See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/259895346
Challenges and Uncertainties Relating to Open Caissons Article · July 2012
CITATIONS
READS
2
149
1 author:
Khaled E. Gaaver Alexandria University 14 PUBLICATIONS 38 CITATIONS
SEE PROFILE
All content following this page was uploaded by Khaled E. Gaaver on Gaaver on 28 January 2014. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately.
Challenges and Uncertainties Relating to Open Caissons Fathi Abdrabbo, Professor, Structural Engineering Department, Faculty of Engineering, Alexandria University, Alexandria, Egypt;
[email protected] Khaled Gaaver , Associate Professor, Structural Engineering Department, Faculty of Engineering, Alexandria University, Alexandria, Egypt;
[email protected] ABSTRACT Open caissons are used for many geotechnical engineering applications. Open caissons may be used as deep foundation elements bypassing weak soils to tip in firm deeper strata, and in rivers and maritime construction to reduce the risk of scour. Open caissons are also used for collecting sewage water through gravity sewer pipe networks or from sewer force mains. In such applications, the design and construction of open caissons require a detailed so il investigation program. In this way, the design and construction plan of an open caisson can be developed with full knowledge of the prevailing subsoil conditions. The engineering and construction techniques are key factors to achieve functional caissons. Based on close observations during construction stages, the current study presents some challenges that were encountered during the construction of two open caissons of internal diameters 20 m and 10 m (65.6 ft and 32.8 ft). This paper describes the procedure followed to alleviate the construction difficulties encountered. Site exploration program and control measures required to satisfy design and construction requirements are crucial aspects. Sinking of open caissons in dense or very dense sands is risky. Incorrect sinking of open caissons may cause extra cost, delay in construction, and harm to nearby structures. Air/water jetting near the cutting edge of an open caisson, outside slurry trench, and/or inside open trench may be used to drive an open caisson downward. Unsymmetrical work around an open caisson may lead to tilting of the caisson. If this occurs, the tilt should be immediately corrected before resuming the sinking process. Improper cleaning of fine materials on the caisson’s excavation bed, and/or inappropriate pouring o f underwater concrete may result in a defective concrete seal. The paper contains a series of practical guidelines to assist those intending to use open caissons, and shares good caisson sinking practice with practitioners. Finally, the study aims to understand the difficulties encountered and to anticipate future problems.
INTRODUCTION Sinking of open caissons is appropriate where the prevailing soil consists of soft to medium clays, silty sands, or loose sands. These soils can be readily excavated using grab buckets within the open caisson and do not offer high skin friction along caisson-soil interface. Open caissons can feasibly extend to great depth at relatively low cost; however, they have some disadvantages, Tomlinson (1986). For example, construction may be halted if obstructions, such as large boulders or tree trunks, are encountered. The available literature is scant regarding the sinking of open caissons because diaphragm trenches and large-diameter secant piles are implemented in the construction of open caissons (Puller, 1996). Moreover, the use of suction caissons is considered as an alternative construction method. Contrary to open caissons, the penetration of suction caissons into soil is due to self-weight of shaft in addition to suction pressure created inside the caisson. Therefore,
adequate seal provided by caisson’s self-weight penetration is essential to apply suction pressure inside the caisson. Suction caissons attracted the attention of many authors, including Chen & Randolph (2007), Byrne & Houlsby (2004), Clukey et al. (2004), Iskander et al. (2002), Randolph & House (2002), and El-Gharbawy & Olson (1999). The construction procedure of open caissons may differ between countries (Allenby et al. 2009). Depending on experience gained in Egypt, open caissons are constructed using consecutive lifts of reinforced concrete cast-in-situ walls. The caisson walls sink in place successively while the soil inside the caisson is excavated using grab buckets. Upon reaching the design depth level, a concrete seal is cast underwater using tremie pipes. After the concrete seal has matured, water inside the caisson is pumped out. The caisson can be used either as is for collecting sewage water or filled with concrete as a deep foundation (Nonveiller 1987). DFI JOURNAL Vol. 6 No. 1 July 2012 [21]
Open caissons have two essential stages to be carefully studied, construction stage and inservice stage. During the in-service stage, most of the applied loads on the caissons are transferred to the soil via end bearing at their bases. The contribution of skin friction developed along the caisson-soil interface may be ignored to support superstructure loads in certain circumstances. In contrast, skin friction along the caisson-soil interface should be precisely estimated during the construction stage to check for feasibility of sinking the caisson through the soil. The uncertainties arising during the calculation of skin friction along the caisson-soil interface may be attributed to the disturbance of adjacent soil due to the construction process. Ranges of skin friction values have been provided by Terzaghi and Peck (1967) for each type of soil. Similarly, values of estimated skin friction during sinking of both pneumatic and open caissons in different soil conditions have been reported by Tomlinson (1986). Puller (1996) mentioned that a comparison of the values recommended by Terzaghi & Peck (1967) with those given by Tomlinson (1986) shows a considerable scatter of skin friction in similarly described soils. The methodology used in this study is based on the observational procedure, which is one of the design approaches listed in Eurocode 7 (Glass & Powderham 1994). Peck (1969) pioneered the applications of the observational method to geotechnical engineering. The philosophy behind the observational technique is to initially base the design on whatever information can be obtained and then to examine all conceivable differences between assumptions and reality. The observational method saves cost and time, and limits construction risks. Nowadays, the observational method is well known to the geotechnical profession (Wu 2011). The method was used to explore some challenges during the construction stages of two open caissons. In this paper, two case studies were considered.
