A neural network approach for attenuation relationships: An application using strong ground motion data from Turkey

This paper presents an application of neural network approach for the prediction of peak ground acceleration (PGA) using the strong motion data from Turkey, as a soft computing technique to remove uncertainties in attenuation equations. A training algorithm based on the Fletcher–Reeves conjugate gradient back-propagation was developed and employed for three sample sets of strong ground motion. The input variables in the constructed artificial neural network (ANN) model were the magnitude, the source-to-site distance and the site conditions, and the output was the PGA. The generalization capability of ANN algorithms was tested with the same training data. To demonstrate the authenticity of this approach, the network predictions were compared with the ones from regressions for the corresponding attenuation equations. The results indicated that the fitting between the predicted PGA values by the networks and the observed ones yielded high correlation coefficients (R²). In addition, comparisons of the correlations by the ANN and the regression method showed that the ANN approach performed better than the regression. Even though the developed ANN models suffered from optimal configuration about the generalization capability, they can be conservatively used to well understand the influence of input parameters for the PGA predictions.     

Critical parameter investigation for earthquake hazard assessment in the Sannio-Matese area of Southern Italy

Probabilistic methods have been applied for the assessment of seismic hazard in a selected region of Southern Italy (Sannio-Matese). This method is mainly suitable for engineering and planning purposes and was first introduced in 1968 by Cornell and efficiently codified into a FORTRAN computer program by McGuire. Special attention is paid in this paper to the specific input parameters, i.e. completeness of data catalogues, time variability of seismic activity, different forms of frequency distributions, and regional attenuation characteristics.     

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。     

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.     

Optimized height of noise barrier for non-urban highway using artificial neural network

This study applies artificial neural network (ANN) for the determination of optimized height of a highway noise barrier. Field measurements were carried out to collect traffic volume, vehicle speed, noise level, and site geometry data. Barrier height was varied from 2 to 5 m in increments of 0.1 m for each measured data set to generate theoretical data for network design. Barrier attenuation was calculated for each height increment using Federal Highway Administration model. For neural network design purpose, classified traffic volume, corresponding traffic speed, and barrier attenuation data have been taken as input parameters, while barrier height was considered as output. ANNs with different architectures were trained, cross validated, and tested using this theoretical data. Results indicate that ANN can be useful to determine the height of noise barrier accurately, which can effectively achieve the desired noise level reduction, for a given set of traffic volume, vehicular speed, highway geometry, and site conditions.     

Some applications of a backpropagation neural network in geo-engineering

A MATLAB based backpropagation neural network (BPNN) model has been developed. Two major geo-engineering applications, namely, earth slope movement and ground movement around tunnels, are identified. Data obtained from case studies are used to train and test the developed model and the ground movement is predicted with the help of input variables that have direct physical significance. A new approach is adopted by introducing an infiltration coefficient in the network architecture apart from antecedent rainfall, slope profile, groundwater level and strength parameters to predict the slope movement. The input variables for settlement around underground excavations are taken from literature. The neural network models demonstrate a promising result predicting fairly successfully the ground behavior in both cases. If input variables influencing output goals are clearly identified and if a decent number of quality data are available, backpropagation neural network can be successfully applied as mapping and prediction tools in geotechnical investigations.     

A new probabilistic approach to analysis of compressive strength test data of rock specimens

In this paper, a new probabilistic approach has been proposed for dealing with the wide scatter in laboratory values of compressive strength test (uni-axial and tri-axial compressive strength tests) data of rock specimens. This wide scatter is essentially due to randomness in number as well as orientation of micro-cracks. In the proposed methodology, Stanley's approach, which uses Weibull's theory based on the weakest link model, has been modified to analyse the compressive strength test data. Stanley's approach is applicable to poly-axial tensile stress conditions. Design of all underground excavations requires, as input data, uni-axial compressive strength and the strength under poly-axial stress conditions. Data from compressive strength tests have been analysed using Weibull's theory and the proposed approach. Corresponding cumulative distribution functions of the state variable, i.e., the applied stress level, have been obtained and goodness-of-fit tests performed to check the fitness of test data to these statistical distributions. These cumulative distribution functions have been subsequently invoked to correlate the applied stress level at failure and the associated risk of failure. The analysis finds its application in specifying the design strength of rocks or rock masses for a permissible probability of failure.     

