th IMEKO TC International Symposium and 1th International Workshop on ADC Modelling and Testing Benevento, Italy, September 15-17, 1 Field measurements of the settlements induced by preloading and vertical drains on a clayey deposit Ernesto Cascone 1, Giovanni Biondi 1 University of Messina, Contrada di Dio, S. Agata, Messina, ecascone@unime.it University of Messina, Contrada di Dio, S. Agata, Messina, gbiondi@unime.it Abstract The paper describes the results of an almost two-year long monitoring period of the field performance of the foundation soils of two tanks founded on a clay soil deposit improved using the technique of preloading and vertical drains. During the preloading period settlements were monitored measuring the vertical displacements of settlement platforms by topographic survey. During the hydraulic leakage tests of the tanks their settlements were measured using a level probe along a flexible pipe placed in a shallow excavation under each tank. The measurement results are discussed showing their usefulness in the application of the Asaoka procedure for settlements prediction. Despite the significant influence of the lithological and mechanical heterogeneity of the foundation soils, absolute and differential settlements of the tanks, as well as angular distortions, resulted consistent with allowable limits envisaging a satisfactory performance of the tanks. I. INTRODUCTION The use of the observational approach in geotechnical engineering projects allows savings in cost or programme [1]. Accordingly, it was widely adopted with reference to full-scale geotechnical systems or reduced-scale physical models, to monitor or predict the static [e.g. -] and the seismic [e.g. 5-7] performance. In this framework, the paper describes the results of an almost two-year long monitoring period of the field performance of the foundation soils of two large steel tanks (diameter of 19 m and height of 15. m) founded on a medium to stiff clayey and silty soil deposit improved using the technique of preloading associated with m long prefabricated vertical drains. Figure 1 shows the plan view of the construction site together with some relevant sections describing the main preloading embankment (H = 9. m) and the placement of the two tanks whose foundation consist of a reinforced concrete ring filled with compacted sand. In-situ investigations, including boreholes, cone penetration tests and dissipation tests, and laboratory tests were carried out to define the geotechnical profile of the construction site and the mechanical properties of the foundation soils of the two tanks. Furthermore, an extensive field monitoring of the site was carried out during the preloading period and during the hydraulic leakage test of the tanks. A detailed description of the geotechnical characterization and of the work sequence is provided in []. In the present paper some of the measurement results are presented highlighting the significant influence of the lithological and mechanical heterogeneity of the foundation soils on the tanks settlement response. The monitoring results are discussed showing their usefulness in the application of the Asaoka procedure for settlements prediction and in the detection of possible occurrence of serviceability limit states of the tanks during their service life. II. MONITORING DURING THE PRELOADING PERIOD During the preloading period and the preloading embankment removal settlements were monitored measuring, by topographic survey, the vertical displacements of settlement platforms (SP in Fig. 1). Measurements started some days before the beginning of the embankment construction (t = ) and were taken about every 3 days for a period of days, until the embankment was removed. Figure a,b and 3a show the time-settlement curves observed at the platforms in the areas where the two tanks had to be built and in the centre of the main preloading embankment. The curves exhibit the typical concave shape in the first 9 days (embankment construction) and then become convex during the preloading period. A sudden increment of settlements was measured at about 1 days from the beginning of the embankment construction, due to the achievement of the full saturation in the drainage system []. Differences in the values of final settlements are apparent depending on the location of the settlement plates. Maximum settlements under the embankment were in the range 3-39 cm and, as expected, occurred along the longitudinal axis (SP, SP1, SP1); points located north of this axis (i.e. SP5 and SP17) showed a stiffer response than the corresponding points located south of the axis (i.e. SP7 and SP19), with differences in measured settlements of about. cm between SP5 and SP7 and of about 7 cm between SP17 and SP19. This result may be ascribed to the lithological and mechanical heterogeneity of the soil deposit. ISBN-1: 97-9-9973--7
