土质边坡稳定性影响因素的研究

边坡稳定性涉及到诸多因素,引入人工神经网络预测边坡稳定性的方法--误差逆传播学习算法效果显著.边坡稳定性预测系统的输入信息包括岩土体参数、几何参数等,而输出信息则是网络预测的稳定系数和稳定状态.土质边坡主要以圆弧滑移破坏为主,通过人工神经网络预测的结果与实际监测结果的对比分析,证实了BP神经网络在评价土质边坡稳定性方面的效果显著;并在此基础上分析了土质边坡影响因素对边坡稳定性的影响程度.     

A hybrid method for grade estimation using genetic algorithm and neural networks

In the present paper, a new hybrid method is proposed for grade estimation. In this method, the multilayer perceptron (MLP) network is trained using the combination of the Levenberg–Marquardt (LM) method and genetic algorithm (GA). Having a few samples for grade estimation, it is difficult to get a proper result using some function approximation methods like neural networks or geostatistical methods. The neural network training methods are very sensitive to initial weight values when there are a few samples as a training dataset. The main objective of the proposed method is to resolve this problem. Here, our method finds the optimal initial weights by combining GA and LM method. Having the optimal initial values for weights, the local minima are avoided in the training phase and subsequently the neural network sustainability is trained optimally. Furthermore, the hybrid method is applied for grade estimation of Gol-e-Gohar iron ore in south Iran. The proposed method shows significant improvements compared to both conventional MLP and Kriging method. The efficiency of the proposed method gets more highlighted when the training data set is small.     

基于神经网络的地质勘测反分析研究

针对地质勘查中,土的力学参数的确定及土的分类这两类复杂问题,根据反问题理论的基本原理,提出了一种基于回归分析与RBF神经网络结合的新型智能方法,建立了从土的力学参数估计到模型分类的完整智能化分析系统。考虑到土的物理参数测定方法比较简单,且实测变异性小,而力学参数实测变异性大的特点,利用RBF神经网络的数值逼近的特性,建立了神经网络模型来逼近两者之间的函数关系,可以有效地反演力学参数。同时,利用RBF神经网络所具有的模式识别功能,为地质勘察中土层划分提供依据。通过对黄石地区岩土勘查资料的分析与预测表明,该方法简捷有效。     

巷道围岩参数的人工神经网络预测

应用人工智能方法解决地下工程问题,提出了预测巷道围岩参数的人工神经网络预测法,构造了预测围岩参数的神经网络模型。预测结果证明,该模型具有很高的预测精度。提出的方法有一定的实用价值和参考价值。     

Prediction of surface crown pillar stability using artificial neural networks

A relatively novel technique, artificial neural networks (ANN), is used in predicting the stability of crown pillars left over large excavations. Data for the training and verification of the networks were obtained from the literature. Four artificial networks, based on two different architectures, were used. The networks used different numbers of input parameters to predict the stability or failure of crown pillars. Multi‐layer perceptron networks using mine type, dip of orebody, overburden thickness, pillar thickness, pillar length, stope height, backfill height, Rock Mass Rating (RMR) of the host rock and RMR of the orebody showed excellent performance in training and verification. Adding three more variables, namely pillar width, rock density and pillar thickness to width ratio, showed symptoms of over‐learning without degrading performance significantly. Radial basis function networks were capable of predicting crown pillar behaviour on the basis of few input functions. It was shown that mine type, dip and pillar thickness to width ratio can be used for a preliminary estimation of stability. Copyright © 2006 John Wiley & Sons, Ltd.     

An intelligent neural network model for evaluating performance of immobilized cell biofilter treating hydrogen sulphide vapors

Biofiltration has shown to be a promising technique for handling malodours arising from process industries. The present investigation pertains to the removal of hydrogen sulphide in a lab scale biofilter packed with biomedia, encapsulated by sodium alginate and poly vinyl alcohol. The experimental data obtained under both steady state and shock loaded conditions were modelled using the basic principles of artificial neural networks. Artificial neural networks are powerful data driven modelling tools which has the potential to approximate and interpret complex input/output relationships based on the given sets of data matrix. A predictive computerised approach has been proposed to predict the performance parameters namely, removal efficiency and elimination capacity using inlet concentration, loading rate, flow rate and pressure drop as the input parameters to the artificial neural network model. Earlier, experiments from continuous operation in the biofilter showed removal efficiencies from 50 to 100 % at inlet loading rates varying up to 13 g H₂S/m³h. The internal network parameter of the artificial neural network model during simulation was selected using the 2^k factorial design and the best network topology for the model was thus estimated. The results showed that a multilayer network (4-4-2) with a back propagation algorithm was able to predict biofilter performance effectively with R² values of 0.9157 and 0.9965 for removal efficiency and elimination capacity in the test data. The proposed artificial neural network model for biofilter operation could be used as a potential alternative for knowledge based models through proper training and testing of the state variables.     

