首页 | 本学科首页   官方微博 | 高级检索  
相似文献
 共查询到20条相似文献，搜索用时 282 毫秒
1.
Artificial neural networks (ANN) have been used in a variety of problems in the fields of science and engineering. Applications of ANN to the geophysical problems have increased within the last decade. In particular, it has been used to solve such inversion problems as seismic, electromagnetic, resistivity. There are also some other applications such as parameter estimation, prediction, and classification. In this study, multilayer perceptron neural networks (MLPNN) and radial basis function neural networks (RBFNN) were applied to synthetic gravity data and Seferihisar gravity data to investigate the applicability and performance of these networks for the method of gravity. Additionally performance of MLPNN and RBFNN were tested by adding random noise to the same synthetic test data. The structure parameters, such as the depths, the density contrasts, and the locations of the structures were obtained closely for different signal-to-noise ratios (S/N). Bouguer data of Seferihisar area were analyzed by MLPNN and RBFNN to estimate depth, density contrast, and location of the structure. The results of MLPNN, RBFNN, and classical inversion method were compared for real data obtained from Seferihisar Geothermal area and similar structure parameters were obtained. The experiments show that in general RBFNN not only increases the speed of the training stage enormously, but also provides slightly better performance.  相似文献   

2.
This study proposes the use of multi-layer perceptron neural networks (MLPNN) to invert dispersion curves obtained via multi-channel analysis of surface waves (MASW) for shear S-wave velocity profile. The dispersion curve used in inversion includes the fundamental-mode dispersion data. In order to investigate the applicability and performance of the proposed MLPNN algorithm, test studies were performed using both synthetic and field examples. Gaussian random noise with a standard deviation of 4 and 8% was added to the noise-free test data to make the synthetic test more realistic. The model parameters, such as S-wave velocities and thicknesses of the synthetic layered-earth model, were obtained for different S/N ratios and noise-free data. The field survey was performed over the natural gas pipeline, located in the Germencik district of Ayd?n city, western Turkey. The results show that depth, velocity, and location of the embedded natural gas pipe are successfully estimated with reasonably good approximation.  相似文献   

3.
The gravity gradient tensor (GGT) is deduced from products of second-order derivatives of the gravitational potential. A new method based on the invariants of the GGT has been proposed in this research to interpret gravity data due to sphere, infinite horizontal cylinder and semi-infinite vertical cylinder. The method estimates the depth of these simple causative sources from the multiplication of the maximum of the gravity vertical component by the maximum value of the invariants I 1 to I 2 ratio. To show the reliability and correctness of the estimated depths on 3-D models, the method has been tested using theoretical data with and without random noise. In addition, I have applied the method to a field-data example in Texas, USA and the depth obtained by the present method is compared with those published in the literature.  相似文献   

4.
—We have developed a least-squares minimization approach to depth determination from magnetic data. By defining the anomaly value T(0) at the origin and the anomaly value T(N) at any other distance (N) on the profile, the problem of depth determination from magnetic data has been transformed into finding a solution to a nonlinear equation of the form f(z)=0. Formulas have been derived for a sphere, horizontal cylinder, dike, and for a geologic contact. Procedures are also formulated to estimate the effective magnetization intensity and the effective magnetization inclination. A scheme for analyzing the magnetic data has been formulated for determining the model parameters of the causative sources. The method is applied to synthetic data with and without random errors. Finally, the method is applied to two field examples from Canada and Arizona. In all cases examined, the estimated depths are found to be in goodagreement with actual values.  相似文献   