20.00 m (65.6 ft) internal diameter. (Caisson PS4). The caisson was designed to collect sewage water through a sewer network at El-Agamy district, west of Alexandria, Egypt. Caisson PS4 is located approximately 400 m (1312 ft) from the shore of the Mediterranean Sea. Service and residential buildings are located near the caisson site. Five pumps were to be installed in the caisson to pump sewage water at a rate of 5000 liter/sec (1320 gal/sec) to a treatment plant located 20 km (12.4 miles) away. Caisson PS4 is the main pump station of the sewer network of the district, see schematic diagram, Fig. 1. The total height of the caisson was 33.31 m (109.3 ft). It is common practice to reduce the wall thickness above the cutting shoe by 30 to 100 mm (1.2 to 3.9 in) from the outside to reduce skin friction along the caisson walls. Therefore, the wall thickness is 1.60 m (5.25 ft) for the top 27.31 m (89.6 ft) height and 1.70 m (5.6 ft) for the bottom 6.00 m (19.7 ft), as shown in 2. The elevation of the ground surface at the construction site is +2.30 m (+7.6 ft) above sea level. The designed upper floor level of the caisson is at +3.15 m (+10.3 ft), while the tip level is designed to be at -30.16 m (-99.0 ft). The wall is reinforced vertically using 10 bars of 22 mm/m (#7 bars at 4 in centers) each side and hoop reinforcement using 8 bars of 22 mm/m (#7 bars at 5 in centers). At the designed sinking level, the dry weight of the caisson’s walls is 90.48 MN (10,170 ton), while the buoyant weight is 56.32 MN (6,330 ton).
CASE STUDY NO. 1 This case study presents some difficulties encountered during sinking of an open caisson of [22] DFI JOURNAL Vol. 6 No. 1 July 2012
[FIG. 1] Schematic diagram of caisson PS4 and emergency caisson
struction method. The engineer’s decision was based on engineering judgment and constructability of the structure. The caisson was sunk successfully under its own weight according to the construction plan, while grabbing the soil inside to an elevation of -24.60 m (-80.7 ft). At this level, it was very difficult to further advance the caisson. The caisson was obstructed; nevertheless the periphery walls were constructed to the full design height. Therefore, part of the caisson walls, 5.56 m (18.2 ft), remained above the ground surface. At this stage, the buoyant weight of the caisson walls was 63.58 MN (7,147 ton). Thus, the average skin friction developed along caisson-soil interface was 32.29 kPa (4.68 psi). The estimated value is in good agreement with the lower limit value recommended by Terzaghi & Peck (1967) and 1.42 times the value reported by Tomlinson (1986). Terzaghi & Peck (1967) mentioned values that vary from 33.50 to 67.00 kPa (4.86 to 9.72 psi) for dense sand, while Tomlinson (1986) repo rted a value of 22.80 kPa (3.31 psi) for sand. After the caisson had been obstructed, four boreholes 50.00 m (164 ft) deep were drilled to further explore the subsoil difficulties encountered. The recovered soil samples from the boreholes were classified according to ASTM D 2487 for soil and ASTM D 6032 for rock. [FIG. 2] Design details of open caisson PS4 Fig. 3 illustrates a typical borehole log along with the corresponding standard penetration Prior to the start of construction, geotechnical tests (SPT), N-values. Retrieved soil samples investigations were conducted at the site by from the boreholes revealed that the soil at the drilling three boreholes up to 50.00 m (164.0 ft) site consists of a top layer derived from oodepth. The retrieved soil samples revealed six litic very poor, weak limestone extending from successive soil strata of similar thicknesses, ground surface to a level that varied from -0.20 as shown in Table 1. The subsoil exploration to -4.70 m (-0.66 to -15.4 ft) . This limestone showed groundwater table at elevation -0.30 m was underlain by a layer of poorly graded very (-1 ft). The data collected from the geotechnical dense sand intermixed with pieces of sandstone investigation was used to design the caisson. of various sizes. At a level that varied from The open caisson technique was recommended -15.40 to -16.70 m (-50.5 to -54.8 ft), a layer by the engineer as the most appropriate concomprising different soils [TABLE 1] Soil strata revealed from boreholes drilled prior to c onstruc tion of sandy silt with clay and silty clay with sand was encountered. This layer exLayer Top level (m) Bottom level (m) Soil classification tended to a level that varied 1 + 2.30 -12.70 Sandy silt from -33.30 to -34.70 m 2 -12.70 -15.70 Sandstone (-109.3 to -113.8 ft) and it contained two relatively 3 -15.70 -22.70 Sandy silt thin layers of very poor 4 -22.70 -33.70 Silty clay weak sandstone. The thick5 -33.70 -36.70 Sandstone ness of the top sandstone 6 -36.70 -47.70 Silty sand layer varied from 0.35 to
DFI JOURNAL Vol. 6 No. 1 July 2012 [23]
[TABLE 2] Summary of soil properties obtained from post-constr uction explorations Layer
First
Second
Third
+ 2.30
