Geodetic Monitoring of the Deformation of a 50,000 t Sugar Storage Tank Founded on 124 Long Bored Piles P. Savvaidis; I. Ifadis Thessaloniki, Greece Abstract A large storage tank capable of storing 50,000 t of sugar was constructed in the installations of the Hellenic Sugar Industry S.A. at Plati, about 50 Km from the City of Thessaloniki. The tank, which is the largest structure of this category in Europe, is founded on 124 long bored piles. Due to the fact that possible deformation of the tank may result into major damage, geodetic measurements were used for monitoring the displacements of the structure both in the vertical and the horizontal sense. These measurements included a GPS high precision control network, 3D intersections and leveling. The recorded values of deformation gave important information about the pile group settlement distribution and horizontal displacements of the tank walls with time and tank loading history. Finally, a comparison of the behavior of the tank to the behavior of similar structures and theoretical deformation analysis was done and the corresponding conclusions were derived. 1. Introduction Monitoring of the behavior of large technical structures has been a reality for the last few decades. Deformations that occur may cause severe damages to structures or even loss of life and injury to people. These deformations may be attributed to several reasons, such as incomplete investigation of foundation soil properties, improper construction of the foundation system, insufficient knowledge of the operating conditions, earthquakes etc. In all cases, the continuous monitoring of the behavior of the structure can detect deformations and displacements authenticating study and construction theories and proving the safety of the operation of the technical work. So, the observations and recordings of deformation donnot present only scientific interest for the geotechnical engineers, but they are also indications of the long-term behavior of the construction . In this paper, the deformation observations of a heavy storage tank for storing 50,000 t of sugar founded on a large group of bored piles are presented. The tank is located in the installations of the Hellenic Sugar Industry S.A. at Plati, about 50 Km from the City of Thessaloniki. It is the largest structure of this category in Europe and its construction was completed in 1995 with a cost of about 2,500,000 USD (fig. 1). The complex consists of a cylindrical shell of a 350 mm thick circular, prestressed concrete wall. The concrete shell has 46 m external diameter, it is 33.36 m high and it is covered by a wooden roof. The superstructure is founded on a 1200 mm thick circular concrete raft, 2,50 m above the ground surface, which is supported by 124 bored piles. The dead weight of the whole structure is approximately 12,000 t. Although numerous articles have been published about piles during the last years and several design methods have been proposed, based on theoretical considerations or results from model tests, knowledge about the bearing capacity and settlements of pile groups is limited. This is due to lack of data from full-scale field tests. So, any new case history related to pile foundation settlements offers useful information on the understanding of the real mechanism and the role of the factors affecting pile group behavior . The same applies also for the detection of horizontal deformations of the superstructure walls, where even less information from field tests has been collected since now. As a result, horizontal and vertical deformation monitoring of the sugar tank can significantly contribute to the knowledge of the behavior of structures of this category. Fig. 1. The sugar storage tank at Plati 2. Foundations consideration and design Storage tanks are relatively flexible structures and they can tolerate greater settlements than other engineering structures. However, there is a limit to the settlements expected to take without distress. The most important undesirable effects of settlements to avoid in designing tank foundations are as follows : Overall settlement of the tank. Differential settlement across the diameter, which may overstress internal piping connections. Differential settlement along the periphery, which may overstress the superstructure. Differential settlement between the tank and the external connection pipework. Due to the fact that the deposit of weak soils was found very deep (aprox. 