Go directly to:
|Abstract||List of contents||graduation||download|
Application of computerized tomography to investigate strain fields caused by cone penetration in sand.
This MSc. thesis study forms the final part of my study Engineering Geology at the department of Applied Earth Science of Delft University of Technology, The Netherlands. it started at January 2003 and was finished in Februari 2004. The project was started on January 2nd 2003 with the literature review. During the months January till May the laboratory work was prepared. The execution of the tests were started in June and lasted until end of November. Parallel with the execution of the tests, test results were processed and this thesis was written. The graduation project was finished with a colloquium on February 27th 2004. The full report (180 pages) can be downloaded from this site.
Members of exam comittee: Dr.ir. D.J.M. Ngan-Tillard, dr. P.L.J. Zitha, dr. R. Ghose, dr L. Gareau, ir. W. Broere and R. Ephraim.
Abstract of Thesis
Everywhere in the Netherlands, where a structure is built, a study of the shallow subsurface is done. It is of major importance to know the geotechnical properties of the soils on which the structure is to be founded. One of the most important methods to explore the subsurface is the cone penetration test (CPT).A large number of correlations between cone penetration test parameters and several geotechnical soil properties exist but have a low degree of repeatability. Empirical relationships require local tuning. Analytical relationships make strong assumptions on failure structures, which are believed to be created by the CPT test but have not been really investigated. Numerical codes have problems in simulating cone penetration correctly and their results are only partially validated with cone resistance measurements.The objective of this thesis is to visualize and characterize the deformation patterns taking place during cone penetration using X-ray Computerized Tomography (CT). Its outcome can be used to improve hypotheses made in analytical models or validate new numerical models, either finite element models or discrete element models. For this purpose, homogenous dry sand samples of various densities are prepared with dry sand pluviation. Baskarp sand is selected due to the finesses, well sorting and small amount of contamination by deviate minerals and materials. Sand samples are created with a density varying between 1,51 g/cm3 and 1,68g/cm3. Small-scale penetration tests are conducted with a 6 mm diameter aluminium cone in calibration chambers with varying density, confining pressures and boundary properties. Measured cone penetration resistances range from 5 MPa in loose sand under low confining pressure (40 kPa) up to18 MPa in dense sand under high confining pressures (80 kPa) (fig. 1). The measured lateral strains ranges from almost zero in loose sand, up to 0,0035 in dense sand. Taking in consideration the measurements of the cone penetration resistance and lateral strain measurements it can be concluded that the repeatability of the small scale CPT test is adequate.To investigate the deformation patterns and failure structures, computerized tomography, a none-destructive investigation method, is used to analyse the sand sample before (fig. 2) and at several stages during the CPT tests (fig. 3). By obtaining the linear relationship between the Hounsfield Values observed in the CT-analyses and the sand density it is possible to define the density changes and volumetric strain in the sand samples.Several scanning parameters are to be set before any CT-scanning can begin: the X-ray intensity X-ray beam strength and slice width. Also the diameter of the sample influences the image quality obtained by the CT-scanner. Several tests show that the optimum scanning parameters are a high X-ray intensity (to decrease noise), a moderate high X-ray beam strength (to decrease beam-hardening effects and to keep the loss of beam sensitivity to a minimum), and a slice thickness of 1 mm (to obtain an adequate resolution in the direction perpendicular to the slices). The diameter of the sand samples used during this thesis study is 10 cm but it is also possible to use larger diameters up to 15 cm. In general the test results showed that a dilating area is formed ahead of the cone tip during the penetration, creating a dilating zone around the probe. A very dilating zone was found at 1-2 mm from the probe and shear bands are believed to be present in these zones. Outside this dilating area an area is observed which is not influenced or slightly compacted by the penetration test. In some cases some compaction takes place in front of the cone. Decreasing the confining pressure result in, besides a decrease of the cone penetration resistance, large differences in size of the area affected and degree of density changes. A larger difference is observed by lowering the sand density. The size of the area dilating is significantly smaller and the degree of dilation is much lower. The use of rigid boundaries only resulted in a small decrease of the dilating area. It was not possible to observe shear bands in the sand sample with the CT- scanner setup present in the Faculty of applied Earth science of the Technical University in Delft. Reasons for this are the not perfect homogeneity of the sand samples and the noise in the CT-scanner.
Figure 1: CT-scan inage before cone penetration test Figure 2: CT-scan image before cone penetration test
Figure 3: CT-scan image after cone penetration test
At the end of the MSc. Thesis study a succesfull presentation was given. During this presentation the problem, objectives, test procedure results and the conclusions were sumarized. The presentation lasted for 30 min and in the end questions were asked by the audience. After the presentation the defence of the MSc. Thesis study took place. The marks for the report and presentation were both an 8. The Presentation (Microsoft Powerpoint presentation) can be dowloaded. During the presentation a small movie was shown to illustrate the creation of the 3D representation of the sand sample with Computerized Tomography. Due to the size (1, 6 Mbyte) this file is only available upon request.