5.
The spatial moments of a contaminant plume undergoing bio-attenuation are coupled to the moments of microbial populations effecting that attenuation. In this paper, a scalable inverse method is developed for estimating field-scale Monod parameters such as the maximum microbial growth rate (μmax), the contaminant half saturation coefficient (Ks), and the contaminant yield coefficient (Ys). The method uses spatial moments that characterize the distribution of dissolved contaminant and active microbial biomass in the aquifer. A finite element model is used to generate hypothetical field-scale data to test the method under both homogeneous and heterogeneous aquifer conditions. Two general cases are examined. In the first, Monod parameters are estimated where it is assumed a microbial population comprised of a single bacterial species is attenuating one contaminant (e.g., an electron donor and an electron acceptor). In a second case, contaminant attenuation is attributed to a microbial consortium comprised of two microbial species, and Monod parameters for both species are estimated. Results indicate the inverse method is only slightly sensitive to aquifer heterogeneity and that estimation errors decrease as the sampling time interval decreases with respect to the groundwater travel time between sample locations. Optimum conditions for applying the scalable inverse method in both space and time are investigated under both homogeneous and heterogeneous aquifer conditions.  相似文献   

6.
A genetic algorithm (GA) is an artificial intelligence method used for optimization. We applied a GA to the inversion of magnetic anomalies over a thick dike. Inversion of nonlinear geophysical problems using a GA has advantages because it does not require model gradients or well-defined initial model parameters. The evolution process consists of selection, crossover, and mutation genetic operators that look for the best fit to the observed data and a solution consisting of plausible compact sources. The efficiency of a GA on both synthetic and real magnetic anomalies of dikes by estimating model parameters, such as depth to the top of the dike (H), the half-width of the dike (B), the distance from the origin to the reference point (D), the dip of the thick dike (δ), and the susceptibility contrast (k), has been shown. For the synthetic anomaly case, it has been considered for both noise-free and noisy magnetic data. In the real case, the vertical magnetic anomaly from the Pima copper mine in Arizona, USA, and the vertical magnetic anomaly in the Bayburt–Sar?han skarn zone in northeastern Turkey have been inverted and interpreted. We compared the estimated parameters with the results of conventional inversion methods used in previous studies. We can conclude that the GA method used in this study is a useful tool for evaluating magnetic anomalies for dike models.  相似文献   

7.
Generalized inversion of the S-wave amplitude spectra from the strong-motion network data in the East-Central Iran has been used to estimate simultaneously source parameters, site response and the S-wave attenuation (Qs). In this regard, 190 three-component records were used corresponded to 40 earthquakes with the magnitudes M3.5–M7.3. These earthquakes were recorded at 42 stations in the hypocentral distance range from 9 to 200 km. The inverse problem was solved in 20 logarithmically equally spaced points in the frequency band from 0.4 to 15 Hz. The frequency-dependent site amplification was imposed, as a constraint, on two reference site responses in order to remove the undetermined degree of freedom in the inversion and obtain a unique inverse solution. Also, a geometrical spreading factor was assumed for removing the trade-off between geometrical spreading and anelastic attenuation. Different source parameters, such as seismic moment (M0), seismic energy (Es), corner frequency (fc) and Brune stress drop (Δσ), were estimated for each event by fitting an ω2 model to the spectra obtained from the inversion. The stress drop values of earthquakes, obtained in this research, are in good agreement with those of other studies. Also average site response values were correlated to the average shear wave velocities in the uppermost 30 m, in high and low frequency bands. The peak frequencies of site amplifications, estimated by the generalized inversion method, where in good agreement with those of horizontal to vertical (H/V) spectral ratios for the S-wave portion of records. However, no perfect matching in amplitude was obtained due to the deficiencies of the H/V ratio technique. By supposing a free shape for Q factor, a frequency dependent function was found, the logarithm of which could be approximated by a linear function, Q(f)=151f0.75. The uncertainties of model parameters have been evaluated by covariance matrix of least-square fit. The residuals were also analyzed in order to assess the validity of the model. The analysis of residuals with respect to magnitude and distance indicates that they are distributed normally with approximately zero mean. The robustness of the results has been studied concerning their sensitivities to the omission of different datasets, selected randomly from original database. The results obtained here can be used in predicting ground-motion parameters applying stochastic methods.  相似文献   

