Implementation of the OM

Initial Surveys

Precise surveys of the pile caps commenced about three months before commencement of arch raising. Precise surveys were initially very difficult because of the plethora of construction activities being carried out across the site. Following discussions on site, construction activities were rescheduled to create a better environment for precise surveying to be implemented. Survey methods were also progressively improved. During a six-week period, significant improvements in surveying accuracy were

Figure 10.13 Flow chart for implementation of the OM.

achieved, for example for total station surveying from about ±5 mm initially (which was inadequate) to ±1.5 mm. Hence, the prolonged time period available for improving the survey accuracy prior to arch raising was invaluable. It proved to be a crucial element in the success of the OM at this site.

Observed Pile Group Deformation

The critical observations for the pile groups were the rotation and horizontal deformation of the pile caps. Before and after each incremental lift, the six components of movement (three displacements and three rotations) were calculated from the surveying data and compared to the amber and red trigger values. For the temporary jacking and turning strut bases, the most critical load combinations were applied at arch angles of close to 30°, and for the main arch bases at angles of nearly 40°. During this phase, seventy-two sets of data had to be reviewed (six sets of data for each of the ten temporary bases and the two main arch bases) immediately after each stage, to determine if adverse trends were developing. The measured horizontal movement and rotation for the eastern arch base arc compared with the calculations from the pile group analyses (Figures 10.14 and 10.15, respectively). Observations during the first couple of arch raising stages indicated that conditions were considerably better than worst credible. The vast majority of the observations were close to the most probable calculated values and occasionally were between most probable and moderately conservative calculated values, as indicated in Figures 10.14 and 10.15 (note that one reading is on the amber trigger, a subsequent check survey indicated readings below the amber trigger). It should be noted that worst credible, moderately conservative and most probable conditions were based on back-analysis of the preliminary pile tests, as mentioned in Section 10.4; the analysis strategy has been discussed in detail by O’Brien and Hardy (2006) and O’Brien (2007). Throughout arch raising, weather conditions were good and wind speeds were low, hence the most onerous load combinations (associated with high winds) were not imposed on the arch. The jacks performed consistently well throughout arch raising and the arch distortion was maintained well within acceptable limits. As a result, it was not necessary to apply significant torsional loads at the jacking bases, which in turn reduced risks to the arch foundations.

During arch raising, amber trigger levels for the foundations were occasionally breached. Resurveys indicated that the high measurements were erroneous due to survey errors or problems with datum points. The amber and red trigger levels for a pile group were dependent on the load combination considered (e.g. combined horizontal and moment, or combined horizontal, moment and torsion). The amber/red trigger levels used during arch raising were based on the most critical load combination (i.e. the load combination which generated the lowest set of deformations to

Figure 10.14 Calculated and observed horizontal movement in the у direction, eastern Arch base.

Figure 10.15 Calculated and observed rotation, x axis, eastern Arch base.

cause structural over-stress in a pile). It was not necessary to modify these during arch raising. If there had been an adverse trend, it may have been appropriate (depending upon the actual conditions at that stage of arch raising) to have re-assessed the trigger levels based on the relevant load combination (e.g. based on known weather conditions). Using a single set of charts kept the process as simple as possible and provided an extra safety margin throughout the OM application.

Once the arch reached 40°, the applied loads on the arch foundations reduced, although recorded movements did not reduce until an arch angle of about 60° was passed. This apparent anomaly was considered to be due to high ground stiffness being mobilised as loads reduced (i.e. similar to observations during pile tests when a high stiffness is mobilised at the beginning of unloading, following initial loading). Once observed movements decreased, the OM phase was reduced in intensity, although the OM was not completed until the restraining lines were operational (when the arch was at an angle of 100°). Figure 10.16 shows the arch in its final position being held in position by the restraining lines. Completion of the arch raising marked the successful conclusion to a remarkable feat of civil engineering. The iconic arch now forms one of the United Kingdom’s most spectacular landmarks.

