

R21 FREEWAY SINKHOLE
CSW was used to evaluate the effectiveness of dynamic compaction which had been undertaken of a large dolomite subsidence.
It was also used as a quality assurance measure to confirm the compactive effort.

STELLENBOSCH LIBRARY
CSW test was used to derive geotechnical design parameters for a boulder alluvial soil on which a new four-storey library was constructed at Stellenbosch University.
The test assisted in the geotechnical soil characterisation of the site, derivation of engineering parameters and design.

STEYN CITY FOUNDATIONS
CSW was used to derive the stiffness profile for a friable granite which was difficult to sample and test in a laboratory.
Based on the testing it was possible to derive a much stiffer profile for the soils and subsequently conventional pad foundations could be utilised.
Overview
Over the past 30 years or so the role of geophysical methods in characterising sites and materials has been steadily increasing internationally (Stokoe et al., 2004). Locally, this is probably less so of the case, except for two exceptions; the use of geophysical methods for dolomite investigations and the CSW test; this article focuses on the latter which has found applicability on a range of typical geotechnical problems.
CSW is a geophysical exploration technique which is used to evaluate the subsurface stiffness using a mechanical vibrator and receivers (or geo-phones) which are placed in linear array. The test involves measuring Rayleigh Wave velocities as they propagate through the soil mass. The velocities measured by the geophones are then converted, by an experienced analyst, to a corresponding stiffness profile with depth at the position of the test. Depending on the size of the shaker used and the specific ground profile, the CSW is generally limited to measuring the very near surface profile (typically to depths of 6m −12m). For A detailed discussion on the stress waves and CSW the reader is referred to the works by Stokoe (2004) and Heymann (2007).
Soil stiffness depends on complex interactions of state (i.e. bonding, fabric and so on), strain level, stress history, and type of loading. A key concept to understand in soil and structure interaction is that the stiffness-strain relationship is very strongly non-linear and that different structures are designed to accommodate different strains. The results of the CSW tests are given as G0 (shear stiffness at small strain) with depth, this is much higher than the stiffness used for say the design of foundations for a large bridge,. The shear stiffness can also be related back to an equivalent Young’s Modulus through the Poisson’s Ratio.
Once the problem has essentially been framed (probably by use of other investigative techniques or previous experience), the CSW can be used to target specific questions related to the ground conditions or to supplement other methods to provide greater confidence in the interpretation of the ground conditions. Three practical examples are provided below, each highlighting a particular useful trait of the test.
Jukskei Bridge at Steyn City, Gauteng
Consultants: Begin Africa (main consultants), Vela VKE (bridge and geotechnical engineers),
Project value (bridge): R55million Contractor: Stefanutti Stocks Client: Steyn City Developments/Begin
The first example, is the case of a new prominent bridge over the Jukskei River (Photo 1). The bridge is a 5 span bridge of equal spans of 30m with central pier heights of up to 15m. Early in the development of the project the quality of the previous geotechnical investigation was highlighted as a concern and an additional geotechnical investigation was undertaken.
The bridge is situated on the Halfway House Granites which are notoriously variable and, given this variably and the quality of the previous investigations (a quality which the client appears have become accustomed to), it is understandable that the client was hesitant to invest in additional geotechnical investigations and design. Previous investigations essentially indicated that all piers and abutments would need to be founded on piles of 10-15m based on very crude percussion drilled boreholes. Percussion drilling involves air drilling boreholes and achieves limited sample recovery, and is probably the least suitable investigative technique to be used on a granite profile.
The subsequent investigations which included core-drilling and CSW, indicated that only one of the abutments and the two central piers immediately adjacent to the river would require piles; with the recommendation of piles for the central piers largely been driven by concerns on scour and constructability rather than the consistency of the soils.
In this instance the CSW was used to derive a stiffness profile of the granite at some of the piers where the granite gradually graded with depth to competent rock. The original assumption of founding only once solid bedrock was encountered ignored the favourable weathered granitic dense sands and even soft rock overlying the solid bedrock at depth.
Pier foundations were designed for bearing pressures of 400kPa and 10mm of settlement. Settlements measured during construction were within the expected range for the loads applied at that stage, and retrospectively the decision to found only some of the piers on piles justified and realised through a R2million savings at a nominal cost of some R450,000 for the drilling and CSW tests.
Engineering Building Extensions, University of Stellenbosch
Consultants: civil and geotechnics -Vela VKE; structural engineers – Ekcon, Bart Senekal and BKS. Project value: R75million Client: University of Stellenbosch
The second example is the case of extensions to the existing engineering building at the University of Stellenbosch. A number of additions including a library, laboratory and workshops were proposed adjacent to the existing engineering building.
The significant feature of the geology of the area is the presence of the wide paleofluvial plain of the Eerste River on which Stellenbosch is situated. Locally the University is on a plain of coarse boulder clay alluvium over 3m thick which was essentially formed after large sandstone blocks slid gradually from the adjacent mountains into the valley floor, due to deep weathering of the underlying phyllite, and transported through fluvial processes, Söhnge and Greeff (1985). What must be highlighted from an engineering point of view, is that the combination of deep weathering of the phyllite and the formation and weathering of the boulder alluvium results in highly variable ground conditions with weathered and poor soils.
