Reliability analysis using radial basis function networks and support vector machines

To reduce computational costs in structural reliability analysis, utilising approximate response surface functions for reliability assessment has been suggested. Based on the similarities of two adaptive and flexible models, the radial basis function neural network (RBFN) and support vector machine (SVM), the derivatives of the approximate functions of RBFN and SVM models with respect to basic variables are given, and two RBFN-based RSMs (RBFN-RSM1, RBFN-RSM2) and two SVM-based RSMs (SVM-RSM1, SVM-RSM2) are studied. The similarities and differences of these methods are reviewed, and the applicability of these methods is illustrated using five examples. It is shown that there is no obvious difference between RBFN-based RSMs and SVM-based RSMs, and the number of samples needed in RBFN/SVM-RSM2 is smaller than that of RBFN/SVM-RSM1.     

Intensity forecast of tropical cyclones over North Indian Ocean using multilayer perceptron model: skill and performance verification

The coastal regions of India are profoundly affected by tropical cyclones during both pre- and post-monsoon seasons with enormous loss of life and property leading to natural disasters. The endeavour of the present study is to forecast the intensity of the tropical cyclones that prevail over Arabian Sea and Bay of Bengal of North Indian Ocean (NIO). A multilayer perceptron (MLP) model is developed for the purpose and compared the forecast through MLP model with other neural network and statistical models to assess the forecast skill and performances of MLP model. The central pressure, maximum sustained surface wind speed, pressure drop, total ozone column and sea surface temperature are taken to form the input matrix of the models. The target output is the intensity of the tropical cyclones as per the T??number. The result of the study reveals that the forecast error with MLP model is minimum (4.70?%) whereas the forecast error with radial basis function network (RBFN) is observed to be 14.62?%. The prediction with statistical multiple linear regression and ordinary linear regression are observed to be 9.15 and 9.8?%, respectively. The models provide the forecast beyond 72?h taking care of the change in intensity at every 3-h interval. The performance of MLP model is tested for severe and very severe cyclonic storms like Mala (2006), Sidr (2007), Nargis (2008), Aila (2009), Laila (2010) and Phet (2010). The forecast errors with MLP model for the said cyclones are also observed to be considerably less. Thus, MLP model in forecasting the intensity of tropical cyclones over NIOs may thus be considered to be an alternative of the conventional operational forecast models.     

Modeling of the Shale Volume in the Hendijan Oil Field Using Seismic Attributes and Artificial Neural Networks

Petrophysical properties have played an important and definitive role in the study of oil and gas reservoirs, necessitating that diverse kinds of information are used to infer these properties. In this study, the seismic data related to the Hendijan oil field were utilised, along with the available logs of 7 wells of this field, in order to use the extracted relationships between seismic attributes and the values of the shale volume in the wells to estimate the shale volume in wells intervals. After the overall survey of data, a seismic line was selected and seismic inversion methods (model-based, band limited and sparse spike inversion) were applied to it. Amongst all of these techniques, the model-based method presented the better results. By using seismic attributes and artificial neural networks, the shale volume was then estimated using three types of neural networks, namely the probabilistic neural network (PNN), multi-layer feed-forward network (MLFN) and radial basic function network (RBFN).     

ANN-based model for predicting the bearing capacity of strip footing on multi-layered cohesive soil

In reality, footings are most likely to be founded on multi-layered soils. The existing methods for predicting the bearing capacity of 4-layer up to 10-layer cohesive soil are inaccurate. This paper aims to develop a more accurate bearing capacity prediction method based on multiple regression methods and multi-layer perceptrons (MLPs), one type of artificial neural networks (ANNs). Predictions of bearing capacity from the developed multiple regression models and MLP in tractable equations form are obtained and compared with the value predicted using traditional methods. The results indicate ANNs are able to predict accurately the bearing capacity of strip footing and outperform the existing methods.     

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.     

Machine learning for pore-water pressure time-series prediction: Application of recurrent neural networks

Knowledge of pore-water pressure(PWP)variation is fundamental for slope stability.A precise prediction of PWP is difficult due to complex physical mechanisms and in situ natural variability.To explore the applicability and advantages of recurrent neural networks(RNNs)on PWP prediction,three variants of RNNs,i.e.,standard RNN,long short-term memory(LSTM)and gated recurrent unit(GRU)are adopted and compared with a traditional static artificial neural network(ANN),i.e.,multi-layer perceptron(MLP).Measurements of rainfall and PWP of representative piezometers from a fully instrumented natural slope in Hong Kong are used to establish the prediction models.The coefficient of determination(R^2)and root mean square error(RMSE)are used for model evaluations.The influence of input time series length on the model performance is investigated.The results reveal that MLP can provide acceptable performance but is not robust.The uncertainty bounds of RMSE of the MLP model range from 0.24 kPa to 1.12 k Pa for the selected two piezometers.The standard RNN can perform better but the robustness is slightly affected when there are significant time lags between PWP changes and rainfall.The GRU and LSTM models can provide more precise and robust predictions than the standard RNN.The effects of the hidden layer structure and the dropout technique are investigated.The single-layer GRU is accurate enough for PWP prediction,whereas a double-layer GRU brings extra time cost with little accuracy improvement.The dropout technique is essential to overfitting prevention and improvement of accuracy.     

