B－P网络在地表水水质综合评价中的应用

Ｂ－Ｐ网络在地表水水质综合评价中的应用梁丽明，孙建星，崔宝波（太原市环境保护监测站）（太原工业大学）Ｂ－Ｐ网络（Ｂａｃｋ－ｐｒａｐａｇａｔｌｏｎｎｅｔｗｏｒｋ）是一种由非线性变换单元组成的前馈网络。其输入输出可采用单调上升的非线性变换，它们的联接权和...     

神经网络敏感性分析及其在遥感影像分类中的应用

针对基于连接权的神经网络敏感性分析方法中求取敏感性系数的不稳定性,提出一种优化连接权的神经网络敏感性分析方法。首先采用遗传算法根据误差最小化原则对神经网络进行优化,在优化的神经网络模型上进行基于连接权的敏感性分析。以1个数值模拟实例和华盛顿广场地区的遥感图像地物分类为例,验证所提方法的有效性。实验结果表明,所提方法求取输入变量的敏感性系数是稳定有效的,能有效筛选出遥感图像中对分类贡献较大的特征波段,达到降维的同时提高分类精度。     

基于遗传算法和BP神经网络的矿井涌水量预测

为避免BP神经网络极易陷人局部解的问题,针对遗传算法具有全局寻优的特点,提出了用遗传算法优化BP神经网络预测方法,并以刘桥二矿为例,对其矿井涌水量进行了预测。首先选取刘桥二矿区的2005年3月至2006年12月的矿井涌水量数据进行分析,然后使用遗传算法优化BP网络的初始权值和阈值,最后使用BP神经网络进行训练。将其成果与纯BP网络算法进行比较,结果表明：遗传算法优化BP神经网络的预测方法的预测精度高于纯BP网络算法,将其应用于矿井涌水量预测是有效可行的。     

模糊神经网络用于储层预测

利用模糊理论和ＢＰ网络相结合组成的模糊神经网络系统,能够克服ＢＰ网络单独使用的局限性,可在地质条件较复杂地区进行储层及油气预测。通过实际资料应用表明,应用单个ＢＰ网络进行储层及油气预测效果较差的地区,采用模糊神经网络能取得较好的效果。     

用神经网络模型预测济宁市地下水水位变化规律

本文通过应用人工神经网络模型中的ＢＰ网络模型,对济宁市地下水水位变化规律了预测,并与线顺归模型的计算结果进行比较,证明ＢＰ网络模型的精度较高。     

遗传神经网络算法改进研究

对传统的遗传算法编码方式进行了改进并与前馈神经网络ＢＰ算法相结合,从而克服了ＢＰ算法收敛速度慢的缺点,算例表明,改进后算法是有效的。     

智能预测在三峡永久船闸中隔墩时效变形分析中的应用

人工神经网络通过调整网络中的连接权值、网络结构，实现输入输出非线性映射关系，对三峡永久船闸中隔墩的变形进行预测。具有全局最优化遗传算法可使连接权值、网络结构得到最优解。将遗传算法与神经网络结合起来进行变形演化特征识别，对三峡永久船闸中隔墩时效变形进行智能预测，应用结果表明该方法具有较高的预测精度。     

改进的BP神经网络在地震砂土液化预测中的应用

地震砂土液化的影响因素具有非线性关系,至今没有形成规范的预测标准。人工神经网络在砂土液化预测中有较好的应用,尤其是BP神经网络,但由于其本身存在缺陷:学习收敛速度慢,易陷入局部极小;遗传算法具有良好的搜索全局最优解的能力。探讨利用遗传算法优化BP神经网络权值和初始阈值来预测地震砂土液化,其效果比传统的BP网络有显著提高。     

神经网络方法在煤层气测井资料解释中的应用

根据煤层气储层及其测井响应特征,提出应用ＢＰ神经网络（简称ＢＰ网络） 方法进行煤层气测井资料的定量解释。通过对华北某地煤层气测井资料的处理,获得了满意的解释结果,证明该方法是有效和实用的。并对应用过程中的若干问题进行了分析讨论     

改进的快速BP算法

ＢＰ网络现在得到了广泛的应用。然而对于学习样本较多、输入输出映射关系复杂的情况，学习速度较慢。我们分析了网络误差曲面特征，提出了快速ＢＰ算法，它可以大幅度地提高学习速度。     

金沙江流域泥石流的组合赋权法危险度评价

针对目前地质灾害危险性评价中对指标权重确定方法的缺陷,使用改进的层次分析法和熵值法确定金沙江流域泥石流的各影响因素权重。在此基础上,提出一种新的组合赋权规则,使两类权重间的差异程度和分配系数间的差异程度相一致,从而排除人为因素及数据奇异点的影响,得到相对客观准确的结果。以金沙江的典型泥石流作为计算实例,对现场数据进行了提取之后,使用上述方法计算出影响因子的权重,将权重代入经典的危险性分析模型进行评估。计算结果与当地泥石流的实际情况和发展趋势相符合,从而证明了该方法的合理性。     