8.
A very fast and efficient approach for gravity data inversion based on the regularized conjugate gradient method has been developed. This approach simultaneously inverts for the depth (z), and the amplitude coefficient (A) of a buried anomalous body from the gravity data measured along a profile. The developed algorithm fits the observed data by a class of some geometrically simple anomalous bodies, including the semi-infinite vertical cylinder, infinitely long horizontal cylinder, and sphere models using the logarithms of the model parameters [log(z) and log(|A|)] rather than the parameters themselves in its iterative minimization scheme. The presented numerical experiments have shown that the original (non-logarithmed) minimization scheme, which uses the parameters themselves (z and |A|) instead of their logarithms, encountered a variety of convergence problems. The aforementioned transformation of the objective functional subjected to minimization into the space of logarithms of z and |A| overcomes these convergence problems. The reliability and the applicability of the developed algorithm have been demonstrated on several synthetic data sets with and without noise. It is then successfully and carefully applied to seven real data examples with bodies buried in different complex geologic settings and at various depths inside the earth. The method is shown to be highly applicable for mineral exploration, and for both shallow and deep earth imaging, and is of particular value in cases where the observed gravity data is due to an isolated body embedded in the subsurface.  相似文献   

9.
A simple and fast determination of the limiting depth to the sources may represent a significant help to the data interpretation. To this end we explore the possibility of determining those source parameters shared by all the classes of models fitting the data. One approach is to determine the maximum depth-to-source compatible with the measured data, by using for example the well-known Bott–Smith rules. These rules involve only the knowledge of the field and its horizontal gradient maxima, and are independent from the density contrast.Thanks to the direct relationship between structural index and depth to sources we work out a simple and fast strategy to obtain the maximum depth by using the semi-automated methods, such as Euler deconvolution or depth-from-extreme-points method (DEXP).The proposed method consists in estimating the maximum depth as the one obtained for the highest allowable value of the structural index (Nmax). Nmax may be easily determined, since it depends only on the dimensionality of the problem (2D/3D) and on the nature of the analyzed field (e.g., gravity field or magnetic field). We tested our approach on synthetic models against the results obtained by the classical Bott–Smith formulas and the results are in fact very similar, confirming the validity of this method. However, while Bott–Smith formulas are restricted to the gravity field only, our method is applicable also to the magnetic field and to any derivative of the gravity and magnetic field. Our method yields a useful criterion to assess the source model based on the (∂f/∂x)max/fmax ratio.The usefulness of the method in real cases is demonstrated for a salt wall in the Mississippi basin, where the estimation of the maximum depth agrees with the seismic information.  相似文献   

10.
An inverse method is developed to simultaneously estimate multiple hydraulic conductivities, source/sink strengths, and boundary conditions, for two-dimensional confined and unconfined aquifers under non-pumping or pumping conditions. The method incorporates noisy observed data (hydraulic heads, groundwater fluxes, or well rates) at measurement locations. With a set of hybrid formulations, given sufficient measurement data, the method yields well-posed systems of equations that can be solved efficiently via nonlinear optimization. The solution is stable when measurement errors are increased. The method is successfully tested on problems with regular and irregular geometries, different heterogeneity patterns and variances (maximum Kmax/Kmin tested is 10,000), and error magnitudes. Under non-pumping conditions, when error-free observed data are used, the estimated conductivities and recharge rates are accurate within 8% of the true values. When data contain increasing errors, the estimated parameters become less accurate, as expected. For problems where the underlying parameter variation is unknown, equivalent conductivities and average recharge rates can be estimated. Under pumping (and/or injection) conditions, a hybrid formulation is developed to address these local source/sink effects, while different types of boundary conditions can also exert significant influences on drawdowns. Local grid refinement near wells is not needed to obtain accurate results, thus inversion is successful with coarse inverse grids, leading to high computation efficiency. Furthermore, flux measurements are not needed for the inversion to succeed; data requirement of the method is thus not much different from that of interpreting classic well tests. Finally, inversion accuracy is not sensitive to the degree of nonlinearity of the flow equations. Performance of the inverse method for confined and unconfined aquifer problems is similar in terms of the accuracy of the estimated parameters, the recovered head fields, and the solver speed.  相似文献   