D-D transversal section original ground surface m long PVDs D 9. 53. 19 9. D. 3 7 15. B-B, C-C transversal section original ground surface m long PVDs B C. TANK 1, 19 B C th IMEKO TC International Symposium and 1th International Workshop on ADC Modelling and Testing Benevento, Italy, September 15-17, 1 A-A longitudinal section 3 19 3 9 1. 19 TANK 1 TANK preloading embankment.. tank foundation embankment original ground surface m long PVDs 3 SP9 SP1 SP1 A SP5 SP11 TANK SP17 A 7 SP SP SP SP SP1 SP1 SP1 SP 11 13 SP7 TANK 1 SP13 SP19 SP3 19 3 19 SP1 SP15 15. 3 9 9. N 9. m high (main) embankment 5. m high embankment 1 3 5 m Fig. 1. Plan view, transversal and longitudinal sections and location of settlements platforms (SP). After 1 days from the beginning of the embankment construction in most of the monitored platforms the rate of settlements reduced. The application of the Asaoka observational procedure allowed plotting the diagrams shown in Figure c,d where, for the platforms placed in the area of Tank 1 (SP-SP) and Tank (SP1-SP), settlements -1, measured at time step k-1, and, measured at time k, are compared. Similarly, the plots of Figure 3b were derived for the settlements measured at the platforms (SP11- SP13) placed along the transversal section of the main embankment. In all the plots of Figure c,d and 3b the final points are almost aligned to the lines bisecting the diagrams; this suggested that only small settlement increments should be expected under the loading transmitted by the preloading embankment which, then, could be removed. Accordingly, the embankment removal started at t = days from the construction and was completed in 1 days (t = 35 days). III. MONITORING DURING THE HYDRAULIC LEAKAGE TESTS The hydraulic leakage tests started at t = 5 days for Tank 1 and at t = 1 days for Tank. The tanks were filled of water in (Tank 1) / 1 (Tank ) days, were kept full for (Tank 1) / 35 (Tank ) days and were finally emptied in 13 (Tank 1) / 1 (Tank ) days. During the tests tanks settlements were measured using a level probe along a flexible pipe placed (before construction) in a shallow excavation under each tank along the direction of the transversal axis of the preloading embankment. Settlements were measured every -3 days at five aligned locations under the tanks: at the edges under the foundation ring, at ¼ and ¾ of the diameter and at the centre of the tanks. Figure a,b shows the settlement histories at the selected five locations under the tanks. At all measurement locations, settlement increased to a maximum and then kept constant while the tanks were full.
w w th IMEKO TC International Symposium and 1th International Workshop on ADC Modelling and Testing Benevento, Italy, September 15-17, 1. 5 1 15 5 3 -. -.3 SP 5 -.5 -.1 -.15 SP SP 5 SP SP 7 SP -. -.3 SP -. -.1 SP SP -. -. -.1 -.5 -.3 -.1 -. -.3-1 -.3 SP 7 -.1 -. -.3 -. -1 -.35 -. -.1 -. a) area of Tank 1 c) area of Tank 1. -.5 SP 1 SP 17 SP 1 -. -.3 -. SP 17 -.1 SP 19 SP -.1 -.15 -. -. -.3 -. SP 1 SP 1 SP -.5 -.1 -.3 -.35 -.1 -. -.3-1 -.3 -. SP 19 -.1 -. -.3 -. -1 -. b) area of Tank -.1 -.1 -. -.3 -. -1 d) area of Tank Fig.. Preloading period: settlement time-histories (a,b) and Asaoka plots (c,d) observed at the platforms in the areas where the tanks had to be built. During the emptying phase a significant fraction of the settlements was recovered. Maximum settlements w max under the centre of the tanks are about -7.5 cm. Figure c,d displays the Asaoka s plots for the settlement measurements carried out at the centre and at one of the tank edges: points in the plots exhibit, especially for the case of Tank 1, a trend to align along the bisector. At the edges of the two tanks the settlement variation, due to filling and emptying, resulted always less than 3 cm, consistently with the prescribed design limit. This confirmed the effectiveness of the improvement technique adopted at the site and allows predicting a satisfactory performance of the tanks.