Conclusions

The application of the OM provided an effective means of minimizing risk during this unique and challenging civil engineering operation. The arch raising was completed as originally planned. The application of the OM through progressive modification enabled risks to be minimized during arch raising. It was possible to demonstrate that the risk of damage to the arch and the arch foundations was kept as low as reasonably practical. The monitoring and analysis effort, although considerable, was kept as simple as possible. This was important to ensure effective practical implementation of the OM and to maintain effective communication across the project team (which is a key aspect of the implementation of the OM). The application of the OM for pile group foundations is unusual, and it requires a consideration of the following:

• a) The trigger levels (for implementation of contingency measures) are dependent on the applied load combinations; hence, a wide range of potential scenarios need to be considered. Nevertheless, once the full range of conditions have been considered, this potential complexity needs to be simplified as far as possible (when e.g. defining trigger levels, movement monitoring and appropriate contingency measures) to facilitate a simple and practical risk management process.
• b) Pile groups are intrinsically stiff; hence, the monitoring system has to be capable of reliably measuring small deformations.

Figure 10.16 The arch in its final position. (Credit: DBURKE/Alamy Stock Photo.)

In common with other applications of the OM, it is essential that potential sudden and progressive failure mechanisms are avoided. For this case history, a programme of pile tests (subject to compression, tension and horizontal loading) taken to failure demonstrated ductile behaviour at deformations well in excess of those which would be mobilized during arch raising. The pile test data also facilitated the calibration of pile group analysis models and the assessment of moderately conservative, most probable and worst credible conditions. Contingency measures need to be quick and practical to implement. Crucially for this case history, a simple contingency (application of a kentledge at the rear of a pile group) was identified which would mitigate the most likely critical conditions.

References

Bolton, M. D. and Whittle, R. W. (1999). A non-linear elastic/perfectly plastic analysis for plane strain undrained expansion tests. Geotechnique, 49, No.l, 133-141.

Burland, J. B. and Kalra, J. C. (1986). Queen Elizabeth second conference centre, geotechnical aspects. Proceedings of ICE, Part 1, Design and Construction, 80, 1479-1503.

Chandler, R. J. (2000). The 3rd Glossop lecture, clay sediments in depositional basins: the geotechnical cycle. QJEGH, 33, 7-39.

England, M. G. (1999). A pile behaviour model, PhD Thesis, Imperial College, University of London.

Fleming, W. G. K. (1992). A new method for single pile settlement prediction and analysis. Geotechnique, 42, No. 3, 411-425.

Geocentrix. (2002). Repute User’s Manual.

Jardine, R. J. (1992). Non-linear stiffness parameters from undrained pressuremeter tests. Canadian Geotechnical Journal, 29, 436-447.

Mandolini, A. and Viggiani, C. (1997). Settlement of piled foundations. Geotechnique, 47, No. 3, 791-816.

NCHRP Report 507. (2004). Load and Resistance Factor Design for Deep Foundations, Transportation Research Board, Washington, DC.

O’Brien, A. S. (2007). Raising the 133 m high triumphal arch at the New Wembley Stadium, risk management via the observational method, Proc 14th European Conf on SMGE, vol. 2, Madrid, Spain, pp. 365-370.

O’Brien, A. S. (2012). Pile-group design. ICE Manual of Geotechnical Engineering, 2, chapter 55, 823-851.

O’Brien, A. S. and Hardy, S. (2006). Non-linear analysis of large pile groups for the new Wembley stadium, Proc 10th Int conf on Piling and Deep Foundations, Amsterdam, pp. 303-310.

O’Brien, A. S., Hardy, S., Farooq, I. and Ellis, E. A. (2005). Foundation engineering for the UK’s new national stadium at Wembley, Proc 16th Int Conf SMGE, vol. 2, Osaka, Japan, pp. 1533-1536.

Poulos, H. G. (1989). Pile behaviour - theory and application. Geotechnique, 39, No. 3, 365—415.

Powderham, A. J. (1998). The observational method - application through progressive modification. Proceedings Journal of ASCE/BSCE, 13, No. 2, 87-110.

Randolph, M. F. (2003). Science and empiricism in pile foundation design. Geotechnique, 53, No. 10, 847-875.