PSO-RBFNN模型及其在岩土工程非线性时间序列预测中的应用

岩土工程受力变形演化是一个典型的非线性问题,其演化的高度非线性和复杂性,很难用简单的力学、数学模型描述,但可用粒子群优化径向基神经网络对岩土工程应力、位移非线性时间序列进行动态实时预测。网络径向基层的单元数通过均值聚类法确定后,所有其它参数：中心位置、形状参数、网络权值,均通过粒子群优化算法在全局空间优化确定。工程实例应用表明,该模型预测结果准确、精度高,有良好的应用前景。     

台湾海峡及其西边地区正常地震动态及危险性特征

本文从地震的时，空分布特征，区域应力场动态，能量释放方式，ｂ值及震群特征６个方面研究了台湾海峡及其西边地区地震活动的正常动态及异常特征。结果表明，具有前兆意义的变化模式表现为区域地震活动在时间，空间及功能方面的有序性变化，即：（ａ）地震空间分布由分散转为集中，形成条带或空区；（ｂ）断裂活动由多组转为单一，应力场趋向一致；（display status     

基于免疫粒子群优化算法的影像纹理分类

粒子群优化算法是基于群智能的随机全局优化方法,它源于对鸟群简化社会系统的模拟。为了提高标准粒子群优化算法的收敛性能,将生物免疫系统的记忆能力和多样性引入标准粒子群优化算法,提出一种免疫粒子群优化算法。在提取纹理样本Laws纹理能量模板特征、小波特征等纹理特征的基础上,提出针对分类问题的粒子表达方法和群体寻优策略,实现了基于免疫粒子群算法的纹理分类。实验结果表明,与标准粒子群优化算法相比,免疫粒子群优化算法在获取训练样本类别中心时具有较好的收敛性能,并且基于该算法的影像纹理分类具有较高的分类精度。     

A combined study of macroseismic data and focal mechanisms applied to the West-Bohemian earthquake,Czechoslovakia

The main shock of the West-Bohemian earthquake swarm, Czechoslovakia, (magnitudem=4.5, depthh=10 km) exhibits an irregular areal distribution of macroseismic intensities 6° to 7° MSK-64. Four lobes of the 6° isoseismal are found and the maximum observed intensity is located at a distance of 8 km from the instrumentally determined epicentre. This distribution can be explained by the energy flux of the directS wave generated by a circular source, the hypocentral location and focal mechanism of which are taken from independent instrumental studies. The theoretical intensity, which is assumed to be logarithmically proportional to the integrated squared ground-motion velocity (i.e.,I=const+log v ² (t)dt), fits the observed intensity with an overall root-mean-square error less than 0.5°. It is important that the present intensity data can also be equally well explained by the isotropic source. The fit was attained by means of a horizontally layered model though large fault zones and an extended sedimentary basin suggest a significant lateral heterogeneity of the epicentral region. The results encourage a broader application of the simple modelling technique used.     

The seismic hazard assessment of the Dead Sea rift, Jordan

The Dead Sea fault system and its branching faults represent one of the most tectonically active regions in the Middle East. The aim of this study is to highlight the degree of hazards related to the earthquake activities associated with the Dead Sea rift, in terms of speculating the possible future earthquakes. The present investigation mainly is based on available data and vertical crustal modeling of Jordan and the Dead Sea model for the Dead Sea basin with particular emphasis of the recent earthquake activities, which occurred on December 31st, 2003 (Mc = 3.7), February 11th, 2004 (strongest Mc = 4.9 R), and March 15th, 2004 (Mc = 4). The present research examines the location of the strong events and correlates them with the various tectonic elements in the area. The source mechanism of the main shock and the aftershock events is also examined. The analyses were based on the available short period seismogram data, which was recorded at the Natural Resources Authority of Jordan, Seismological Observatory. The seismic energy appears to have migrated from the south to the north during the period from December 31st up to March 12th, where the released seismic energy showed a migration character to the southern block of the eastern side of the Dead Sea, which led the seismic event to occur on March 15th.     

A laboratory experiment to monitor the contact state of a fault by transmission waves