11.
Stochastic modelling is applied to the analysis of local earthquake recordings, which are usually extremely rich in random incident-wave trains that are chaotically superimposed because of scattering effects in the Earth's crust. The presence in the seismic signal of effects connected with the scale of inhomogeneity in the lithosphere cannot be deterministically described in detail. The application of a stochastic second-order autoregressive model to accelerometric records for the higher magnitude (ML ? 6) Friuli earthquakes and to short-period seismometric records for the aftershocks of the strong earthquake of 6 May 1976 has allowed inferences to be drawn about the spectral properties of seismic signals and the propagation mechanisms of seismic waves. These inferences are based on an extremely small number of parameters of a mathematical model suitable for simultaneously describing the random sequence of scattered wave trains in the time and frequency domains. Useful physical information has been obtained about the dynamic characteristic correlation times and the predominant frequency of the seismic signals; moreover, the strength, σ2e(t), of the innovation of the stochastic process fitting the real digital data set has been estimated. From the envelopes of σ2e(t), the quantity heuristically used in the stochastic approach to describe seismic excitation, the·mean free-path between successive scatterings (l), or the equivalent diffusivity coefficient (d) and turbidity (g), and their dependence on seismic wave frequency have been investigated. For Friuli, using seismometric data at an epicentral distance of ～ 20 km and earthquakes with a magnitude just under 2, mean free-path estimates obtained by means of autoregressive parameters vary from ～ 5 km for the strong interaction model to ～ 30 km for the single scattering model. Furthermore, by means of accelerometric records for the strongest earthquakes in Friuli during May and September 1976, the dependence for the maximum of the seismic excitation on the epicentral distance R was estimated as (σ2e)max ～ R?ν (with ν 1.94 ± 0.13), which is in good agreement with results obtained for the same region using standard methods by means of acceleration peaks versus R. Lastly, stochastic modelling provides a method of estimating change versus time for the predominant frequency and characteristic correlation time of narrow band digital recordings. These two parameters were computed by means of autoregressive parameters in different physical situations and were found to be functions of the earthquake source, the instrumentation frequency response, and the Earth's filtering effects.  相似文献   

12.
In this paper, I introduce a novel approach to modelling the individual random component (also called the intra-event uncertainty) of a ground-motion relation (GMR), as well as a novel approach to estimating the corresponding parameters. In essence, I contend that the individual random component is reproduced adequately by a simple stochastic mechanism of random impulses acting in the horizontal plane, with random directions. The random number of impulses was Poisson distributed. The parameters of the model were estimated according to a proposal by Raschke J Seismol 17(4):1157–1182, (2013a), with the sample of random difference ξ?=?ln(Y 1 )-ln(Y 2 ), in which Y 1 and Y 2 are the horizontal components of local ground-motion intensity. Any GMR element was eliminated by subtraction, except the individual random components. In the estimation procedure, the distribution of difference ξ was approximated by combining a large Monte Carlo simulated sample and Kernel smoothing. The estimated model satisfactorily fitted the difference ξ of the sample of peak ground accelerations, and the variance of the individual random components was considerably smaller than that of conventional GMRs. In addition, the dependence of variance on the epicentre distance was considered; however, a dependence of variance on the magnitude was not detected. Finally, the influence of the novel model and the corresponding approximations on PSHA was researched. The applied approximations of distribution of the individual random component were satisfactory for the researched example of PSHA.  相似文献   

