Jeroen Dijkstra

With the construction of the North/south metro line underneath the center of Amsterdam the properties of the Eem clay have become of interest. The site investigation revealed that the Eem clay is overconsolidated to such an extent that ageing is not the sole cause for this overconsolidation. Besides the slight overconsolidation there is also a large deviation in the data set, causing ambiguous interpretations of the data cluster and of the preconsolidation pressure profiles with depth. There are several processes that could cause the overconsolidation ratios as seen in the OCR profile with depth. Little is known about one of these processes: what is the influence of permafrost on the OCR profile with depth? To generate more insight into the overconsolidation properties of the Eem clay a research project was proposed to investigate to what degree historic freezing and thawing affected the preconsolidation pressures of the Eem clay. To answer this question an extensive literature study on the theory of freezing and thawing and on the geology was conducted. From literature it was found that the Eem clay could have been frozen in the Weichselian and other processes like erosion had no effect on the present preconsolidation pressures. The effect of the lowering of the ground water table is not known. A special laboratory test program was set up to look at the effects of freezing and thawing on the measured apparent preconsolidation pressures of reconstituted samples of the Eem clay. The testing program was designed to assess freeze-thaw behavior at effective confining pressures that are similar to the in-situ stresses at the time of freezing (i.e. 50, 100 and 150 kPa or about 5, 10 and 15m below ground), under alternating cycles of freezing and thawing (i.e. 1,3 and 5 cycles). The natural Eem clay sample was dried, crushed and reconstituted at a water content of 1.75 times the liquid limit. The reconstituted slurry was consolidated inside the Rowe cell to the desired confining pressures and samples, cut from slices of the Rowe cell clay sample, were subjected to one- dimensional freezing inside the freezer under the same confining pressure as in the Rowe cell. Meanwhile a reference sample was aged under the same confining pressure. The freezing cell was specially developed for this research and subjected the samples to a closed-system freezing environment with limited access to water. Upon completion of the freezing and thawing program the samples were tested to determine their consolidation properties inside an oedometer. The samples that were frozen showed different consolidation behavior from the reference samples. As expected, freezing and thawing resulted in consolidation of the sample (drop in void ratio). In addition to the drop in the void ratio, the consolidation coefficient (i.e. the slope of the compression curve in a e- 10 log P plot) reduced by an average of 0.075 after several freeze-thaw cycles independent of the confining pressure. The change of the Cc value is possibly due to the reorientation of the clay particles or packets after the ice has broken the bonds between the particles/packets. The change in the void ratio is explained by the expulsion of water upon thawing. In-between 20 and 25% of the moisture drains from the sample after 3 cycles of freezing and thawing. The overconsolidation profile with depth of samples that have been frozen is similar to that of a sample that has been dried although less pronounced. It decreases with depth. The apparent preconsolidation pressure tends to increase after freezing between 0 kPa and 100 kPa, independent of the confining pressure. This is less than would be expected based on the observed change in void ratio, but the decrease in the slope of the virgin compression curve after freezing tends to reduce the apparent preconsolidation effect. The method of determining the apparent preconsolidation pressure was investigated. Six methods were compared and it was found that the Butterfield and Becker method were more consistent and accurate XII than the Casagrande method. The Koppejan method tends to underestimate the preconsolidation pressures. Two methods, the Jacobsen method and the Modulus method were found to be poor indicators of the preconsolidation pressure. It is advised to encourage the use of the Butterfield method as well as the Casagrande method in practice. The hypothesis that the permafrost from the Weichselian was of some influence to overconsolidation ratio measured is not supported by the results of this study. The research indicates that historic freezing and thawing is likely not the cause of the present pre-consolidation pressures of the Eem clay. The effect on the preconsolidation pressure is small in comparison with the present day vertical effective pressure. Probably drying in the Weichselian together with ageing are the causes of the OCR profile obtained by the site investigation.

Small important details, or lack of details, of the current testing procedures for the testing of the buckled discharge capacity of prefabricated vertical drains can have a large influence on the measured discharge capacities. It is shown that for specific types of prefabricated vertical drains there is large difference between the discharge capacity of a single sharp kink and three sharp kinks. This is partly due to the shape of the core, but mainly addressed to the quality of the filter in relation to the fixation of the filter to the core, leading to a particular collapse and thereby blocking the channels. The practical problem of air blocks is also addressed as currently there is no standardized method of removal. The statements are supported with test results of various drain types.

