速度倒数法滑坡预警模型流变试验研究

滑坡失稳的预测预报研究是地质工程领域中的一项重要课题,准确地确定预测预报理论模型的参数是实际应用中的难点。在实际滑坡监测中通常可以观察到位移曲线呈现阶梯形,这些阶梯形位移变化点就是滑坡的变形突变点。为研究滑坡变形突变点的变形特征,进行了不同荷载作用下的天然试样以及不同荷载、不同含水率作用下的浸水试样的流变试验,得到了累计位移-时间曲线以及变形速度-时间曲线。依据秦四清的锁固段理论以及速度倒数法滑坡预警模型对试验结果进行分析。研究结果表明：荷载和含水率的变化对模型参数没有影响,模型参数是关于材料属性的函数;变形过程中突变点的变形特征与破坏时的变形特征相似,并且速度倒数法预警模型在突变点和破坏点确定的模型参数基本一致。因此,滑坡监测曲线中早期位移突变点确定的模型参数可以用于确定滑坡破坏时的预警模型。     

基于Verhulst模型的滑坡位移预测研究及其程序化实现

以甘肃省黄茨滑坡位移时间预测为例,在滑坡工程地质条件、成因、发生与发展过程分析的基础上,结合地面监测桩以及位移计监测的位移时间数据,运用Verhulst预测模型建立了该滑坡位移预测研究的思路.在此基础上,运用Ex-cel内嵌的VBA语言编写了相应的位移时间预测预报程序,解决了笔算困难问题.通过具体实例分析,将Verhulst模型、灰色GM(1,1)模型预测结果与实际监测结果进行对比分析,验证了该模型在滑坡位移时间预测中的适用性以及程序的可靠性.研究结果表明,Verhulst预测模型适宜于滑坡临滑预报,而灰色GM(1,1)预测模型适宜于滑坡中短期预测预报,通过Ver-hulst模型预测黄茨滑坡的临滑时间在1995-01-26至1995-01-27之间,预测结果与滑坡实际滑动时间较为一致,由此说明运用Verhulst预测模型对滑坡进行临滑预报是可行的.     

滑坡的变形破坏行为与内在机理

自20世纪60年代日本学者斋藤借助于蠕变试验成果进行滑坡预测预报以来,人们就一直不停地对斜坡变形破坏行为和滑坡预报方法进行研究和探索,先后提出了数十种滑坡预测预报模型和方法。随着滑坡变形监测实例的不断增多,其变形监测资料越来越丰富,各式各样的变形-时间曲线相继产生。斜坡变形-时间曲线的类型、特征以及形成这些变形-时间曲线的力学条件等诸多问题都是滑坡预警预报必须查明的最基本科学问题。本文通过对各类滑坡变形破坏行为和变形-时间曲线的分析总结,结合岩土体流变试验成果,根据斜坡变形-时间曲线特征,将滑坡分为稳定型、渐变型、突发型3类,并给出了产生这3类变形行为的力学条件。同时,从细观力学的角度分析认为,斜坡产生宏观变形破坏行为的主要原因是岩土体细观尺度颗粒的"流动"和"微破裂",但在不同岩性组成的斜坡和同一斜坡的不同变形阶段,"流动"和"微破裂"将分别发挥不同的作用。     

滑坡预报的多元回归分析方法

在回顾了滑坡预报的概念、类型及已有滑坡预报的几种主要模型的基础上，利用多元回归分析方法，根据最小二乘法原理，建立了非线性回归预报模型，提出滑坡破坏的时间为位移．时间曲线的拐点，此方法一般适合于临滑预报。结合实例，利用黄龙西村滑坡位移．时间监测资料，采用该模型进行滑坡时间预报，和其他滑坡预报模型预报的结果接近，说明该模型具有一定的合理性。     

滑坡预测预报研究现状与发展趋势

文章在对国内外滑坡预测预报研究成果深入分析的基础上，从滑坡发生时间、滑坡活动强度、滑坡危害三方面系统总结了滑坡预测预报研究的现状，评述了滑坡预测预报研究中的主要问题、各种研究方法的优缺点、适用范围和有效性及各国研究的特点。进而，分析预测了滑坡预测预报研究的发展趋势和未来研究的突破口。     