The effect of the 2 to 3m thick boulder layer is that most buildings in Stellenbosch (generally 4 storeys, but up to 7 storeys), have been founded at shallow depth (2m or less) on this interlocking boulder layer, albeit using low bearing pressures (in the order of 300kPa) and are according to Brink (1985) “remarkably free of cracking”. This includes the existing engineering building which is reported to be founded at 1,8m depth, on the boulder clay layer with a bearing pressure of 225kPa.
The problem with the above profile is that there is no means of testing the layer to derive a modulus of compressibility for the boulder-clay layer; the existing buildings possibly being constructed on a trial-error or experiential bases. A requirement for laboratory testing would be to obtain good quality samples with minimal disturbance, unfortunately this is not practically possible and any type of penetration test would simply “refuse” on the boulders.
Brink highlights this difficulty by providing a broad range of bearing pressures which have been previously used on the boulder clay, from as little as 175kPa to upwards of 400kPa. CSW tests were subsequently undertaken to evaluate the overall stiffness of the boulder layer in a more sophisticated manner. The CSW showed the stiffness for the boulder clay and residual phyllite to be surprisingly consistent with no significant change in stiffness and a gradual improvement with depth.
Foundations were eventually placed at a nominal depth of 1m on the boulder clay layer and designed for a bearing pressure of 200kPa. Although not very different from that used previously for the engineering building, the bearing pressure was at the lowest end of the range provided by Brink. No untoward settlement was recorded during construction or the subsequent year or so following construction.
Subsidence on R21-Freeway, Olifantsfontein – Dynamic compaction of stone columns
Consultants: Vela VKE Project value (remedial works contract): R15million Contractor: Raubex (and Franki as specialist Subcontractor)
In the final example Vela VKE were appointed design engineers for the R3.6 billion 45km upgrade of the R21 freeway as part of phase 1 of SANRAL’s visionary network improvement project – GFIP. Over half of the R21 route is located on dolomitic land, and for a large portion actually follows the contact between the Timeball Hill Shales and Malmani Dolomite along which some of the poorest dolomite occurs. Towards the end of construction a subsidence formed just south of the slip-lane to the off-ramp of the Olifantsfontein Interchange.
Geotechnical investigations of the depression followed, with a remedial contract let shortly after. The depression was ascribed to a deeply weathered dolomitic profile. With no significant cavities identified, the remedial work was thus aimed at stiffening the ground profile where the subsidence occurred. This was done by pounding in stone columns by means of dynamic compaction (Photo 3), placing a geotextile over the compacted area as well as a 1m thick granular engineered soil mattress (here a Colto G6 material) on top of the geotextile, where after pavement layerworks were reinstated.
In this case the CSW was used to evaluate the effectiveness of the dynamic compaction and of the G6 raft. For the dynamic compaction and particularly the stone columns the CSW is probably on the limit of its practicality. But for the G6 layer clear correlations could be drawn between the CSW and the CBR and Plate Load Tests which were conducted.
One would typically expect a G6 material with a CBR of 25 – 45% to have a small strain stiffness in the range of G0 = 110 – 210MPa. CSW test small strain stiffnesses were in the range of G0 = 180 – 260MPa. The plate load test, although a crude test, can provide a good estimate of the soil stiffness, the plate load test derived Young’s modulus on top of the G6 mattress was in the range of E = 97 -225 MPa, corresponding to a G0 = 125 – 290MPa. The soil stiffnesses thus derived from the plate load tests correspond closely to those determined from the CSW test and what is expected for a G6 material.
Conclusions
The geotechnical engineer’s role on a project to characterise the near surface soils and derive engineering parameters in order to design the structures which are to be founded on them is greatly enhanced by having a varied toolbox of investigative techniques and test methods. Ultimately no single method is globally applicable to all ground profiles and projects. Nor, can reliance be placed on the outcome of one single method of investigation, even on a small project.
The CSW adds to geotechnical engineer’s armoury. It provides the stiffness profile for the near surface soils; in one of our examples it provided the overall stiffness of a difficult alluvial profile. It is non-intrusive which makes it cost effective and, in the case of a construction project, would not disrupt the works.
There are a number of limitations to the use of such geophysical methods and certainly on its own its applicability is limited. For example the presence of near surface buried structures or extremely stiff layers can affect the interpretation of the results, and even reduce the depth of penetration of the test. It can also not be extended to interpreting soil behaviour such as heave or collapse.
However, the above examples show how CSW tests have been successfully utilised in geotechnical investigation, design, and construction assurance. Beyond that, this method has proven useful in saving clients’ money when used appropriately by experienced geotechnical engineers. Given the multitude of factors controlling the interpretation of ground conditions the CSW will increasingly play an important role in geotechnical engineering as the test gains familiarity and given that many of the soils encountered across the country are actually difficult to sample and/or test in a laboratory.
Acknowledgements and references
The authors would like to thank Professor Gerhard Heymann (University of Pretoria) and Mr Alan Parrock (ARQ Consulting) for their assistance during the execution of the projects discussed.
A list of works cited can be obtained by contacting the authors.