基于神经网络方法的地下管道系统地震可靠性分析

基于反向传播的多层前馈网络（BP网络）理论，建立了管道单体地震反应预测模型和管道系统连通性预测模型。模拟数值计算提供的样本，两个模型通过离线学习，得到训练后的神经网络，以此执行实时模拟，两模型充分利用了BP网络较好的泛化能力，达到了迅速评估管网系统震后运行状态的目的，神经网络方法克服了传统方法需要大量计算时间的缺点，为地下管道的震后评估提供了一条新的途径。     

Short term streamflow forecasting using artificial neural networks

The research described in this article investigates the utility of Artificial Neural Networks (ANNs) for short term forecasting of streamflow. The work explores the capabilities of ANNs and compares the performance of this tool to conventional approaches used to forecast streamflow. Several issues associated with the use of an ANN are examined including the type of input data and the number, and the size of hidden layer(s) to be included in the network. Perceived strengths of ANNs are the capability for representing complex, non-linear relationships as well as being able to model interaction effects. The application of the ANN approach is to a portion of the Winnipeg River system in Northwest Ontario, Canada. Forecasting was conducted on a catchment area of approximately 20 000 km². using quarter monthly time intervals. The results were most promising. A very close fit was obtained during the calibration (training) phase and the ANNs developed consistently outperformed a conventional model during the verification (testing) phase for all of the four forecast lead-times. The average improvement in the root mean squared error (RMSE) for the 8 years of test data varied from 5 cms in the four time step ahead forecasts to 12.1 cms in the two time step ahead forecasts.     

Fragility functions of blockwork wharves using artificial neural networks

The use of artificial neural networks in the general framework of a performance-based seismic vulnerability evaluation for earth retaining structures is presented. A blockwork wharf-foundation-backfill complex is modeled with advanced nonlinear 2D finite difference software, wherein liquefaction occurrence is explicitly accounted for. A simulation algorithm is adopted to sample geotechnical input parameters according to their statistical distribution, and extensive time histories analyses are then performed for several earthquake intensity levels. In the process, the seismic input is also considered as a random variable. A large dataset of virtual realizations of the behavior of different configurations under recorded ground motions is thus obtained, and an artificial neural network is implemented in order to find the unknown nonlinear relationships between seismic and geotechnical input data versus the expected performance of the facility. After this process, fragility curves are systematically derived by applying Monte Carlo simulation on the obtained correlations. The novel fragility functions herein proposed for blockwork wharves take into account different geometries, liquefaction occurrence and type of failure mechanism. Results confirm that the detrimental effects of liquefaction increase the probability of failure at all damage states. Moreover, it is also demonstrated that increasing the base width/height ratio results in higher failure probabilities for the horizontal sliding than for the tilting towards the sea.     

3D inversion of DC data using artificial neural networks

In this paper, we investigate the applicability of artificial neural networks in inverting three-dimensional DC resistivity imaging data. The model used to produce synthetic data for training the artificial neural network (ANN) system was a homogeneous medium of resistivity 100 Ωm with an embedded anomalous body of resistivity 1000 Ωm. The different sizes for anomalous body were selected and their location was changed to different positions within the homogeneous model mesh elements. The 3D data set was generated using a finite element forward modeling code through standard 3D modeling software. We investigated different learning paradigms in the training process of the neural network. Resilient propagation was more efficient than any other paradigm. We studied the effect of the data type used on neural network inversion and found that the use of location and the apparent resistivity of data points as the input and corresponding true resistivity as the output of networks produces satisfactory results. We also investigated the effect of the training data pool volume on the inversion properties. We created several synthetic data sets to study the interpolation and extrapolation properties of the ANN. The range of 100–1000 Ωm was divided into six resistivity values as the background resistivity and different resistivity values were also used for the anomalous body. Results from numerous neural network tests indicate that the neural network possesses sufficient interpolation and extrapolation abilities with the selected volume of training data. The trained network was also applied on a real field dataset, collected by a pole-pole array using a square grid (8 ×8) with a 2-m electrode spacing. The inversion results demonstrate that the trained network was able to invert three-dimensional electrical resistivity imaging data. The interpreted results of neural network also agree with the known information about the investigation area.     