基于数值模拟的三峡库区典型堆积层滑坡变形预测方法

目前对堆积层滑坡的变形预测大多基于数学模型或方法,忽略了引起滑坡位移显著变化的动力外因及滑坡自身的地质特征,因此,预报准确度和可信度较低。以三峡库区典型堆积层滑坡--鹤峰场镇滑坡为例,通过4组主要控制因素科学组合构建了滑坡的基本地质模型;以此为基础,重点考虑引起滑坡发生变形的库水作用动力因素,建立滑坡的数值-力学模型。通过实际监测点的变形监测结果与数值-力学模型中模型监测点的变形进行拟合分析,获取了实际时间与数值-力学模型中时步的等效关系;基于时间-时步等效关系及三峡水库设计水位调度曲线,得到了不同时步水位的波动特征;通过时步的外延,并在相应的时步段对数值-力学模型施加等效时间的库水作用,预测了滑坡在未来库水位变动条件下的变形。该预测方法既考虑了滑坡的工程地质模型又考虑了地下水作用效应,克服了纯数学方法预测的不足。     

滑坡预测预报中改进的Pearl模型及其应用

针对选点法和变换法在求取Pearl模型参数时误差较大的缺陷,且预报判据缺乏足够论证,可能导致预测预报准确度较低问题,将非线性拟合确定参数的方法应用到Pearl模型的滑坡预测预报中,推导了Pearl模型速度和加速度最大值判据预报滑坡发生时间的计算公式。将非线性拟合确定参数的方法和推导的判据计算公式应用于一些滑坡实例的预报中,结果表明:滑坡实际发生时间介于Pearl模型速度最大值判据和加速度最大值判据的预报时间之间,用Pearl模型加速度最大值判据能起到提前预报的作用。     

三峡库区滑坡综合预报系统的设计与实现

由于滑坡变形演化过程的复杂性、随机性和不确定性，大大增如了滑坡预报的难度。文章提出的预报系统是将20多种滑坡预报模型集成滑坡预报库，利用专家系统将定量、定性预报有机结合实现综合预报，并根据更新的监测资料对预报参数、预报结果进行修正和完善。随着时间的延伸和监测资料不断更新，预报准确度越来越高，实现动态追踪预报。整个系统基于GIS平台开发，利用了GIS本身所具有的强大空间数据和图形管理功能，同时还具有数据维护与数据处理等功能。     

陕西泾河南岸大堡子高速远程黄土滑坡运动过程模拟

高速远程滑坡具有速度快、运程远的特点,一旦发生,人员逃离十分困难,往往造成严重灾难,因此对这类滑坡展开研究具有重要意义。对已发生滑坡运动过程进行参数反演和模拟可为潜在高速远程滑坡的预测提供借鉴。因此本文以陕西泾河南岸大堡子高速远程黄土滑坡为例,在野外调查测绘和室内试验的基础上,利用Sassa的滑坡运动模型对该滑坡运动过程进行模拟分析。结果表明,滑坡体最大平均速度为9.56m·s-1,具有高速运动特点,运动持时为24.5s,滑坡运动过程可分为启动加速阶段（0~5.7s）和运动减速阶段（5.7~24.5s）,加速阶段平均加速度为1.68m·s-2,减速阶段加速度为-0.51m·s-2,加速过程比减速快3倍左右。滑坡区高陡地形和黄土高结构强度为滑坡高速运动提供条件,开阔阶地地形以及阶地砂砾石层近饱和含水条件决定滑坡远程运动特点。     

基于滑坡等速变形速率的临滑预报判据研究

前人根据滑坡的变形监测曲线,分别提出了以位移速率、位移切线角、改进切线角等参数作为滑坡预警判据的方法,但因滑坡个性特征差异明显,采用统一固定判据是不充分的。在改进切线角和岩土体材料蠕变强度特征点的基础上,提出了T-t曲线转换的适用范围。以西原模型为基础,研究了蠕变参数在等速变形阶段发展至恒定值或稳定值,等速变形速率可看作滑坡蠕变过程中各参数的综合外在表现,与临滑切线角存在相关性,并呈反比关系。通过对16个典型蠕变滑坡进行阶段划分和T-t曲线转换,获得滑坡等速变形速率和临滑切线角,并建立两者之间关系,相关系数较高,同时结合未破坏滑坡变形过程的最大切线角,以下限作为滑坡预警判据,保证了预测预报的可靠度。     

滑坡动态力学监测及破坏过程案例分析

边坡临滑预警一直是地灾研究的难点与热点问题。本文采用一种力学监测方法(牛顿力监测)对雅安宝兴县某傍山公路边坡进行监测,该边坡位于唐包滑坡老滑坡体下缘边界处。经过4个月的连续监测,获取了大量监测数据,并成功预报了两次局部滑坡。本文首先整合牛顿力监测数据和降雨量监测数据,再将监测曲线与滑坡演化过程进行对比分析,揭示滑坡过程中的力学演化规律,对降雨条件下诱发滑坡的原因进行了分析。然后对牛顿力监测预警成功的案例其临滑预警时长与滑坡体量间的关系进行拟合,发现存在明显的正相关关系。最后讨论了牛顿力监测方法与斋藤模型之间的关联性以及优劣势,并根据各自的特点提出了由面到点的监测预警思路。通过分析,牛顿力监测曲线与滑坡演化过程能较好对应,并可将土质滑坡分为3阶段:(1)牛顿力上升阶段; (2)牛顿力突降阶段; (3)滑坡阶段。本文为牛顿力监测系统的推广提供了实践经验,并为力学监测与位移监测结合的研究提供一个新的思路。     