We performed a series of laboratory experiments in which elastic waves were transmitted across a simulated fault. Two types of experiments were carried out: (1) Normal Stress Holding Test (NSHT): normal stress was kept constant for about 3 h without shear stress and transmission waves were observed. (2) Shear Stress Increasing Test (SSIT): shear stress was gradually increased until a stick-slip event occurred. Transmission waves were continuously observed throughout the process of stress accumulation. We focused on the change in transmission waves during the application of shear stress and especially during precursory slips.It was found in NSHT that the amplitude of transmission waves linearly increased with the logarithm of stationary contact time. The increase amounted to a few percent after about 3 h. Creep at asperity contacts is responsible for this phenomenon. From a theoretical consideration, it was concluded that the real contact area increased with the logarithm of stationary contact time.We observed in SSIT a significant increase in wave amplitude with shear stress application. This phenomenon cannot be attributed to the time effect observed in NSHT. Instead, it can be explained by the mechanism of “junction growth” proposed by Tabor. Junction growth yields an increase in real contact area. It is required for junction growth to occur that the material in contact is already plastic under a purely normal loading condition. A computer simulation confirmed that this requirement was satisfied in our experiments. We also found that the rate at which the amplitude increased was slightly reduced prior to a stick-slip event. The onset time of the reduction well coincides with the onset of precursory slip. The cause of the reduction is attributed to the reset of stationary contact time due to displacement. This interpretation is supported by the result of NSHT. Taking the time of stationary contact in SSIT into account, we may expect the change in wave amplitude to be, at most, only a few percent. The observed slight reduction in increasing rate is, in this sense, reasonable. The static stiffness of the fault also decreases with precursory slip. It was also found that low frequency waves are a better indicator of precursory slip than high frequency waves. This might suggest that low frequency waves with longer wavelength are a better indicator of average behavior of faults. The problem, however, merits a further investigation. The shifts in phase were also found to be a good indicator of the change in contact state of the fault. The changes in both amplitude and phase of transmission waves are unifyingly understood through the theory of transmission coefficient presented by Pyrak-Nolte et al. Rough surfaces have a tendency to give larger stick-slips than smooth surfaces. The amount of precursory slip is larger for rough surfaces than for smooth surfaces. Although it was confirmed by a computer simulation that rough surfaces have larger contact diameters than smooth surfaces, the rigorous relationship between the surface roughness (contact diameter) and the amount of precursory slips was not established.     

江西瑞昌1995年ML4.9级地震前后九江地震台倾斜固体潮汐变化分析

通过月整点值数据文件和滑动值数据文件这两种数据处理形式。对九江台地倾斜资料进行调和分析计算。分析了M2潮汐因子与瑞昌ML4.9级地震的关系。认为存在震前短临异常或震后异常，并与趋势异常得到相互印证。     

新疆伽师强震群的震源破裂特征

利用国内外短周期P波初动符号及全球数字地震台网 (GDSN)的宽频带数字记录资料 ,研究了 1997年 1月 2 3日至 1997年 11月 4日伽师强震群的震源机制解和伽师强震群中的 5个强震、后续地震及其周边地震的震源破裂过程。从震源机制解来看 ,伽师强震群主要有走滑和正倾两种破裂类型 ,主压应力轴方向主要为NNE向和近垂直 ,而主张应力轴为NW向并近水平 ,与区域构造应力场方向存在差异 ,具有明显的局部特征。从破裂过程看 ,伽师强震群的破裂过程相对简单 ,破裂面积不大 ,上升时间较短 ,是由一点向四周快速扩散的脆性破裂 ,无明显的伸展方向 ,与阿图什地震完全不同。研究结果表明 ,伽师强震群与震源附近的地壳结构在垂向和横向上的非均匀变化有着密切关系 ,而阿图什地震与塔里木盆地的现今构造运动关系密切。伽师强震群是在震源区附近地壳上部垂向和水平向力共同作用下发生的多次沿NNE向的快速脆性破裂 ,从而形成了以张性破裂和左旋走滑为主的震源特征     

喀什"奇异震群"与构造地震的对比研究

20 0 1年 11月 14、 15日新疆喀什台记录到 86次振动事件 (“奇异震群”事件 )。为识别其性质 ,应用波谱分析、偏振分析等方法 ,对该事件和同范围构造地震的地震波特征进行了对比研究 ,发现 :(1)喀什“震群”的全波列衰减比构造地震的全波列衰减快很多 ;(2 )与构造地震的速度谱相比 ,“震群”的频谱相对较为“集中”,而构造地震的则较为“发散”;(3)“震群”位移谱的拐角频率和卓越频率均较构造地震的相应量高出一倍多 ,且离散较小。通过对喀什“奇异震群”这些特征进行分析研究 ,认为这一事件极有可能是人工事件     

Temporal variations of shear-wave splitting in field and laboratory studies in China

Observations of shear-wave splitting at seismic stations above a swarm of small earthquakes on Hainan Island, China, and other examples world-wide, suggest that the time-delays of split shear-waves monitor the build up of stress before earthquakes and the stress release as earthquakes occur. Rock physics experiments on marble specimens also show variations of shear-wave time-delays with uniaxial pressure analogous to the field observations. The rock experiments show an abrupt decrease in time-delays immediately before fracturing occurs. Similar precursory behaviour has been observed before earthquakes elsewhere, and is believed to be important for two reasons. Precursory changes in shear-wave splitting could be used for short-term forecasting, but of greater importance may be the information such behaviour provides about the source processes in earthquake preparation zones.