VIEWS: 4 PAGES: 7 POSTED ON: 9/23/2011
FIELD AND LABORATORY TESTS INVESTIGATING SETTLEMENTS OF FOUNDATIONS ON WEATHERED KEUPER MARL Hornig, E.-D. Bergakademie Technical University Freiberg, Institute of Geotechnical Engineering, Germany ABSTRACT: Measured settlements of buildings on the weathered Keuper Marl appeared to be much smaller than calculated settlements, which were based on stiffness modules from standard oedometer tests. Therefore both special triaxial K0-tests and oedometer tests were carried out for an accurate determination of stiffness moduli. Modules obtained in the triaxial K0-tests were at least two to three times the values obtained in the oedometer tests. To verify observations from the laboratory tests, the loads and the settlements of two single footings on weathered Keuper mudstone have been measured during construction of a building during one year. Also a large scale footing load test with measurements of deformations was conducted on the weathered Keuper mudstone. The measured settlements of the two single footings and the tested foundation were compared with the settlements based on conventional calculations with moduli from oedometer tests and triaxial K0-tests. Up to a foundation pressure of σv=500 kN/m² the calculated settlement with ES-modules from triaxial K0-tests was found to correspond well to the measured deformation. For foundation pressure beyond 500 kN/m², the foundation response was highly non-linear and it could not be described any more with the linear-elastic model. Therefore the footing load test was also simulated by FEM analyses. Keywords: stiffness; mudstone; oedometer; K0-Triaxial; field-measurements 1. INTRODUCTION Observations and measurements at actual constructions prove that calculated settlements of foundations in the Keuper mudstone, based on oedometer tests, are usually too large. By the way, the German standard DIN 4019 therefore provides the possibility of reducing the calculated settlements by a factor down to ½. Figure 1. Natural structure of Keuper Marl One of the main problems of using standard Several series of such “K0-tests” have been oedometer tests for getting a module of deformation performed for different soil samples. The first is the disturbance of the soil samples when they are results by Illner [1] and Schmidt/Illner [2] proved put into the oedometer ring, as these soils are often that the modules of deformation Es from K0 triaxial laminated and brittle (Figure 1). tests are three times higher than those gained by conventional oedometer tests. A significant improvement can be gained by conducting lab tests in a special triaxial apparatus, In the research described in this paper, two parallel namely a one-dimensional compression under tests were conducted on the same soil specimen in oedometer conditions but with a free lateral surface order to compare the stress strain behavior: one in (without ring), which means that stress is controlled the oedometer test and one in the triaxial K0-test. It in such a way that lateral strain remains zero (K0 appeared that modules obtained in the triaxial K0- condition). tests were at least two to three times the values obtained in the oedometer tests. 1 A large scale footing test was performed in order to After preparation an oedometer test was performed. verify settlement predictions based on parameters The other, bigger, part of the drill core sample (h/d obtained by K0-triaxial tests. In this context, = 12-20cm/10cm) was fitted without much settlement analyses have been made using preparation into the Triaxial apparatus for a conventional methods as well as numerical by using compression test under K0 conditions. the Finite Element Method (FEM). The oedometer tests were performed on a Taking into account the load-displacement behavior conventional basis by increasing the load step by beyond the regular working load of the footing, the step. The load was doubled after 24 hours. The FEM analyses have been performed using different triaxial compression test was performed in a fully types of constitutive models. computer controlled testing unit described by Menzies [3]. In both tests a first loading and several 2. LABORATORY TESTS unloading- reloading steps were carried out. It is very difficult, if not impossible, to put a representative soil sample out of the weathered Keuper mudstone without disturbance of the structure into the oedometer ring (Figure 2 left). The reason is