用BP网络求解土体的导热系数

以人工神经网络为基本工具,利用其强大的非线性映射能力,综合考虑土体的物理性质指标对其导热性能的影响,为求解土体的工程参数提供了一条新的途径。结果表明,ＢＰ网络能够充分体现土体物理指标之间的非线性关系和隐含关系,具有较高的求解能力,所求出的土体的导热系数与实测值接近,其精度远高于线性回归,且简单实用。     

基于权重反分析方法的加权距离判别法及应用

距离判别分析法在应用过程中常将各评价指标的重要性视为一致,实际上各指标重要性是不同的,故需对各指标进行加权处理。权重的合理取值对最终的评判结果有着很大的影响,目前尽管确定权重的方法有很多,但都有其不足之处。鉴于此,在马氏距离的基础上,推导了加权马氏距离的求解公式,并采用一对多的分类方法建立了多总体加权马氏距离判别准则。在确定各指标权重时采用了基于优化理论权重反分析的新思路,把直接给权重问题转化为通过实际工程样本反求各影响因素权重问题,该方法不需先知经验,求解过程完全由实际样本驱动,具有很强的可操作性。将改进后的加权距离判别分析方法应用于实际工程中的围岩分类问题,结果表明,该方法的判别结果是可靠的。距离加权的方法完善了距离判别分析法理论,权重反分析方法是一种通用方法,可用于任何求解权重的问题,为权重的求解提供了一种新思路。     

Application of Weights of Evidence to Mineral Potential Mapping of Yujiacun Ore Field in Northwest Yunnan Province, China

The multivariate information conprehensive processing technique is especially important at present to the digital mineral prospecting. However, the GIS-based weights of evidence have provided us with powerful tool for the quantitative assessment of mineral resource potential. In this paper, the mineralization model is established, based on the achievements made by previous researchers, to mend such deficiencies ad few references on ore fields in Yujiacun, Yunnan Province and the shortage of quantitative prediction and assessment of mineral resources. In addition, the weights of evidence are used to make a systematic quantitative prediction and assessment of mineral resources there, so that 2 mineral prospecting target areas of grade Ⅰ and 8 mineral prospecting target areas of grade Ⅱ are delineated, providing the further mineral resource exploration with the basis for the selection of mineral deposits.     

基于独立信息数据波动赋权的泥石流危险度评价

针对经典泥石流危险度评价中以灰色关联度作为次要危险因子筛选依据的不足,采用复相关系数作为泥石流次要危险因子筛选的依据,从而建立了泥石流危险度评价的主要危险因子与次要危险因子组合的多因子体系。在此基础上,针对次要危险与主要危险因子之间存在显著相关性,同时为了体现次要危险因子对泥石流危险度的贡献作用,采用独立信息数据波动赋权法计算泥石流次要危险因子的权重,建立了泥石流危险度评价的计算公式。以此计算东川市12条泥石流沟的危险度评价结果,计算结果比经典方法更符合次要因子的补充地位,从而证明了该方法的合理性。     

藏东南艰险地区灾害地质调查思路与方法论述

地质灾害调查自20世纪90年代以来经历了多轮发展,目前演变为以孕灾背景调查和隐患点识别的灾害地质调查。“藏东南重要城镇和交通干线地质灾害调查”项目工作区位于西藏东南部,有工作程度落后、地质环境脆弱和工作条件恶劣等特点。项目主要从详实基础地质数据、地质灾害规律和孕灾背景调查评价和地质灾害隐患点识别等3个方面重点进行灾害地质调查工作。为了在藏东南特殊的环境中实现这些调查要求,应用了野外调查数据采集系统、InSAR、三维激光扫描、无人机航测遥感系统和简易观测等技术方法。并且总结了各类调查技术方法的适用性。通过总结经验和找到不足,从而为今后该地区的调查评价提供借鉴和参考。     

Calculation of the CIPW norm: New formulas

A completely new set of formulas, based on matrix algebra, has been suggested for the calculation of the CIPW norm for igneous rocks to achieve highly consistent and accurate norms. The suggested sequence of derivation of the normative minerals greatly deviates from the sequence followed in the classical scheme. The formulas are presented in a form convenient for error-free implementation in computer programs. Accurate formulas along with the use of variable molecular weights for CaO and FeO; corrected formula weights for apatite, pyrite and fluorite; and suggested measures to avoid significant rounding off errors to achieve absolute match between the sum of the input weights of the oxides and the sum of the weights of the estimated normative minerals. Using an analogous procedure for determining the oxidation ratios of igneous rocks as used in the SINCLAS system of Verma et al (2002, 2003), the suggested calculation scheme exactly reproduces their results except for apatite for reasons explained in the text, but with a superior match between the totals for about 11,200 analyses representing rocks of a wide range of composition     

Kriging without negative weights

Under a constant drift, the linear kriging estimator is considered as a weighted average ofn available sample values. Kriging weights are determined such that the estimator is unbiased and optimal. To meet these requirements, negative kriging weights are sometimes found. Use of negative weights can produce negative block grades, which makes no practical sense. In some applications, all kriging weights may be required to be nonnegative. In this paper, a derivation of a set of nonlinear equations with the nonnegative constraint is presented. A numerical algorithm also is developed for the solution of the new set of kriging equations.     

