The fourth Dennis Waite Lecture was given by Tony Biddle, manager, civil engineering, at the Steel Construction Institute. The lecture appraised the application, design and testing of steel bearing piles, and centred on the SCI's new Steel bearing pile guide. Biddle explained that this report drew heavily on research carried out by the offshore industry and had been adapted for onshore use. It supersedes the earlier SCI Steel bearing pile guide, produced by GM Cornfield.
Biddle claimed that the UK construction industry was slow in taking advantage of recent developments and that codes of practice used by designers did not adequately reflect recent research.
The advantages of steel bearing piles over concrete bored piles during construction were fast installation and prefabrication, resulting in shorter on-site programme periods, said Biddle. Dynamic testing aided this, allowing rapid handover of the works to the next contractor on site. It was thought that these advantages more than offset the slightly higher basic material cost of steel piles on many types of structure.
He then made particular reference to the use of steel bearing piles as the foundations to integral bridges (or frame bridges). Integral bridges are defined as bridges that have no expansion joints at deck level (Figure 1).
Not using expansion joints and designing a bridge as a rigid frame allows longer spans and slimmer decks, but often with the added cost of a moment transfer connection on to the piled foundation. This was seen as an area where steel had a particular advantage over reinforced concrete, since steel is 'moment friendly', being more ductile than concrete, and readily accommodates the tensile bending stress from moments. Biddle thought this represented sound practice and efficient design but stressed the importance of good detailing in the reinforced concrete connection between the steel piles and the bridge decks.
Design guidance is given in the new Steel bearing piles guide. Biddle explained limit state design to Eurocode 7 (EC7) using the partial factors method as it applies to piling (Figure 2).
For example, Working load = pile ultimate resistance (Rc) (fl x m1 x m2)
where Rc = load at an uncontrollable settlement or say 40mm pile head settlement.
fl = 1.35 from BS5400 (load factor)
m1 = 1.5 from EC7 Part 1 (factor based on amount of testing)
m2 = 1.3 from EC7 Part 1 (factor based on pile material)
The product of the two EC7 factors gave a multiple of 1.95 for a single pile test result, similar to the usual lumped factor of 2.0 but derived by a more rigorous approach. Biddle noted that this would be easier to justify to a client than the rounded figure. When calculated by the above method the pile head movement under working load had been shown to be of the order of 7mm to 10mm for friction piles and pile elastic compression of only 2mm to 3mm for endbearing piles (dependent on their length).
Biddle explained the differing mechanism of load transfer between friction piles and end bearing piles (Figures 3 and 4).
He added that shaft friction for driven steel piles was generated in a different manner to that for bored concrete piles and that lower unit adhesion () values had been found than were currently used in clay soils. However, higher unit end bearing could be achieved in granular material and particularly in rocks than was recommended in BS8004:1986 for bored piles.
A controversial aspect of steel tube pile design is the phenomenon of plugging, where the latest research was in conflict with cur- rent codes of practice. Plugging is defined as the case where the internal shaft friction between the soil and steel tube is greater than the gross end bearing of the pile, such that a plug is formed within the open tube (Figure 5).
An example of this is BS6349:Part 2, which recommends that a plugged pile be taken as the design case, whereas BS8004 only asks that a check on this be carried out. The SCI accordingly states that plugging is very rare and recommends that BS6849 be revised. Evidence for this claim were the site trials carried out on the piling for the A46 Newark bypass bridges. On this contract the significance of whether the piles would plug or not arose due to an argument between the designer and the checker and as a result a 'target' penetration into the underlying Mercia mudstone could not be agreed upon. Site pile testing yielded the graph shown in Figure 6, demonstrating that the unplugged calculation method was the most valid prediction for this site.
As a result, the SCI recommends that the bearing capacity of a steel tube pile be calculated on the skin friction developed over the total embedded shaft area plus the pile wall end bearing.
In addition to the 'plugging' debate, Biddle said an argument was taking place over whether the SPT N method, or the pseudo effective stress method used offshore was most applicable to granular soils. Testing indicated the SPT method to be most accurate. It was felt that more shearbox testing data was needed on sites to refine the effective stress method as used offshore.
For piles bearing on to cohesive soils, there was debate over what value of the adhesion factor () should be used. Shaft friction is calculated using:
fs = Cu
(where Cu = triaxial test undrained cohesion)
The code of practice for foundations BS8004 recommends an value of 0.5 for bored piles and values of between 0.3 and 0.6 for driven steel piles (H or sheet). SCI research suggests that a value of =0.25 over the full shaft area will be developed in the short term, ie during a load test, and that this will double to =0.5 over a period of weeks or months. This is because short term tests do not allow for the dissipation of excess pore water pressures set up during the driving of steel piles. It was acknowledged that it would be difficult to get site specific proof of this long term build up in strength, so therefore =0.25 should be used to derive a conservative minimum design pile capacity.
