Section 4 Pile design

4.1.1 Generally speaking, the methodology for driven pile presented in ISO 19901-4 Petroleum and natural gas industries – Specific requirements for offshore structures – Part 4: Geotechnical and foundation design considerations presents a suitable basis for driven pile design.

4.1.2 Further guidance on driven pile design is given in the following Sections.

4.2.1 It should be noted that the reliability of the ‘main text method’ for axial capacity in sands in Section 8.1, ISO 19901-4 Petroleum and natural gas industries – Specific requirements for offshore structures – Part 4: Geotechnical and foundation design considerations is below that of the CPT-based methods. Whilst the ‘main text method’ may be suitable for a preliminary assessment of pile capacity, the CPT-based design methods, such as those presented for sand in the Annex A of ISO 19901-4 Petroleum and natural gas industries – Specific requirements for offshore structures – Part 4: Geotechnical and foundation design considerations, generally present a more reliable method for assessment of axial capacity. Lehane et al. (2017) make a comparison of the different methods by examining the measured values of capacity to calculated values and determination of the coefficient of variation of the different methods. However, it should be noted that there are occasions when the CPT-based methods for sand give a wide range of capacity estimates; therefore, it may be prudent to calculate capacity using more than one method before committing to a design.

4.2.2 Aside from the main text method for clay in Section 8.1, ISO 19901-4 Petroleum and natural gas industries – Specific requirements for offshore structures – Part 4: Geotechnical and foundation design considerations, other methods for clays that may be applied include the Imperial College Pile Design method, NGI-05 and Fugro-96 method, as referred to by Lehane et al. (2017).

4.2.3 When selecting a method for design it is important to ensure that it is applicable and calibrated to the soil types under question. For example, if the soil type is unconventional, such as silt or carbonate, then it may be the case that a pile capacity method cannot be directly applied, or that more than one method should be considered for the same soil type to ensure that the range of possible soil and pile behaviours is examined.

4.2.4 It is important that sufficient site investigation data is collected to allow a confident application of a particular pile design method, both in terms of quantity and quality.

4.2.5 For assessment of pile axial capacity, the use of a reliable pile design method is a preferred approach, rather than an alternative that may place reliance upon proof of pile capacity during installation. Use of dynamic testing during installation requires significant experience in similar soil types and requires conversion to an equivalent static capacity and application of various corrections (e.g. due to scour or cyclic loading). This leads to uncertainty in pile length required and may create longer piles with other impacts such as increased installation time and additional noise, see for example Jardine et al. (2015).

4.2.6 Ageing of skin friction in sands and clays is a well-recognised concept. It may be used but would require careful demonstration that it is valid for the soil conditions and piles in question, taking into account various factors such as cyclic loading, timing of maximum loads after installation and any pile brittleness effects on the structure. It is important to differentiate between setup effects that occur relatively quickly after driving is complete (i.e. days, weeks or months), versus ageing effects that occur over a much longer time (i.e. months or years).

4.3.1 There are various PY curve models available for the analyses of lateral stiffness and capacity in clays. The PY curve models tend to have an empirical, or semi empirical, basis and were developed taking into account a range of different conditions – either intentional, or coincidental. When selecting a PY curve model the following conditions should be considered.

4.3.2 Cyclic or static loading: The load condition being analysed and whether it tends towards cyclic or static when considering the soil types and their likely behaviour. For example, for a 50 or 100 year in-place condition it is common to assume that cyclic degradation has occurred and therefore a cyclic PY curve should be used. However, under other conditions, such as fatigue loading, the majority of fatigue damage may occur when the soil is not experiencing significant effects of cyclic degradation (either due to lack of degradation or recovery from previous degradation) and in this case it may be more appropriate to use static PY curves. A final example is for pushover loading where the displacements involved are higher than the zone typically affected by cyclic degradation and therefore at displacements associated with ultimate pushover state, static resistance without cyclic degradation is more appropriate (Gilbert et al., 2010).

4.3.3 Implementing the model as developed: Many PY curve models do not consider the effect of axial loading on lateral response, or vice versa. In this case it is appropriate to consider if the two loading types may interact and whether adjustment to the model should be made to account for this. For example, the Jeanjean (2017) model in clays would require axial loading to be accounted for.

