In going through some papers, I noticed this one, Friction anlysis of large diameter steel cylinder penetration process using 3D-DEM, whose abstract is as follows:
Large open-ended cylinder piles have been widely used for engineering foundation of port. The penetration process of the large-diameter steel cylinder exhibit complex behaviors, which is difficult to be measured by test and reproduced in numerical models. This study presents a friction analysis of large diameter steel penetration process by using the discrete element method (DEM), which can simulate large deformation and nonlinearity well. Centrifugal model and full-scale model were developed to analyze the sliding friction of the cylinder during installation and the contact force chain of soil particles. The validity of the DEM model was examined by comparing with theoretical values and published studies. Parametric studies were carried out to study the effects of contact parameters on side friction. Simulation results showed that, unlike pile penetration, there is no obvious soil-plug effect during the penetration process of large-diameter steel cylinder. Besides, the inside friction is smaller than the outside friction for large-diameter steel cylinder. What’s more, the computational cost of full-scale model based on the upscale theory was less than the centrifugal model. There is a close relationship between the side friction and micro contact parameters, which provides a reference for the follow-up study of cylinder or pile penetration using DEM.Friction anlysis of large diameter steel cylinder penetration process using 3D-DEM
Put into simple terms, DEM models the soil by a grain-by-grain analysis of its response to load, which in this case came from 22 m O.D. open-ended pipe piles for the Hong Kong-Zhuhai-Macau Bridge. Believe it or not, I seriously considered using something like this for the STADYN project, but was dissuaded from doing so because of complexity and stability issues. That’s probably for the best, because the program I was in encouraged a “roll your own” approach to computer code, and to be honest at my stage in life and background I wasn’t up to such a development.
One thing that helped the authors of this paper was the fact that the soils they were driving into was cohesionless, without the chemical bonding that comes with cohesive soils. It’s also interesting to note that these piles were vibrated into place; the whole subject of vibratory driving, its performance prediction and static capacity determination, is another long-running subject in this business.
Although the issues of cohesion and verification (always the fun part of geotechincal engineering) need further resolution, DEM is, in my opinion, the ultimate solution of the soil interaction question, and needs to be disseminated further in our industry.