滑坡位移-时间曲线特征的物理模拟试验研究

位移-时间曲线表征了滑坡变形状态及相对的稳定性程度,是监测滑坡变形和失稳预测至关重要的数据资料。目前所获取的滑坡变形全过程监测资料并不多,大多数是临滑前的短期监测资料,而一些突发型滑坡的变形监测资料更是难以收集。由此,滑坡位移-时间曲线簇的类型及分布特征的研究受到了数据资料不足的约束而难以开展,一些滑坡预警理论都是基于推断和理论分析的基础上而得出的,缺少数据支撑。本文通过物理模拟的办法,自行设计制作了变角度的物理模拟试验框架,开发完成了位移数据的自动采集系统,经一年多的试验测试,对试验设备软硬件及模型材料不断调整、优化,完成了7组12条位移-时间曲线的采集工作,从而获取了不同受力条件下斜坡的变形破坏的时间-位移曲线簇,为进一步的理论研究提供了科学的数据资料。     

Characteristics of a rapid landsliding area along Jinsha River revealed by multi-temporal remote sensing and its risks to Sichuan-Tibet railway

The Sichuan-Tibet railway goes across the Upper Jinsha River, along which a large number of large historical landslides have occurred and dammed the river. Therefore, it is of great significance to investigate large potential landslides along the Jinsha River. In this paper, we inspect the deformation characteristics of a rapid landsliding area along the Jinsha River by using multi-temporal remote sensing, and analyzed its future development and risk to the Sichuan-Tibet railway. Surface deformations and damage features between January 2016 and October 2020 were obtained using multi-temporal InSAR and multi-temporal correlations of optical images, respectively. Deformation and failure signs obtained from the field investigation were highly consistent. Results showed that cumulative deformation of the landsliding area is more than 50 cm, and the landsliding area is undergoing an accelerated deformation stage. The external rainfall condition, water level, and water flow rate are important factors controlling the deformation. The increase of rainfall, the rise of water level, and faster flow rate will accelerate the deformation of slope. The geological conditions of the slope itself affect the deformation of landslide. Due to the enrichment of gently dipping gneiss and groundwater, the slope is more likely to slide along the slope. The Jinsha River continuously scours the concave bank of the slope, causing local collapses and forming local free surfaces. Numerical simulation results show that once the landsliding area fails, the landslide body may form a 4-km-long dammed lake, and the water level could rise about 200 m; the historic data shows that landslide dam may burst in 2–8 days after sliding. Therefore, strategies of landslide hazard mitigation in the study area should be particularly made for the coming rainy seasons to mitigate risks from the landsliding area.

     

Analysing the spatio-temporal evolution of an active debris slide in Eastern Himalaya,India

Landslide is one of the prominent geohazards in the Himalayas where loss of lives and property are common. Owing to the complicated geomorphic and tectono-stratigraphic setting of this active Fold-thrust belt (FTB), landsliding of all possible types and spatial scales observed exhibit conspicuous spatio-temporal signatures and evolution. This evolution of landslides is commonly studied by regional assessment and by examining the multi-temporal landslide inventories of a particular area. The success of creating such multi-temporal landslide inventory depends on (i) the availability of relevant past source data (e.g., images, post event maps, air photos etc.) of suitable resolution, scale and quality, (ii) time of generation of source data with respect to the time of landsliding event, (iii) skill of the investigators in interpreting the old images, air photos etc. However, this method is of restricted use in studying the spatio-temporal evolution of a single landslide which is perennially active in the Himalayan terrain, where rapid changes in land use and land cover patterns readily obliterate the signatures of past landsliding. Moreover because of scale constraints, subtle and frequent changes in the spatial dimensions of these individual landslides, and their temporal activity become difficult to identify in such regional assessment carried out over a larger area. In this study therefore, a different approach is adopted whereby the spatio-temporal activity and style of Lanta Khola landslide, a perennially active and large (0.25 km²) debris flow in the Eastern Himalayas, has been studied in detail through detailed scale (1:1000) site-specific geological mapping in phases during the last 28 years (1983–2011). Such site-specific geological observations coupled with numerical slope stability analysis utilising the limit equilibrium method facilitate in detailed understanding of the temporal and spatial evolution and inherent mechanism of this perennial landslide.     