利用人工神经网络预测电离层foF2参数

利用人工神经网络技术实现了电离层foF2参数提前1小时预测.从foF2时间序列本身的变化特征出发,根据时间序列相关分析结果确定网络输入参数.选用当前时刻foF2值,预测时刻前一天的foF2值,预测时刻前7天foF2平均值,当前时刻前7天foF2平均值,foF2的一阶差分及表示当前时刻t的变量共六个参数作为神经网络输入,下一时刻值作为神经网络输出.对于太阳活动高年平均预测相对误差小于6%,均方根误差小于0.6 MHz,太阳活动低年平均预测相对误差小于10%,均方根误差小于0.5 MHz     

建筑结构利用TLCD减振的神经网络智能控制

本文提出了建筑结构利用调谐液体柱型阻尼器（ＴＬＣＤ）减振的神经网络智能控制方法。首先阐述了确定ＴＬＣＤ半主动控制策略;然后利用ＢＰ人工神经网络方法计算并控制ＴＬＣＤ隔板孔洞的面积,以调节和控制阻尼比＆Ｔ,实现对建筑结构的智能控制。地震作用下的数值分析表明,本文所述的方法是十分有效的。     

Fracture-frequency prediction from borehole wireline logs using artificial neural networks

Borehole-wall imaging is currently the most reliable means of mapping discontinuities within boreholes. As these imaging techniques are expensive and thus not always included in a logging run, a method of predicting fracture frequency directly from traditional logging tool responses would be very useful and cost effective. Artificial neural networks (ANNs) show great potential in this area. ANNs are computational systems that attempt to mimic natural biological neural networks. They have the ability to recognize patterns and develop their own generalizations about a given data set. Neural networks are trained on data sets for which the solution is known and tested on data not previously seen in order to validate the network result. We show that artificial neural networks, due to their pattern recognition capabilities, are able to assess the signal strength of fracture-related heterogeneity in a borehole log and thus fracture frequency within a borehole. A combination of wireline logs (neutron porosity, bulk density, P-sonic, S-sonic, deep resistivity and shallow resistivity) were used as input parameters to the ANN. Fracture frequency calculated from borehole televiewer data was used as the single output parameter. The ANN was trained using a back-propagation algorithm with a momentum learning function. In addition to fracture frequency within a single borehole, an ANN trained on a subset of boreholes in an area could be used for prediction over the entire set of boreholes, thus allowing the lateral correlation of fracture zones.     

Estimation of soil water content in watershed using artificial neural networks

Soil water content (SWC) is an important factor in transfer processes between soil and air, contributing to water and energy balances, and quantifying it remains a challenge. This study uses artificial neural networks (ANNs) to analyse spatial and temporal variation of SWC in a Brazilian watershed, based on climate information, soil physical properties and topographic variables. Thirty eight input variables were tested in 200 models. The outputs were compared with 650 gravimetric moisture measurements collected at 26 points (25 field studies). The results show that it is possible to estimate SWC efficiently (Nash-Sutcliffe statistic, NS = 0.77) using topographic data, soil physical properties and rainfall. If only climate information is considered, modelling is less efficient (NS = 0.28). Using many variables does not necessarily improve performance. Alternatively, SWC can be estimated by simplified models using rainfall and topographic maps information, although the performance is less good (NS = 0.65).     

Modelling suspended sediment concentration using artificial neural networks for Gangotri glacier