In the recent years a sound, GPS based, logging system has been developed by Cofra B.V. The logging system is used for the registration of, amongst others, the drain coordinates, push forces at selected intervals and maximum depth of the vertical drains. The system has proven itself to be of added value for both the contractor as well as the client and can be used on projects upon the request of the client. Examples are given showing the advantages of this system over the traditional logging system without coordinates and advanced screen for the operator. As the GPS system is capable to load and show AutoCAD drawings to the operator, valuable information can be provided to the operators. These include the locations of underground infrastructure or levels of installation. When strict levels of installation are requested, the operator is given the ability, on specific base units, to choose an option to automatically stop installation at chart datum levels reducing the specific risk for the client. When unforeseen deviations are encountered, the data is shared with the client to investigate the effect on the consolidation process and mark the locations for predrilling. The registered data is also loaded into a GIS system to generate overviews or perform data analysis.

In the last two years, several large infrastructure projects were successfully constructed in the Netherlands with the use of HDPE vertical barriers. The executed schemes used HDPE panels made of 2mm thick HDPE liner welded to patented HDPE Geolock profiles. The panels have a width of 2.5 meter and lengths ranging from 8 meter to 26 meter. During the installation, a hydrophilic swelling cord is incorporated into the locks to create a complete seal. This article describes the application of the HDPE barrier on six different projects. It includes the design aspects, installation methods as well as laboratory measurements with regards to the durability of the material used, water tightness of the system and swelling capacity of the swelling cord when used in salt water. In the projects described, the barriers were used to prevent horizontal flow of water. On the projects the panels were installed by different contractors using equipment supplied by the manufacturer of the HDPE screens. In four projects (three near the city of Leeuwarden and one at the N381), the screens were used in the construction of aqueducts and underpasses. The screens were installed on these projects with a special installation plate in narrow trenches, made using partial predrilling and displacement, as well as in ‘mixed-in-place’ walls up to a depth of 18 meters. Near Nijmegen the screens were used in a river widening project to prevent the flow of water to the low lying land during high river discharges. The panels were installed using a reclaimable weight down to a depth of 26m in a ‘cement-bentonite’ trench made by a diaphragm wall grab. In a project near Rotterdam the panels were also installed in a diaphragm wall. They were installed to prevent ground water inflow into a long low lying section of a new highway.

With the construction of the North/south metro line underneath the center of Amsterdam the properties
of  the  Eem  clay  have  become  of  interest.  The  site  investigation  revealed  that  the  Eem  clay  is
overconsolidated to such an extent that ageing is not the sole cause for this overconsolidation. Besides
the  slight  overconsolidation  there  is  also  a  large  deviation  in  the  data  set,  causing  ambiguous
interpretations of the data cluster and of the preconsolidation pressure profiles with depth. There are
several processes that could cause the overconsolidation ratios as seen in the OCR profile with depth. 
Little is known about one of these processes: what is the influence of permafrost on the OCR profile
with depth? 

To generate more insight into the overconsolidation properties of the Eem clay a research project was
proposed  to  investigate  to  what  degree  historic  freezing  and  thawing  affected  the  preconsolidation
pressures  of  the  Eem  clay.  To  answer  this  question  an  extensive  literature  study  on  the  theory  of
freezing and thawing and on the geology was conducted. From literature it was found that the Eem
clay could have been frozen in the Weichselian and other processes like erosion had no effect on the
present preconsolidation pressures. The effect of the lowering of the ground water table is not known. 

A  special  laboratory  test  program  was  set  up  to  look  at  the  effects  of  freezing  and  thawing  on  the
measured apparent preconsolidation pressures of reconstituted samples of the Eem clay. The testing
program was designed to assess freeze-thaw behavior at effective confining pressures that are similar
to the in-situ stresses at the time of freezing (i.e. 50, 100 and 150 kPa or about 5, 10 and 15m below
ground), under alternating cycles of freezing and thawing (i.e. 1,3 and 5 cycles). 
The natural Eem clay sample was dried, crushed and reconstituted at a water content of 1.75 times the
liquid limit. The reconstituted slurry was consolidated inside the Rowe cell to the desired confining
pressures  and  samples,  cut  from  slices  of  the  Rowe  cell  clay  sample,  were  subjected  to  one-
dimensional  freezing  inside  the  freezer  under  the  same  confining  pressure  as  in  the  Rowe  cell.
Meanwhile  a  reference  sample  was  aged  under  the  same  confining  pressure.  The  freezing  cell  was
specially  developed  for  this  research  and  subjected  the  samples  to  a  closed-system  freezing
environment with limited access to water. Upon completion of the freezing and thawing program the
samples were tested to determine their consolidation properties inside an oedometer.

