Preliminary inspections Anchor heads and caps were generally found to be in good condition and with no evidence of significant corrosion of the protected anchor heads.
Original design loads for the anchors were typically specified as 550kN. Results of pre-piling checklifts some twenty years later revealed load levels lower than this, with the majority being in the range 200kN to 400kN.
It was suspected, as often occurred on such works, that the lock-off loads had been set at a lower level than the design working loads.
This might have been because the original design loads in the anchors caused the sheet piles to be pulled back beyond acceptable tolerances, or, alternatively, to allow for some outward movement of the piles if the full design loads should finally come on to the wall.
Records were found later indicating that movements in the sheet piles had occurred during initial stressing and that, as a result, many anchors had not been stressed to their full design load.
Monitoring will be covered in detail in a further paper. It is worth noting, however, that the present phase of monitoring of the ground anchors at the Erith site typically showed variations in their loads before and after piling. These fluctuations appeared unrelated to the works, and seemed attributable to other tidal, seasonal and diurnal influences.
Similarly, fluctuating movements of the river wall of up to 15mm were recorded over the duration of the works: again apparently unrelated to the works themselves. They again appeared to represent the response of the wall to tidal, seasonal and diurnal influences.
Piling Pile loads varied between 300kN and 675kN. Where no constraints existed, 350mm diameter CFA piles were used, with only the most heavily loaded piles needing to be 500mm diameter. However where piles were located above the anchors (Figure 9), where constraints on depth applied, many of them - even those comparatively lightly loaded - were increased to 500mm diameter to reduce length and keep their toe levels above the exclusion zone above the ground anchors.
As noted earlier, where replacement tie-bar and deadman anchors were installed, it was anticipated that horizontal loads would be imposed on piles in the permanent condition.
Piles close to the deadman anchors of the tie bars, therefore also had to be designed to resist these loads.
Where it was deemed that a potentially laterally loaded pile could safely be installed between anchors, or terminated a safe distance above an anchor, these were constructed prior to stressing the replacement tie.
Stressing of the tie took place only after the concrete in the piles had reached their 28 day strength. With replacement ties stressed and corresponding ground anchors decommissioned, remaining piles could then be installed.
For those piles which would be laterally loaded but could not be installed before stressing of the ties, steel casings were driven which would act as permanent 'external reinforcement' over the depth at which lateral loads were assumed to act. In this way, fresh concrete in the piles would not be exposed to lateral loads. These laterally loaded piles were typically 600mm in diameter.
Failure/damage to ground anchors During post-piling checks of the ground anchors, three out of the total of 81 were found to have been damaged by the piling operations and to have failed.
The remains of a punctured anchor cap indicated that one of these had failed explosively, though as the heads were only visible from the river, it is not known when this occurred. The remaining two showed excessive movement at the head during post piling check-lifts, when only minimal loads had been applied. On both these anchors however, the cap and head assembly had remained intact.
The first task was to try to establish reasons for these failures. Cab logs were obtained for piles adjacent to each of the failed anchors, as well as several which would act as control datums.
Where relevant, depth records on the logs were also checked to ensure that maximum toe levels had not been exceeded in error. It should be noted that the provision of onboard monitoring and logging of the pile construction process had been a stipulation of the contract.
Several of the pile logs adjacent to the anchors showed anomalies in speed and torque. The depths at which these anomalies occurred were plotted on to pile elevations on cross sections through the anchors.
It could be seen that the anomalies all corresponded with the depth of the failed anchor, suggesting that the auger had cut through one or more of the tendons. When plotted on plan, the piles with anomalies proved to be those nearest to, and on the same side of, the failed anchor.
From this it was concluded that each of the failed anchors must have originally been installed approximately 6¦ out of alignment on plan or that the piles had exceeded their envisaged tolerances by a similar degree. This was considerably outside the pile and anchor tolerances which had been assumed.
It would have been of interest to extract the failed tendons to confirm and examine the position at which they failed and their mode of failure. However, because of the diffi cult access, the cost involved was not commercially justified for the likely benefi t to the project that such removal and inspection might bring.
Survey monitoring of the wall indicated that no abnormal movements were occurring in the area around the failed anchors. Checklifting also showed that anchors on either side of the failed anchors had not experienced any significant changes in load.
It was known that the anchors had originally been designed to be capable of supporting a 50% overload, as an allowance for possible failures. Replacement of the failed anchors was therefore important so as to maintain long term integrity of the wall and reinstate a degree of redundancy but was not considered an emergency, given the ongoing monitoring of wall behaviour.
With piling and substructure complete and, in some cases, construction of the concrete superstructure frame underway, it was not feasible to install more sheet piled deadman tie bar anchors to replace the failed ground anchors, nor was it cost-effective to design and install replacement ground anchors: both options which had been considered during the design risk assessment stage.
Instead, the building substructure was used to provide the required resistance. In each case the complex, substantial and interlinked pile caps provided the necessary anchor capacity based solely on passive resistance on the leading face.