A Comparison of Modified Fuzzy Weights of Evidence,Fuzzy Weights of Evidence,and Logistic Regression for Mapping Mineral Prospectivity

Weights of evidence and logistic regression are two of the most popular methods for mapping mineral prospectivity. The logistic regression model always produces unbiased estimates, whether or not the evidence variables are conditionally independent with respect to the target variable, while the weights of evidence model features an easy to explain and implement modeling process. It has been shown that there exists a model combining weights of evidence and logistic regression that has both of these advantages. In this study, three models consisting of modified fuzzy weights of evidence, fuzzy weights of evidence, and logistic regression are compared with each other for mapping mineral prospectivity. The modified fuzzy weights of the evidence model retains the advantages of both the fuzzy weights of the evidence model and the logistic regression model; the advantages being (1) the predicted number of deposits estimated by the modified fuzzy weights of evidence model is nearly equal to that of the logistic regression model, and (2) it can deal with missing data. This method is shown to be an effective tool for mapping iron prospectivity in Fujian Province, China.     

Assessment of aquifer vulnerability to contamination in Khanyounis Governorate, Gaza Strip—Palestine, using the DRASTIC model within GIS environment

Groundwater is a very important natural resource in Khanyounis Governorate (the study area) for water supply and development. Historically, the exploitation of aquifers in Khanyounis Governorate has been undertaken without proper concern for environmental impact. In view of the importance of quality groundwater, it might be expected that aquifer protection to prevent groundwater quality deterioration would have received due attention. In the long term, however, protection of groundwater resources is of direct practical importance because, once pollution of groundwater has been allowed to occur, the scale and persistence of such pollution makes restoration technically difficult and costly. In order to maintain basin aquifer as a source of water for the area, it is necessary to find out, whether certain locations in this groundwater basin are susceptible to receive and transmit contamination. This study aims to: (1) assess the vulnerability of the aquifer to contamination in Khanyounis governorate, (2) find out the groundwater vulnerable zones to contamination in the aquifer of the study area, and (3) provide a spatial analysis of the parameters and conditions under which groundwater may become contaminate. To achieve that, DRASTIC model within geographic information system (GIS) environment was applied. The model uses seven environmental parameters: depth of water table, net recharge, aquifer media, soil media, topography, impact of vadose zone, and hydraulic conductivity to evaluate aquifer vulnerability. Based on this model and by using ArcGIS 9.3 software, an attempt was made to create vulnerability maps for the study area. According to the DRASTIC model index, the study has shown that in the western part of the study area the vulnerability to contamination ranges between high and very high due to the relatively shallow water table with moderate to high recharge potential, and permeable soils. To the east of the previous part and in the south-eastern part, vulnerability to contamination is moderate. In the central and the eastern part, vulnerability to contamination is low due to depth of water table. Vulnerability analysis of the DRASTIC Model indicates that the highest risk of contamination of groundwater in the study area originates from the soil media. The impact of vadose zone, depth to water level, and hydraulic conductivity imply moderate risks of contamination, while net recharge, aquifer media, and topography impose a low risk of aquifer contamination. The coefficient of variation indicates that a high contribution to the variation of vulnerability index is made by the topography. Moderate contribution is made by the depth to water level, and net recharge, while impact of vadose zone, hydraulic conductivity, soil media, and Aquifer media are the least variable parameters. The low variability of the parameters implies a smaller contribution to the variation of the vulnerability index across the study area. Moreover, the “effective” weights of the DRASTIC parameters obtained in this study exhibited some deviation from that of the “theoretical” weights. Soil media and the impact of vadose zone were the most effective parameters in the vulnerability assessment because their mean “effective” weights were higher than their respective “theoretical” weights. The depth of water table showed that both “effective” and “theoretical” weights were equal. The rest of the parameters exhibit lower “effective” weights compared with the “theoretical” weights. This explains the importance of soil media and vadose layers in the DRASTIC model. Therefore, it is important to get the accurate and detailed information of these two specific parameters. The GIS technique has provided an efficient environment for analysis and high capabilities of handling large spatial data. Considering these results, DRASTIC model highlights as a useful tool that can be used by national authorities and decision makers especially in the agricultural areas applying chemicals and pesticides which are most likely to contaminate groundwater resources.