13.
Field determined hydraulic and chemical transport properties can be useful for the protection of groundwater resources from land-applied chemicals. Most field methods to determine flow and transport parameters are either time or energy consuming and/or they provide a single measurement for a given time period. In this study, we present a dripper-TDR field method that allows measurement of hydraulic conductivity and chemical transport parameters at multiple field locations within a short time period. Specifically, the dripper-TDR determines saturated hydraulic conductivity (Ks), macroscopic capillary length (λc), immobile water fraction (θim/θ), mass exchange coefficient (α) and dispersion coefficient (Dm). Multiple dripper lines were positioned over five crop rows in a field. Background and step solutions were applied through drippers to determine surface hydraulic conductivity parameters at 44 locations and surface transport properties at 38 locations. The hydraulic conductivity parameters (Ks, λc) were determined by application of three discharge rates from the drippers and measurements of the resultant steady-state flux densities at the soil surface beneath each dripper. Time domain reflectometry (TDR) was used to measure the bulk electrical conductivity of the soil during steady infiltration of a salt solution. Breakthrough curves (BTCs) for all sites were determined from the TDR measurements. The Ks and λc values were found to be lognormally distributed with average values of 31.4 cm h−1 and 6.0 cm, respectively. BTC analysis produced chemical properties, θim/θ, α, and Dm with average values of 0.23, 0.0036 h−1, and 1220 cm2 h−1, respectively. The estimated values of the flow and transport parameters were found to be within the ranges of values reported by previous studies conducted at nearby field locations. The dripper TDR method is a rapid and useful technique for in situ measurements of hydraulic conductivity and solute transport properties. The measurements reported in this study give clear evidence to the occurrence of non-equilibrium water and chemical movement in surface soil. The method allows for quantification of non-equilibrium model parameters and preferential flow. Quantifying the parameters is a necessary step toward determining the influences of surface properties on infiltration, runoff, and vadose zone transport.  相似文献   

14.
—We have developed a least-squares minimization approach to determine the shape (shape-factor) of a buried polarized body from a residual self-potential anomaly profile. By defining the zero anomaly distance and the anomaly value at the origin on the profile, the problem of the shape-factor determination is transformed into the problem of finding a solution of a nonlinear equation of the form f(q) = 0. Procedures are also formulated to estimate the depth of polarization angle, and the electric dipole moment. The method is applied to synthetic data with and without random noise. The obtained shape-factor agrees very well with the model shape-factor when using synthetic data. After adding ± 2 percent random error in the synthetic data, the shape factor obtained is within ± 4 percent. Finally the validity of the method is tested on a field example from the Ergani copper district, Turkey.  相似文献   

15.
A transient model, hereafter referred to as ROM-TM, was developed to quantify river ecosystem metabolic rates and reaeration rates from field observation of changes in dissolved O2 (DO) and the ratio of 18O to 16O in DO (δ18O-DO). ROM-TM applies an inverse modeling approach and is programmed using MATLAB. Parameters describing photosynthesis, ecosystem respiration, gas exchange, and isotopic fractionation, such as maximum photosynthetic rate (P m ), photosynthetic efficiency parameter (a), respiration rate at 20 °C (R 20 ), gas exchange coefficient (K), respiration isotopic fractionation factor (a R ), and photorespiration coefficient (β R ), can be abstracted by minimizing the sum of square errors between the fitted data and the observed field data. Then DO and δ18O-DO time series can be reconstructed using estimated parameters and input variables. Besides being capable of teasing apart metabolic processes and gas exchange to provide daily average estimates of metabolic parameters at the ecosystem scale, ROM-TM can be used to address issues related to light including light saturation phenomena at the ecosystem level, the effect of cloud cover on the metabolic balance, and photorespiration. Error and uncertainty analysis demonstrates that ROM-TM is stable and robust for the random errors of DO time series. The photosynthetic parameters P m and a are more sensitive than other parameters to lower-resolution time series data.  相似文献   

