一种基于M估计的Robust-ELM滑坡位移预测方法

采用传统ELM算法进行滑坡位移预测时,其网络输出权值由最小二乘估计得出,导致ELM抗差能力较差,从而造成网络训练参数不准确。为此,将M估计与ELM相结合,提出一种基于M估计的RobustELM滑坡变形预测方法。该方法利用加权最小二乘方法来取代最小二乘法计算ELM输出权值,以减少滑坡监测数据中粗差对ELM预测的干扰。分别以链子崖、古树屋滑坡体为例,将Robust-ELM进行了单维、多维粗差的抵御性验证。结果表明,该方法能够有效降低粗差对预测的影响,具有良好的抗差能力。     

Causes of rockfalls in Al-Huwayshah area, Yemen

According to topography of Yemen, most areas and villages are located at obligated crest, toe of mountain and under cliffs. Therefore Al-Huwayshah consisting of Tawilah sandstone group is characterized by steep slope reach to 90° in some areas. This area is affected by strong tectonic movements and faults that occurred during the geological epochs. This effect enhances to find out fractures and joints as well as the rocks become brittle and ready to slide depending on the position of area. And there are some fractures and joints on the surface of the areas due to tectonic movement associated with opening the Red Sea and Gulf of Aden. The rainfalls, fractures, joints, earthquakes, gravity, vegetations, temperatures and human activities play a big role in the processes of rockfalls and landslides in this area. All those factors are considered as the basic causes and catalyzed factors for occurrence of rockfall in the studide area. In addition, the authors carried out laboratory test for many rock samples to get the physical properties of the rocks.     

Causes of rockfalls in Al-Huwayshah area,Yemen

According to topography of Yemen,most areas and villages are located at obligated crest,toe of mountain and under cliffs.Therefore Al-Huwayshah consisting of Tawilah sandstone group is characterized by steep slope reach to 90° in some areas.This area is affected by strong tectonic movements and faults that occurred during the geological epochs.This effect enhances to find out fractures and joints as well as the rocks become brittle and ready to slide depending on the position of area.And there are some frac...     

基于TLS的滑坡形变分析方法研究

以四川理县一地质滑坡为例,分析扫描点空间位置的理论精度,探讨点云数据获取以及点云过滤、点云拼接、地理参考等数据预处理的关键技术,给出一套比较直观、全面的滑坡形变分析方法,包括基于点云比较的形变分析、基于TIN比较的形变分析、基于特征点的形变分析以及基于DEM的形变分析。为了充分利用非地面点云数据,提出一种提取滑坡区域电线杆、树干的几何特征来分析滑坡水平变化的新方法。通过分析得到的滑坡形变量与GNSS远程实时监测结果基本一致。     

滑坡变形预测灰色神经网络耦合模型的构建及适用性分析

结合灰色模型和神经网络的数据处理特点,提出串联、并联和混联式3种结构的灰色神经网络滑坡变形预测模型。串联式将滑坡变形位移时序分解为趋势项和随机项,采用灰色模型提取滑坡位移时序趋势,利用神经网络逼近随机波动;并联式以灰色模型和神经网络分别对滑坡预测,采用智能非线性组合,按照预测目标精度动态调整权重,从而获取最终组合预测结果;混联式通过增加灰白化层及灰模型群,对神经网络拓扑结构进行优化,达到弱化滑坡原始监测数据随机性、提高预测模型稳健性的目的。将3种模型应用于古树屋滑坡变形预测,并对其适用性进行讨论。结果表明,3种结构的灰色神经网络耦合模型均提高了预测精度,适用于复杂状况下滑坡体的变形预测。     

变形预测的逆多元统计模型

测量和预测目标点的空间变形是掌握变形体稳定性的重要手段,本文论证了逆多元统计预测模型用于滑坡的稳定性分析,利用变形点的空间三维移动量模拟时间的变化特征,与正常的时间序列进行对比,检验变形体的稳定性。并结合实例说明了在滑坡变形分析中逆多元统计模型的实用性和可靠性。     

基于无人机倾斜摄影的黄土滑坡信息多维提取与应用分析

本文针对黄土地区地质灾害调查与评估需求,探索了一种基于无人机倾斜摄影影像和飞控数据的滑坡单体信息多维提取的方法.通过影像处理获得的DEM、DOM和实景三维模型等多维数据,对贾家村滑坡坡形、坡度、相对高差、植被覆盖率、纵坡比降、表面积与物源体积进行定量提取,结合区域气象数据分析了贾家村滑坡年均降雨量时空分布规律,利用相关...     

A temporal model for landslide risk based on historical precipitation

Of the recognized nonsteady-state factors that influence slope stability, probably most critical in many field situations is the character of precipitation and infiltration activity. A groundwater response model used in conjunction with precipitation records can provide a historical catalog of estimated maximum groundwater levels in a particular study area. An extreme-value statistical analysis of this catalog is linked with geotechnical slope stability analyses to provide a landslide hazard model for estimating the probability of slope failure within a given time. This modeling approach can provide meaningful input to risk assessments for landslide mitigation programs and to decision analyses and cost-benefit studies important for land-use planning and resource management.This paper was presented at Emerging Concepts, MGUS 87 Conference, Redwood City, California, 13–15 April 1987.     

An approach to mass movements in the Jiuzhaigou catchment

This paper presents a preliminary study of the main types, features, genesis, and countermeasures of mass movements in the Jiuzhaigou catchment, which is one of the most scenic regions in China. The main features are debris flows, rock-segment flows, landslides, and avalanches. Rough relief, fractured rocks, drifts, steep slopes, intensive rainstorms, and earthquakes produced their geneses. Countermeasures should include the integration of engineering and biological projects.     

滑坡灾害预报的非线性动力学方法

基于现代反演理论,建立了滑坡灾害预测的非线性动力学模型,给出了可预报时间尺度的确定方法及稳定性判断准则。通过实例分析表明,非线性动力学分析方法是行之有效的。