that for these structured soils the dimensions of the oedometer ring are too small. Figure 4 + 5. oedometer testing unit (left) and K0-Triax testing unit (right) The horizontal deformations of the sample were measured by radial sensors with Hall Effect Semiconductors (Clayton [4]) (Fig. 4). The K0 controlled compression tests were performed on Figure 2 + 3. Disturbed sample in oedometer ring (left) and undisturbed sample for K0-Triax (right) unsaturated, drained samples (corresponding to oedometer tests). They were programmed in such a Alternatively triaxial K0-tests on samples with a way that during a certain time period the cell diameter of 100 mm (size of the drilling core) and a pressure increases continuously to a given height of 120 - 200 mm can be performed with maximum or minimum. relative ease (Figure 2 right). Table 1. Index values of the tested weathered Keuper mudstone: moisture content w [%] 13 - 14 liquid limit wL [%] 33,6 – 38,3 plastic limit wP [%] 13 – 17,8 plasticity index IP [-] 14,2 – 16,6 Figure 6 + 7. sample for K0-Triax with Radial sensors consistency index IC [-] 1,2 – 1,4 2.2. Test Results consistency semi-solid – solid The measured deformations from the oedometer dry density ρd [g/cm³] 1,85 – 1,95 were compared with the results gained from K0 triaxial compression tests. Figure 8 and 9 show results of two typical samples of the weathered 2.1. Sample Preparation and Testing Procedure Keuper mudstone. Out of the same drill core sample, one part was fitted into the oedometer ring (h/d = 2,0cm/7,1cm). 2 10 100 1000 σ 0 [kN/m2] 10000 For the initial loading of sample a mean ratio H* of 0,00 2.11 between stiffness modules of K0-Triax and of K0-Triax sample 1 / Keuper Oedometer was calculated. Over whole stress range 0,01 the ratio did not vary much. At reloading a mean 0,02 ratio H* of 2.46 was gained, at which the ratio placement data Oedometer decreased from 3.45 to 1.31 during the area of stress Oedometer 0,03 Wn = 13.0 % increase. ρ d = 1,91g/cm³ K 0-Triax 0,04 Wn = 13,0 % Table 3. Comparison Stiffness modules ES of Keuper 2 ρ d = 1,96 g/cm³ between Oedometer and K0-Triaxial: ε[- 0,05 ] Figure 8. Test results Keuper mudstone sample 1 stress range first loading Oedometer K0-Triax ratio σ0 ES (EB) ES (EB) H* 10 100 1000 σ 0 [kN/m²] 10000 [ kN/m2 ] [ MN/m2 ] [ MN/m2 ] [-] 0 100-200 9 33 3,7 0,01 sample 2 / Keuper K0-Triax 200-400 15 44 2,9 0,02 400-800 20 80 4,0 0,03 800-1200 33 90 2,7 placement data 0,04 Oedometer Oedometer mean 3,3 Wn = 14,5 % 0,05 ρ d = 1,84 g/cm³ reloading 0,06 K 0-Triax Oedometer K0-Triax ratio σ0 Wn = 13,8 % 0,07 ρ d = 1,96 g/cm³ ES (WB) ES (WB) H* ε [-] [ kN/m2 ] [ MN/m2 ] [ MN/m2 ] [-] 0,08 50-100 Figure 9. Test results Keuper mudstone sample 2 100-200 41 73 1,8 200-400 55 84 1,5 Table 2. Comparison Stiffness modulus ES of Keuper 1 400-800 67 129 1,9 between Oedometer and K0-Triaxial: mean 1,7 stress range first loading For sample Keuper 2, which was got from a Oedometer K0-Triax ratio different site, the corresponding mean value of H* σ0 ES (EB) ES (EB) H* for initial loading was 3.3. Thereby H* did not vary [ kN/m2 ] [ MN/m2 ] [ MN/m2 ] [-] much, too. At reloading, however, a mean ratio H* 50-100 13 25 1,92 was 1.7, the variation was small, the amount of H* 100-200 17 40 2,35 200-400 29 67 2,31 was nearly the half the value of initial loading. 400-800 42 80 1,90 800-1600 58 120 2,07 3. IN SITU MEASUREMNTS mean 2,11 reloading To verify observations from the laboratory tests, the Oedometer K0-Triax ratio loads and the settlements of two single footings on σ0 ES (WB) ES (WB) H* weathered Keuper mudstone have been measured [ kN/m2 ] [ MN/m2 ] [ MN/m2 ] [-] during construction of a building in the City of 50-100 36 Stuttgart during one year. Figure 10 and 11 show 100-200 29 100 3,45 the placed indirect load measuring device in one of 200-400 38 100 2,63 the two columns. Parallel to the measurements on 400-800 105 138 1,31 site, samples taken close to the foundations have 800-1600 been analysed to obtain soil parameters, especially mean 2,46 stiffness moduli from Triaxial K0-tests and H* = ratio of stiffness module gained from K0-Triax Oedometer tests. Using these characteristic values and from Oedometer resp. H* = ES(K)/ ES(O) ES=stiffness module, EB=initial loading, WB=reloading, K=K0-Triax, the settlements of the foundations have been O=oedometer calculated. 