16.
Several parameters are needed to describe the converted-wave （C-wave） moveout in processing multi-component seismic data, because of asymmetric raypaths and anisotropy. As the number of parameters increases, the converted wave data processing and analysis becomes more complex. This paper develops a new moveout equation with two parameters for C-waves in vertical transverse isotropy （VTI） media. The two parameters are the C-wave stacking velocity （Vc2） and the squared velocity ratio （7v,i） between the horizontal P-wave velocity and C-wave stacking velocity. The new equation has fewer parameters, but retains the same applicability as previous ones. The applicability of the new equation and the accuracy of the parameter estimation are checked using model and real data. The form of the new equation is the same as that for layered isotropic media. The new equation can simplify the procedure for C-wave processing and parameter estimation in VTI media, and can be applied to real C-wave processing and interpretation. Accurate Vc2 and Yvti can be deduced from C-wave data alone using the double-scanning method, and the velocity ratio model is suitable for event matching between P- and C-wave data.  相似文献   

17.
  相似文献   

18.
This paper presents a new clustering procedure based on K-means and self-organizing map (SOM) network algorithms for classification of earthquake ground-motion records. Six scalar indicators are used in data analysis for describing the frequency content features of earthquake ground motions, named as the average spectral period (T avg ), the mean period (T m ), the smoothed spectral predominant period (T 0), the characteristic period (T 4.3), the predominant period based on velocity spectrum (T gSv ), and the shape factor (Ω). Different clustering validity indexes were applied to determine the best estimates of the number of clusters on real and synthetic data. Results showed the high performance of proposed procedure to reveal salient features of complex seismic data. The comparison between the results of clustering analyses recommend the smoothed spectral predominant period as an effective indicator to describe ground-motion classes. The results also showed that K-means algorithm has better performance than SOM algorithm in identification and classification procedure of ground-motion records.  相似文献   

19.
We have developed a least‐squares minimization approach to depth determination using numerical second horizontal derivative anomalies obtained from magnetic data with filters of successive window lengths (graticule spacings). The problem of depth determination from second‐derivative magnetic anomalies has been transformed into finding a solution to a non‐linear equation of the form, f(z) = 0. Formulae have been derived for a sphere, a horizontal cylinder, a dike and a geological contact. Procedures are also formulated to estimate the magnetic angle and the amplitude coefficient. We have also developed a simple method to define simultaneously the shape (shape factor) and the depth of a buried structure from magnetic data. The method is based on computing the variance of depths determined from all second‐derivative anomaly profiles using the above method. The variance is considered a criterion for determining the correct shape and depth of the buried structure. When the correct shape factor is used, the variance of depths is less than the variances computed using incorrect shape factors. The method is applied to synthetic data with and without random errors, complicated regionals, and interference from neighbouring magnetic rocks. Finally, the method is tested on a field example from India. In all the cases examined, the depth and the shape parameters are found to be in good agreement with the actual parameters.  相似文献   

20.
《水文科学杂志》2013,58(5):896-916
Abstract

The performances of three artificial neural network (NN) methods for combining simulated river flows, based on three different neural network structures, are compared. These network structures are: the simple neural network (SNN), the radial basis function neural network (RBFNN) and the multi-layer perceptron neural network (MLPNN). Daily data of eight catchments, located in different parts of the world, and having different hydrological and climatic conditions, are used to enable comparisons of the performances of these three methods to be made. In the case of each catchment, each neural network combination method synchronously uses the simulated river flows of four rainfall—runoff models operating in design non-updating mode to produce the combined river flows. Two of these four models are black-box, the other two being conceptual models. The results of the study show that the performances of all three combination methods are, on average, better than that of the best individual rainfall—runoff model utilized in the combination, i.e. that the combination concept works. In terms of the Nash-Sutcliffe model efficiency index, the MLPNN combination method generally performs better than the other two combination methods tested. For most of the catchments, the differences in the efficiency index values of the SNN and the RBFNN combination methods are not significant but, on average, the SNN form performs marginally better than the more complex RBFNN alternative. Based on the results obtained for the three NN combination methods, the use of the multi-layer perceptron neural network (MLPNN) is recommended as the appropriate NN form for use in the context of combining simulated river flows.  相似文献   

设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号