Mechanism Of Penetrating Of Penetrometers In Soft Clay With Large Deformation And Determination Of Undrained Shear Strength

Soft clay distributes widely throughout the world. With the development of marine resources and marine industry in various countries, massive offshore structures such as levee and port engineering, pile foundations of offshore wind turbines, oil and gas platform, submarine pipelines, mooring system, etc., are under construction. It poses a challenge to the design of offshore geotechnical engineering as to how to determine the soil parameters, mainly the shear strength of soft clay, in order to ensure the safety of the structures and meanwhile save construction costs. Currently, cone penetration test is the most widely used in-situ testing method of marine clay. And an improved form of cone penetrometer (i.e. the full flow penetrometers of T-bar and Ball) has become increasingly popular in marine soil characterization. However, the interpretation of the above mentioned three penetrometer tests relies on theoretical or empirical formulations to obtain the undrained shear strength of soft clay. Considering the siginificance of the three penetrometers in soil characterisation, it is important and challenging to establish rational relationships between the penetrometer test results and soil properties based on the knowledge of the failure mechanism during penetration.Due to the changing boundary conditions and large deformation of soil occurring during penetration of the three penetrometers, large deformation finite element analysis is demanded. In this paper, a coupled Eulerian-Lagrangian (CEL) method built in the commercial software Abaqus is utilized to simulate the penetration process of the three penetrometers in uniform single soft clay layer from ground surface, and penetration from an embedded depth was also performed for full-flow penetormeters. The failure mechanism and influential factors on the penetration resistance including the friction ratio of the interface between penetrometers and soil, the soil rigidity index, were studied. The results show that:(1) Both the horizontal and vertical stresses in the adjacent soil increases obviously during the penetration of the three penetrometers. The maximum stresses occurs around the cone shoulder, while it is below the T-bar or Ball. (2) The shear strain contour is generally parallel to the shaft of the cone penetrometer, while it appers as concentric cicles around the cone probe, which shares some similarity with the cylindrical and spherical cavity expansion theories respectively. (3) Two different failure mechanisms can be found for the three penetrometers advancing from the ground surface, i.e. the shallow failure mechanism and the deep failure mechanism. (4) The normalized penetration resistance of steady state for the prebored Ball is higher than that for prebored T-bar by nearly 18.9%. It is also found that the value of full-flow penetrometers moving from the ground surface is significantly smaller than from an embedded depth, which is related to the cavity formded above the penetrometers in the former case. Therefore, it is recommended to use the prebored case which is more appropriate to achieve a steady state condition for full-flow penetrometers. The cone penetration resistance increases with the interface friction ratio and the soil rigidty index increasing. The steady resistance of T-bar increases nearly linearly with the interface friction ratio. The soil rigidty index has no effect on the steady state penetration resistance of the full-flow penetrometers, but it influences the displacement to achieve the steady state as concluded by existing references.The in-situ test rates for the three penetrometers are generally 20 mm/s, which generates as much as 5-7 orders of magnitudes of that in conventional geotechnical laboratory tests. Natural soft clay generally has certain structure with a common sensitivity among 2-6, while the accumulated plastic shear strain is commonly 10-25 (i.e.1000%～2500%). Therefore, it may lead to significant discrepancy negelecting the effects of strain rate and strain softening. While in the graphical user interface of Abaqus, the aboved mentioned behaviors of soil can not be modelled by the built-in soil constitutive models, neither can the linear distribution of initial undrianed shear strength along depth be modelled. In this paper, an empirical formulation considering the effects of strain rate and strain softening on soil shear strength (Einav and Randolph, 2005) is used to modify the elastic-perfectly plastic Tresca constitutive model, and the modified Tresca model is developed based on the subroutines in Abaqus. Then, using the modified Tresca model and CEL method, parametric studies are carried out to investigate the influence of strain softenging, strain rate and soil strength on the penetration resistance of the three penetrometers. The results show that as the gradient of soil strength along depth increases, the penetration resistance of the three penetrometers all increase obviously and continuously with depth. The steady state penetration resistance almost increases by a factor of 1.3 as the rate parameter μ increases from 0.1 to 0.2 for all the three penetrometers, and decreases with the brittleness parameter δ95 decreasing. The stable value of the normalized penetration resistance of the three penetrometers decreases with the strain-softening parameter δrem, but the peak values of full-flow penetrometers are not affected by δrem.The undrainded shear strength of remoulded soft clay is frequently required in the design of offshore geotechnical engineering. The cycling penetration tests of full-flow penetrometers are important means of obtaining the remoulded shear strength of soft clay. However, the conversion coefficient between the penetration resistance and the. undrained shear strength of intact and remoulded soft clay is indeed different, which can not be misused. In this paper, the cyclic penetration processes of T-bar and Ball penetrometers are analysed, investigating the cyclic degradation regulation of the penetration resistance and the conversion coefficient of the remoulded shear strength of soil. It shows that the patterns of the cyclic penetration resistance of T-bar and Ball are similar, which declines sharply within the first few cycles, and then gradually flattens. The initial degradation rate of the penetration resistance is related to the sensitivity of the soft clay, which increases with the sensitivity increasing. The ratio between the initial and the steady-state penetration resistances qin/qrem is constantly smaller than the sensitivity St of the soft clay. The extraction ratio of the first cycle essentially decreases with the sensitivity St of the soft clay, and the conversion coefficient of remoulded soft clay increases with St. The formulas between qrem/qin, St and the first extraction ratio qext/qin are given through curve fitting respectively.Finally, the existing theoretical and empirical methods calculating the undrained shear strength of clay, as well as the experimental methods are summarized and compared. The recommended coefficients relating the penetration resistance to undrained shear strength of intact and remoulded soft clay are given based on the large deformation analysis in this paper.