The physical model test of landslides is important for studying landslide structural damage, and parameter measurement is key in this process. To meet the measurement requirements for deep displacement in landslide physical models, an automatic flexible inclinometer probe with good coupling and large deformation capacity was designed. The flexible inclinometer probe consists of several gravity acceleration sensing units that are protected and positioned by silicon encapsulation, all the units are connected to a 485-comunication bus. By sensing the two-axis tilt angle, the direction and magnitude of the displacement for a measurement unit can be calculated, then the overall displacement is accumulated according to all units, integrated from bottom to top in turn. In the conversion from angle to displacement, two spline interpolation methods are introduced to correct and resample the data; one is to interpolate the displacement after conversion, and the other is to interpolate the angle before conversion; compared with the result read from checkered paper, the latter is proved to have a better effect, with an additional condition that the displacement curve move up half the length of the unit. The flexible inclinometer is verified with respect to its principle and arrangement by a laboratory physical model test, and the test results are highly consistent with the actual deformation of the landslide model.

This paper presents a new model of the load sharing law with a three-stage load sharing pattern via a representative case study in the Three Gorges reservoir region, China. A definitive new three-stage load transfer pattern is presented, including end-bearing soil arching, friction soil arching and the sliding mass in front of the pile. By means of the soil arching effect between the anti-sliding pile and landslide mass, the law of the load sharing ratio under different cases, including different intervals, section dimensions, driving forces, and shearing parameters of the sliding mass and the pile-soil interface, is presented by using the explicit finite-difference numerical modelling method. The results show that (a) the effect scale of the soil arching effect is within the scale of four times of the width of the pile; (b) the soil arching only exits within a certain pile interval, and it will become inefficiency beyond the maximum pile interval; (c) there is a threshold value for the cohesion strength parameters of the sliding mass, beyond which the load sharing ratios of soil arching keep in a steady level.