The samples that were frozen showed different consolidation behavior from the reference samples. As
expected, freezing and thawing resulted in consolidation of the sample (drop in void ratio). In addition
to the drop in the void ratio, the consolidation coefficient (i.e. the slope of the compression curve in a
e-
10
log  P  plot)  reduced  by  an  average  of  0.075  after  several  freeze-thaw  cycles  independent  of  the
confining pressure. The change of the Cc value is possibly due to the reorientation of the clay particles
or packets after the ice has broken the bonds between the particles/packets. The change in the void
ratio is explained by the expulsion of water upon thawing. In-between 20 and 25% of the moisture
drains from the sample after 3 cycles of freezing and thawing.
The  overconsolidation  profile  with  depth  of  samples  that  have  been  frozen  is  similar  to  that  of  a
sample  that  has  been  dried  although  less  pronounced.  It  decreases  with  depth.  The  apparent
preconsolidation pressure tends to increase after freezing between 0 kPa and 100 kPa, independent of
the confining pressure. This is less than would be expected based on the observed change in void ratio,
but  the  decrease  in  the  slope  of  the  virgin  compression  curve  after  freezing  tends  to  reduce  the
apparent preconsolidation effect.

The method of determining the apparent preconsolidation pressure was investigated. Six methods were
compared and it was found that the Butterfield and Becker method were more consistent and accurate


                                                                                                                                                                XII
than  the  Casagrande  method.  The  Koppejan  method  tends  to  underestimate  the  preconsolidation
pressures.  Two  methods,  the  Jacobsen  method  and  the  Modulus  method  were  found  to  be  poor
indicators of the preconsolidation pressure. It is advised to encourage the use of the Butterfield method
as well as the Casagrande method in practice.

The hypothesis that the permafrost from the Weichselian was of some influence to overconsolidation
ratio  measured  is  not  supported  by  the  results  of  this  study.  The  research  indicates  that  historic
freezing and thawing is likely not the cause of the present pre-consolidation pressures of the Eem clay.
The  effect  on  the  preconsolidation  pressure  is  small  in  comparison  with  the  present  day  vertical
effective pressure. Probably drying in the Weichselian together with ageing are the causes of the OCR
profile obtained by the site investigation.

BeauDrain vacuum consolidation has been used for the accelerated consolidation of a 7.5m thick very soft soil layer consisting of peat and clay underneath a new 10.5m high embankment near Amsterdam, the Netherlands. The requirement for the consolidation was a strict 20cm residual settlement over 30 years after construction. The tight construction schedule of the project led to filling rates of up to 0.75m of sand per week. During weekly meetings between the main contractor, the engineer of the client and the specialists of Cofra, the filling rate was evaluated using up to date monitoring data. This led to the safe construction of the maximum 18m thick preloading embankment in a staggering 8 months between September 2011 and April 2012. Due to unforeseen soil replacement sections, within the area to be consolidated, and the very soft peat, which caused major problems for the contractor to construct the working platform, the design was altered during the installation phase. In the altered design the BeauDrain system was combined with BeauDrain-S and Prefabricated Vertical Drains to prevent air leakage and pumping of water from the soil replacement of the neighboring highway. The embankment was handed over to the client within 10 to 12 months after the start of the pumps with settlements of up to 4 meters. This article presents the design, the solutions chosen and monitoring data obtained during the consolidation phase of the project.

Small important details, or lack of details, of the current testing procedures for the testing of the buckled discharge capacity of prefabricated vertical drains can have a large influence on the measured discharge capacities. It is shown that for specific types of prefabricated vertical drains there is large difference between the discharge capacity of a single sharp kink and three sharp kinks. This is partly due to the shape of the core, but mainly addressed to the quality of the filter in relation to the fixation of the filter to the core, leading to a particular collapse and thereby blocking the channels. The practical problem of air blocks is also addressed as currently there is no standardized method of removal. The statements are supported with test results of various drain types.
Keywords: Prefabricated vertical drain, Discharge capacity, Buckled testing, EN-ISO 12958

In the western part of the Port of Amsterdam a new storage terminal of oil products is being built. The site investigation revealed that underneath a single tank the thickness of the compressible layers could differ up to 3 meters. It was concluded by the client that ground improvement was required to avoid excessive differential settlements of the storage tanks and associated maintenance costs. The initial ground improvement design proposed by the client consisted of the application of dynamic replacement (DR). A trial showed that the traditional DR method as well as the CDC technique did not achieve sufficient improvement. Therefore, a full ground improvement was made, with large excavations up to a depth of 8 meters below the surface, removing more than 1,000,000m3 of material. The excavations were backfilled with sand. This very loose sand was compacted in one phase using the CDC technique. This paper presents an overview of the initial trial results and the final work method with a focus on the method of compaction and the compaction results.