4.3.4 It should be ensured that it is appropriate to apply the PY curve model to the soil type(s) under question. For example, the pile tests used by Reese and Van Impe (2005) to develop the PY curve model in stiff clay were performed in high plasticity clay that shows a very large reduction in post peak resistance when compared to more typical North Sea clays.

4.3.5 The concept of conservatism does not necessarily apply to the assessment of lateral pile stiffness and capacity. For instance, it is not necessarily conservative for a structure if the foundation model used is 'soft' or 'weak'. Hence, the assessment of lateral pile capacity and stiffness should be as accurate as possible, giving due attention to uncertainties where they exist.

4.4.1 Jeanjean (2017) presents a comprehensive discussion of the main text PY curve model for clays as presented in ISO 19901-4 Petroleum and natural gas industries – Specific requirements for offshore structures – Part 4: Geotechnical and foundation design considerations and propose an updated version that may be more appropriate. Through an examination of the theoretical background the development of the method is explained and they further refine the method and definition of input parameters and assumptions generally leading to a significant increase in lateral resistance and stiffness for monotonic loading.

4.4.2 Zhang et al. (2017) further expand on the method for monotonic PY curves, presented by Jeanjean (2017), to give a methodology for cyclic PY curves. The method can account for non-symmetrical loading and a simplified approach can be derived to allow easier integration of the PY curves into PSI modelling for structural analyses. In order to use the simplified method, it would need appropriate calibration for local soil conditions and the nature of cyclic loading applicable to the structure in question.

4.4.3 The method proposed by Jeanjean (2017) and Zhang et al. (2017) requires the input of site-specific direct simple shear strength test results for monotonic and cyclic loading in order to derive the PY curve shape (i.e. P/P_u with y/D). However, in cases where such data are not available, for example in preliminary design, recommendations are given for different shear strength ranges.

4.4.4 Jeanjean (2017) concludes that diameter effects are appropriately captured within the method and as such the proposed method can be applied without correction. The alpha value can be calculated using the main text method for axial capacity in clays presented in ISO 19901-4 Petroleum and natural gas industries – Specific requirements for offshore structures – Part 4: Geotechnical and foundation design considerations. However, the final alpha value selected for calculation of PY curves may also need adjustment to account for the combined effect of axial and lateral loading.

4.5.1 It is generally recognised that the PY curve methods included in common International Standards such as ISO 19901-4 Petroleum and natural gas Industries – Specific requirements for offshore structures – Part 4: Geotechnical and foundation design considerations are not appropriate for large diameter piles except, perhaps, as a first approach to preliminary design. Byrne et al. (2015) and Byrne (2017) present the results of the pile soil analyses (PISA) project which aimed to provide a suitable process to design large diameter piles as well as to provide a more accurate assessment for smaller diameter piles if required. This process considered a rule-based method which takes basic strength and stiffness parameters to derive soil reaction curves and a numerical based method which takes more detailed strength and stiffness parameters combined with the use of a suite of detailed three-dimensional finite element analyses to calibrate a simple one-dimensional (1D) finite element model (i.e. PY curves). The rule-based method is considered suitable for initial feasibility or concept designs, whereas the numerical based method is generally considered necessary further into the detailed design process.

4.5.2 As shown in Figure 3.4.1 Components of soil capacity considered under PISA methodology (reproduced from Byrne, 2017), the PISA methodology also considers additional sources of soil resistance to resist lateral and moment loading including vertical stresses due to external wall friction, shear force at the pile base and a base moment. These additional components of capacity are usually combined with traditional PY curves used in a 1D finite element analyses and in PSI models through the structural design process. However, caution should be exercised with a more refined assessment to ensure that components of axial resistance (e.g. friction) are not double counted as lateral resistance (or vice versa) without an assessment of interaction to ensure the design does not become unconservative.

4.5.3 Care should be taken with selection of soil properties given their differing influence on ULS, SLS and FLS design cases and especially the influence of cyclic loading on pile response and capacity. Byrne (2017) also recognises that engineering judgement is required in using the process to ensure that factors such as installation effects or cyclic loading are appropriately considered in use of the methodology that has currently only been developed for static loading. See Ch 3, 4.4 PY curve models for further discussion on PY curve methods.