-0.20 / -4.70
-15.40 / -16.70
Bottom level (m)
-0.20 / -4.70
-15.40 / -16.70
-33.30 / -34.70
Soil classification
Limestone
Sand/Sandstone
Sandy silt/Silty clay
18.20
18.00
15.80 – 16.90
1.94
0.45 – 1.37
-
Undrained shear strength (kN/m2)
-
-
18.00 – 26.00
Undrained angle of shearing resistance (Degrees)
-
40.00
11.00 – 16.00
Top level (m)
Natural unit weight (kN/m3) Unconfined compressive strength (MPa)
3.00 m (-1.1 to -9.8 ft), while the bottom sandstone was a discontinuous layer of 2.00 m (6.6 ft) maximum thickness. Groundwater table was measured at a level of -0.20 m (-0.66 ft). The properties of sand in the second layer were interpreted based on the standard penetration test results and visual classification of the retrieved soil samples. Accordingly, the relative density of the sand was about 75%, and the unit weight was 18 kN/m 3 (3093 lb/yd3). The corresponding angle of shearing resistance of sand is 40 degrees. In this situation, it is important to note that the interpreted properties of the second layer should be used with caution due to the effect of sandstone pieces on the SPT results. Core samples recovered from limestone in the first layer revealed that the recovery values varied from 25 to 40% while the rock quality designation was zero. Laboratory tests on rock samples including compressive strength were conducted. Undisturbed samples of cohesive soil in the third layer were tested in direct shear using shear box apparatus. The achieved results o f the laboratory tests are presented in Table 2 and Fig. 3. The average value of undrained shear strength (Cu) of the cohesive soil is 22 kPa (3.2 psi) and the corresponding average undrained angle of shearing resistance is 13.60º. The average values were implemented in the stability analysis.
[FIG.3] A typical borehole in case study # 1
[24] DFI JOURNAL Vol. 6 No. 1 July 2012
During the excavation process inside the caisson, the second layer was visually observed and found to be weak sandstone, contrary to the laboratory classification of soil samples retrieved from the boreholes. The contradiction may be attributed to the disturbance of the retrieved samples by the sampling process. Block samples were recovered from the second layer during excavation inside the caisson. The unconfined compressive strength of the tested
samples varied between 0.45 and 1.37 MPa (65.3 and 198.7 psi). Therefore, it is important to carefully consider soil types that reveal anomalously high N-values in SPT. Double or triple cores may be used to retrieve relatively undistributed rock samples. Appreciable difference of the subsoil condition was observed when comparing the post-construction and the pre-construction explorations, especially above the level of -16.70 m (-54.8 ft). Soil investigations prior to construction showed a layer of sandy silt to elevation -12.70 m (-41.7 ft) overlaid sandstone of 3.00 m (9.8 ft) thick. In contrast, post-construction geotechnical explorations indicated weak limestone overlaid sandstone extending up to an elevation o f -15.40 to -16.70 m (-50.5 to -54.8 ft). The difference in soil types may be attributed to the procedure of soil sampling. Washed soil samples using tricone do not represent the in-place state of soil formations, especially in rock layers. Also soil sampling using a split barrel in weak rocks leads to unrealistic soil descriptions and properties compared to the real state conditions. The post-construction boreholes clearly showed that the adopted method of sinking the open caisson was not the proper construction method due to the following reasons: 1. There are high values of standard penetration test results recorded with the geomaterial in the second layer. Also, there are relatively high values of unconfined compressive strength of both limestone in the first layer and sandstone in the second layer as real state geomaterial conditions. The higher values of SPT are sufficient to provide an early indication that sinking of the caisson is not the proper construction technique. 2. There are irregularities in the thickness of the sandstone layers, which adds difficulties to the sinking process and extends the excavation time. The existence of sandstone increases the frictional resisting force along the caisson-soil interface during the sinking process. Also, sandstone underneath the cutting edge of the caisson will not slump towards excavation inside the caisson and offers high bearing resistance. 3. Pieces of sandstone interbedded in sand cause difficulties in excavation using a clam-