An attention-constrained neural network with overall cognition for landslide spatial prediction

The occurrence of landslides is affected by various environmental factors. When predicting landslides, conventional neural networks optimize parameters using global connectivity, which limits their efficiency in extracting features of contributing factors. In this study, we developed an attention-constrained neural network with overall cognition (OC-ACNN) to focus on important features from the complex data. The method has four steps: (1) extract the overall cognition as the prior input based on historical landslide distribution and contributing factors, (2) embed an attention mechanism in hidden layers to allocate more weight to noteworthy features, (3) update weights and fit the nonlinear relationship by the back-propagation neural network (BPNN), and (4) generate prediction results using a classifier. This model was applied to the Sichuan-Tibet Highway, considering 10 predisposing factors and 1449 historical landslides. The evaluation results indicate that OC-ACNN (0.822) had a higher predictive capability than multiple linear regression (MLR, 0.734) and BPNN (0.789) in terms of the area under the receiver operating characteristic curve (AUC). Further, we compared different attention patterns and score functions for use with the proposed model. The results show that OC-ACNN offered greater predictive performance than Self-ACNN (without OC, 0.803) and that the improved cosine (0.822) score function had better results and stability than others (0.819 highest).

     

Modeling and Interpretation of Pressuremeter Test Results with Artificial Neural Networks

In this paper, three types of artificial neural network (ANN) are employed to prediction and interpretation of pressuremeter test results. First, multi layer perceptron neural network is used. Then, neuro-fuzzy network is employed and finally radial basis function is applied. All applied networks have shown favorable performance. Finally, different models have been compared and network with the most outstanding performance in two stages is determined. Contrary to conventional behavioral models, models based neural network do not demonstrate the effect of input parameters on output parameters. This research is response to this need through conducting sensitivity analysis on the optimal structure of proposed models.     

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.     

Comparison of surrogate models with different methods in groundwater remediation process

Surrogate modelling is an effective tool for reducing computational burden of simulation optimization. In this article, polynomial regression (PR), radial basis function artificial neural network (RBFANN), and kriging methods were compared for building surrogate models of a multiphase flow simulation model in a simplified nitrobenzene contaminated aquifer remediation problem. In the model accuracy analysis process, a 10-fold cross validation method was adopted to evaluate the approximation accuracy of the three surrogate models. The results demonstrated that: RBFANN surrogate model and kriging surrogate model had acceptable approximation accuracy, and further that kriging model’s approximation accuracy was slightly higher than RBFANN model. However, the PR model demonstrated unacceptably poor approximation accuracy. Therefore, the RBFANN and kriging surrogates were selected and used in the optimization process to identify the most cost-effective remediation strategy at a nitrobenzene-contaminated site. The optimal remediation costs obtained with the two surrogate-based optimization models were similar, and had similar computational burden. These two surrogate-based optimization models are efficient tools for optimal groundwater remediation strategy identification.     

用人工神经网络方法分析油藏压裂效果——以靖安油田塞A井区长2油藏为例

通过已有的少数井的试井资料分析得出压裂裂缝参数,以现有的参数为样本建立人工神经网络系统.以影响压裂结果的地层厚度、孔隙度、泥质含量、压裂施工参数、工作压力加砂排量为输入参数,以裂缝导流能力和裂缝半长为输出参数,用BP神经网络训练,推断出所有井的压裂裂缝参数,从而得到整个油藏的压裂裂缝分布特征,对压裂措施的效果有了直观的评价.     