冬小麦水肥生产函数的Jensen模型和人工神经网络模型及其应用

作物水肥生产函数研究是非充分灌溉理论的重要内容,也是提高农田水、肥利用效率的基础.在作物水分生产函数Jensen模型的基础上,引入肥料因子构造了水肥生产函数的Jensen模型;同时构造了作物水肥生产函数的人工神经网络模型.利用北京地区冬小麦田间试验资料对以上2个模型进行了分析,表明以上模型均可用于描述水分、肥料等因素对作物产量的影响,进而可对作物产量进行预测,且模型都具备一定的精度.     

Prediction of Elastic Modulus of Jointed Rock Mass Using Artificial Neural Networks

Two artificial neural network models for the prediction of elastic modulus of jointed rock mass from the elastic modulus of corresponding intact rock and joint parameters have been demonstrated in this paper. The data collected from uniaxial and triaxial compression tests on different rocks with different joint configurations and different confining pressure conditions, reported in the literature are used as input for training the networks. Important joint properties like joint frequency, joint inclination and roughness of joints are considered separately for making the network more versatile. Two different techniques of artificial neural networks namely feed forward back propagation (FFBP) and radial basis function (RBF) are used to predict the elastic modulus ratio.     

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.     

Landslide susceptibility mapping at Vaz Watershed (Iran) using an artificial neural network model: a comparison between multilayer perceptron (MLP) and radial basic function (RBF) algorithms

Landslide susceptibility and hazard assessments are the most important steps in landslide risk mapping. The main objective of this study was to investigate and compare the results of two artificial neural network (ANN) algorithms, i.e., multilayer perceptron (MLP) and radial basic function (RBF) for spatial prediction of landslide susceptibility in Vaz Watershed, Iran. At first, landslide locations were identified by aerial photographs and field surveys, and a total of 136 landside locations were constructed from various sources. Then the landslide inventory map was randomly split into a training dataset 70 % (95 landslide locations) for training the ANN model and the remaining 30 % (41 landslides locations) was used for validation purpose. Nine landslide conditioning factors such as slope, slope aspect, altitude, land use, lithology, distance from rivers, distance from roads, distance from faults, and rainfall were constructed in geographical information system. In this study, both MLP and RBF algorithms were used in artificial neural network model. The results showed that MLP with Broyden–Fletcher–Goldfarb–Shanno learning algorithm is more efficient than RBF in landslide susceptibility mapping for the study area. Finally the landslide susceptibility maps were validated using the validation data (i.e., 30 % landslide location data that was not used during the model construction) using area under the curve (AUC) method. The success rate curve showed that the area under the curve for RBF and MLP was 0.9085 (90.85 %) and 0.9193 (91.93 %) accuracy, respectively. Similarly, the validation result showed that the area under the curve for MLP and RBF models were 0.881 (88.1 %) and 0.8724 (87.24 %), respectively. The results of this study showed that landslide susceptibility mapping in the Vaz Watershed of Iran using the ANN approach is viable and can be used for land use planning.     

典型降水预报ANN多指标优选——以太湖湖西区为例

为提高多种典型人工神经网络应用于降水预报的精度与稳定性并做出优选,对太湖流域湖西区丹徒、丹阳、金坛、溧阳、宜兴5站的年降水量时间序列建立基于组成成分分析的人工神经网络模型,并通过平均相对误差、平均绝对误差、均方根误差及合格率4项评价指标对比分析预报效果。该模型采用Mann-Kendall法、秩和检验法、谱分析法进行组成成分分析;建立BP网络、小波神经网络、RBF网络、GRNN网络及Elman网络模拟并预测随机成分,与确定性成分叠加得年降水量预报结果。在湖西区的研究结果表明,基于组成成分分析的人工神经网络模型的拟合及预测精度高于原始人工神经网络和线性自回归模型,GRNN网络的预测精度与稳定性高于其他4类神经网络。     

Artificial neural network model for prediction of rock properties from sound level produced during drilling