shell. Excavation of this type of geomaterial requires special trenching equipment. 4. The high level of groundwater table presents additional difficulties to the underwater excavation process since excavation must be controlled by divers. Due to these difficulties, the caisson stuck at the level of -24.60 m (-80.7 ft) for approximately 31 months. During this period, traditional methods of sinking the caisson were used including pumping out water from the caisson using powerful surface pumps, and excavating geomaterial inside the caisson using vibratory excavators. Also excavation of soil outside the caisson up to 3.00 m (9.8 ft) depth was carried out to reduce skin friction on caisson-soil interface. The excavation depth of soil outside the caisson was restricted to ensure stability of nearby structures. Secant piles were designed and installed to maintain the safety of neighboring structures. Moreover, excavation beneath the cutting edge of the caisson walls was carried out by divers. All of these methods failed to move the caisson downward. It was a challenge to complete sinking the caisson by developing a reliable procedure to complete the sewage project. In such circumstances, many factors control the adopted restoration procedure such as: 1. Legal liability and responsibility of the owner/engineer or the contractor. 2. Expected cost of the proposed rehabilitation procedure. 3. Expected time of the proposed restoration technique. 4. The free space between the caisson wall and the existing nearby structures is limited, thus the capability of this space to accommodate the machinery involved in the restoration procedure must be considered. 5. Soil stratification and groundwater table. 6. The risk measurement of buildings nearby the caisson. In conclusion, due to the critical function of the caisson PS4 and the legal situation, it was impossible to backfill the obstructed caisson and to construct a new one nearby implementing a proper construction procedure. In addition, relocation of the caisson would require rerouting of the approach inlet and outlet sewer DFI JOURNAL Vol. 6 No. 1 July 2012 [25]
pipes, which had been installed already using a pipe jacking technique. Thus, complete sinking of the obstructed caisson by using a remedial technique was essential. A restoration proposal was developed to construct a circular slurry trench outside the caisson and as close to the wall as possible such that the drilling machine could accomplish the drilling process. The outside slurry trench was formed by constructing slurry piles, each 600 mm (23.6 in) in diameter, at a spacing of 1.00 m (3.28 ft) and extending to level -33.50 m (-109.9 ft). see Fig. 4. Each slurry pile incorporated a steel tube of 50 mm (2 in) in diameter provided with an end nozzle. The tubes inserted into the slurry piles are used to jet air at a pressure of 20 bars (290 psi), if needed. The slurry trench acts as a separator between the walls of the caisson and the geomaterial extending further away from the caisson. The slurry trench reduces the skin friction along caissonsoil interface. The construction of the slurry trench faced difficulties due to drilling through subsoil hard formation and the presence of saltwater from the nearby sea. The saltwater affected the performance of the slurry mud pumped inside the drilling holes for base and side stability. To overcome the effect of saltwater, the slurry mud mixture comprised 1.00 kg (2.2 lb) of sodium
carbonate, 40.00 kg (88 lb) of bentonite, and 1,000 liters (264 gal) of water. Sodium carbonate was added to the slurry mud to overcome the effect of the sodium chloride and sulfate present in groundwater. An interior open trench inside the caisson and close to the caisson walls was excavated. The width of the inside trench is 3.00 m (9.8 ft) and extending to a level of -33.50 m (-109.9 ft). Thus it is believed that the outside slurry trench and the inside open trench bounded a geomaterial wall extending down from the cutting edge of the obstructed caisson up to a level of -33.50 m (-109.9 ft), see Fig. 5. To study the stability of the geomaterial wall, which extended from level -24.60 to level -33.50 m (-80.7 to -109.9 ft), first we considered section (a-a) at level -24.60 m (-80.7 ft), see Fig. 5. A trial plane of failure was considered to be inclined at an angle (where = 45 – /2 = 25°). The buoyant weight of 1.00 m (3.28 ft) length along the periphery of the caisson wall is 868 kN/m (59,478 lb/ft). Therefore, the driving force on the trial plane of failure is 366 kN/m (25,079 lb/ft) and the resisting force is 660 kN/m (45,225 lb/ft). This means that the geomaterial under the tip of the caisson was stable up to a level of -27.70 m (-90.9 ft), the top surface of the sandy silt/silty clay layer. The geomaterial wall sustained the imposed loads resulting from the caisson walls with-
All dimensions are in meters [FIG. 4] Layout of the caisson and slurry piles
[26] DFI JOURNAL Vol. 6 No. 1 July 2012
caisson. This observation was a positive sign for the initiation of caisson sinking. Movement of the caisson was observed during injection of the compressed air. It was observed that the caisson moved vertically by 20, 50, and 100 mm (0.8, 1.0 and 2.0 in) in the three successive days respectively. Overnight, the caisson sank by 7.90 m (26.0 ft) and stayed at a level of -32.50 m (-106.6 ft).