Lateral resistance of the piles (most of which were heavily reinforced 500mm or 600mm diameter piles and some of which were also steel cased) and friction on the underside of the pile cap were ignored in the calculations but acknowledged as further enhancing the overall factor of safety.
Following installation and fixing of the tie bars and completion of the superstructure to at least first or second fl oor level (to provide some precompression in the piles), the replacement tie bars were prestressed to a load similar to the horizontal component of the relevant anchor prior to failure.
Throughout these procedures, survey monitoring of the wall continued. These observations continued to record no more than the normal range of movements, indicating that the wall had not been adversely affected by these isolated anchor failures.
In overall project terms, the loss and subsequent replacement of the three failed ground anchors was considered not unreasonable given the size of the development and the nature of 'as-built' information which was available.
The perceived benefits from the considerable engineering analysis and effort, was that it enabled the successful construction of all of the apartment blocks on the project even though the majority were located above a potential construction nightmare.
In other circumstances, the presence of the anchors would have sterilised the most desirable parts of the site or rendered it uneconomic to develop.
Summary At first sight, the development had appeared to present insuperable problems to an economic solution.
The 1970s phase of bank raising and flood prevention works appeared to have effectively sterilised the site for its proposed regeneration with residential housing.
However, by piecing together the history of the bank-raising works, it became possible to see that the heavy foundation loads could be transferred successfully into the ground above, beyond and between the prestressed ground anchors.
In particularly congested or heavily loaded areas, judicious and controlled detensioning of the anchors, and their replacement by suitable horizontal tie bars allowed larger piling 'windows' to be created.
With the available information, a logical risk appraisal could be developed for the site, which even allowed a proportion of the anchors to be damaged by the pile installation process without jeopardising the fulfilment of the project. Measures had already been laid down in the preliminary assessments as to how such problems would be met and overcome.
It is hoped that some of the lessons learned on this project may serve as a guide to other similar future developments.
An interesting further point that has emerged from this project is that an increasing number of the people who were involved in the 1970s bank-raising works (and other similar high technology projects that came to fruition around that time) are now nearing the end of their careers, have already retired, or otherwise left the industry.
A valuable store of knowledge on techniques and methodologies is very close to the point of being lost, and hard-learned lessons from quite an innovative period of the industry are on the verge of having to be learnt all over again.
It was also somewhat surprising, and, perhaps, depressing, how little information of this important project had been preserved over a relatively short time span.
Conclusion The key to this project has been the assessment of risks associated with piling in the proximity of highly stressed ground anchors.
A rationale was developed, based upon a knowledge of the construction processes and techniques associated with the ground anchor system, to evaluate how closely piles could be installed to the anchors with a reasonable assumption that there would be minimal interference to them.
An essential element of the design was the desk study.
The surprising paucity of records for a relatively recent, technically demanding and highly supervised bank raising project highlights how important it is to keep clear and accurate records of construction for all projects, and to ensure their safe keeping for posterity.
Preservation of such important records should not just be left to chance, particularly in the case of works below ground which cannot be readily inspected.
This will be increasingly important as engineering solutions, procurement routes and ownership of projects become more and more complex. In this particular case, the engineering solution was aided by the available records and the knowledge of individuals connected with the original bank-raising project.
These enabled the client to realise, in full, its expectations for this site, with a scheme which, despite its design complexity, has remained economically viable.
Acknowledgements The authors would like to pay tribute to the tenacity of Lucky Wehalle, of the Environment Agency, who, during the transfer of responsibilities from the former Greater London Council, managed to rescue and ensure the preservation of at least some of the contract drawings and information on the bank-raising works at the Erith site. Without his foresight, the design of the Chandlers Wharf redevelopment would have been even more difficult.
The authors would also wish to thank George Wimpey Central London, the client and main contractor for the Chandlers Wharf project, for assistance with and permission to publish the information contained within this paper.
For the record, other participants in this aspect of the project were:
Anderson Construction (groundwork and structures subcontractor), Westpile (piling contractor) and Dew Pitchmastic (replacement deadman anchors).
References 1 Turner MJ (1980). Rock anchors: An outline of some current design, construction and testing practices in the United Kingdom.
International Conference on Structural Foundations in Rock. Sydney, 1980.
2 DD81 (1983) British Standard draft code of practice for ground anchorages.
3 BS8081 (1989). British Standard code of practice for ground anchorages.
4 Turner MJ, Selley PJ and Cooper GM (1992). Ground anchorages to the retaining walls at Pen-y-Clip. International Conference on Retaining Structures, ICE, Cambridge, 20-23 July 1992.
5 Raison CA (1987). Ground anchorages: drill hole alignment determination at the British Library, Euston. Proc ICE Part 1, 1987.
6 Institution of Civil Engineers (1996).
Specification for piling and embedded retaining walls. Thomas Telford.
7 Bruce DA (1989). Methods of overburden drilling in geotechnical construction - A generic classification'. Ground Engineering. Vol 22, No 7. pp25-32.