Supporting data for "Learning More Lessons from a Catastrophic Man-made Slope Failure Disaster in Shenzhen by Physical Experiments"

AbstractLandslide on 20 December 2015 in Shenzhen is essentially fill soil from excavation activities. The soil fills were inside gourd-shaped bedrock. It buried 33 buildings with a minimum distance of 300 m to the fill slope toe within 13 minutes. The conventional theory underlying the mechanism of the cause of Shenzhen landslide are rainfall, liquefaction, geometry, absence of compaction, and excess pore water pressure. The catastrophes continuously take places even recent days. Hence, Yue Z.Q. [2008] proposed an alternative modeling theory for this source of the mechanism. Gas, which contributes the source of mechanism remains among the least studied.<br>Three peculiar elements in the Shenzhen landslide accounted. First, gentle slope angle of ~17.5°. Second, rainfall intensity in multilevel timeframe was low. Annual rainfall from 2013 to 2015 on average was 1965.7 mm. The total daily rainfall from October 7 to December 20, 2015 was 104.8 mm. Daily rainfall from December 15–19, 2015 was zero. Third, the amount of liquefied soil observed in the field was tremendously little compare to the slope volume. It suspected that adjoining leaked gas pipe spark the Shenzhen landslide [Yue Z.Q. 2018]. Present research introduced new experimental setups to investigate the possible conditions of the landslide. At first, observation carried out without gas involvement and according to pre-determined water content. Then, source of energy in form of gas are established. Two types of bedrocks created, 2D and 3D models with scale 1:900 and 1:1000, respectively. Ten different experiment conditions presented (A1, B1 – B3, C1 – C6) under three major types of tests. First and second are 2D and 3D model bedrock with soil. Third, 3D model bedrock with hydrogen peroxide solution (H2O2(aq)) and cement. Initial applied air pressure were 2 ~ 4 kPa for the first two types.<br>Bottom soil layers designed as impermeable as possible by full saturation. While top soil layers designed more than its liquid limit. Entire experiments revealed there is no deformation with the absence of gas pressure. Non-linear pressure of air applied. Maximum pressure were 85kPa (A1), 112kPa (B1), 235kPa (B2) and 240kPa (B3). Numerous deformation detected. Estimated 6.2% (A1), 29.47% (B1), 155% (B2), and 167% (B3) from total initial area. The results are reasonable compared to 137% of the same ratio from (Xu et al., 2017) and 169% from (Wang et al., 2017). Comparatively, supporting tests performed by ejection of piston from steel cylinder. The result show that the maximum height of ejecting piston (0.27 – 0.95 m) and velocity (2.29 – 4.32 m/s). The test results indicated pressurized gas contributed in the magmatic explosive volcanoes concede with available theory of gas by Yue Z.Q. [2008]. One example is in the case of 8 gram H2O2(aq). Total mass of 1.17 gram Oxygen could rocket up 285.90 gram steel piston. It flew for 34cm high.<br>Like water, gas is a fluid material and has no shear strength. Therefore, gas material and gas pressure infilled in soils can reduce soil effective strength, which can induce soil slope instability. Thus, modeling of geo-hazard phenomena shall include the useful studies [Yue Z.Q., 2008 – 2018].

2015年12月20日深圳滑坡本质上是开挖活动产生的填土方。这些填土方位于葫芦形基岩内部，在13分钟内掩埋了33栋建筑，这些建筑与填方坡脚的最小距离为300米。
解释深圳滑坡成因机制的传统理论包括降雨、液化（liquefaction）、几何形态、压实不足及超孔隙水压力（excess pore water pressure）。此类灾害直至近期仍时有发生，因此Yue Z.Q. [2008]提出了一种解释该成因机制的替代建模理论，而作为成因机制来源之一的气体，其研究仍十分匮乏。
深圳滑坡存在三个特殊特征：其一，坡角平缓，约为17.5°；其二，多时间尺度下的降雨强度均较低——2013至2015年平均年降雨量为1965.7毫米，2015年10月7日至12月20日总降雨量为104.8毫米，且2015年12月15日至19日期间无降雨；其三，现场观测到的液化土体（liquefied soil）量相较于边坡体积微乎其微。Yue Z.Q. [2018]推测，邻近泄漏的燃气管道是引发此次滑坡的原因。本研究采用新型实验装置探究滑坡的可能发生条件：首先在无气体参与且按预设含水量的条件下观测，随后引入气体形式的能量源。实验构建了两种基岩模型——二维（2D）模型（比例1:900）与三维（3D）模型（比例1:1000），并在三类试验下设置十种条件（A1、B1–B3、C1–C6）：第一类为含土体的2D基岩试验，第二类为含土体的3D基岩试验，第三类为含过氧化氢溶液（H₂O₂(aq)）与水泥的3D基岩试验，前两类初始气压为2~4kPa。
底部土层通过完全饱和设计为尽可能不透水，顶部土层设计为超过其液限（liquid limit）。实验表明无气压时无变形；施加非线性气压后，A1、B1、B2、B3的最大气压分别为85kPa、112kPa、235kPa、240kPa，对应初始总面积变形率为6.2%、29.47%、155%、167%，与Xu等[2017]的137%及Wang等[2017]的169%具有合理性。辅助试验中钢筒活塞最大弹射高度0.27~0.95m、速度2.29~4.32m/s，结果验证了加压气体对岩浆爆发式火山的贡献与Yue Z.Q. [2008]理论一致——如8克H₂O₂(aq)产生1.17克氧气，可推动285.90克钢活塞上升34cm。
与水类似，气体是流体物质且无剪切强度（shear strength），填充于土体中的气体及气压会降低土体有效强度（effective strength），进而诱发边坡失稳。故而地质灾害（geo-hazard）现象的建模应纳入此类研究[Yue Z.Q., 2008–2018]。