Reducing carbon emissions of deep mixing through various design, laboratory and field measures, Doctoral Thesis Juha Forsman

In the Nordics, column and mass deep mixing are performed using the so-called Nordic deep mixing method, in which dry binder is injected into soil by compressed air (dry mix method, DDM). Although the functionality of deep mixed soil (soilmix) is based on the stiffness (modulus) of the soilmix, in practice, most of the design parameters used in Finland are based on the shear strength of soilmix. It is important to determine the shear strength of soilmix as reliably as possible, regardless of the purpose of DDM or what kind of low carbon (or traditional carbon intensive) binder will be used. A reliable determination of the shear strength of soilmix helps choose deep mixing as a ground improvement method and optimize the amount of soilmix needed and the binder recipe (binder quality and quantity). The shear strength of soilmix (lab) has been examined in terms of, e.g. the long-term
durability of soilmix, receiving traffic loads to DDM columns, stabilization tests, QC soundings and field to laboratory strength ratios (kL/F). Based on the laboratory tests, QC soundings and settlement measures, it can be concluded that the long-term behavior of deep mixed clay and peat are satisfactory from the engineering perspective and controlled with the shear strength of soilmix. Due to Eurocode recommendations, there was a major increase in the intensity of the traffic loading from 10 kPa to 9+31 kPa in the 2010s in the Finnish guideline (FTA 2018). The 3D FEM simulations and parameter development were undertaken to better understand the interaction between column and soft clay under a rapid traffic load. It appears that the previous design with a traffic load of 10 kPa leads to a similar number of columns as with the new simulated traffic load of 9+31 kPa and parameters—
derived from the shear strength of soilmix.

The stabilization test series performed as a round-robin study was examined via the internal (within laboratory) variation of the individual series (COVin) and via the external variation of the parallel stabilization test series (COVex). When looking at the variation of the σUCS of the individual stabilization test specimens in the test series, the results show that COVin≤10% is achievable but some development in guidance and uniform quality control for laboratories are needed. Based on the test columns’ QC soundings, clay or organic clay deep mixed with the low-carbon gypsum, slag or fly ash containing binders have achieved strength levels like the reference binder lime-cement. The effect of various factors (curing time, water content and shear strength of natural clay) on the shear strength variation (COV) was analyzed: none visibly affected the COV for any of the binder types considered. The strength variation determined by QC soundings was found to follow the log-normal distribution. In general, the mode value (the greatest likelihood) is in the scale of COV = 0.2-0.4. Based on statistical analyses, lab – kL/F graphs that consider the underlying uncertainties were determined
using two approaches: cautious trendlines based on determined transformation uncertainty
and simulation-based validation. Simulation-based validation of the defined characteristic trendlines was performed using the average COV values determined based on the stabilization test and QC sounding results. For all the studied binders, the 31% percentile calculated from the simulated values corresponded well with the characteristic “cautious” curve. The newly derived nomograms will be used in the update of the FTA design guideline to replace the outdated kF/L nomogram in use since 2001, which lacks, among other things, low-carbon binders.

The appendices to this Doctoral Thesis are available upon request: juha.forsman@ramboll.fi