3 of 900 kN each were used in such a way that an influence on measurements could be excluded. During the whole test the settlements were measured at three points on top of the foundation, using a measuring scaffold, with bearings far away from the calculated influence area of the footing. Four (re)loading – unloading steps were performed up to the failure of the strip. Figure 14 shows the measured settlements under loading cycles. Large scale footing load test 10 100 1000 σ 0 [kN/m²] 10000 0 50 100 150 Figure 10 + 11*. Placed rebar strain meter in column before 200 placing of concrete *(Martin [5], Hornig et al. [6]) s [mm] 250 4. LARGE SCALE FOOTING LOAD TEST Figure 14. Loading cycles and measured settlements The large scale footing load test (Fig. 12 + 13) with 5. CALCULATIONS measurements of deformations taken at the test foundation itself and at several depths, using 5.1. Conventional Settlement Analyses extensometers, was carried out on the weathered The measured settlements of the two single footings Keuper mudstone. and the footing load test were compared with the settlements based on conventional calculations with Also parallel to the test on site, samples taken modules from oedometer tests and triaxial K0-tests. beneath and close to the test foundation have been analysed to obtain soil parameters. Using stiffness Figure 15 shows the measured and the calculated moduli from Triaxial K0-tests and Oedometer tests settlements of one single footing under construction the settlements of the foundation have been of the building up to foundation pressure of σ0 = calculated. 300 kN/m². comparison: measurements - analytcal calculation / foundation 2 0 50 100 150 200 σ 0 [kN/m²] 250 300 0 5 10 15 20 25 Figure 12 + 13*. Photo (left) and schematic view (right) of the 30 calculation / K0-Triax first-loading module calculation / Oedometer first-loading module large scale footing load test *(Schnürch [7], Hornig [8], 35 calculation / Oedometer reloading module measurement building foundation 2 Hornig et al. [6]) s [mm] 40 Figure 15. Measured settlements – calculated settlements For the test a cylindrical, reinforced concrete footing with a diameter of 1,8 m and a height of 0,8 A comparison of the measured settlements in Fig. m was used. The loading was done by three 3 MN 15 with the calculated based on triaxial K0-test- hydraulic presses. Ten grouted injection anchors modules and reloading modules out of Oedometer with a length of 10.5 m and a max. bearing capacity 4 tests shows a very good agreement. The calculated Table 4. Parameters for the calculations with the FE-Code settlements, which were based on first loading PLAXIS modules out of Oedometer tests, are much too large. Hardening- Mohr- Soil Coulomb The measured and the calculated settlements of the φ’ [°] 30 30 footing load test under loading up to foundation c’ [kN/m²] 35 30 pressure of σ0 = 800 kN/m² are shown in Figure 16. E50 [MN/m²] 38 - EOed 30 - footing load test: measurements - analytical calculation / linear elastic Eur [MN/m²] 80 - Eref [MN/m²] - 46 0 100 200 300 400 500 600 700 σ 0 [kN/m2 ] 800 0 νur [-] 0,2 - ν [-] - 0,315 10 pref [kN/m²] 150 - 20 m [-] 0,5 - 30 To describe the geometrical situation of the 40 measurement cylindrical foundation, the finite element mesh 50 calculation / K0-Triax calculation / Oedometer shown in Figure 17 with 15 node elements under s [ mm ] 60 axial symmetrical conditions was used. After preliminary examination for the influence of the Figure 16. Measured settlements – calculated settlements anchors and the necessary vertical and horizontal expansion, calculations were made with a fine mesh For stresses up to σ0 = 500 kN/m² a comparison of in the central zone and a coarse mesh in the outer the measured settlements in Fig. 16 with the zone. calculated based on triaxial K0-test-modules shows -2.000 0.000 2.000 4.000 6.000 8.000 10.000 12.000 a very good agreement. Since the foundation 10.000 response was highly non-linear for foundation A A pressure beyond 500 kN/m², this response could no 8.000 longer be described with the linear-elastic model. The calculated settlements with a working pressure 6.000 up to σ0 = 500 kN/m², which were based on oedometer test modules, are much too large. 