The dynamics of suspended sediment involves inherent non‐linearity and complexity because of existence of both spatial variability of the basin characteristics and temporal climatic patterns. This complexity, therefore, leads to inaccurate prediction by the conventional sediment rating curve (SRC) and other empirical methods. Over past few decades, artificial neural networks (ANNs) have emerged as one of the advanced modelling techniques capable of addressing inherent non‐linearity in the hydrological processes. In the present study, feed‐forward back propagation (FFBP) algorithm of ANNs is used to model stage–discharge–suspended sediment relationship for ablation season (May–September) for melt runoff released from Gangotri glacier, one of the largest glaciers in Himalaya. The simulations have been carried out on primary data of suspended sediment concentration (SSC) discharge and stage for ablation season of 11‐year period (1999–2009). Combinations of different input vectors (viz. stage, discharge and SSC) for present and previous days are considered for development of the ANN models and examining the effects of input vectors. Further, based on model performance indices for training and testing phase, a suitable modelling approach with appropriate model input structure is suggested. The conventional SRC method is also used for modelling discharge–sediment relationship and performance of developed models is evaluated by statistical indices, namely; root mean square error (RMSE), mean absolute error (MAE) and coefficient of determination (R²). Statistically, the performance of ANN‐based models is found to be superior as compared to SRC method in terms of the selected performance indices in simulating the daily SSC. The study reveals suitability of ANN approach for simulation and estimation of daily SSC in glacier melt runoff and, therefore, opens new avenues of research for application of hybrid soft computing models in glacier hydrology. Copyright © 2015 John Wiley & Sons, Ltd.     

Rainfall‐induced landslide hazard assessment using artificial neural networks

In Japan, landslides triggered by heavy rainfall tend to occur during the annual rainy season from early June until the middle of July; these landslides constitute a major hazard causing significant property damage and loss of life. This paper proposes the use of back propagation neural networks (BPNN) to predict the probability of landslide occurrence for a scenario of heavy rainfall in the Minamata area of southern Kyushu Island, Japan. All of the landslides were detected from aerial photographs taken in 1999, 2001 and 2003, and a geospatial database of lithology, topography, soil characteristics, land use and precipitation was constructed using geographical information systems (GIS). The training sample consists of 602 cells that include landslide activity and 1600 cells in stable areas. Using the trained BPNN with 49 input nodes, three hidden layers, and one output node, 239 589 cells were processed to produce a map of landslide probability for a maximum daily precipitation of 329 mm and a maximum cumulative precipitation of 581 mm for an incessant, intense rainfall event in the future. The resultant hazard map was classified into four hazard levels; it can be referenced for land‐use planning and decision‐making for community development. Copyright © 2005 John Wiley & Sons, Ltd.     

Effective prediction of scour downstream of ski-jump buckets using artificial neural networks

The main objective of this research was to analyze and quantify the uncertainty of artificial neural network in prediction of scour downstream ski-jump buckets. Hence, at first, three artificial neural network models were developed to predict depth, length, and width of scour hole. Then, Monte-Carlo simulation was applied in the estimates of artificial neural network modeling procedure. The uncertainties were quantified by means of two criteria: 95 percent prediction uncertainty and d-factor. The results of the artificial neural network models showed superior performance of it in comparison with some empirical formulas because of higher correlation coefficient (R ² > 0.95) and lower error (RMSE < 1.63). The obtained result from uncertainty analysis of the models revealed the satisfactory performance of them. In this procedure it was clarified that the artificial neural network model for length prediction was more reliable than the others with d-factor and 95 percent prediction uncertainty equal to 2.53 and 92, respectively.     

Modeling of the river ecological status with macrophytes using artificial neural networks

Biomonitoring methods based on macrophytes have been used mandatorily in the assessment of freshwaters since the implementation of the Water Framework Directive (WFD). The Macrophyte Index for Rivers (MIR) was developed in Poland for the monitoring of running waters under the WFD requirements. This index shows the degree of river degradation under the influence of water pollutants, especially nutrients. The aim of the present study was to determine the relationship between the MIR and various hydrochemical parameters using artificial neural networks (ANNs). Physico-chemical parameters of water (monthly results for the whole year), which were derived from 147 lowland river survey sites, all located in Poland, were applied to model the MIR values. Water quality variables were determined over three timeframes: the annual average; the average for the vegetation period; and the average for the summer period. Quality of the networks was assessed using coefficient of determination (R²), Nash-Sutcliffe efficiency (NSE) and root mean square error (RMSE). The best modeling quality was obtained for yearly average values of water quality parameters. The quality statistics were: R² = 0.722, NSE = 0.721 and RMSE = 0.056 (training dataset); R² = 0.555, NSE = 0.533 and RMSE = 0.101 (validation dataset); R² = 0.650. NSE = 0.600 and RMSE = 0.089 (testing dataset). This indicates that macrophytes reflect the whole year impact of pollution, whereas summer.     