Where piles bear into rock and end bearing predominates, the values for unit end bearing that are given in BS8004 have been found to be uneconomically conservative. For example, inferred bearing values of 29MPa and 38Mpa have been measured from steel pile tests in weathered mudstones and sandstones. This contrasts to the values of 2MPa and 4MPa given in BS8004. Biddle claimed that in some harder rocks, the measured values of 100s of MPa using dynamic testing methods were to be believed. A reason for lower strength being predicted was that soil testing at rockhead was often done by CPTs. CPT equipment typically has maximum allowable pressures of 70kPa whereas rock strength could be greater than 300MPa.Under-prediction was inevitable and the best way to test pile capacity was by use of trial piles driven into the rock.
Biddle turned to the concepts and practice of stress wave analysis and dynamic pile testing. He described the two main software packages, GRLWEAP and CAPWAP, both of which use stress-wave theory. GRLWEAP is used to predict driveability and load capacity and to aid pile section and hammer selection by using a computerised model of the hammer, pile and soil system. CAPWAP analysis of stress wave data from dynamic test results gives driving stresses, soil resistance distributions and allows for a pseudo static capacity prediction both during and after driving. Both packages are highly developed and have been validated worldwide.
Comparisons between the data gained by static testing and CAPWAP analysis were made. Using the Goble and Rausche database, a graph of predicted pile capacity using CAPWAP and measured capacity using static testing was presented (Figure 7). This showed maximum scatter about the predicted - measured line of the order 15%to 20%. This was consistent with the limit state design approach given in Eurocode 7, said Biddle, where the factors m1 and m2 had values of 1.3 and 1.5. This allowed for errors in the pile test results and in the generic calculation prediction methods used.
The practice of using pile driving formulae such as the Hiley Formula was also discussed. Biddle said it was often the case that the 'set' calculated from such formulae was used as the sole judge of a pile's capacity and hence suitability. Such formulae were only relevant to pile driving in the 1920s and were now obsolete and he urged engineers not to use them as they were too imprecise. GRLWEAP and CAPWAP were the tools of today, he said, although caution was advised in clays where the 'phenomenon of set up' could be difficult to model.
Neil Buchanon described a site in Glasgow where steel piles had stopped short of their design penetration into Boulder Clay. At the time this was thought to be due to the piles founding on boulders. He asked whether dynamic testing could give a definitive answer to this problem. Biddle said it could not and the best solution was to gather all the data and look for a pattern to establish the general soil profile.
Fiona Chow of Imperial College challenged some of Biddle's findings. Imperial College research had shown that values as high as 1.0 could be expected for London Clay, compared to the SCI values of 0.25 to 0.5. Caution was urged when using pile stress wave theory for open ended piles.
The issue of what 'long term' and 'short term' values meant was then addressed. Biddle said it depended on pile type, shape and length and it was a complicated problem. In regard to the value, the object of using generic prediction methods was to confidently predict lower bound values of pile resistance for a wide range of soils for the soil condition just after driving. This allowed verification on site. These values would obviously be lower than long term values. Databases were being developed to investigate the phenomenon of set up with time.
Asked to comment on the durability of steel piles, Biddle replied that the relevant codes of practice, BS8004 and BS6349, dealt with this adequately, although further research work was required to determine the value of coatings on buried steel.
The comment that pile head settlement of 7mm to 10mm at working load could be expected was questioned. In particular, the effects of elastic compression on longer piles and the amount of movement required to mobilise full shaft friction was discussed. Biddle said the 7mm to 10mm movement related to the average embedded lengths measured by the SCI of 15m to 20m but that piles with greater free lengths could perform differently. This order of movement was considered enough to mobilise sufficient shaft friction to support the working load. The elastic impression of endbearing piles was generally much less than 5mm.
Len Threadgold commented that while SPT N values could be directly used for the design of piles in granular soils, with cohesive soils it was often the case that N values had to be converted to an undrained cohesion value using the correlation of Stroud and Butler (1974) before design values could be applied to the pile. It was considered that direct SPT N value correlations should be used.
Biddle concluded the meeting with the observation that the science of steel piling had moved a long way in recent years. In particular, the use of stress wave theory enabled the 'factors of ignorance' - in the form of excessively high factors of safety - to be removed. He urged engineers to have faith in this technology to allow progress to continue.
Biddle, AR (1997). Steel bearing piles guide, the Steel Construction Institute.
Cornfield, GM (1989). Steel bearing piles, Fourth edition, the Steel Construction Institute.
British Standards Institution (1986). BS8004: Code of practice for foundations.
British Standards Institution (1988). BS6349: Code of practice for maritime structures, Part 2.
Goble, Rausche, Likins & Associates. GRLWEAP 1997-2, Wave equation software program for pile driveability prediction, Cleveland, Ohio.
Goble GG, Likins G and Rausche F (1991). CAPWAP users manual.