23 m) and foundation conditions could not be effectively improved by strengthening techniques such as replacement with engineering fills, vibration, compaction, stone columns, line columns etc., a piled foundation system was the only viable foundation solution. Furthermore, the loading conditions, the existence of soft clayey strata and the available large amount of surrounding plant were the main reasons for the decision to avoid the displacement piles and to choose the rotary, cast in place, large bored piles with bentonite slurry as the most efficient pile type. Finally the pictured in fig. 2 group of 124 piles was designed and constructed. Each pile has a diameter of 1.20 m and it is 37.17 m long. Taking into account the interaction between piles in a semi-infinite mass, the mean total settlement of the group was evaluated equal to 12 cm by applying Poulos method for piles group settlement computation . Fig. 2. Group of piles layout. Black dots show benchmark locations for settlement monitoring 3. A survey of similar structures Some structures where similar construction techniques were used are: A heavy storage tank capable of storing 15,000 t of liquid ammonia . It is located at the industrial area of the city of Thessaloniki near the seaside. It is founded on a circular concrete raft, which is supported by 112 bored piles. The design and construction of the foundation system is very similar to that used in the sugar storage tank under consideration in this paper. The ammonia tank has been monitored for pile settlements for many years. The measurement techniques were similar to the ones used for the sugar storage tank. In this structure, the maximum vertical displacement occurred at the central piles, with settlements getting smaller values when moving to the periphery of the tank. A heavy storage tank capable of storing 40,000 t of sugar at Forlipopoli, Italy , . The circular concrete raft is founded on 396 piles. It is a structure identical to the one at Plati. Both horizontal and vertical deformation monitoring of the tank have been carried out. The observed vertical displacements obtained maximum values again at the central piles of the foundation system. Also, horizontal deformations were detected at the lower part of the circular, prestressed concrete wall of the cylindrical shell of the tank. 3. Measurement of horizontal deformation The deformations of the tank were observed after the completion of its construction during the first loading with sugar. The tank was gradually filled with 10,000 t of sugar. Then the sugar was removed, the tank remained empty for about five months and, finally, it was again filled with sugar to the maximum load of 50,000 t. The loading history of the tank is shown in fig. 3. 60000 Sugar Load [tons] 50000 40000 30000 20000 10000 0 0 100 200 300 400 500 Days from start Fig. 3. Loading history of the sugar storage tank In order to monitor the horizontal deformation of the cylindrical shell concrete wall, two lines of control points, each of 12 cone-shaped targets, were installed on the wall of the tank from bottom to top (profiles A and B), at the diameter NW-SE, the lower 6 at distance of 3 m from each other and the rest at distance of 6 m. The determination of horizontal displacements of the superstructure of the tank was very important because, following the experience gained during the construction and monitoring of the similar sugar tank at Forlipopoli, a new design for the strengthening of the lower part of the wall of the circular shell of the tank was used and had to be validated. A control network consisting of 6 reference points (pillars) was established around the tank. The network was remeasured with GPS receivers each epoch. The coordinates of the targets on the wall were computed due to the measurement of intersections from the points of the control network . A free network adjustment of the complete network (both reference and control points included) was done for each epoch and a similarity transformation was used to obtain a common reference system between the coordinates of the points of the specific epoch and the coordinates of the points of the zero measurement (as computed also by a free network adjustment) with the use of the NetS software package . Then, the horizontal displacement of each control point was computed as the difference between the zero measurement coordinates and the corresponding transformed last epoch coordinates. The results of these measurements showed no significant horizontal deformation of the wall of the tank, proving the effectiveness of the new design used. 4. Measurement of vertical deformation In order to monitor the vertical deformation of the tank, 16 benchmarks (P1, P2, …, P16) were installed on the upper free part of certain piles (fig.2). The benchmarks can be divided into three groups: P1, P2, P3, and P4 located on the central piles, P5, P6, P7, P8, P9, and P10 located on piles at an intermediate periphery and P11, P12, P13, P14, P15, and P16 located on piles at the external periphery of the slab. High precision geodetic levelling was used for the determination of the possible settlements of the benchmarks with time and load . Three reference elevation points were established at distances of 50, 100 and 400 m from the tank respectively. The measurements were carried out with a Topcon automatic level with optical micrometer. In each epoch, the same levelling network consisting of the reference and the control points was measured. The network was then adjusted . In this way, the elevation of the control points was computed for each epoch. The vertical movement of the benchmarks could then be computed by comparing their elevations between epochs. The vertical movements of the pile benchmarks during the measurements can be seen in fig. 4 (P1, P2, P3, and P4), fig. 5 (P5, P6, P7, P8, P9, and P10) and fig. 6 (P11, P12, P13, P14, P15, and P16). For a more thorough investigation, settlements of piles P1, P2, …, P16 were also taken into consideration, resulting in a clear deformation of the slab (fig.7). 