SWAT manual calibration and parameters sensitivity analysis in a semi-arid watershed in North-western Morocco

Being a laborious approach, manual calibration of hydrologic model in a semi-arid context requires in-depth knowledge of the watershed and as much as possible field input data to obtain reliable simulations. In this study, manual calibration and relative sensitivity analysis approaches of the SWAT model (Soil and Water Assessment Tool) were applied for water balance in a 1993 km² watershed (on the R’dom river) located in North-western Morocco. The watershed is located in a semi-arid area dominated by agro-forestry activities. The objectives of this study were (i) to perform a local sensitivity analysis of the SWAT model taking into consideration the watershed characteristics and (ii) to implement a detailed methodology of manual calibration and validation of the model in a semi-arid context. Sensitivity analysis has been carried out on 12 different SWAT input parameters, and has revealed that 4 input parameters only were the most influential ones on flow components of the R’dom watershed. Model manual calibration was conducted along 2006 and 2007 by comparing measured and predicted monthly and daily discharges and taking Nash-Sutcliffe coefficient (NSE), determination coefficient (R ²), and percent bias (PBIAS) as goodness-of-fit indicators. Validation has been performed by the same approach through 2008 and 2009 period. All final NSE values were above 0.5, R ² values exceeded 0.7, and PBIAS lower than 25% demonstrating satisfactory model performances over the study watershed conditions. The SWAT model set-up with measured input data, manually calibrated and validated, reflects well the real hydrologic processes occurring in the R’dom watershed and can be used to assess current and future conditions and to evaluate alternative management practices.     

Forecasting Spontaneous Heating Susceptibility of Indian Coals Using Neuro Fuzzy System

Coal mine fires due to spontaneous heating have been a great concern both for the industry and researchers worldwide. They start in a very small scale and gradually expand in size resulting in environmental, safety and economic concerns. Most of these fires could be averted if suitable preventive measures are taken. Since the spontaneous heating potential of all types of coals are not the same, its accurate forecasting is essential to plan efficient preventive measures, production schedule and storage capabilities of a mine. This study presents a neuro-fuzzy approach for forecasting the auto-oxidation of coals based on laboratory test results on fifty cone coal samples collected from different coalfields of India. Using the constituents of proximate analysis as input parameters, crossing point temperature (CPT) is predicted using adaptive neuro-fuzzy inference system (ANFIS) model. Gaussian membership function has been adopted for describing the input variables. The results indicate that ANFIS model can predict CPT values with reasonable accuracy.     

Global Attenuation Relationship for Estimating Peak Ground Acceleration

Peak Ground Acceleration (PGA) is a very important ground motion parameter which is used to define the degree of ground shaking during an earthquake. It is also very helpful for designing earthquake resistant structure. The PGA can be estimated by attenuation relationships using magnitude, distance, source type etc of a ground motion. In the past, several researchers have developed over 450 attenuation relationships for predicting PGA for a specific region. In the present study an attempt has been made to develop an attenuation relationship on the basis of these available previous relationships in rock site which will be applicable for any region of the world. In the present study, PGA has been expressed as a function of moment magnitude and hypo-central distance in rock site. Chi-square test have also been performed with available earthquake data in American and Indian region for verifying the accuracy of the generated attenuation relationship. Using multiple regression and Genetic Algorithm (GA) the attenuation relationship equations have also been generated. These equations will be very helpful for performing seismic hazard analysis and predicting earthquake force in any region of the world.     

Advanced seismic slope stability analysis

The objective of this study was to present an advanced methodology for assessing seismic slope stability by taking into account the uncertainties related to the main input parameters. The methodology was applied on a real landslide in order to show the advantages of using the proposed procedure and establish the baseline trends of dynamic response and calculated permanent seismic displacements. It involves the following steps: preliminary analysis, probabilistic static and seismic factor of safety analysis, and permanent seismic displacement analysis. Estimating post-failure maximum seismic deformation of landslide mass and sounding properties is the most important part of this study. It involves both Newmark sliding block method and continuum mechanics approach, applied for characteristic set of input values in order to have more accurate assessment of slope performance and determine the relative importance of input parameters. The results of the analysis showed the benefits of using the proposed step-by-step methodology. The obtained difference in the results between the two methods depends strongly on the set input data for a particular analysis.     