4.6.2 When designing monopiles it is important to take accumulated displacements and rotations into account as they may impact the ability of the foundation to function properly over the entire service life. Accumulated displacements have been the subject of various research including Leblanc et al. (2010) and Rudolph et al. (2014). In the current approaches, such as those presented in ISO 19901-4 Petroleum and natural gas industries – Specific requirements for offshore structures – Part 4: Geotechnical and foundation design considerations, the direction of cyclic loading is assumed to be constant, however, as presented by Rudolph et al. (2014), this is unlikely to be the case with various short-term, seasonal and long-term variation in load direction and magnitude. There are various relationships for predicting accumulated rotation, including that in DGGT (2013) where a logarithmic approach is suggested.

4.6.3 For monopiles it generally appears to be the case that variation in loading direction causes displacements that may be significantly higher than that caused by uni-directional loading. Therefore, it is important to assess the potential accumulated displacement and rotation taking into account the soil type and loading characteristics combined with appropriate methodology such as that presented by Rudolph et al. (2014) and Leblanc et al. (2010). Where accumulated rotations or displacements could be critical, it may be appropriate to ensure that an appropriate safety margin is included either in the specified allowable rotation or displacement or by application of load and material factors to the design methods.

Figure 3.4.1 Components of soil capacity considered under PISA methodology (reproduced from Byrne, 2017)

4.7.1 Chalk is found extensively across north-west Europe and is a fine-grained material consisting of calcite debris with a typical unconfined compressive strength (UCS) of 3–5 MPa and cone resistance in the order of 4–50 MPa (Carrington et al., 2011).

4.7.2 According to Lord et al. (2002) axial shaft friction for driven open-ended steel tubular piles should be limited to 20 kPa in low and medium density chalk and shaft resistances in the order of 120 kPa may be expected in high-density grade A chalk. Shaft resistance in chalk is dependent upon a number of mechanisms including installation method, remoulding, excess pore pressures, ‘tightening’ of discontinuities, strength of the chalk itself and the degree of porosity. Large set-up may be expected after driving. For example, Buckley et al. (2017) found that the remoulded zone varied in a range 0,59–1,64 times the pile wall thickness.

4.7.3 The range of frictions presented by Lord et al. (2002) did not allow an optimised design for low to medium density chalk and in recent years different investigators have sought to improve the methodology.

4.7.4 Carrington et al. (2011) presented an assessment method for shaft resistance based upon laboratory testing including cyclic direct simple shear tests and presented an improvement for shaft friction such that a range of 20–50 kPa could be used for low to medium density chalk. Ciavaglia et al. (2017) conducted a series of onshore pile tests in low to medium density chalk and the results showed that shaft friction a few days after driving was an average of 23 kPa and that, for an aged pile, this increased sevenfold to 168 kPa. Buckley et al. (2017) found similar results for driven piles where resistance at the end of driving was in the range 15–17 kPa and a setup factor of up to 5,3 was found 246 days after driving. Buckley et al. (2017) also tested the instrumented Imperial College piles that were installed by pushing and they found that these piles had a setup factor of less than 1,0: this has significant implications for some piles, such as large diameter piles or those with significant structural weight applied during installation, that are often installed to a greater or complete extent by self-penetration and may not experience the same friction enhancing effects that driven piles do.

4.7.5 When extrapolating the use of field tests into design, care should be taken to include the effect of lateral loading on shaft friction. For example, Ciavaglia et al. (2017) found that if a pile was subjected to lateral loads up to 50 per cent of the lateral capacity, the effect of setup was reduced such that shaft resistance was up to 65 per cent lower. Efforts should therefore be made to assess the degree to which lateral capacity will be mobilised and what influence this may have on shaft friction and this will depend upon pile flexibility. In addition, where using field tests to support offshore pile design it should be ensured that the chalk characteristics and nature of the applied loading are similar; for example, to take account of any cyclic effects. These effects will have different impacts on monopiles or driven piles.