4.000 5.2. Finite Element Method (FEM) Analyses As described in 5.1, beyond a working load of σ0 = 2.000 500 kN/m² nonlinear effects increase. Therefore the footing load test was simulated by FEM analyses -0.000 using the Finite Element Code PLAXIS (Vermeer and Brinkgreve [9+10]) which applies two non- Figure 17. Cross section of the axial symmetrical loaded mesh linear models: the Mohr-Coulomb model and the Hardening-Soil model. The needed model The excavation and the test construction were parameters in table 4 were mainly gained through simulated by stage constructions. The (re)loading – the lab-tests, additionally some were chosen within unloading steps were described by multipliers. FEM simulations. The module EOed from triaxial K0-tests was used. 5.3. Results and Comparison Figure 18 shows the measured and the calculated deformations under loading, unloading and reloading up to the ultimate bearing pressure of σ0 = 1550 kN/m². As described, for the calculations the Mohr-Coulomb model and the Hardening-Soil model were used. 5 comparison: measurements - numerical simulations ACKNOWLEDGEMENTS 0 200 400 600 800 1000 1200 σ 0 [kN/m²] 1400 0 The in situ measurements have been supported by: measurements 20 FE-calculation: Hardening Soil 40 FE-calculation: Mohr Coulomb Hallesche Nationale Krankenversicherung Stuttgart 60 Ed. Züblin AG – Niederlassung Stuttgart 80 Terrasond Gesellschaft für 100 Baugrunduntersuchungen mbH & Co. KG 120 Dr. G. Hafner, Büro für Ingenieurgeologie, Erd- 140 und Grundbau, Stuttgart 160 FH Stuttgart – FB Vermessung und Geoinformatik 180 s [mm] 200 The large scale footing load test has been supported Figure 18. Measured and calculated settlements by: Only the Hardening-Soil model could describe the Daimler-Chrysler AG – Werk Sindelfingen deformations under loading, unloading and Ed. Züblin AG – Niederlassung Stuttgart reloading realistically. Simulations using the Mohr- Bauer Spezialtiefbau GmbH Coulomb model only represent the initial loading Terrasond Gesellschaft für well. Baugrunduntersuchungen mbH & Co. KG Bauunternehmen Hämmerle GmbH & Co. KG 6. CONCLUSIONS FMPA Baden-Württemberg – Abteilung 4 – Geotechnik For base pressures of up to working load, the Smoltczyk & Partner GmbH settlements for the measured single footings and for FH Stuttgart – FB Vermessung und Geoinformatik the tested large scale footing can be described very well with the linear-elastic model within REFERENCES conventional settlement analyses. An accurate determination of the deformation parameter is very 1. Illner, C. (1997). Verformungsverhalten von Gipskeuperböden. Forschungsbericht Fachhochschule important for the investigated soils. The K0-test in Stuttgart – Hochschule für Technik, Labor für the triaxial apparatus is therefore very useful. Geotechnik, unveröffentlicht Alternatively the use of reloading-modules from Oedometer tests can also give practical results. 2. Schmidt, H.-H., Illner, C. (1998). Bestimmung von Verformungsmoduln an Gipskeuperproben. Tagungsband der 25. Baugrundtagung, Stuttgart. Deformations beyond base pressures of σ0 = 500 kN/m² should be simulated numerically by using 3. Menzies, B. K. (1988). A Computer Controlled non-linear models. In this case, the stiffness Hydraulic Triaxial Testing System. American Society modules out of triaxial K0-tests should be used as for Testing and Materials. STP 977, Philadelphia, 82- well. Calculations with the Finite Element Code 94. PLAXIS (Vermeer and Brinkgreve [9+10]), which 4. Clayton, C. (1989). The Use of Hall Effect are based on the Mohr-Coulomb and the Hardening- Semiconductors in Geotechnical Instrumentation. Soil models, achieve good results. Geotechnical Testing Journal. Vol. 12, No. 1, 69-76. 5. Maritn, S. (2001). Messung und Berechnung von Setzungen an Einzelfundamenten im Gipskeuper. Diplomarbeit, Institut für Geotechnik, Universität Stuttgart, unveröffentlicht 6. Hornig, E.-D., Buchmaier, R. F., Schmidt, H.-H. (2002). Eindimensionale Kompression überkonsolidierter bindiger Böden am Beispiel des Gipskeupers. Forschungsbericht, Fachhochschule Stuttgart – Hochschule für Technik. 6 7. Schnürch, R. (1998). Experimentelle Ermittlung geotechnischer Kenngrößen für die Setzungsberechnung bei überkonsolidierten Böden am Beispiel des Gipskeupers. Diplomarbeit, Fachhochschule Stuttgart – Hochschule für Technik, unveröffentlicht 8. Hornig, E.-D. (2000). Fundamentprobebelastung auf verwitterten Gipskeuperböden, Spezialsitzung „Forum für junge Geotechnik-Ingenieure“. 26. Baugrundtagung, Hannover, 60-61. 9. Vermeer, P. A., Brinkgreve R.B.J. (1998). PLAXIS Bedienungshandbuch Version 7. A.A. Balkema, Rotterdam. 10. Vermeer, P. A., Brinkgreve R.B.J. (1998). PLAXIS Material Model Manual Version 7. A.A. Balkema, Rotterdam. 7