[FIG. 5] Stability analysis in case study # 1
out any slumping toward the excavated open trench inside the caisson. Then, consideration was given to the stability of a trial section (b-b) through the sandy silt/silty clay layer at a lev el of -31.00 m (-101.7 ft), see Fig. 5. The weight of the caisson wall and the enclosed geomaterial up to a level of -31.00 m (-101.7 ft) was 933 kN/m (32 ton/ft). Thus the driving force on the trial plane of failure was 574 kN/m (19.7 ton/ft), whereas the resisting force was 225 kN/m (7.7 ton/ft). This analysis indicated that the geomaterial wall was unstable through the sandy silt/silty clay layer. Consequently, it was expected that shear failures occur due to excessive shear stresses imposed at a level of -31.00 m (-101.7 ft). The process of sinking the caisson was resumed by blowing compressed air through the steel tubes installed in the slurry trench via a header pipe over a three day period. Air bubbles were observed to rise through the water inside the
According to the stability analysis of the geomaterial wall, it was evident that failure of the geomaterial wall was expected at a level of -31.00 m (-101.7 ft). To reduce the risk level in the remedial procedure, the compressed air system was designed and installed in the outside slurry trench. The objective of the compressed air system was to overcome any shortcoming resulting from probable deviation of predicted outcome from actual performance. The prediction procedure may be deficient if one or more of the following is missed or deficiently predicted; soil stratigraphy, soil properties, subsoil heterogeneity, prediction method and capability, stress history, and stress path (Focht 1994). The result of the completion of sinking the caisson while implementing the above procedure was that the caisson moved down to 2.34 m (7.7 ft) below the designed level. Therefore the top level of the caisson walls was at a level of +0.81 m (+2.7 ft), instead of +3.15 m (+10.3 ft). The caisson walls were extended to a level of +3.15 m (+10.3 ft) by pouring concrete inside shuttering and scaffolding. It is important to note that sudden sinking of the caisson had no side effect on the buildings located at a distance of 4.00 to 6.00 m (13 to 20 ft) from the caisson. This may be due to the gentle movement of the caisson. The caisson moved downward through a cohesive geomaterial, which exhibited neither strain softening nor strain hardening. In other words, the geomaterial behaved as an elastoplastic material such that the caisson sank smoothly. The movement of the caisson was monitored after complete sinking for 30 days and no movement was observed. At this stage, the dominant resisting force was the bearing stress developed at the caisson’s tip. The tip is at a distance between 0.80 to 2.20 m (2.62 to 7.2 ft) above sandstone layer. Thus sandstone contributed to the bearing capacity at the caisson tip. If friction along caisson-soil interface is ignored, the imposed bearing stress at the tip of the caisson is about DFI JOURNAL Vol. 6 No. 1 July 2012 [27]
518 kPa (75 psi). If half of the friction at the caisson-soil interface is considered, the imposed bearing stress at the tip of the caisson is 162 kPa (23.5 psi). Simple analysis, using bearing capacity theories, shows that the caisson is stable at the level of -32.50 m (-106.6 ft). The contractor resumed the work to complete the excavation inside the caisson and to execute the concrete seal. Tremie pipes of 200 mm (8 in) diameter were used to deliver the concrete from concrete trucks via a concrete pump to construct the concrete seal. After the concrete seal had been matured, it was discovered that the concrete seal did not function properly. Groundwater seeped through the seal when water inside the caisson was pumped out. The concrete seal was constructed without cleaning debris and soft deposits on the excavation bed. A layer of soft deposits was trapped under the concrete seal, and pockets of soft deposits intervened into the seal. As a result, the cast concrete seal was of poor quality and was thinner than the designed thickness. Accordingly, the concrete seal became unable to resist stresses induced by uplift water pressure. Furthermore, it was observed by divers that cracks in the concrete seal developed more and more as pumping of water inside the caisson continued. Inadequate planning and improper execution of tremie concreting led to a defective concrete seal. Unfortunately, another remedy to an unexpected problem was needed. The goal of the remedial work was to lower the groundwater table to beneath the toe level of the caisson to demolish the defective concrete seal and construct a new seal in dry conditions. The design of the remedial work required another borehole to explore soil stratification up to 100 m (328 ft) depth. The soil samples obtained revealed very poor weak sandstone bed extending from depth 50 m up to 100 m (164 ft to 328 ft) below the ground surface. The sandstone is moderately weathered. As the caisson was 400 m (1300 ft) from the sea, additional impacts and uncertainties arose in the calculation of water seepage into the caisson. Moreover, another uncertainty arose from the permeability coefficient of the rock mass. The sandstone contained cracks and joints that were filled with fine material. Water channels were developed through cracks and joints in the sandstone with the progress of pumping water. Dewatering calculations were performed utilizing deep [28] DFI JOURNAL Vol. 6 No. 1 July 2012
