A novel and generalized approach in the inversion of geoelectrical resistivity data using Artificial Neural Networks (ANN)

The non-linear apparent resistivity problem in the subsurface study of the earth takes into account the model parameters in terms of resistivity and thickness of individual subsurface layers using the trained synthetic data by means of Artificial Neural Networks (ANN). Here we used a single layer feed-forward neural network with fast back propagation learning algorithm. So on proper training of back propagation networks it tends to give the resistivity and thickness of the subsurface layer model of the field resistivity data with reference to the synthetic data trained in the appropriate network. During training, the weights and biases of the network are iteratively adjusted to make network performance function level more efficient. On adequate training, errors are minimized and the best result is obtained using the artificial neural networks. The network is trained with more number of VES data and this trained network is demonstrated by the field data. The accuracy of inversion depends upon the number of data trained. In this novel and specially designed algorithm, the interpretation of the vertical electrical sounding has been done successfully with the more accurate layer model.     

Evaluation of methods for determination of hydraulic properties in an aquifer–aquitard system hydrologically connected to a river

Field and laboratory methods have been used to determine the hydraulic properties in a multiple-layer aquifer–aquitard system that is hydrologically connected to a river. First, hypothetical pumping tests in aquifer–aquitard systems were performed to evaluate the feasibility of MODFLOW-PEST in determining these parameters. Sensitivity analyses showed that: the horizontal hydraulic conductivity in the aquifer has the highest composite sensitivity; the vertical hydraulic conductivity has higher composite sensitivity than the horizontal hydraulic conductivity in the aquitard; and a partial penetration pumping well in an aquifer layer can improve the quality of the estimated parameters. This inverse approach was then used to analyze a pumping-recovery test conducted near the Platte River in southeastern Nebraska, USA. The hydraulic conductivities and specific yield were calculated for the aquitard and aquifer. The direct-push technique was used to generate sediment columns; permeameter tests on these columns produced the vertical hydraulic conductivities that are compatible with those obtained from the pumping-recovery test. Thus, the combination of the direct-push technique with permeameter tests provides a new method for estimation of vertical hydraulic conductivity. The hydraulic conductivity, determined from grain-size analysis, is smaller than the horizontal one but larger than the vertical one determined by the pumping-recovery test.     

小波神经网络在重磁资料反演中的应用前景

对BP神经网络在重力密度界面反演以及小波分析在位场分离上的应用进行了深入的研究,进而对小波神经网络在重磁资料反演中的应用前景进行了分析、评价。     

最小结构磁化率成像反演方法在水下目标探测中的应用

磁测技术对发现和确定水下目标十分有效,已被广泛用于探测水下沉船、水下掩埋管线、电缆和其他(电)磁性物体.然而,水下目标往往具有较强的剩磁与退磁特点,影响了定性解释结果和定量计算精度.为了更精准地确定和刻画水下目标的位置与形状,以磁异常模量数据为基础,采用最小结构模型进行磁化率成像反演;在模型试验中采用了L₁范数和L₂范数分别进行反演,并且对两者结果进行了对比分析,结果显示L₁范数反演结果具有更为规则与清晰的边界,而L₂范数反演结果则相对更为平滑.因此,对于常见水下目标磁测,L₁范数反演更适合.以港珠澳大桥沉管隧道磁测数据为例,基于磁异常模量的最小结构模型反演了水下沉管的埋设状态,其平面位置、宽度和埋深具有较好的精准度.因此,本方法能够基于磁测数据更加精细地计算水下掩埋目标的位置和规模,具有较强的实际应用价值.      

浅谈2．5D磁法反演估算铁矿资源量的多解性

2．5D磁法反演估算铁矿资源量,关键是在垂直磁异常走向上求拟合模型的截面积。该文利用一种常见模型拟合4种实测磁场,发现铁矿截面积与埋深和磁性大小有一定的规律：磁化强度一定时,截面积与埋深呈线性正相关;埋深一定时,截面积与磁化强度呈近似指数形式的反相关。固定深部或外围磁性体参数,单独考虑浅部或中问铁矿模型时也有这种规律。在无钻孔资料及磁性资料不清楚时,磁化强度100000×10^-3～150000×10^-3A／m为估算铁矿资源量的最佳磁参数。因磁法反演的多解性,此类方法适用于钻孔验证前的资源量估计。     

Estimation of Hardgrove grindability index of Turkish coals by neural networks

In this research, different techniques for the estimation of coal HGI values are studied. Data from 163 sub-bituminous coals from Turkey are used by featuring 11 coal parameters, which include proximate analysis, group maceral analysis and rank. Non-linear regression and neural network techniques are used for predicting the HGI values for the specified coal parameters. Results indicate that a hybrid network which is a combination of 4 separate neural networks gave the most accurate HGI prediction and all of the neural network models outperformed non-linear regression in the estimation process.     

TEM surveys for magnetic viscosity of rocks in situ

We discuss the results of a field experiment in the Malaya Botuobiya area (West Yakutia) at a site where earlier surveys revealed slowly decaying transient responses. That time-dependent voltage decay indicated magnetic viscosity effects associated with magnetic relaxation of superparamagnetic grains in rocks. In this study, we have applied a high-resolution array TEM survey to contour the anomaly and parametric soundings with systems of different configurations to explore the vertical pattern of magnetic viscosity. The parametric data have been inverted, by means of manual and automated fitting, with a reference model of a layered magnetically viscous earth, using, respectively, analytical formulas and simulation based on a forward solution by separation of variables. According to both automated and manual inversion, the section at the center of the anomalous site fits a three-layer earth model with an intermediate magnetically viscous layer between two nonmagnetic layers. This model is consistent with a priori evidence of local geology and may provide more details of the latter. The inversion results have been further used to estimate the volumetric percentage of superparamagnetic grains in the magnetically viscous layer, assuming magnetite to be the main ferrimagnetic phase.