800 780 Elevation [mm] 760 P1 740 P2 720 P3 700 P4 680 660 0 37 93 190 353 408 Days from start Fig. 4. Vertical movements of piles P1, P2, P3, and P4 during the first loading of the tank with sugar As it can be seen from the above figures, the vertical displacements of pile benchmarks generally follow the loading and unloading of the tank. The maximum settlements of the slab occurred at central piles P1, P2, P3, and P4 with a mean value of approximately 63.5 mm. The settlements observed were smaller with increasing distance from the center of the circular slab. 820 800 Elevation [mm] 780 P5 760 P6 740 720 P7 700 P8 680 P9 660 0 37 93 190 353 408 P10 Days from start Fig. 5. Vertical movements of piles P5, P6, P7, P8, P9 and P10 during the first loading of the tank with sugar 820 800 Elevation [mm] 780 P11 760 P12 740 P13 720 P14 700 P15 680 0 37 93 190 353 408 P16 Days from start Fig. 6. Vertical movements of piles P11, P12, P13, P14, P15 and P16 during the first loading of the tank with sugar 5. Conclusion In this paper, the recorded values of the actual deformations of a sugar storage tank capable of storing up to 50,000 t of sugar were presented, along with the measurement methods used. From the results of the measurements and the analysis of displacements following conclusions could be drawn: 1. During the first loading of the tank with 10,000 t of sugar, the mean value of the observed settlement was equal to the 25% of the predicted mean final settlement. After sugar removal the mean remaining settlement was 9% of the final predicted value. The differential settlements between the center of the foundation slab and the peripheral points were fractions of the overall slab settlement, quite close to the ratio estimated in the computations. Fig. 7. Slab deformation under maximum load during the first loading of the tank with sugar (curved lines are lines of equal vertical displacement, in mm) 2. During the loading test with the maximum load of 50,000 t, the mean observed settlement seems to be approximately the 50% of the expected final settlements. 3. The differential settlements between successive piles, which may be attributed to soil profile variability and to small differentiations in pile construction, were kept less than several mm. 4. The observed horizontal displacements of the tank wall were insignificant, proving the effectiveness of the design and construction method used. 5. The model of vertical deformation of the sugar tank at Plati was very similar to the behavior observed in the other two tank structures mentioned above (the ammonia storage tank in Thessalonici and the sugar storage tank in Forlipopoli). References 1. Badellas A., Savvaidis P. Monitoring of Deformation of Technical Works and Ground Landslides with Geodetic Methods. Papageorgiou Publ. Co., 1990, p. 257 + XV, Thessaloniki, 1990. 2. Badellas A., Savvaidis P., Tsotsos S. Settlements Measurement of a Liquid Storage Tank Founded on 112 Long Bored Piles. Proc. of 2nd Inter. Symp. on Field Measurements in Geomechanics, Kobe, Japan, Balkema Publ., pp. 435-442, 1988. 3. Poulos H.G. Group Factors for Pile Deflection Estimation. Journal GE, ASCE, Vol. 105:1489-1509, 1979. 4. Vosdou S., Michaelidis P., Savvaidis Ch. Measurement of horizontal displacements of a Sugar Storage Tank at Plati. Diploma Thesis, Dept. of Civil Engineering, AUTH, 1997. 5. Bakavelou D., Panda E., Stefoulis A. Measurement of Settlements of a Sugar Storage Tank at Plati. Diploma Thesis, Dept. of Civil Engineering, AUTH, 1997. 6. Savvaidis P. Program NetS for the Adjustment of Geodetic Networks. Scientific Papers from the School of Technology, Department of Civil Engineering, A.U.Th., No. 13, Thessaloniki 1995. Dr. Eng. Paraskevas D. Savvaidis is Professor in Engineering Surveying and Director of the Laboratory of Geodesy, Department of Civil Engineering, Aristotle University of Thessaloniki (AUTH), Univ. Box 465, 54006 Thessaloniki, Greece. Tel. +30 310 995724, Fax +30 310 996159, e-mail: email@example.com. Dr. Eng. Ioannis M. Ifadis is Associate Professor in the Laboratory of Geodesy, Department of Civil Engineering, AUTH, Univ. Box 465, 54006 Thessaloniki, Greece. Tel. +30 310 995745, Fax +30 310 996159, e-mail: firstname.lastname@example.org.