淮河息县站流量概率预报模型研究

应用美国天气局采用的由Roman Krzysztofowicz开发的贝叶斯统计理论建立概率水文预报理论框架，即以分布函数形式定量地描述水文预报不确定度，研究了淮河息县站流量概率预报模型。理论和经验表明，概率预报至少与确定性预报一样有价值，特别当预报不确定度较大时，概率预报比现行确定性预报具有更高的经济价值。     

Local Modeling模式及其在月径流预测中的应用

Local Modeling方法是一种动力系统预测方法，将其应用于河西内陆区黑河干流出山口莺落峡水文站月平均流量的中长期预测预报，取得了较为理想的成果。预测试验的结果表明，该预测模型有较高计算精度，尤其适用于非主汛期各月的月平均流量的预测；对于主汛期6～9月的月平均流量的预测，在考虑前期来水与预见期内降水的影响后，亦可获到较为理想的预测结果。可以认为，该方法的预报精度达到了水文情报预报规范的要求，Local Modeling方法的应用，将为西北干旱地区河川径流的中长期预报提供了一个新的途径。     

基于数据融合的侧扫声纳图像预处理

论述了基于数据融合的侧扫声纳图像处理的特殊过程。根据现代侧扫声纳系统的特点，提出了利用声线跟踪法进行斜距改正，同时对目标阴影区进行了处理；对文献[1]的灰度不均衡的改正算法进行了改进。利用小波变换检测出灰度突变区，根据剔除突变区后计算的灰度改正系数进行航向上的灰度改正，并用模拟数据和实际数据进行了验证；鉴于波束展宽效应对远场目标造成的拖尾效应，论述了对拖尾效应使用的去卷积的改正算法。     

2004年陕西春季沙尘天气环流特征分析

应用我们已经建立的《阳泉市短期冰雹预报系统》分析预报阳泉市2004年6月份出现的连续冰雹天气,特别是该系统中的指标叠加方法,效果较好。     

2003年9月19日陕西暴雨分析

利用太原地区发生的798个地质灾害个例和汛期降水资料,从地质环境背景着手,在地理信息系统(MAPGIS)支持下,对太原地质灾害危险性区划和诱发因素进行了系统研究。提出了精细的地质环境概率量化评价方法,得出了地质灾害影响因素的重要结论。建立了地质灾害预报模型,制定了预报等级标准和预报规则。经业务试验,预报结果与实况基本吻合。     

指数平滑技术在斜坡位移预测中的应用

应用趋势型数据指数平滑模型预测了链子崖危岩体监测点GA的位移量.根据其观测数据呈线性趋势的特点,选取趋势型二次指数平滑的线性预测公式和合适的平滑常数值进行了计算预测.预测结果表明位移预测值与实际观测值之间的误差很小,说明该模型可很好地应用于斜坡变形位移的预测.     

A model for moisture and heat flow in fractured unsaturated porous media

The present study investigates propagation of a cohesive crack in non‐isothermal unsaturated porous medium under mode I conditions. Basic points of skeleton deformation, moisture, and heat transfer for unsaturated porous medium are presented. Boundary conditions on the crack surface that consist of mechanical interaction of the crack and the porous medium, water, and heat flows through the crack are taken into consideration. For spatial discretization, the extended finite element method is used. This method uses enriched shape functions in addition to ordinary shape functions for approximation of displacement, pressure, and temperature fields. The Heaviside step function and the distance function are exploited as enrichment functions for representing the crack surfaces displacement and the discontinuous vertical gradients of the pressure and temperature fields along the crack, respectively. For temporal discretization, backward finite difference scheme is applied. Problems solved from the literature show the validity of the model as well as the dependency of structural response on the material properties and loading. Copyright © 2016 John Wiley & Sons, Ltd.     

Extended finite element framework for fault rupture dynamics including bulk plasticity

We present an explicit extended finite element framework for fault rupture dynamics accommodating bulk plasticity near the fault. The technique is more robust than the standard split‐node method because it can accommodate a fault propagating freely through the interior of finite elements. To fully exploit the explicit algorithmic framework, we perform mass lumping on the enriched finite elements that preserve the kinetic energy of the rigid body and enrichment modes. We show that with this technique, the extended FE solution reproduces the standard split‐node solution, but with the added advantage that it can also accommodate randomly propagating faults. We use different elastoplastic constitutive models appropriate for geomaterials, including the Mohr–Coulomb, Drucker–Prager, modified Cam‐Clay, and a conical plasticity model with a compression cap, to capture off‐fault bulk plasticity. More specifically, the cap model adds robustness to the framework because it can accommodate various modes of deformation, including compaction, dilatation, and shearing. Copyright © 2013 John Wiley & Sons, Ltd.     

L1范数与IQR统计量组合的GNSS坐标序列粗差探测算法

GNSS坐标时间序列中不可避免地含有粗差,未剔除的粗差将会导致参数估计有偏。因此,粗差探测与剔除是GNSS坐标序列分析中一项重要的数据预处理工作。针对GNSS坐标时间序列特点,提出了一种将L1范数(L1-norm)估计与四分位距统计量IQR(interquartile range)组合的移动开窗粗差探测算法,称之为L1_Mod IQR。该方法的主要思想是,首先利用L1范数估计得到较"真实"的残差,然后再对残差采用IQR统计量进行粗差探测。将L1_Mod IQR法与"3σ"法、基于最小二乘的τ检验法等粗差探测算法进行了模拟计算与对比,验证了该算法的有效性。进一步采用L1_Mod IQR算法对中国区域10个IGS站的高程时间序列进行了分析,结果表明中国区域IGS站高程序列的粗差剔除率最小为0.1%,最大为2.6%。并且以WUHN站为例与SOPAC提供的结果进行了对比,结果表明SOPAC提供的"Clean"数据仍含有大量的粗差,而L1_Mod IQR算法能够有效地剔除粗差。