Reliability-based robust design for reinforcement of jointed rock slope

Traditional reliability-based design methodologies often involve selection of design which is of lowest cost and satisfies safety requirements. But, this design is sensitive to variation in statistics of input parameters (noise parameters) and might become unsatisfactory if an underestimation of coefficient of variation of input parameters is made. A relatively new design methodology known as robust geotechnical design (RGD) is applied for the case of reinforcement of rock slope using end-anchored rock bolts. This ensures selection of a cost-effective and safe design for which probability of failure (P_f) of reinforced rock slope is least sensitive to the noise parameters. Reliability-based RGD approach involves evaluation of P_f for each design with different possible noise parameters. Finding P_f for the complex geotechnical structure is computationally expensive, and thus an augmented radial basis function-based response surface is used as a surrogate to the finite element model of rock slope. This response surface, being very efficient, also performs well for a range of values of noise parameters. Later, minimum distance algorithm is applied to obtain a cost-effective and robust design. Finally, a comparison is made in the costs between two robust designs obtained for different target probability of failure for the same rock slope.     

Rock–support interaction analysis based on numerical modelling

The use of yield in supports to control the final loading that develops upon a support system has been one of the most important deformation control techniques used by tunnelling engineers, both historically and currently. Successful use of this approach requires a thorough understanding of the process of rock–support interaction as it is an approach that can fail dramatically if incorrectly applied. There is a fine line between the yield support technique improving the conditions, and the approach resulting in the development of a large area of failed rock, which could ultimately be detrimental. The relationship between the support action and the rock has historically been studied using analytical approaches with the application of significant simplifying assumptions.This paper presents a new approach, where a state-of-the-art numerical model is run repeatedly to develop rock–support interaction curves. This has the advantage of allowing more realistic tunnel geometry, stress states and ground conditions to be simulated. It does, however, use the familiar output form of the relatively simple rock–support interaction curve as opposed to complex and voluminous graphics. Its disadvantage lies in the considerable number of computer runs required to develop the full solutions. Computer software has, however, been written to automate much of this process using a programming language within the modelling package.The analysis approach has been further improved by plotting not one rock–support interaction curve but a whole family of curves representing variations in the rock mass quality of the assumed ground, since this is the most variable of the input parameters for most tunnelling situations. This form of output allows engineers to study the practical range of yield they may require for their rock conditions and also to define at what rock mass quality they can expect the yielding approach to cease to be an effective strategy. This new approach has been presented on a test case history with idealized rock mass properties to illustrate the approach. However, it is an approach that can be specially tailored to any set of rock conditions, tunnel geometry or stress.     

沙牌坝址基岩场地地震动输入参数研究

重大水利水电工程地震动输入参数必须根据专门的地震危险性分析结果来确定。目前由地震危险性分析得到的一致概率反应谱具有包络的意义,不能反映实际地震的频谱特性,输入“一致概率反应谱”可能导致地震作用偏大;拟合设计反应谱人工生成地震动加速度时程的频率非平稳性也没有得到很好解决。为了解决这些问题,得到与坝址地震危险性一致、具体地震的输入参数,结合沙牌大坝提出了一套适用于重大水利水电工程基岩场地地震动输入参数确定方法：通过以有效峰值加速度为参数的概率地震危险性计算分析,确定坝址不同超越概率下的有效峰值加速度及对坝址贡献最大的潜在震源区;在最大贡献潜在震源内利用震级空间联合分布概率最大法确定坝址设定地震,依据加速度反应谱衰减关系确定与坝址设定地震对应的设计反应谱;根据设定地震结果和时变功率谱模型参数衰减关系确定时变功率谱,将时变功率谱和最小相位谱按三角级数叠加法进行强度和频率非平稳地震加速度时程合成。在对沙牌坝址区域的地震活动性及地震构造环境分析评价的基础上,采用上述方法,得到了坝址基岩场地不同超越概率下的有效峰值加速度、设计反应谱、强度和频率非平稳地震加速度时程等地震动输入参数。     

Seismic hazard maps for the U.K.

Past studies of seismic hazard in the U.K. that have used modern probabilistic methods of hazard assessment have been site-specific studies, mostly in connection with nuclear installations. There has been a need for general-purpose maps of seismic hazard to show relative variation of exposure within the U.K. and to give some guidance on absolute values. Such maps have now been produced, incorporating, for the first time, the wealth of new information on historical earthquakes in Britain that has been gathered over the last 15 years. The hazard calculations were undertaken using a new computer code based on the USGS program SEISRISK III, but incorporating a logic tree approach to model variation in the input parameters (e.g. focal depth) or uncertainty in the formulation of the model (e.g. attenuation parameters). An innovative approach was taken to the formulation of seismic source zones, in which two overlapping models were employed. The first of these uses relatively broad source zones based loosely on an interpretation of seismicity and tectonics, while the second uses numerous small zones that reflect the locations of past significant earthquakes. This double approach (using the logic tree methodology) has the merit of both considering the general trend of earthquake activity as well as focusing in on known danger spots. The results show that the areas of highest hazard are western Scotland, north-western England and Wales, where the intensity with 90% probability of non-exceedance in 50 years is 6 EMS.     

