With four large hydropower projects located within the geologically challenging Sichuan and Yunnan provinces, China Three Gorges instituted a monitoring and analysis program to help guard against catastrophic events.
By Endi Zhai, Fan Qixiang and Jin Hua
The China Three Gorges Corp. (CTG) developed four hydropower stations along 600 km of China’s Jinsha River. These plants include 6.4-GW Xiangjiaba, 10.2-GW Wudongde, 13.9-GW Xiluodu and 16-GW Baihetan. Located in the transitional region that stretches from the Tibetan Plateau and Yunnan-Guizhou Plateau to the Sichuan Basin, they are in an area with complex geological conditions that have caused hundreds of landslides.
Xiangjiaba and Xiluodu were completed first, in 2014, and investigations into the reservoirs at each showed potential geohazards including landslides, crumble accumulation and rock deformation areas.
Holocene and late-Pleistocene active faults are quite developed, widely distributed, of immense scale and in multiple directions, and have high amplitudes and complex internal structures. Some of these fault zones also consist of several parallel or intersecting secondary faults.
Main fault zones in the region include the Leshan-Yibin, Mabian-Yanjin, Ebian-Jinyang, Daliangshan, Zemuhe, Xiaojiang, Mopanshan- Yuanmou, and the Lianfeng-Huayingshan, with past earthquakes exhibiting magnitudes between 6.0 and 7.0 and more than 20 greater than 7.0.
Establishing a geohazard monitoring system
During the first years of operation at Xiangjiaba and Xiluodu, old landslides became active, causing adverse modification to the landscape.
These landslides were caused by reservoir impoundment and/or heavy rainfall. To avoid or reduce the loss of life and property, CTG researched the early prediction of landslides, using real-time monitoring and warnings in the influence zones of potential landslides.
For monitoring, an emphasis was placed on surface deformation measurements, while also taking into account what might be occurring at lower depths.
Surface monitoring involved measuring horizontal and vertical displacements, while the rate of deformation was measured using the global navigation satellite system (GNSS), which includes global positioning systems (GPS), the Russian GLObal navigation satellite systems (GLONASS) and other traditional geotechnical monitoring equipment such as inclinometers and surface monuments.
A reference network was created by installing surface deformation monitoring monuments and sub-control centers at areas having the potential for a collapsible reservoir bank or landslide.
To monitor deeper deformations, inclinometers were installed at potentially collapsible banks or landslide sites at depths usually about 5 m deeper than the sliding surfaces.
Automatic rain gauges were also installed to collect precipitation information at or near areas of concern.
A total control center was set up at CTG’s Chengdu regional headquarters to store and analyze data from the affected areas.
Monitoring ground deformation for the Yulin II landslide
The Yulin II Landslide occurred in April 2017 and is located on the right bank of the Xiluodu reservoir area in Yunnan Province, about 39 km away from the dam. Large deformation of more than 10 m occurred, and some areas near the river bank collapsed into the river, but the overall landslide area just deforms. The deformation area is about 200,000 m3 with a dimension of about 700 m perpendicular to the direction of the river. The maximum depth of the sliding surface is about 100 m, and the overall sliding volume is about 12 million m3.
The geological condition at the landslide area has some bedrock outcrops with a formation of Ordovician and Silurian Luau (O3+S (s)). The lithology consists of dark gray, grayish yellow, yellow green sandstone, siltstone, mudstone, shale, sandy shale and marl argillaceous limestone. The bedrock is more overlaid with the by slope residual mulch. The stratum topography is N150 ~ 200 W/SW ∠ 600 ~700, and the attitude is stable that is along the sloping shore slope. The main substance of the landslide is grey yellow ~ brown yellow gravel soil with a gravel content of about 30%, more than 0.5~2.0cm in grain size, more compact structure, lower surface and edge covered by limestone stone collapse, stone size 0.1~0.5m, with the maximum being 2 m.
Nine GNSS monitoring points, three GNSS transfer stations, one automatic Pluviometer and one sub-control center have been installed within the sliding zone in the past two to three years.
Data could be read automatically every second, but to save storage space, the data is typically read on a daily basis during the dry season and hourly in the rainy season. The rainy season in the lower reach of the Jinsha River region typically runs from May to October.
As of April 2017, the cumulative displacements measured by GNSS ranged from 10,416.2 mm to 14,228.2 mm. The settlement value was between 5,692.8 mm and 10,686.2 mm.
The slope deformation area was divided into two parts, called Zones A and B. Zone A, which had more significant movement, represents the local area of a single designated measurement point, and Zone B represents other areas within the overall landslide zone. Data collected was used to measure the relationship between the vertical displacement and reservoir level, also factoring in rainfall intensity.
This is detailed in Table 1, which lists the average and maximum displacement rates for select periods at points designated “TP03” and “TP06” when the rates were so significant that it caused CTG’s management to ready its reconnaissance team for a potential emergency.
After Xiluodu reservoir’s impoundment, geological monitoring showed the landslide was reactivated — the disintegration was serious and the deformation was strong. At present, the landslide movement has decreased, but it is still in the sliding phase. The whole body tends to be unstable, but the possibility of the entire mass sliding at high speed is small.
The accumulative displacement of each measuring is still increasing, but the rate has become insignificant. The horizontal displacement across the river has reached 10 m to 14 m, the overall settlement is roughly 10 m, and the maximum deformation rate was 202 mm/day.