List of thesis contents
Problem statement and research objectives1 Introduction 1.1 Research objectives 2 Literature study 2.1 Sand sample preparation methods 2.1.1 Tamping method 2.1.2 Pepper box method 2.1.3 Sand pluviation method 2.2 Generalities on cone penetration test 2.3 Models and theories for predicting CPT results. 2.3.1 Bearing capacity analysis: 2.3.2 Cavity Expansion theory 2.3.3 Steady State Approach 2.3.4 Incremental finite-element method 2.3.5 Limitations of current models 2.3 Calibration chamber 2.4 Mini Cones 2.5 Deformation patterns 2.6 Parameters studied 3 Small scale CPT equipment. 3.1 Used equipment 3.1.1 Loading apparatus 3.1.2 Probes 3.1.3 Top and bottom plates. 3.1.4 Vacuum system 3.1.5 Points plateau and ball strips to correct to misalignment and eccentricity 3.2 Measuring devices 3.2.1 Vacuum pressure gauge 3.2.2 Load measuring device 3.2.3 Displacement measuring device 3.2.4 Horizontal Strain measuring device 3.2.5 Plate displacement devices 3.2.6 A/D devices 3.3 Accuracy of measurements 3.4 Test Point program 4 Sample preparation for cone penetration testing 4.1 Selection of sand preparation technique 4.2 Sample preparation using pluviation technique 4.3 Pluviation in other studies 4.4 Pluviation tests in this study 4.5 Influence of parameters 4.6 Discussion of the parameters 4.6.1 Falling height 4.6.2 Funnel size 4.6.3 Sand characteristics 4.6.4 Duration of sample preparation 4.7 Density-Falling height relationship for Baskarp sand. 5 Cone penetration testing 5.1 Procedure 5.2 Calibration of the CPT tests. 5.3 CPT test results 5.3.1 Density - cone resistance relation 5.3.2 Density - volumetric strain in the sample. 5.4 Movement of the top plate 5.5 Influence of confining pressure 5.6 Influence of boundary rigidity 5.7 Influence of stopping during testing 5.7.1 Discontinuous tests 5.7.2 Measured penetration resistance during CPT tests scanned by the CT scanner 6 Conclusions of Part A 7 Introduction of Part B 7.1 Research objectives
8 Literature study of X-ray computerized tomography 8.1 CT scanner 8.1.1 System of X-ray scanner 8.1.2 Attenuation 8.1.3 Hounsfield units values 8.1.4 Parameters 8.1.5 Artefacts 8.1.6 Resolution of the CT scanner 8.1.7 Performances of the CT scanner used in this study 8.2 Previous CT-studies on sand deformation by loading. 8.3 Processing of CT-scans 8.3.1 Amira Program 8.3.2 Qwin 9 Optimal parameters 9.1 Optimum sample diameter 9.1.1 Sand sample diameter: 310 mm 9.1.2 Sand sample diameter: 260 mm 9.1.3 Sand sample diameter: 200 mm 9.1.4 Sand sample diameter: 100 mm 9.1.5 Conclusion optimum diameter 9.2 Optimum X-ray energy 9.3 Optimum X-ray intensity 9.4 Optimum slice width 9.5 Optimum size of scanned area. 9.6 Optimum digital filter 9.7 Chosen parameter setting during further testing 10 Density-HU Relation 11 The scanned CPT tests 11.1 Test description 11.2 Data processing 11.3 Test results 11.3.1 Analyse of the absolute-density-zones 11.3.2 Analyse of the density-change-zones 11.3.3 Analyse of the plotlines 11.4 Discussion 11.4.1 General 11.4.2 Noise in the CT images 11.4.3 Comparisons between the four CPT tests analysed with the CT-scanner 11.4.4 Comparison of results with previous studies 12 Conclusions Part B 13 General conclusions and recommendations 13.1 Conclusions 13.2 Recommendations 13.2.1 Test setup 13.2.1 R 13.2.2 Testing program 13.2.3 Practical use of test results REFERENCES ---------- A APPENDIX A A.1 Sand selection A.2 Relative density B APPENDIX B B.1 Appendix B1: strain calculation B.2 Appendix B2: probe strength calculation C APPENDIX C C.1 Specification of the CT scanner D APPENDIX D D.1 Vertical plotlines D.2 Horizontal plotlines D.3 Absolute densities; D.4 Density changes in vertical direction D.5 Density changes at horizontal places E APPENDIX E E.1 Pictures of equipment E.2 Design drawings of top and bottom plate
List of contents
Full text of thesis report available in PDF (4.6 Mbyte)
Website: Jeroen van Nes, The Netherlands
Update: 30 May 2005