Having reviewed several problem pile results and/or carried out back analysis of preliminary pile tests, I believe that procedures for carrying out pile load testing in the UK could be significantly improved.
talking point david illingworth
Fleming’s paper “A new method for single pile prediction and analysis” was published in Geotechnique over 25 years ago and proposed a method of interpreting pile test data. Through the paper, Fleming put forward the hypothesis that hyperbolic functions could be used to separately model both the shaft and base settlement behaviour under load.
The approach mainly considers two parameters related to both shaft friction and base behaviour, with an allowance also included for elastic deformation. Fleming also outlined that the assessment of pile behaviour is based on figures defined as infinite time settlements. It is mainly the use of infinite time figures that the current testing procedure - as given in the ICE Specification for Piling and Embedded Retaining Walls (SPERWall) - does not sensibly reflect.
The currently used test procedure uses varying hold periods, together with limited recording of data. Until last year’s update of SPERWall, the typical test procedure also included cycling although this current version states that ‘unless specified otherwise the single cycle will be used by default’. These three points are not conducive to the assessment of infinite time figures. It may be noted that the use of automatic recording of pile test data was only available shortly before this paper was published and, prior to this, readings were typically taken manually. The development of automated load maintenance and deflection reading was a direct result of this paper and although the method of recording test data has improved over the last 25 years or so, consideration of what could be obtained from preliminary pile testing has not. Consequently, the industry’s database of good quality published data by which to verify our design approaches is limited.
The use of a normal preliminary static pile test as a design tool should be advocated and is in the spirit of Eurocode EC7. Careful consideration needs to be given to such tests, where the test may be extended where the performance is better than expected and also in the use of amended smaller increments where the pile is not performing as well. The design engineer should be allowed direct access so that, whilst the test is being carried out for the prescribed load increments, these can be amended to allow further increments when the pile is showing non elastic characteristics. In this respect, additional increments could be included close to when the shaft resistance is mobilized and where the effect of the base is becoming more dominant. Longer hold increments are required, typically of 2-4 hours duration, to allow sensible prediction of infinite time – with the load kept relatively constant during the period; with readings taken at one minute intervals. These hold periods do not have to be at standard increments.
We often see preliminary tests carried out as proof tests which, together with the typical SPERW procedure, may not allow sensible back analysis. The test procedure often seems to be restricted and this may be due to contractual arrangements, ground conditions and/or time restraints. Undertaking preliminary pile testing has a significant cost and too often the tests are curtailed for various non-technical reasons without consideration given to what additional information can be obtained.
Until both piling and testing industries can accept the idea that preliminary load test need to be carried out with an appropriate testing regime it seems hard to understand how a positive advancement in geotechnical design practices can be made.
This is of particular interest with the Chalk conference coming up in September, there seems to be a lack of appropriate pile test data on chalk within the industry. This was mentioned by the FPS within a Ground Engineering article in November 2004 and also referring to CIRIA publication 1994. We have seen several preliminary pile test results within Chalk, where failure has not been achieved. Yet again this shows the industry’s reluctance to use preliminary pile testing as a design tool.
- David Illingworth is technical director of Piledesigns