Méthodes de conception et étude du comportement sismique des fondations superficielles sur sol naturel et traité, considérant l'interaction sol-structure

Abstract

Current rules and regulations prescribe a static equivalent approach in order to evaluate the seismic bearing capacity of non-liquefiable soils, as does the Canadian Foundation Engineering Manual (CFEM). Newer methods, such as interaction diagrams or pseudodynamic plasticity based approaches could be used in order to improve the seismic design of shallow footings. For liquefiable or poor soil conditions, deep foundations are often preferred. Nevertheless, in some cases it may be economically interesting to choose shallow footings combined with soil reinforcement (such as vibroreplacement). Generally, engineers chose to install stone columns in order to mitigate the liquefaction risk, however, the beneficial effects of such a reinforcement upon the geotechnical properties of the soil reinforced are often neglected. As a consequence, the design adopted might be too conservative, while in fact the real system performance is unknown. Many researchers recently proposed new bearing capacity factors Nc, Nq and Nγ for dynamic conditions, taking into account inertia forces in the soil (Richards, Pecker, Fishman, Choudhury). In order to represent adequately a complex loading on a footing, interaction diagrams are a better option to estimate the bearing capacity instead of classical methods. Experimental interaction diagrams were developed recently. Theoretical interaction diagrams are included in the Eurocode 8 and consider inertial forces in the soil. A comparative analysis indicates that classical methods overestimate the soil’s resistance to moment. Stone column reinforcement has two main advantages: (i) it mitigates the liquefaction risk as the stone columns act as a vertical draining system, (ii) it improves the geotechnical properties of the soil between the columns. Priebe’s method consists in computing the geotechnical properties of an equivalent homogenous soil. Baez Satizabal proposes a method to estimate the geotechnical properties of the soil between the stone columns. An Excel spreadsheet called CaPoDyn (Dynamic Bearing Capacity) has been developed in this thesis to compute the soil bearing capacity for static and seismic conditions, for a nonreinforced soil and for a stone column reinforced soil. Classical, experimental and theoretical methods presented in the previous sections (interaction diagrams, Richards, Priebe, baez Satizabal) are combined when possible. Each of these methods has been validated separately in order to be implemented in CaPoDyn, and results for the static case have been validated with a finite difference code (FLAC). It appears that the approach combining Priebe’s and the CFEM methods gives the most realistic bearing capacities. Two parametric studies were conducted. The first one considers different area replacement ratios and acceleration coefficients. For a soil with good geotechnical properties (ϕ>35«), the stone columns do not have a great impact on the bearing capacity, even for a great area replacement ratio. The second parametric study conducted with FLAC focuses on the effect of the dimensions of the soil reinforcement. It takes into account two case scenarios: (i) “construction” where the reinforcement is installed prior to the footing installation,(ii) “rehabilitation” where the stone columns are installed around the previously built footing. The bearing capacity does not increase significantly after a width of treatment reaching 4B for the construction scenario or 1,5B on each side of the footing for the rehabilitation scenario. It is also observed that it is more efficient to reinforce on a large width than on large depth when the objective is to improve the bearing capacity. The last part of the thesis focuses on the seismic behaviour of a surface footing resting on a stone column reinforced soil. The analyses were made with the finite difference code FLAC. Two cases are considered: (i) water table deep below the surface level, (ii) water table at the surface level. In general, the greater the area reinforced with stone columns and the greater the area replacement ratio, the less settlement is recorded for the footing. There are some exceptions that are difficult to anticipate since the inclusion of a rigid element in the soil can lead to an amplification of the seismic signal. As a consequence it is recommended the use numerical models to optimize the reinforcement area. However, no matter which reinforcement parameters are chosen, the seismic behaviour of the footing is significantly improved compared to the case where the soil is left unreinforced (the footing settlement is twice as low, or even more). It appears that soil reinforcement on large width and depth might be a good option for certain accelerograms. For a large area of soil reinforcement each case should be studied to evaluate the economic benefit of the improvement in seismic behaviour.

Date	17 Sept 2013
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Marie-José Nollet (Supervisor) & Denis Leboeuf (Co-supervisor)