wells outside the caisson to construct a new seal in dry conditions, but a large number of deep wells with large pump capacities were required. Furthermore, dewatering outside the caisson might have affected the stability of the adjacent buildings. Due to difficulties arising from pumping large amount of water to lower water inside the caisson to the level of -32.50 m (-106.6 ft) and side effect of dewatering on the adjacent buildings, a water cut-off wall was designed to be constructed around the caisson walls. The choice of the geometry of the cut-off wall including material, thickness, and length is significant. Cut-off walls should have sufficient thickness to prevent hydraulic fracture. The depth of the cut-off wall should be determined to prevent piping and heave in soil at the excavation bed and to reduce the seepage. One of the following materials was proposed to be used in the construction of the cut-off wall: soil-bentonite, soil-cement-bentonite, cement bentonite, and plain concrete. The selected construction material was controlled by the design of the appropriate thickness to prevent hydraulic fracture in the cut-off wall. U.S. Army Corps of Engineers (USACE 1986) recommended that the minimum width of a soil-bentonite cutoff should be 0.03 m per 0.30 m (0.1 ft per ft) of differential hydraulic head. According to this criterion, the required width of the soil-bentonite trench is excessively thick. Therefore, plain concrete was used to construct the cut-off wall with a thickness of 1.20 m (4 ft). It is known that the flow of water beneath impervious cutoff walls may produce heave or piping in soil. Heave occurs if the uplift force at the sheeting toe exceeds the submerged weight of the overlying soil. When the velocity of water at the exit exceeds the critical velocity of water, piping occurs. To prevent both heave and piping, the cut-off wall should be of sufficient depth below the excavation bed. Design charts provided by NAFAC DM-7 (1982) for two-dimensional flow were used to assess the penetration depth of the proposed cut-off wall. These charts were produced for homogeneous dense sandy soil including safety factor of 2. Interconnecting panels of diaphragm wall, which extended to the level of -87.70 m (-288 ft), were designed to form the cut-off wall. The clear space between the outer surface of the caisson walls and the inner surface of the cut-off wall was about 2.00 m (6.6 ft).
At this stage, the contractor faced another challenge; it was difficult to obtain a diaphragm wall machine capable of operating at depths of up to 90 m (295 ft) below the ground surface. At that time, no suitable machine was locally available; therefore the equipment was imported from Europe. Another difficulty was that the free space between the caisson walls and the adjacent buildings was insufficient to accommodate the imported diaphragm wall machine. Consequently, it was essential to backfill inside the caisson and the area around the caisson using structural fill to prepare a working platform fo r the diaphragm wall machine. After constructing the cut-off wall, the backfill inside the caisson was removed to complete the work. To lower the groundwater table below the level of -32.50 m (-106.6 ft), seven deep wells of diameter 0.60 m (24 in) and extending to level -55.00 m (-180 ft) were designed and installed in the annular space between the caisson and the cut-off wall. An electric submersible pump of capacity 200 m 3/hour (785 yd3/hr) at 60 m (197 ft) head was mounted in each well. During removal of the contaminated concrete seal, the concrete was observed to be inhomogeneous and containing soft spots. Additionally, freefrom-cement aggregates were observed, indicating that the concrete of the seal was washed out during concrete pouring. A large amount of fine material had been deposited at the lower surface of the defective seal. The defective seal was demolished and a new seal and reinforced concrete base of the caisson were constructed under dry conditions. The construction of the cut-off wall, dewatering process, and difficulties from the inaccurate interpretation of soil conditions, prior to construction, doubled the construction cost and increased the construction time to about five times the anticipated time. Based on the presented case study, it can be concluded that improper interpretation of subsurface ground conditions leads to inappropriate design of the caisson. The difficulties arising from inaccurate interpretation of soil conditions cause challenges that may increase the cost and time of construction. Improper cleaning of fine material deposited on the excavation bed, and/ or incorrect procedures in pouring of underwater concrete may produce an inadequate concrete seal. Legal liability may divert the decisions of the engineer from adopting the proper procedure. Therefore, geotechnical engineers
are advised to avoid the use of open caissons in such circumstances. The alternative method is to use interconnecting panels of reinforced concrete diaphragm walls to form the caisson walls and to grout the soil below the excavation bed before excavation inside the caisson to form an impervious blanket.