w w. 5 1 15 5 -.5 -.1 -.15 -. -.5 -.3 -.35 -. a) SP 11 SP 1 SP 13 -. -.3 -. -.1 -. -.3 -. -.1 -. -.3 -. -.1 SP 11 SP 1 SP 13 b) -.1 -. -.3 -. -1 Fig. 3. Preloading period: settlement time-histories (a) and Asaoka plots (b) at the centre of the embankment. In Figure 5 the settlement profiles are shown in a normalized diagram where the settlement w and the radial distance r are divided by the settlement w max and by the tank radius R, respectively. Both tank foundations exhibit the classical dish-shaped settlement profile with larger values near the centre and settlements that decrease smoothly toward the edges. Despite the symmetry in the tank structures and in the applied load, settlement profiles are not symmetric; again, a stiffer response of the northern area of the tank foundation can be attributed to the lithological and mechanical heterogeneity of the alluvial deposit. The effect of the soil heterogeneity on settlement response becomes apparent by comparing the profiles of measured settlements with the profile shapes (shaded area in Fig. 5) suggested by D Orazio and Duncan [] for tanks overlying homogeneous soil deposits. IV. CHECKING SERVICEABILITY LIMIT STATES OF THE TANKS As it is usual in the case of dish-shaped settlement profiles [9,1], the possible occurrence of a serviceability limit state of the tank bottom has been checked referring to threshold values lim of the differential settlement R between the centre and the edge of the tank and to threshold value lim, of the angular distortion [, 9]. a) Tank 1 5 15 5 35 5 55 5 75 tank edge 1/ diameter tank centre 3/ diameter tank edge b) Tank 3 5 7 9 tank edge (d=.3 m) tank centre (d=1.7 m) tank edge (d=.3 m) tank centre (d=1.7 m) -1 c) Tank 1 d) Tank -1 Fig.. Hydraulic tests of the tanks: a,b) settlement time-histories; c,d) Asaoka plots.
R lim = 1/ (Sowers, 19) th IMEKO TC International Symposium and 1th International Workshop on ADC Modelling and Testing Benevento, Italy, September 15-17, 1 R R e) Tank 1 17 days: T1 5% full 3 days: T1 full 5 days: T1 full 7 days: T1 empty South distance 5 1 15 5 North f) Tank 37 days: T 9% full days: T full 7 days: T full days: T empty South r/r North South r/r North -1.5-1 -.5.5 1 1.5-1.5-1 -.5.5 1 1.5.. w /w max.. w /w max... 1 D'Orazio and Duncan (197). Fig. 5. Hydraulic tests of the tanks: settlement profiles. 1 1 D'Orazio and Duncan (197) lim =.5 D D'Orazio and Duncan (197) lim =.5 D 1 Tank 1 d =.3 m d = 3.1 m max =.w max (Coduto, 1) Skempton and MacDonald (195) 1 w max a) start emptying.5.5.75.1 max Fig.. Checking serviceability limit state of the tanks bottom. b) In Figure a the differential settlements R, computed at both edges (d =.3 m, d = 3.1 m) of the tanks (Fig. 5), are plotted against the maximum absolute settlement w max measured, at different dates, at the centre of the tanks. It can be observed that, despite the considerable values of the maximum settlement measured at the tank centre (w max = 7.5 cm) and the rigid rotation ( o =.5 1%), the differential settlements R are always in the range 3-5 cm and, thus, are much smaller than the threshold value lim = 7.5 cm that can be estimated according to the recommendation by D Orazio and Duncan [9] for steel tanks of diameter D: lim =.5 D. Furthermore, for most of the data plotted in Figure a, the design -w max relationship proposed by Coduto [11] represents a conservative upper-bound. In Figure b the values of R are plotted versus the maximum value of the angular distortion max computed, along the north-south axis of the two tanks, from settlement measurements []. Despite max increases with R, its values resulted generally lower than the threshold limit lim = 1/ suggested by Sowers [1]. Moreover, for most of the data in Figure b the empirical relationship = 35 proposed by Skempton and MacDonald [13] for buildings
