To better predict the response of laterally loaded caissons and the displacement distribution around the foundations, three different beam theories [Timoshenko beam theory with vertical (z) displacement in soil, Timoshenko beam theory without z displacement, and rigid beam theory with z displacement] were compared in the energy-based variationally method, and two caissons from real projects under the initial loading state were studied. The results from the three methods are compared with measured data from field tests and three-dimensional differential analysis. From the study, the method of the Timoshenko beam theory that considers z displacement in the soil domain produces the most accurate results. The rigid beam theory produced the worst prediction when estimating caisson displacement, rotation, and bending moment. Therefore, for the structure–soil interaction of a laterally loaded caisson, the suggested structure is a Timoshenko beam, and the vertical soil movement around the beam should be considered.

At present, the interface shear test is mainly used to evaluate the anti-sliding performance of the new foundation base. However, the traditional interface shear test has certain limitations in simulating the load change during the construction process and cannot accurately simulate the interface shear characteristics between the structure and the soil under the continuous change of the normal stress. Based on the self-developed large-scale interface shear equipment, this paper carried out the interface shear test and mechanism research of cement soil concrete, sand concrete, clay concrete and other materials in different curing cycles under the loading and unloading modes of variable normal stress repeated steps and continuous loading modes of variable normal stress steps. In addition, this paper deduced the formula of the minimum interface friction coefficient based on Mohr–Coulomb criterion. The experimental results show that the curing effect of cement soil can significantly improve the shear mechanical properties of the interface, and the friction coefficient of the cement soil concrete interface will also increase step by step with the increase of the curing time of the cement soil. The sliding shear surface can be remolded under the preloading of normal pressure, so that the interface shear characteristics of each shear material under repeated loading and unloading can be approximately equal to the interface shear characteristics of multiple equivalent materials under separate loading. In the case of a continuous change of normal stress, the rapid increase of normal stress will lead to accelerated entry into the limit shear state, resulting in plastic failure of the shear plane as a whole. In the engineering with a continuous change of stress, the interface shear friction coefficient of the material with high cohesion fluctuates greatly. The minimum interface friction coefficient formula and test proposed in this paper can be used to evaluate the interface friction coefficient range, and the sand concrete interface shear performance under the continuous normal stress loading mode showed good consistency. The self-developed large-scale interface shearing equipment and its test data provide theoretical basis and solutions for the improvement of traditional interface shearing equipment.