CASE STUDY NO. 2 Case study No. 2 refers to the construction of a wastewater pump station at a village in El-Behera province, west of the Nile river delta, Egypt. An open caisson of 10.00 m (32.8 ft) internal diameter, 1.00 m (3.28 ft) wall thickness, and 12.00 m (39.4 ft) depth below the ground surface was considered for the station. The caisson is bounded by a store on the eastern side, 4.00 m (13 ft) from the caisson wall. On the southern side, there is a bank building 6.00 m (20 ft) from the caisson. Farmer buildings on the other sides are located approximately 15.00 m (49 ft) from the caisson. Two boreholes were drilled at the site up to 15.00 m (49 ft) below the ground surface. Fig. 6 presents a typical borehole log. The recovered soil samples from the boreholes were classified
[FIG. 6] A typical borehole in case study # 2
DFI JOURNAL Vol. 6 No. 1 July 2012 [29]
in accordance with ASTM D 2487. The subsoil is poorly graded sand with silt extending to 15.0 m (49 ft) depth below the ground surface. Groundwater table was encountered at a depth of 1.30 m (4.3 ft) below ground surface. Standard penetration test results indicated that the sand is medium dense throughout the top 5.00 m (16.4 ft) and became very dense below this level. Prior to the excavation process, the caisson walls were constructed to 12.00 m (39.4 ft) above the cutting edge. After the concrete had matured, excavation inside the caisson walls began using a cablesuspended grab. Due to the excavation process and the body weight of caisson, the caisson sank up to 8.00 m (26 ft) below the ground surface. At this depth, the open caisson was stuck. The dry weight of the caisson walls was 10.37 MN (1165 ton). When the caisson sank 8.00 m (26 ft) below the ground surface, its weight reduced due to buoyancy force to 8.05 MN (905 ton). Thus, the average skin friction developed along caisson-soil interface was 26.70 kPa (3.87 psi). The average calculated value is less than the lower limit value recommended by Terzaghi & Peck (1967) for dense sand by 20%, which varied from 33.50 to 67.00 kPa (4.86 to 9.72 psi). Also, the average calculated skin friction is 17% greater than the value reported by Tomlinson (1986) of 22.80 kPa (3.31 psi). The uncertainties in skin friction at caisson-soil interface produce a high level of risk during caisson sinking. The caisson was observed while its tip was at 8.00 m (26.2 ft) below the ground surface. It was found that the caisson had tilted southward by 0.4%. Ground loss of soil around the caisson was observed in the southern direction. Ground loss that occurred at one side of the caisson may have been due to heterogeneity of the subsoil condition, improper excavation process, and due to inclination of the caisson. The flow of soil inside the caisson can cause more tilting of the caisson. Ground loss may also damage nearby structures; therefore, the excavation process inside the caisson was stopped. At this stage, the soil inside the caisson was 1.00 m (3.3 ft) above the cutting shoe. It was evident that the friction resistance along caisson walls and the bearing resistance at caisson tip were greater than the body weight of the caisson. To sink the caisson to the required depth, friction resistance along the caisson-soil interface [30] DFI JOURNAL Vol. 6 No. 1 July 2012
needed to be deliberately decreased and the vertical orientation of the caisson needed to be corrected. Correction of the caisson verticality became difficult as the embedded depth of the caisson increased. To decrease the friction at the caisson-soil interface, a slurry trench around the caisson walls was installed. Seventy holes were drilled around the caisson, and the drilled holes were filled with slurry mud. The slurry piles of diameter 400 mm (16 in) were extended to 15.00 m (39.2 ft) depth below the ground surface. The slurry mud consisted of 1:2 (bentonite: cement). Steel pipes of 50 mm (2 in) diameter were inserted in the holes for water or air jetting, if required. The pipes were provided with slotted holes along the depth from 8.50 m to 15.00 m (28 ft to 49 ft). The caisson sank 0.40 m (1.3 ft) on completion of the slurry trench without inside excavation. Unfortunately, the caisson tilted by 4% due to unsymmetrical air jetting. Therefore, it was unreasonable to continue using air or water jetting to advance the downward movement of the caisson before adjusting the tilt of the caisson. This was achieved by dewatering and excavating more intensively below the cutting edge at the higher side of the caisson than the lower side. Based on the above discussion, it is not reasonable to sink open caissons in dense or very dense sands (N-Value for 300 mm (1 ft) 20). The values given by Terzaghi & Peck (1967) for skin friction on the caissons may be sensibly used to predict the friction resistance during sinking open caissons in sand. Incorrect sinking of open caissons may cause extra cost, delay in construction, and harm nearby structures. Moreover, uncertainties involved in the subsoil conditions and skin friction along caisson-soil interface impact risk measurements on the construction of open caissons. Unsymmetrical work around the open caisson leads to tilting of the caisson. The tilt should be corrected before resuming sinking process of the caisson. Furthermore, this case study demonstrates that improper construction of an engineering design, due to lack of knowledge and experience, may lead to further engineering problems that need to be rectified.