Evolution of natural risk: analysing changing landslide hazard in Wellington,Aotearoa/New Zealand

This paper addresses the temporal variation of rainfall-triggered landslide hazard within the broader context of natural risk evolution. Analysis of a sequence of aerial photos covering a period of 60 years allowed the establishment of a record of landsliding for a site in the Wellington region, New Zealand. The data show one very dominant peak in the magnitude of landslide occurrence in the late 1970s, followed by a continuous decrease. Landslide hazard can be expressed by the frequency and magnitude of the landslide events, with the total surface area affected used as a surrogate for magnitude. However, the distinct decline of landslide magnitude through time from the 1980s onwards indicates that landslide hazard may change with time. This possibility is further explored by correlating potential landslide triggering storms with the magnitude of the landslide event, using the ‘Antecedent Soil Water Status’ model in combination with daily rainfall. The relation between magnitudes of rainfall and magnitudes of landslide events is found to be weak, suggesting that a given ‘Critical Water Content’ (antecedent soil water status and rainfall on the day) does not produce similar magnitudes of landsliding. Furthermore, the study shows that reactivation of previous landslides before the peak landslide occurrence of the late 1970s is low, while the situation is reversed after this peak and reactivation in the subsequent years plays a larger role. It is concluded that the pattern of landsliding cannot be explained by the pattern of rainfall and other factors are controlling the variation of landslide hazard in time. A possible explanation is a change of the geomorphological system with time, instigated by a massive period of landsliding (the late 1970s peak). Subsequent sediment exhaustion of source areas resulting from this period appears to alter the system’s subsequent reaction to an external trigger such as rainfall. The study demonstrates that landslide hazard analysis in general should not rely on the integral of the frequency–magnitude relationship only, but should include potential non-linear changes of system settings to increase the understanding of future system behaviour, and therefore hazard and risk.

When hazard avoidance is not an option: lessons learned from monitoring the postdisaster Oso landslide,USA

On 22 March 2014, a massive, catastrophic landslide occurred near Oso, Washington, USA, sweeping more than 1 km across the adjacent valley flats and killing 43 people. For the following 5 weeks, hundreds of workers engaged in an exhaustive search, rescue, and recovery effort directly in the landslide runout path. These workers could not avoid the risks posed by additional large-scale slope collapses. In an effort to ensure worker safety, multiple agencies cooperated to swiftly deploy a monitoring and alerting system consisting of sensors, automated data processing and web-based display, along with defined communication protocols and clear calls to action for emergency management and search personnel. Guided by the principle that an accelerating landslide poses a greater threat than a steadily moving or stationary mass, the system was designed to detect ground motion and vibration using complementary monitoring techniques. Near real-time information was provided by continuous GPS, seismometers/geophones, and extensometers. This information was augmented by repeat-assessment techniques such as terrestrial and aerial laser scanning and time-lapse photography. Fortunately, no major additional landsliding occurred. However, we did detect small headscarp failures as well as slow movement of the remaining landslide mass with the monitoring system. This was an exceptional response situation and the lessons learned are applicable to other landslide disaster crises. They underscore the need for cogent landslide expertise and ready-to-deploy monitoring equipment, the value of using redundant monitoring techniques with distinct goals, the benefit of clearly defined communication protocols, and the importance of continued research into forecasting landslide behavior to allow timely warning.

     

基于交互多模算法的Kalman平滑器在滑坡监测中的应用

将滑坡的监测视为机动目标的跟踪问题,基于传统机动目标跟踪算法的不足,首次引入交互多模算法对滑坡监测数据进行处理。最后在此基础上给出了一种卡尔曼滤波平滑算法,实例仿真结果说明了其有效性。     

基于BP神经网络的滑坡监测多源异构数据融合算法研究

针对滑坡监测中的多源异构数据融合问题,论文提出了一种基于BP神经网络的多源异构监测数据融合算法。该算法将影响滑坡变形的温度、湿度、风力、云量、单日降水量和累计降水量等多环境因子变量作为输入变量,以滑坡位移变化量数据作为期望输出数据,并利用各环境因子变量和滑坡位移变化量的相关性及显著性进行环境因子变量筛选,以提高算法的预测精度。论文采用甘肃省永靖县黑方台党川滑坡的实测数据进行了试验,结果表明:反向传播（Back-Propagation,BP）神经网络数据融合算法适用于具有多源异构监测数据的滑坡变形预测;在进行环境变量因子筛选后,BP神经网络数据融合算法的决定系数达到0.985,均方根误差（RMSE）达到0.4787 mm,从而有效提高了变形预测结果的精度。