'Smart Rocks' for Debris-Flow Landslide Research

泥石流是一种特别具有破坏性的滑坡，因为大量潮湿的土壤和岩石可以作为液化体以非常高的速度移动，而且几乎没有预警。泥石流可能是由地震、火山喷发或地下水上涨引发的。由于泥石流运动迅速，可能造成大量人员死亡；一个极端的例子是2001年由地震引发的泥石流摧毁了圣萨尔瓦多？S郊区圣特克拉，造成700多人死亡。美国还没有经历过如此毁灭性的事件，但降雨或地震引发的水流滑坡造成了重大的物质损失，并造成了多人死亡，特别是在加利福尼亚州人口稠密的地区。随着气候变化和滑坡多发地区的城市化进程的继续，预计人员和物质损失都将增加。泥石流具有能够在非常平坦的斜坡上长距离流出的特殊特征。为了确定高危区域并计划可能的防御措施，有必要计算潜在的滑动速度和跳动长度。对这些滑动进行建模是困难的；当滑动被触发时，滑动材料被观察到液化，即高内部水压力迫使土壤颗粒分离，并且质量表现为重粘性流体。然而，当滑道向下移动时，它会分离成一个排干的口鼻？由更粗的材料组成，预计会阻止水流，以及仍处于液化状态的内部，在那里，较细的土块部分会推动滑坡向前推进。这两个区域如何相互作用，使整体质量保持高效率？是什么让它即使在平坦的斜坡上也能长途旅行，尽管它显然与滑梯有关，但人们对此还不太清楚。能够在长时间内保持较高的内部流体压力。在许多情况下，这似乎令人惊讶，因为即使是内部液化的滑动材料的粒度分布也表明，维持土壤处于液化状态所需的水压力相对较快地消散。泥石流运动的复杂数值模型已经开发出来，但一个根本的问题是，它们需要预测滑动过程中的内部水压力。为了阐明这一基本问题，该项目将开发两种尺寸的仪表化智能岩石？使用最新开发的MEMS仪器：智能鹅卵石？高尔夫球大小，以测量内部颗粒如何振动，以及水压力如何在滑动体液化内部发展和消散；以及智能鹅卵石？软球大小，更多的仪器，因此也能够跟踪更粗的颗粒在滑动过程中如何向滑坡口移动。两者都将是椭圆形的，以模拟天然岩石，并有金属外壳，以便使用金属探测器在幻灯片末端更容易找到它们。该项目的头两年将由一个包括机械和土木工程研究生以及土木工程、机械和电气工程本科生在内的团队用于智能岩石的设计和广泛的实验室测试。第三年，智能岩石将被部署在俄勒冈州威拉米特国家森林的USGS实验滑坡水槽的大型人工水流滑道中；运营该水槽的研究团队高度支持这一项目。项目组将利用这些测试的结果，以迄今尚不可能的方式校准和改进现有的泥石流模型。此外，一个与当地初中和高中建立了良好合作关系的项目将寻求让学生和教师作为观察员和参与者参与设计智能岩石并解释结果。

Debris flows are a particularly destructive class of landslide, in that large volumes of wet soil and rock can move as liquefied masses at very high velocity, and with little warning. Debris flows may be triggered by earthquakes, volcanic eruptions, or rising groundwater. Because of their rapid motion, they can result in a large number of fatalities; an extreme example is the destruction of San Salvador?s suburb of Santa Tecla by an earthquake-triggered debris flow in 2001, resulting in over 700 deaths. The US has not experienced such a devastating event, but flow slides triggered by rainfall or earthquakes have caused significant material damage, and a number of fatalities, particularly in densely-populated areas of California. Both human and material losses are expected to rise with climate change, and as urbanization of landslide-prone areas continues.Debris flows have the particular feature of being able to run out for long distances over very flat slopes. In order to define high-hazard zones and plan possible defensive measures, it is necessary to calculate potential slide velocities and run-out lengths. Modeling these slides is difficult; when the slide is triggered, the sliding material is observed to liquefy, i.e. high internal water pressures force the soil grains apart, and the mass behaves as a heavy viscous fluid. However, as the slide moves down-slope, it segregates into a drained ?snout? of coarser material which could be expected to brake the flow, and a still-liquefied interior, where the finer portion of the soil mass pushes the slide forward. How these two zones interact, so that the overall mass maintains a high ?efficiency? which allow it to travel long distances even on flat slopes, is not well understood, although it is obviously related to the slides? capacity to maintain high internal fluid pressures over a long runout. In many cases this seems surprising, since the grain size distribution of even the interior, liquefied, sliding material would suggest a relatively rapid dissipation of the water pressures required to maintain the soil in a liquefied condition. Sophisticated numerical models for debris flow motion have been developed, but a fundamental problem is that they require, as input, a prediction of the internal water pressures during sliding.In order to clarify this basic problem, this project will develop two sizes of instrumented ?smart rocks? using recently-developed MEMS instrumentation: a ?smart pebble? of golf-ball size to measure how the interior particles vibrate, and how water pressure develop and dissipate in the liquefied interior of the sliding mass; and a ?smart cobble? of soft-ball size, more heavily instrumented, so as to also be capable of tracking how coarser particles move towards the landslide snout during sliding. Both will be ovoid in shape, to simulate natural rocks, and have metal casings to make them easier to find at the end of the slide, using a metal detector. The first two years of the project will be spent in design and extensive laboratory testing of the smart rocks, by a team which will include graduate students from mechanical and civil engineering as well as civil, mechanical and electrical engineering undergraduates. In the third year, the smart rocks will be deployed in large-scale artificial flow slides at the USGS experimental landslide flume in the Willamette National Forest, Oregon; the research team which operates the flume is highly supportive of this project. Results from these tests will be used by the project team to calibrate and refine existing models of debris flows in a way which has not been possible to date. Further, a well-established partnership program with local middle and high schools will seek to involve students and teachers as observers and participants in designing the smart rocks and interpreting the results.