Prediction of Uplift Pile Displacement Based on Cone Penetration Tests (CPT)

Accurate prediction of uplift pile displacement is necessary to ensure appropriate structural and serviceability performance of civil projects. On the other hand, in recent years, machine-learning models have been applied to many geotechnical-engineering problems, with some degrees of success. The scope of this research includes three main stages: (1) the compilation of load–displacement data sets, obtained from the published literature, (2) analysis of machine learning models that predict the uplift pile displacement based on the cone penetration test data, and the relative importance of input parameters that have been evaluated using senility analysis by the artificial neural network, In addition, this paper also examines the different selection of input parameters and internal network parameters to obtain the optimum model, (3) A parametric study has also been performed for the input parameters to study the consistency of the suggested model. The statistical parameters and parametric study obtained in this research show the superiority of the current model. It is demonstrated that machine learning models such as ANN and GP models outperform the traditional methods, and provide accurate uplift pile displacement predictions.     

Activity rating of Pliocene–Quaternary faults: a formalized approach (example of the Baikal rift system)

It is suggested to estimate the Pliocene–Quaternary fault activity in a formalized way from synthesis of different data. The respective database consists of two main sections: (i) general information and basic fault parameters and (ii) geomorphic, structural, paleoseismic, seismological, geophysical, geodetic, engineering-geological, hydrological, and meteorological data. The fault characteristics are scored according to their significance, and the cumulative score measures the fault activity. With this approach, the faults in the Barguzin and Tunka rift basins and in the northeastern flank of the Baikal rift system have been divided into five activity classes (low, medium, relatively high, high, and very high activity) and mapped correspondingly. It has been recommended that the concept of an hazardous fault, as updated with regard to the activity rating, refers to faults of relatively high, high, and very high activity. Thus identified hazardous faults within the study area are quite few (4–8%), though this percent may increase slightly as more input data become available. The underestimation cannot be dramatic because all known seismological and structural characteristics of faults essential for the activity rating have been already taken into account. The new approach may be useful in seismic risk assessment and in choice of sites for instrumental monitoring of seismicity.     

Landslide susceptibility deterministic approach using geographic information systems: application to Breaza town, Romania

The study is a deterministic-based approach on landslide susceptibility. The purpose of the paper is to create quantitative susceptibility maps by joining the one-dimension infinite slope stability model with a raster-based GIS (ILWIS) and taking into account the spatial distribution of input parameters. A landslide-prone area, with relative homogeneous geology and geomorphology, located in the Subcarpathian sector of the Prahova River, Romania, was selected for the study. There are frequent problems caused by active landslides in the studied area, especially in years with heavy precipitation, often causing destruction of houses and roads situated on the slopes (1992, 1997, and 2005). Detailed surveys covering a 7-year period provided the necessary input data on slope parameters, hydrological components, and the geotechnical background. Two simulations were used: one on dry soil conditions and one on fully saturated soil conditions. A third test was based on the level of the groundwater table mapped in summer 2008. Detailed analyses were particularly focused on landslides to compare predicted results with actual results using field measurements. The model is very suitable for use in raster GIS because it can calculate slope instability on a pixel basis, each raster cell being considered individually. The drawback of the model is the highly detailed data of input parameters. Despite this disadvantage, in conclusion, the usefulness of slope stability models on a large-scale basis was emphasized under infinitely high failure plain conditions and lithological homogeneity.     

Seismic Hazard for Selected Sites in Greece: A Bayesian Estimate of Seismic Peak Ground Acceleration

A procedure for estimating maximum values of seismic peak ground accelerationat the examined site and quantiles of its probabilistic distribution in a future timeinterval of a given length is considered. The input information for the method areseismic catalog and regression relation between peak seismic acceleration at a givenpoint and magnitude and distance from the site to epicenter (seismic attenuation law).The method is based on Bayesian approach, which simply accounts for influenceof uncertainties of seismic acceleration values. The main assumptions for the method are Poissonian character of seismic events flow and distribution law of Gutenberg-Richter's type. The method is applied to seismic hazard estimation in six selected sitesin Greece.