基于GIS的海域定级系统研究

为提高海域定级工作的效率和成果的精度与质量,在海域定级原理研究的基础上,基于GIS技术和数据库技术设计开发了海域定级系统。该系统的主要功能模块包括:因素因子体系的建立;语言型、域值型、数值型、点扩散型和线扩散型影响因子的量化方法;层次分析法计算影响因素因子权重值;海域定级及结果显示。     

逐步判别分析法在筛选泥石流评价因子中的应用

运用逐步判别分析原理,在多个评价因子中按其对泥石流沟判别能力贡献的大小进行筛选,从而确定泥石流沟的评价因子。应用改进的层次分析法对筛选后的评价因子进行赋权,使求得的判断矩阵满足一致性要求,避免了层次分析法后期检验过程中多次调整判断矩阵带来的过于主观性的问题。在评价因子的筛选和赋权的基础上进行了单沟泥石流危险度评价的实例验证,结果表明,运用逐步判别分析法筛选后的评价因子客观、准确地反映了当地泥石流危险度的影响因素,与传统方法所得危险度结果相比更符合泥石流危险度实际情况,可靠性更强。     

双评分准则逐步回归法

阐明了以双评分准则(CSC)作逐步回归的基本原理,通过实例给出了计算方法与步骤。用此法所建模型能同时报好预报对象的数量和类别。     

基于特征筛选与差分进化算法优化的滑坡危险性评估方法

突发性地质灾害危险性评估对灾害防治与风险管理具有重要意义。由于不同地区影响灾害发生的因子各不相同,实际评估过程中难以全面客观地选取适宜的评估因子。机器学习对处理灾害系统的高维非线性问题独具优势,但因模型难以调优而评估效果有限。本文尝试提出一种双向优化的滑坡危险性评估方法：在构建因子敏感性指数开展定量敏感性分析的基础上,结合重要性分析、相关性分析、共线性分析构建四维(Four-Dimensional, 4D)特征筛选法用于评估因子综合优选;为克服模型难以调优的问题,引入差分进化(Differential Evolution, DE)算法优化支持向量机(Support Vector Machine, SVM)与多层感知机(Multi-Layer Perceptron, MLP) 2种推广能力较强的机器学习模型。最后,以福建省滑坡为例,开展评估方法研究。研究表明：4D特征筛选法能更加客观全面地选取适宜性更高的危险性评估因子,从而降低数据维度、减少信息冗余以提升评估模型性能;DE算法对SVM与MLP具有显著的优化效果,有益于增强模型滑坡危险性的评估准确度,DE-SVM、DE-MLP相较于未优化...     

河北数字遥测地震台站避雷系统

雷电对地震遥测台站的正常运行和仪器设备安全造成很大威胁,雷电灾害被“国际十年减灾”组织列为十种最严重的自然灾害之一。本文就雷电的入侵途径和防护方法以及接地系统和屏蔽措施做一些介绍。     

Characterization and source identification of trace elements in PM2.5 from Mira Loma, Southern California

Concentrations of thirty-five trace elements in ambient fine particulate matter (PM_2.5) were measured from September 2001 to January 2002 in Mira Loma, a semi-urban area in southern California. The most abundant species were found to be sulfur (S; 23% of the total trace element concentration), followed by Si, Fe, Ca, and Al (soil-related elements; 51% of the total). In general, total trace element concentrations were found to be significantly higher for the drier months of September and October, compared to December and January. Factor analysis, enrichment factor (EF) analysis, and ratio analysis (Al/Zn) revealed a significant contribution of soil-related sources to the ambient trace elements for PM_2.5 in the study area. Other important contributors to the trace elements in ambient PM_2.5 in Mira Loma included motor vehicle-related emissions (brake pads, lubricant oils, gasoline, and diesel combustion), secondary sulfates, sea salts, and biomass burning. The influence of sea salts on the study area was identified using a backward trajectory analysis.     

蒸馏分离-苯芴酮比色法测定煤中锗

介绍蒸馏分离-苯芴酮比色法测定煤中锗的含量,对测量过程中的影响因素进行分析,提出解决办法。     

边坡危险滑动面及稳定安全系数的力学解析方法

为了分析边坡的稳定性,提出了一种基于极限平衡理论的分析方法。假定坡体为只受重力作用的弹性体,通过弹性力学解析解求得应力分布,将滑动面用分段二次多项式函数表示,当分段间隔足够小时,这种表示方式可以描述任意曲面形状,根据坡体滑面所满足的Mohr-Coulomb准则,定义边坡稳定安全系数,其为多项式系数的函数。通过混和罚函数优化方法可以求出最危险滑动面对应的多项式系数,获得边坡的稳定安全系数,从而判定边坡的稳定性。所提出的方法与以往方法不同,不用将滑体划分成垂直条块,不用假定滑面上的法向应力分布形式,而直接利用已有的坡体弹性应力解。算例表明:本文方法所得到的结果与简化Bishop法和有限元强度折减法获得的结果非常接近,且基本介于有限元法和简化Bishop法之间。     

Determination of Fracture Toughness of Anisotropic Rocks by Boundary Element Method

Summary  This paper presents a systematic procedure for determining fracture toughness of an anisotropic marble using the diametral compression test (Brazilian test) with a central crack on the discs. Additionally, a novel formulation to increase the accuracy in Stress Intensity Factor (SIF) calculations using Boundary Element Method (BEM) is applied to determine the stress intensity factors and the fracture toughness of anisotropic rocks under mixed-mode loading. The numerical results show that the SIFs for both isotropic and anisotropic problems are in good agreement with those reported by previous authors. The marble with clear white-black foliation from Hualien (in eastern Taiwan), was selected for the Brazilian tests. Diametral loading was conducted on the Cracked Straight Through Brazilian Disc (CSTBD) specimens to evaluate their fracture toughness. In addition, a new fracture criterion was developed to predict pure mode I, pure mode II or mixed mode (I–II) fracture toughness of the anisotropic marble. The new fracture criterion is based on the examination of mode I, mode II and mixed mode (I–II) fracture toughness for different crack angles and anisotropic orientation. Authors’ address: Associate Professor Chao-Shi Chen, Department of Resources Engineering, National Cheng Kung University, No. 1 Dasyue Rd., East District, Tainan 701, Taiwan     

利用人工神经网络预报芜湖的雾

雾对人类的活动有着很大的影响，尤其是对军事活动影响更为突出。由于雾是一种小概率事件，准确地定点预报雾难度更大。本文提出了神经网络B－P算法输入层因子的筛选方法，隐层挑选、位置矩阵等特殊处理方法，它们对于网络的稳定学习是非常有效的。通过对芜湖雾日预报试验，取得了较为满意的试报效果。