is approximately an upper bound. All the computed data are within the shaded area in Figure b delimited by the thresholds limits, lim and lim, suggested by D Orazio and Duncan [9] and by Sowers [1], respectively. The only exception is represented by the measurements carried out at the start of tank emptying which exceed the limit lim = 1/; however, no damage was observed in the tank steel structure and in the piping connections during and after the hydraulic leakage tests. Thus, it is confirmed that the recommended criteria to check the occurrence of a serviceability limit state should be based on maximum and differential settlements rather than on angular distortions. For the considered case study, despite the lithological and mechanical heterogeneity of the foundation soils significantly affected the tanks settlement response, maximum absolute settlements and angular distortions resulted consistent with allowable limits. V. CONCLUDING REMARKS In the paper a well-documented case history concerning the improvement of the foundation soil of two large tanks through the method of preloading and prefabricated vertical drains is described. Data of the observed behaviour during the preloading period and the hydraulic leakage tests of the tanks were collected during an almost two-year long monitoring period. In situ settlement measurements confirmed the effectiveness of the technique adopted in the project even if pointed out that the lithological and mechanical heterogeneity of the foundation soil deposit significantly affects the tanks settlement response. Differential settlements and angular distortions were compared with expected profile shapes for tanks overlying homogeneous compressible soil layers. A general fair agreement was observed even if the heterogeneity of the soil deposit affects the tanks response. Absolute and differential settlements, as well as angular distortions, resulted consistent with allowable limits envisaging a satisfactory performance of the tanks under service conditions. Accordingly, the large set of measurements collected during field monitoring will allow a suitable numerical modelling with the final aim to calibrate a predictive tool for the behaviour of the tanks under service conditions and of future installations at the site. embankment constructed on peat, Canadian Geotechnical Journal, 19, Vol.1, pp. 9 3. [3] R.Berardi, R.Lancellotta, Yielding from field behavior and its influence on oil tank settlements, J. Geotech. Geoenv.,, Vol. 1 (5), pp. - 15. [] E.Cascone, V.Bandini, A.Galletta, G.Biondi, Acceleration of the consolidation process of a clay soil by preloading and vertical drains: field measurements and numerical predictions, CRC Press/Balkema, Glasgow, 9, pp. 379-35. [5] D.E.Hudson, Dynamic tests of full-scale structures, in: Wiegel, R.L., Editor. Earthquake engineering. Englewood Cliffs, NJ. Prentice Hall, 197, pp.17-19. [] G.Biondi, P.Capilleri, M.Maugeri, Dynamic response analysis of earth-retaining walls by means of shaking table tests, th U.S. Nat. Conf. Earth. Enging., San Francisco, CA,. [7] G.Biondi, M.R.Massimino, M.Maugeri, Influence of frequency content and amplitude of input motion in DSSI investigated by shaking table tests, 1, submitted to Bulletin Earthquake Engineering, under review. [] E.Cascone, G.Biondi, A case study on soil settlements induced by preloading and vertical drains, Geotextiles and Geomembranes, Vol. 3, June 13, pp. 51 7. [9] T.B.D'Orazio, J.M.Duncan, Differential settlements in steel tanks. Journal of Geotechnical Engineering Vol. 113 (9), 197, pp. 97-93. [1] W.A.Marr, J.A.Ramos, T.W.Lambe, Criteria for Settlement of Tanks, 19, Journal of Geotechnical Engineering 1(), 117 139. [11] D.P.Coduto, Foundation design. Principle and practice, 1, nd Edition. Prentice Hall. [1] G..F.Sowers, Shallow foundations. Foundation Engineering, 19, G.A.Leonards, ed., McGraw Hill Book Co., Inc., New York, N.Y.. [13] A.W.Skempton, D.H.MacDonald, Allowable settlements of buildings, 195, Proc. Institute of Civil Engineers, Vol. 3, No. 5, 77-7. REFERENCES [1] GeoTechNet Project GTC--3333, WP3: Innovation Design Tools in Geotechnics, Observational Method and Finite Element Method, editor Noel Huybrechts, BBRI. [] R.K.Rowe, M.D.MacLean, A.K.Barsvary, The observed behaviour of a geotextile-reinforced