Modeling the Neuman’s well function by an artificial neural network for the determination of unconfined aquifer parameters

An artificial neural network is designed as an improved alternative approach to the conventional type-curve matching technique for the determination of unconfined aquifer parameters. The network is implemented in a six-step protocol consisted of input selection, data splitting, design of network architecture, determination of network structure, network training, and network validation. The network is trained for the well function of unconfined aquifers by the back-propagation technique, adopting the Levenberg-Marquardt optimization algorithm. By applying a principal component analysis (PCA) on the training input data and through a trial-and-error procedure, the structure of the network is optimized with the topology of (3 × 6 × 3). The replicative, predictive, and structural validity of the developed network are evaluated with synthetic and real field data. The network eliminates graphical error inherent in the type-curve matching technique and provides an automatic and fast procedure for aquifer parameter estimation, particularly when analyzing many alternative pumping tests routinely obtained from continuous data loggers/data collection systems.     

土体渗流固结参数识别方法

根据土体固结过程中超孔隙水压力观测资料，建立了基于遗传算法的土体渗流团结参数非线性识别方法，解决了经典高斯－牛顿极小化问题所存在的局部极小问题和最小二乘法所存在的当初始值选择不合适时迭代过程发散的问题，提出了根据观测仪器的精度，建立 工终止条件的方法，数值计算结果表明，本文所提出的非线性反演方法适合于土体团结参数识别等类似的反问题。     

基于改进BP网络算法的边坡稳定性评价

影响边坡稳定性的因素复杂且具有随机性和不确定性。由于神经网络方法不仅能考虑定量因素,而且能考虑定性因素的影响,因而神经网络适用于解决非确定性的边坡稳定性评价问题。结合遗传算法的并行搜索结构和模拟退火的概率突跳特性,提出了一种用于BP网络权值学习的GASA混合策略,并综合考虑影响边坡稳定性的各方面因素,建立了基于GASA混合策略的神经网络模型,并利用大量工程资料对网络进行训练和测试,得出了一些有意义的结论。     

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.     

Classification of slopes and prediction of factor of safety using differential evolution neural networks

Slope stability analysis is one of the most important problems in geotechnical engineering. The development in slope stability analysis has followed the development in computational geotechnical engineering. This paper discusses the application of different recently developed artificial neural network models to slope stability analysis based on the actual slope failure database available in the literature. Different ANN models are developed to classify the slope as stable or unstable (failed) and to predict the factor of safety. The developed ANN model is found to be efficient compared with other methods like support vector machine and genetic programming available in literature. Prediction models are presented based on the developed ANN model parameters. Different sensitivity analyses are made to identify the important input parameters.     

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.     

Fast computation of Hankel Transform using orthonormal exponential approximation of complex kernel function

The computation of electromagnetic (EM) fields, for 1-D layered earth model, requires evaluation of Hankel Transform (HT) of the EM kernel function. The digital filtering is the most widely used technique to evaluate HT integrals. However, it has some obvious shortcomings. We present an alternative scheme, based on an orthonormal exponential approximation of the kernel function, for evaluating HT integrals. This approximation of the kernel function was chosen because the analytical solution of HT of an exponential function is readily available in literature. This expansion reduces the integral to a simple algebraic sum. The implementation of such a scheme requires that the weights and the exponents of the exponential function be estimated. The exponents were estimated through a guided search algorithm while the weights were obtained using Marquardt matrix inversion method. The algorithm was tested on analytical HT pairs available in literature. The results are compared with those obtained using the digital filtering technique with Anderson filters. The field curves for four types (A-, K-, H-and Q-type) of 3-layer earth models are generated using the present scheme and compared with the corresponding curves obtained using the Anderson sc heme. It is concluded that the present scheme is more accurate than the Anderson scheme     

土石坝地震永久变形参数反演方法研究

提出了一种基于径向基网络的土石坝永久变形参数反演分析模型。该模型充分利用了径向基神经网络的非线性映射能力,只需要进行少量的样本设计,即可反演坝体永久变形参数,可以解决土石坝动力参数反演计算耗时长的问题。同时在对永久变形参数进行灵敏度分析的基础上,建立考虑参数灵敏度的网络训练目标函数,进一步提高了反演精度。将所建立的模型用于紫平铺面板堆石坝地震永久变形参数反演,采用三维有限元法进行静动力分析,并采用改进的沈珠江模型计算坝体地震永久变形。结果表明,反演参数计算的大坝地震永久变形和坝体实测永久变形数值接近,趋势一致,因而所建立的模型能够有效地反演坝体地震永久变形参数,为土石坝的动力参数反演提供了一种简便、有效的方法。