岩性储集层的混沌识别技术研究

本文从动力学系统原理出发,论述了混沌理论中的李指数法,并从地震剖面上提取李指数,并应用李指数来区分和识别储集层的岩性变化特征和含油气性。计算和应用研究表明,李指数随着相空间维数的增加容易趋于稳定值,这就使它预测地震序列横向变化的可靠性增大,这正是李指数预测岩性和含油气性的一个优越性。成功地应用实例表明,混沌识别技术是岩性储集层的一种有效识别技术,并具有广阔的应用前景。     

周期基数插值对偶尺度函数和对偶小波

在对二维周期基数插值小波的尺度函数和小波函数研究的基础上,对二维周期基数插值小波的对偶尺度函数和对偶小波进行了研究,得到了一系列重要的结论。     

Integrated evaluation of urban groundwater hydrogeochemistry in context of fractal behaviour of groundwater level fluctuations

The present study was carried out as part of a complex survey of urban groundwater quality and quantity in Szeged, southeast Hungary. The concentrations of 12 inorganic contaminants in 28 shallow groundwater monitoring wells were determined over a 2-year period (2010–2012). The evaluation of concentrations indicates remarkable contamination all over the city. Discriminant analysis (DA) was used to evaluate the spatial changes of groundwater quality. The groundwater levels were measured over a 14-year period (2000–2013). The fractal properties of water level fluctuations time series characterizing the groundwater system and Tisza River were investigated using rescaled range (R/S) analysis. The resulting Hurst exponents clearly showed the persistency and thus long memory effects of both the groundwater and the river flow. Comparison of the results of DA with the results of R/S analysis thus implies that the geological conditions and the changing groundwater quantities are not related to groundwater quality.

EDITOR A. Castellarin

ASSOCIATE EDITOR A. Fiori     

云南地区中小地震静力学和动力学参数定标关系

Source spectra for moderate and small earthquakes are obtained after removing the path effect, site effect, and instrument response, etc. in the observed S-wave spectra. Based on the Brune source model...     

��������������ʱ�մԼ���������μ�R/S����

??????????κ??????R/S????????????????????????????M??4.5????1970??2010??????M??1.0?????з???????????????????????????????????????????????????????е?H???H=0.79????????????е?H?(H=0.87)???????????????????????????ν???     

Multifractal characteristic analysis of near-fault earthquake ground motions

This study aims to reveal the multi-scaling behavior and quantify the irregularity of near-fault earthquake ground motions from a new perspective of multifractal theory. Based on multifractal detrended fluctuation analysis, the multifractal characteristic parameters of acceleration time series for typical near-fault ground motions are calculated, and their correlations with two period parameters (i.e., mean period T_m and characteristic period T_c) and box-counting fractal dimensions are analyzed. Numerical results of strong nonlinear dependence of generalized Hurst exponents h(q) upon the fluctuation orders q indicate that near-fault ground motions present the multifractal properties and long-range correlation obviously. Furthermore, the scaling exponent h(2) of near-fault records has a strong correlation with their periods T_m and T_c, and strongly negative correlation with their box dimension. Moreover, h(2) can be regarded as a measure of frequency content and irregularity degree of strong earthquake ground motions. Finally, it is pointed out that the long-range correlation of small and large fluctuation is the major source of multifractality of near-fault ground motions.     

Evaluation of codified elastic design spectrum models for regions of low-to-moderate seismicity

Design spectrum (DS) model is typically specified in a seismic code of practice for structural design. In a region of low-to-moderate seismicity where seismic code does not exist, a DS model in a well established code of practice is usually adopted, while the suitability of such model has seldom been evaluated. In this article, the elastic DS models for reference (rock) site stipulated in six major codes of practice (AS1170.4–2007, EN1998-1:2004, GB50011–2010, IBC–2012, NBCC–2010 and NZS1170.5:2004) have been compared and scrutinized. Three cities of low-to-moderate seismicity, namely, Melbourne (Australia), Hong Kong (China) and Karlsruhe (Germany), have been selected for illustrative purposes. Particular emphasis has been put on the parameterization scheme for DS model. It is found that huge discrepancies (over 100%) exist among the models, especially at the long period range, due to differences in spectral shapes and the recommended corner periods, which would lead to undesirable effects on the use of the displacement-based seismic design approach. It is urged that the values of corner periods should be determined specifically and cautiously based on the regional seismicity pattern and local geological conditions.     

Dimensionless input parameters in discrete element modeling and assessment of scaling techniques

A set of dimensionless input parameters were defined for DEM using a characteristic time which is a function of density and elastic modulus of particles and an arbitrary characteristic length. Dimensionless strain rate and mass damping ratio are inversely proportional to the characteristic time, and stress is normalized by elastic modulus to give dimensionless stress. It was demonstrated that the response of a model in the dimensionless scale is invariant with the choice of density, elastic modulus and the characteristic length if dimensionless strain rate and mass damping ratio are kept constant. Small time step is a prohibitive aspect of DEM. Scaling techniques are widely employed to enlarge the time step. Using the dimensionless scheme, it was learned that density scaling is equivalent to the use of a higher strain rate, and stiffness scaling results in a higher strain rate and an elevated stress state in the dimensionless scale.     

基于杨赤中推估法空间插值的三维地质隐式建模

杨赤中推估法是一种对空间域复合变量通过连续的几何滤波过程来建立核函数的最小二乘推估法,建模过程简便且能基于少量已知数据点取得好的建模效果。针对地质勘查中离散、稀疏而不规则分布的地质特征点数据难以构建地质体三维模型的难点问题,提出了一种基于杨赤中推估法的三维地质空间插值与模型建立的自动化方法。该方法首先以地质特征点数据库为基础,选用负幂指数函数模型建立适合三维地质空间插值的杨赤中推估法估值数学模型;在此基础上,构建一套基于杨赤中推估法的三维地质空间插值计算和地质体隐式建模的自动化实现流程与程序;最后以实例矿化插值数据为基础,采用基于移动立方体算法的三维隐式建模方法,快速构建实例矿体三维模型。与人工交互圈定地质体边界和进行矿体推断的三维地质显式建模相比,这种方法能快速直观地分析地质特征并处理样品分析数据,方法可行且高效。     

北京地区气候变量的多分形特征研究

在气候系统的变量时间序列中往往存在着不同尺度间的自相似结构,这种自相似结构,又称为分形的特征.本文运用多分形去趋势波动分析方法(MF-DFA),分析北京市近50年来不同气象变量的逐日序列,并用一个扩展的二项式串级模式来分别估计其多分形谱.结果表明,平均气温等变量表现出多重分形的特征,并且多分形谱宽一致.而气温日较差和日照时数则表现出单分形的特征.且这种分形特征与长程记忆性相关,为中长期气候预测提供了理论基础.