Daily reservoir inflow forecasting using artificial neural networks with stopped training approach

In this paper, an early stopped training approach (STA) is introduced to train multi-layer feed-forward neural networks (FNN) for real-time reservoir inflow forecasting. The proposed method takes advantage of both Levenberg–Marquardt Backpropagation (LMBP) and cross-validation technique to avoid underfitting or overfitting on FNN training and enhances generalization performance. The methodology is assessed using multivariate hydrological time series from Chute-du-Diable hydrosystem in northern Quebec (Canada). The performance of the model is compared to benchmarks from a statistical model and an operational conceptual model. Since the ultimate goal concerns the real-time forecast accuracy, overall the results show that the proposed method is effective for improving prediction accuracy. Moreover it offers an alternative when dynamic adaptive forecasting is desired.     

Modelling groundwater levels in an urban coastal aquifer using artificial neural networks

The prediction of groundwater levels in a basin is of immense importance for the management of groundwater resources, especially in coastal regions where the water table fluctuations are to be limited to avoid seawater intrusion. In this paper, an Artificial Neural Network (ANN) methodology is presented to predict groundwater levels in individual wells with one month lead. Groundwater levels were also predicted in neighboring wells using model parameters from the best network of a well. This methodology is applied to an urban coastal aquifer in Andhra Pradesh state, India. The results suggest that the feed forward neural network with Levenberg Marquardt (LM) algorithm is a good choice for predicting groundwater levels in individual wells. Bayesian Regularization (BR) model parameters of Balaji Nagar well are also used successfully to predict groundwater levels in the study area. It was observed that the ANN‐based algorithms were a better choice for the prediction of groundwater levels with limited hydrological parameters. Copyright © 2007 John Wiley & Sons, Ltd.     

Evaporation estimation using artificial neural networks and adaptive neuro-fuzzy inference system techniques

Evaporation, as a major component of the hydrologic cycle, plays a key role in water resources development and management in arid and semi-arid climatic regions. Although there are empirical formulas available, their performances are not all satisfactory due to the complicated nature of the evaporation process and the data availability. This paper explores evaporation estimation methods based on artificial neural networks (ANN) and adaptive neuro-fuzzy inference system (ANFIS) techniques. It has been found that ANN and ANFIS techniques have much better performances than the empirical formulas (for the test data set, ANN R² = 0.97, ANFIS R² = 0.92 and Marciano R² = 0.54). Between ANN and ANFIS, ANN model is slightly better albeit the difference is small. Although ANN and ANFIS techniques seem to be powerful, their data input selection process is quite complicated. In this research, the Gamma test (GT) has been used to tackle the problem of the best input data combination and how many data points should be used in the model calibration. More studies are needed to gain wider experience about this data selection tool and how it could be used in assessing the validation data.     

Stormwater quantity and quality response to climate change using artificial neural networks

This research investigates the potential impacts of climate change on stormwater quantity and quality generated by urban residential areas on an event basis in the rainy season. An urban residential stormwater drainage area in southeast Calgary, Alberta, Canada is the focus of future climate projections from general circulation models (GCMs). A regression‐based statistical downscaling tool was employed to conduct spatial downscaling of daily precipitation and daily mean temperature using projection outputs from the coupled GCM. Projected changes in precipitation and temperature were applied to current climate scenarios to generate future climate scenarios. Artificial neural networks (ANNs) developed for modelling stormwater runoff quantity and quality used projected climate scenarios as network inputs. The hydrological response to climate change was investigated through stormwater runoff volume and peak flow, while the water quality responses were investigated through the event mean value (EMV) of five parameters: turbidity, conductivity, water temperature, dissolved oxygen (DO) and pH. First flush (FF) effects were also noted. Under future climate scenarios, the EMVs of turbidity increased in all storms except for three events of short duration. The EMVs of conductivity were found to decline in small and frequent storms (return period < 5 years); but conductivity EMVs were observed to increase in intensive events (return period ≥ 5 years). In general, an increasing EMV was observed for water temperature, whereas a decreasing trend was found for DO EMV. No clear trend was found in the EMV of pH. In addition, projected future climate scenarios do not produce a stronger FF effect on dissolved solids and suspended solids compared to that produced by the current climate scenario. Copyright © 2010 John Wiley & Sons, Ltd.     