4.7.6 According to Lord et al. (2002) driving in grade A high density chalk can be difficult, whereas driving in lower grade B, C and D chalk will be easier. Set-up effects can be very significant and although this is beneficial for shaft resistance it can have a significant impact if any driving delays are encountered, with a higher risk of premature refusal.

4.8.1 High shaft frictions can be generated for driven piles in rock, for example, as reported by Rodway and Rowe (1980) and Long (1991). The driven pile is sometimes used with relief drilling to ensure that further pile penetration can be achieved. As yet, there is little practical experience on shaft capacity in rock and the impact of relief drilling on shaft capacity.

4.8.2 If driven piles in rock, with or without relief drilling, are critical to the pile design (i.e. a significant proportion of the pile capacity is due to a driven element in rock), then it may be prudent to perform appropriate pile testing to optimise pile lengths. A non-optimal pile design may result in longer piles that become too difficult to install with higher potential for premature refusal.

4.9.1 Drilled and grouted piles, or rock sockets, may prove to be most suitable under certain conditions where hard soils or rocks are encountered and typical configuration is shown in Figure 3.4.2 Typical drilled and grouted pile configuration.

4.9.2 CIRIA (2004) gives a useful overview of issues of pile design in weak rock and expansion of some key issues is provided in the following Sections.

4.9.4 In weak rock, an approach such as that by Kulhawy and Phoon (1993) may be suitable for defining shaft resistance. It may be appropriate to apply a limit to the frictions determined. Deliberate roughening may be used to enhance the shaft friction; for example, by the inclusion of grooves in the drilled hole.

4.9.5 Once the shaft friction has been determined, it is necessary to confirm that grout-steel friction will not limit the amount of shaft friction that can be mobilised. The methodology presented in ISO 19902 Petroleum and natural gas industries – Fixed steel offshore structures for calculating the capacity of a grouted connection may be used; making appropriate assumptions for the smoothness of the pile steel surface.

4.9.6 If excessive debris has collected near to the hole base and remains present then end bearing may only be mobilised at excessive displacements; unless the debris is removed. Furthermore, unless the pile is plugged (e.g. by grout) the end bearing will be restricted to that on the pile annulus although this may be substantial. It should also be noted that the displacements required to mobilise end bearing may be such that significant degradation is caused on shaft friction and in this case it is common practice to only count shaft friction or end bearing in design.

4.10.1 Where suction caissons (also called suction piles or suction buckets) are considered, it should be first established that the concept is applicable.

4.10.3 It is common to practise perform sizing of a suction caisson using simplified methodology such as the work by Suryasentana et al. (2017) or alternative formulations of upper bound methods, see for example Hamilton and Murff (1995). Following this the suitability of design should then be confirmed suitable by the use of finite element analyses that can more precisely take into account aspects such as soil layering or other soil properties.

4.10.4 Where a structural analysis requires stiffness input it should be ensured that the stiffness matrix in the structural analyses is compatible with the level of loads used. Many structural analyses packages only accept a linear stiffness matrix input and in this case it may be necessary to iterate the stiffness matrix until compatibility between loads and stiffness is achieved. It may be necessary to vary the stiffness matrix depending on what is critical to the analyses and associated load levels. For example, the load levels and impact of cyclic degradation may be quite different for ULS loading (where fewer bigger waves occur) when compared to FLS loading.

4.10.5 An initial estimate of suction pile stiffness may be made using the methodology published by Suryasentana et al. (2017) or Doherty et al. (2005) and validated by finite element analyses as appropriate.

4.10.6 Detailed design will also require an estimate of the long-term performance of the foundation to ensure that rotations and displacements remain within acceptable limits. This aspect was investigated by Zhu et al. (2017) for response of suction caissons in single-layer and layered seabeds and an expression for accumulated rotation is provided along with the required calibration parameters depending on soil conditions and relative layer thicknesses. The current work has focussed on uni-directional loading and further development or validation is required before applying the methodology to multi-directional loading.

4.10.8 If re-spudding of the caisson is required, due to refusal or some other installation issue, the required offset distance from the refusal location should be considered such that there is no significant impact upon the installed foundation.

4.10.9 Where experience in local or comparable soil or seabed conditions is not available, then it may be prudent to perform field trials of suction caisson installation before committing to the concept.