CONCLUSIONS This paper presents some challenges that were encountered during the construction of two open caissons under two different subsoil
conditions. Procedures used to overcome the encountered construction challenges were described. The following conclusions may be drawn: 1. Proper interpretation of subsurface soil conditions is a crucial aspect during design and selection of the proper technique for the construction of open caissons. Difficulties arising from the erroneous interpretation of subsoil conditions cause extra cost and delay in construction. 2. Sinking of open caissons in dense or very dense sands (N-Value for 300 mm (1 ft) 20) is risky due to high friction resistance on caisson-soil interface. 3. Incorrect sinking of open caissons may cause extra cost, delay in construction, and harm to nearby structures. 4. Air/water jetting near the cutting edge of an open caisson, outside slurry trench, and/or inside open trench may be used to drive an open caisson downward. 5. A unique procedure to calculate skin friction along soil-caisson interface does not exist, and the values recommended by Terzaghi and Peck (1967) may be sensibly used. 6. Improper cleaning of fine material deposited on the excavation bed, improper pouring of underwater concrete, and improper interpretation of subsurface soil conditions caused some challenges to the open caisson in case study No. 1. These challenges doubled the construction cost and increased the construction time to approximately five times the anticipated time. 7. Unsymmetrical work around the open caisson in case study No. 2 led to tilting of the caisson by 4% from the vertical. The tilt should be immediately corrected before resuming sinking the caisson. 8. Improper construction of an engineering design may lead to further engineering problems that need to be remedied.
REFERENCES 1. Allenby, D., Waley, G., and Kilburn, D. (2009), Examples of open caisson sinking in Scotland, Proceedings of the ICE Geotechnical Engineering , 162(1), 59-70.
2. Byrne, W. B., and Houlsby, G. T. (2004), Experimental investigations of the response of suction caissons to transient combined loading, Journal of the Geotechnical and Geoenvironmental Engineering , ASCE, 130(3), pp. 240-253. 3. Chen, W., and Randolph M. F. (2007), Uplift capacity of suction caissons under sustained and cyclic loading in soft clay, Journal of the Geotechnical and Geoenvironmental Engineering , ASCE, 133(11), pp. 1352-1363. 4. Clukey, E. C., Templeton, J. S., Randolph, M. F., and Phillips, R. (2004). Suction caisson response under sustained loop current loads, Proceedings. 36th Annual Offshore Technology Conference, Houston, Paper No. OTC 16843. 5. US Army Corps of Engineers (USACE 1986), Engineering manual EM 11110-2-1901. USA 6. El-Gharbawy, S. L., and Olson, R. E. (1999), The cyclic pullout capacity of suction caisson foundations, Proceedings 9th International Offshore and Polar Engineering Conference, ISOPE’99 , vol.2, France, pp 660667. 7. Focht, J. A. (1994), Lessons learned from missed predictions, Journal of the Geotechnical and Geoenvironmental Engineering , ASCE, 120(10), pp. 1653-1683. 8. Glass, P. R. and Powderham, A. J. (1994), Application of the observational method at Limehouse Link, Geotechnique (44), 665679. 9. Iskander, M., El-Gharbawy, S. L., and Olson, R. E. (2002), Performance of suction caissons in sand and clay, Canadian Geotechnical Journal , 39(3), pp 576-584. 10. NAFAC DM-7 (1982), Design manual 7.1, Department of the Navy, Naval Facilities Engineering Command. 11. Nonveiller, E. (1987), Open Caissons for Deep Foundations, Journal of the Geotechnical and Geoenvironmental Engineering , ASCE, 113(5), pp. 424-439. 12. Peck, R. B. (1969), Advantages and limitation of the observational method in applied soil mechanics, Geotechnique, 19 (2), pp. 171187. 13. Puller, M. (1996), Deep excavations a practical manual, Thomas Telford Publishing, London. DFI JOURNAL Vol. 6 No. 1 July 2012 [31]
14. Terzaghi, K. and Peck, R. B. (1967), Soil mechanics in engineering practice, 2nd edition, John Wiley, New York. 15. Tomlinson, M. J. (1986), Foundation design and construction, Longman, Harlow. 16. Randolph, M. F., and House, A. R. (2002), Analysis of suction caisson capacity in clay, Proceedings 34th Annual Offshore Technology Conference, Houston, Paper No. OTC 14236.
[32] DFI JOURNAL Vol. 6 No. 1 July 2012
17. Wu, T. H. (2011), 2008 Peck lecture: The observational method: case history and models, Journal of the Geotechnical and Geoenvironmental Engineering , ASCE, 137(10), pp. 862-873.
