B2020-13 System response of lime-cement columns

The powerful numerical tools developed in Part 1 of the PhD project (BIG A2020-13 Deep mixing in urban environment) based on Volume Averaging Technique (VAT) were verified against equivalent 3D simulations. Subsequently, VAT was validated against field data from Paimio test embankment (Vepsäläinen & Arkima 1992, Abed & Karstunen 2023ab, Abed et al., 2026) and Västlänken project (E02 Centralen, see Bozkurt 2023, Bozkurt et. al 2023a,b, Bozkurt et al. 2025, Bozkurt 2025). The advantage of VAT simulations is that the complexity of the material response of the natural clay and the columns, respectively, can be accounted for by adopting appropriate constitutive models for two constituents, soft clay and deep mixed columns. This facilitates exploring the real system performance, following the effective stresses and strains in the columns and the clay, respectively.

In Part 2 of the PhD project, the VAT techniques were further exploited in simulations using statistical methods to quantify what is the permissible range of the column and soil properties, to the system performance. Thus, we can identify the main drivers to the forecasted magnitude and extent of ground movements, which control the performance of geostructures on deep mixed columns, enabling to optimise the columns stiffness.

User defined models, compatible with Plaxis FE code, were developed, verified and validated for the two main uses of deep mixing in Sweden: embankment on soft soils and deep excavations on soft soils with deep mixing as support on the passive side of the excavation. The VAT formulation enables modelling 3D problems with periodic ground improvement in computationally effective way, thus enabling systematic sensitivity studies and optimisation of columns stiffness. The results suggest that for embankment problems, much more compressible and weaker columns could be used, resulting both economic and environmental savings.

The results suggest that for embankment problems, much more compressible and weaker columns could be used without affecting the function, resulting both economic and environmental savings.

The project has resulted in many high quality publications in scientific journals, and a two new plug-ins for Plaxis FE code that enable modelling the complex 3D soil-structure interaction with coupled 2D FE analyses. The approach has already been used in a number of MSc thesis at Chalmers, and the method is introduced in level 5 MSc course ACE150 Soil modelling and numerical analyses. Furthermore, project funding enabled to completion of an excellent PhD project at Chalmers.

Until now, there has been no effective way of predict deformation in 2D for embankments and excavations on deep mixed columns. The approach can equally apply, provided the bulk permeability in the stabilised area is changed, to embankment on stone columns. Note, however, that the technique is based on soil and the columns working together (with equal vertical settlement in the case of the embankments), and thus should not be applied to piled embankments.

Bozkurt, S., Abed, A., & Karstunen, M. (2025a). Alternative ways of modelling stabilised excavations. Proceedings of the 19th Nordic Geotechnical Meeting– Göteborg 2024 (SGF Report 1:2025), 1, 160–173. Swedish Geotechnical Society.

Bozkurt, S. (2025). Numerical analysis of geostructures stabilised with lime-cement columns in soft clays [Doctoral dissertation, Chalmers University of Technology].

Bozkurt S., Abed A., & Karstunen M. (2025). Homogenisation method for braced excavations stabilised with deep-mixed columns. Computers and Geotechnics, 181, 107095.

Bozkurt, S., Dijkstra, J., & Karstunen, M. (2025c). Influential factors on the performance of embankments stabilised on deep-mixed columns. Transportation Geotechnics, 55, 101643.

Ayman Abed, Urs Vogler, Minna Karstunen (2026) Volume Averaging Technique for Deep-Mixed Columns under Embankments: Verification and Validation, International Journal of Geomechanics