It is the largest landslide deformation rate measured by CTG for a landslide that has not yet collapsed. As such, a continuous monitoring and warning system are in place, and a risk assessment for the collapse condition is ongoing. A retrofit might be necessary if the risk is found to be intolerable.
Safety analysis for a deformation body surge in Xiluodu Reservoir
Due to continuous rainfall during the wet season in 2016, a deformation body located about 7 km from Xiluodo Dam was designated for emergency examination. The deformation body is located about 400 m from the Jinsha River’s edge and about 500 m above it.
The deformation body began to skew and produced cracks on August 2. Multiple cracks appeared on a highway and lateral slope surface. The biggest crack paralleled the Jinsha River, with a length of about 200 m, an open width of 1 cm to 5 cm, and a vertical distortion ranging from 10 cm to 80 cm. A series of small sliding phenomenon happened at the front edge of the deformation body near the bedrock steep and a small gully.
CTG installed a number of monitoring points to obtain real-time deformation conditions, which began Aug. 18, 2016. As of April 24, 2017, the edge closest to the river had an accumulative horizontal displacement of 36 mm to 67 mm, and the back edge displacement has reached up to 46 mm.
It appeared the deformation belonged to the local superficial sliding due to insufficient drainage along the roadway. Given the current deformation situation, it is less likely that the ancient landslide could cause large-scale integration sliding. But, in the case of continuous rainfall, small sliding and deformations might happen.
Because of the close distance of the deformation body to Xiluodu Dam (7 km) and the nearby Majiahe Dam (4 km) and Huangjuebao Wharf (1 km), secondary disasters, such as surge, could occur. Thus, an urgent safety assessment was performed.
If the deformation body were to slide into the Jinsha River as a whole with the reservoir level at 574 m, the surge height would be 113.2 m at the entry point, producing a maximum surge of 50.4 m on the opposite riverbank. The surge would decay quickly downstream, peaking at 4 m at Majiahe and 2.2 m at Xiluodu.
The assessment showed the deformation body consists primarily of gravel with poor cementation, so it would not be possible for it to slide into the river as a whole. And even if it were to slide simultaneously, collisions with the bedrock slope surface would cause it to disintegrate. The speed of entry into the water would also be greatly reduced due to the slope’s topography.
Therefore, it was estimated that the deformation body, at the current scale, would not pose a significant safety issue for Xiluodu.
Monitoring and stability assessment for mega landslide near hydropower station
The Jinpingzhi Landslide is located about 900 m downstream of Wudongde Dam. The overall volume of the landslide is more than 1 billion m3. Its stability is crucial for Wudongde.
According to its topography, this landslide can be divided into five zones. Zones 1, 2 and 3 are quarternary accumulation bodies, and 4 and 5 are in-situ bedrock with good stability. Zone 2 is a strong creep deformation zone, with a maximum deformation speed of 0.5 m/year at its leading edge.
Therefore, the failure mode and damage scale for Zone 2 have been monitored and assessed with 17 monitoring points.
The general characteristics of Zone 2’s deformations showed a gradual increase in the displacement rate at each monitoring point from its trailing edge to leading edge, with the direction pointing toward the Jinsha River, making it a pull-type landslide. The failure mode would be multistage traction sliding.
CTG’s design consultant performed numerous field investigations, lab testing and numerical analysis for the landslide. CTG’s review team reviewed monitoring data and performed a check against the consultant’s work, using the limit equilibrium and Morgenstern-Price methods to calculate the safety factor in different loading conditions of the landslide’s state.
CTG also performed a check for mitigation measures. These include the construction of five parallel drainage tunnels with several branch tunnels for the quick dissipation of ground water.
From the landslide’s leading to trailing edge, the safety factor increased gradually, presenting the characteristics of traction-type damage. It is consistent with the deformation law. In a natural state, the leading edge safety factor is smaller than 0.9, putting it in phase with a small-scale collapse at the leading edge.
To enhance the stability of the landslide, one intercepting tunnel, five drainage tunnels and branches were built in 2016. Chinese codes for a minimum factor of safety were met by reducing the water table fall to 60% of its natural state.
Due to the mega scale of the hydropower projects and the cascading reservoirs, CTG adopted real-time monitoring and emergency preparedness of reservoir-induced landslides as an important part of its risk assessment system. Often, the landslide areas have a very steep slope, as high as a few hundred to more than 1,000 m, and traditional geotechnical exploration and an engineering retrofit program may not be technically or economically feasible. Thus, CTG used GNSS systems (including GLONASS, Galileo and China’s Baidou System) for monitoring. The landslide sliding rate reached more than 200 mm/day during times the reservoir level was dropping quickly but later decreased to less than 11 mm/day during the dry season, with no significant reservoir level change. This exceeded any warning threshold ever published.
CTS’s hazard management team continues to examine the data collected and to perform analysis. Regular review meetings have been held by nationally and/or internationally well-known geotechnical experts to diagnose the risk. So far, the risk is manageable.
CTG is considering similar monitoring programs for its other facilities, although the monitoring points and equipment types may differ.
Zhai, Endi, Fan Qixiang and Jin Hua, “Monitoring and Assessment of Reservoir-Induced Landslides for the Jiansha River Hydropower Project in China,” Proceedings of HydroVision International 2017, PennWell Corp., Tulsa, Okla., 2017.