In many rock engineering applications such as foundations, slopes and tunnels, the intact rock properties are not actually determined by laboratory tests, due to the requirements of high quality core samples and sophisticated test equipments. Thus, predicting the rock properties by using empirical equations has been an attractive research topic relating to rock engineering practice for many years. Soft computing techniques are now being used as alternative statistical tools. In this study, artificial neural network models were developed to predict the rock properties of the intact rock, by using sound level produced during rock drilling. A database of 832 datasets, including drill bit diameter, drill bit speed, penetration rate of the drill bit and equivalent sound level (L_eq) produced during drilling for input parameters, and uniaxial compressive strength (UCS), Schmidt rebound number (SRN), dry density (ρ), P-wave velocity (V_p), tensile strength (TS), modulus of elasticity (E) and percentage porosity (n) of intact rock for output, was established. The constructed models were checked using various prediction performance indices. Goodness of the fit measures revealed that recommended ANN model fitted the data as accurately as experimental results, indicating the usefulness of artificial neural networks in predicting rock properties.     

Evaluating distribution pattern of petrophysical properties and their monitoring under a hybrid intelligent based method in southwest oil field of Iran

This study deals with reservoir characterization based on well log data using an unsupervised self-organizing map (SOM) and supervised neural network algorithms with the aim of clustering log responses into reservoir facies of an oil field located in southwest of Iran. In order to promote and justify the quality control and quantify spatial relationships for petrophysical properties, some of neural network-based approaches were introduced such as the SOMs as the intelligent clustering method compared with other hybrid methods, principal component analysis networks (PCANs) and multilayer perceptron (MLP) and statistical clustering (CA) methods. The results obtained from all the abovementioned methods are compared to each other, and the best option is selected based on accuracy and capabilities of clustering and estimation of the petrophysical data, concluding that for predicting any characteristic of the reservoirs, the appropriate network should be chosen and a unique network cannot be convenient for all of them. Accordingly, the SOM clustering technique was employed to classify the reservoir rocks. Based on the SOM visualization, the reservoir rocks were classified into six facies associated with specific petrophysical properties; among them, F6 expressed the best reservoir quality which is characterized by the low amount of density, highest DT, high amount of neutron porosity (NPHI), and lowest GR response. Ultimately, the performance of all the methods was compared to estimate the porosity and permeability within each facies. The results revealed the preference and reliability of PCAN in predicting porosity and confirmed the capability of MLP in permeability prediction. This study also indicates that neuro-prediction of formation properties using well log data is a feasible methodology for optimization of exploration programs and reduction of expenditure by delineating potentially oil-bearing strata with higher accuracy and lower expenses. The resulting neural net-based model can be used as a powerful and distributive system to reduce the high impact of risk in similar fields.     

基于遗传神经网络的克钦湖叶绿素反演研究

叶绿素a浓度能够在一定程度上反映内陆湖泊水质情况。为实现对克钦湖水体叶绿素a浓度的监测,于2010年8月15日对克钦湖进行了现场光谱测量和同步采样。通过分析叶绿素a浓度和光谱数据之间的关系,建立基于反射比、人工神经网络和遗传神经网络的叶绿素a浓度估测模型。结果表明:利用R700nm/R670nm反射比建立的模型估测精度为R2=0.67;人工神经网络模型的估测精度较高,R2=0.882;将遗传算法引入神经网络之后,模型的估测精度进一步提高,R2达到0.956,将模型预测的结果与克里格内插法相结合对研究区的叶绿素a空间分布情况进行定量估测,发现北湖的叶绿素a浓度明显高于南湖,有由北向南逐渐递减的趋势,这为今后利用高光谱数据对克钦湖叶绿素a浓度大面积遥感反演提供了研究基础。     

BP神经网络在地质样品测定中的应用

反向传播人工网络（Ｂ ａｃｋ ｐｒｏｐａｇａｔｉｏｎ ａｒｔｉｆｉｃｉａｌ ｎｅｔｗｏｒｋｓ）是一种动态信息处理系统,它具有联想记忆、自组织、自适应、自学习和容错性等特征。该模型可应用于地质样品测定,以及从地质样品数据中提取信息,实现对地质样品的分类识别及对矿产资源预测评价。ＢＰ神经网络在地质样品测定中将会有很好的应用前景。     

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.     

钻头配方设计神经网络专家系统

利用人工神经网络中的ＢＰ算法,设计的金刚石钻头配方神经网络专家系统具有知识规则较少、界面友好等特点,用于开发钻头配方设计是可行的。实践中经与相同设计条件的钻头相比,其使用效率和寿命均有所提高。