Landslide susceptibility analysis using GIS and artificial neural network

The purpose of this study is to develop landslide susceptibility analysis techniques using an arti?cial neural network and to apply the newly developed techniques to the study area of Yongin in Korea. Landslide locations were identi?ed in the study area from interpretation of aerial photographs, ?eld survey data, and a spatial database of the topography, soil type and timber cover. The landslide‐related factors (slope, curvature, soil texture, soil drainage, soil effective thickness, timber age, and timber diameter) were extracted from the spatial database. Using those factors, landslide susceptibility was analysed by arti?cial neural network methods. The landslide susceptibility index was calculated by the back‐propagation method, which is a type of arti?cial neural network method, and the susceptibility map was made with a geographic information system (GIS) program. The results of the landslide susceptibility analysis were veri?ed using landslide location data. The validation results showed satisfactory agreement between the susceptibility map and the existing data on landslide location. A GIS was used to ef?ciently analyse the vast amount of data, and an arti?cial neural network to be an effective tool to maintain precision and accuracy. The results can be used to reduce hazards associated with landslides and to plan land use and construction. Copyright © 2003 John Wiley & Sons, Ltd.     

Structural damage detection using the optimal weights of the approximating artificial neural networks

This work presents a novel neural network‐based approach to detect structural damage. The proposed approach comprises two steps. The first step, system identification, involves using neural system identification networks (NSINs) to identify the undamaged and damaged states of a structural system. The partial derivatives of the outputs with respect to the inputs of the NSIN, which identifies the system in a certain undamaged or damaged state, have a negligible variation with different system errors. This loosely defined unique property enables these partial derivatives to quantitatively indicate system damage from the model parameters. The second step, structural damage detection, involves using the neural damage detection network (NDDN) to detect the location and extent of the structural damage. The input to the NDDN is taken as the aforementioned partial derivatives of NSIN, and the output of the NDDN identifies the damage level for each member in the structure. Moreover, SDOF and MDOF examples are presented to demonstrate the feasibility of using the proposed method for damage detection of linear structures. Copyright © 2001 John Wiley & Sons, Ltd.     

应用人工神经网络技术的大型斜拉桥子结构损伤识别研究

本文应用人工神经网络技术对大型斜拉桥结构进行了子结构损伤识别研究。文中首先介绍了子结构损伤识别的基本方法，然后应用自组织竞争神经网络建立了对于大型桥梁结构识别子结构损伤情况的子结构损伤识别方法，并且应用BP网络进一步建立了大型桥梁结构各子结构内部的损伤位置和损伤程度的识别方法，数值模拟了一大跨度斜拉桥子结构损伤以及子结构内部损伤的识别过程，最后得出结论：(1)基于自组织竞争网络的子结构损伤识别方法能迅速准确地识别大型结构的损伤情况；(2)基于BP网络所建立的结构损伤识别方法，能对子结构中结构损伤的位置和程度进行进一步的识别；(3)基于人工神经网络技术的结构损伤识别方法是大型土木工程结构损伤识别的有效方法，可在工程结构损伤识别中广泛应用。     

基于模态分析和神经网络的裂缝损伤识别

提出了裂缝损伤诊断的神经网络方法,探讨了用模态技术和神经网络对混凝土结构裂缝损伤进行识别与定位的方法。文中以一简支矩形截面梁为研究对象,通过完好结构和损伤结构的有限元分析,获取两者的损伤标识量,输入BP神经网络训练。以损伤位置和裂缝高度作为输出参数,对其进行单处损伤定位的研究。数值仿真结果表明,采用神经网络方法可以对裂缝做出较好的诊断。