CPT-based lateral displacement analysis using p-y method for offshore mono-piles in clays

Transcription

1 Geomechanics and Engineering, Vol. 7, No. 4 (2014) DOI: CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee Department of Civil and Environmental Engineering, Yonsei University, 134 Shinchon-dong, Seodaemn-g, Seol , Repblic of Korea (Received Agst 27, 2013, Revised May 17, 2014, Accepted Jly 03, 2014) Abstract. In this stdy, a CPT-based p-y analysis method was proposed for the displacement analysis of laterally loaded piles. Key consideration was the continos soil profiling capability of CPT and cone resistance profiles that do not reqire artificial assmption or simplification for inpt parameter selection. The focs is on the application into offshore mono-piles embedded in clays. The correlations of p-y fnction components to the effective cone resistance were proposed, which can flly tilize CPT measrements. A case example was selected from the literatre and sed to validate the proposed method. Varios parametric stdies were performed to examine the effectiveness of the proposed method and investigate the effect of property profile and its depth resoltion on the p-y analysis. It was fond that the calclation cold be largely misleading if wrongly interpreted sb-layer condition or inappropriate resoltion of inpt soil profile was involved in the analyses. It was also fond that there is a significant inflence depth that dominates overall load response of pile. The soil profile and properties within this depth range affect most significantly calclated load responses, confirming that the soil profile within this depth range shold be identified in more detail. Keywords: p-y analysis; cone penetration tests; laterally loaded piles; offshore wind trbines; displacement analysis; ndrained shear strength 1. Introdction For offshore wind trbine strctres, the installation of fondation costs arond 21% of the total constrction cost (IEA 2008), which is mch higher than for inland wind trbines. Optimized fondation design and constrction are particlarly desired for offshore strctres as great amont of constrction cost can be saved. Among several, mono-piles are a common fondation type that is often adopted for offshore wind trbines for water depths shallower than arond 30 to 40 m. The design of offshore mono-piles is similar to that of inland piles, while the lateral load response and lateral load carrying capacity are key design consideration as waves and winds are predominant load components. The lateral load capacity of piles can be defined in two different aspects of the allowable load capacity (H all ) and the ltimate load capacity (H ). H refers to failre of either soil or pile itself while H all is associated with the level of displacement that is tolerable for a given strctre (Broms Corresponding athor, Professor, jnlee@yonsei.ac.kr Copyright 2014 Techno-Press, Ltd. ISSN: X (Print), (Online)

2 Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee 1964, Dncan et al. 1994, Zhang et al. 2005). For the displacement analysis of laterally loaded piles for H all, the beam-on-elastic fondation (BEF) approach based on the p-y analysis is often sed (Matlock 1970, Reese et al. 1975, Dncan et al. 1994). The p-y analysis may be less rigoros than the continm-based fll nmerical analysis as it is based on simplified soil springs and assmed load response crve. The p-y method has been however widely adopted in practice mainly de to the simplicity and reasonability of calclated reslts. As for other fondation design and analysis, the adeqacy of inpt parameters is crcially important for the sccessfl implementation of the p-y analysis. When offshore environment is involved, the soil characterization sing the conventional sampling- and testing-based approach is sbjected to varios experimental ncertainties with limited reliability of estimated soil parameters. For this reason, in-sit testing methods are preferred and regarded more effective for offshore cases (Titi et al. 2000, Tmay and Krp 2001, Lee and Randolph 2011). There have been several cases for the applications of in-sit test reslts into the p-y analysis (Briad et al. 1985, Robertson et al. 1989, Gabr et al. 1994, Haldar and Bab 2009). Briad et al. (1985) and Robertson et al. (1989) proposed the p-y analysis methods based on the pressre metertest (PMT) and dilatometer test (DMT) reslts, respectively. Gabr et al. (1994) also proposed a DMT-based p-y method sing the hyperbolically defined load response crve. PMT and DMT were freqently adopted becase the lateral loading mechanisms of the tests were similar to those of laterally loaded piles. Less attention has been given to the cone penetration test (CPT) mainly de to the different loading direction of the vertical cone penetration from the lateral pile loading process. However, it has been well recognized that the cone resistance is essentially governed by the horizontal effective stress rather than the vertical effective stress (Schnaid and Holsby 1991). Moreover, it was recently fond that the cone resistance is closely related to lateral pile load response as both are governed by the horizontal effective stress (Lee et al. 2010). In the present stdy, a penetrometer-based p-y analysis method sing CPT reslts is proposed for offshore piles embedded in clays. CPT is adopted to take advantage of the cost effectiveness and poplarity in offshore soil investigation. The continos profiling capability of CPT is an important consideration for the proposed method. The proposed CPT-based p-y analysis can take into accont the depth variation of soil and layer profiles in detail by directly sing CPT reslts. The effects of soil layering and property variation on the p-y analysis are examined in comparison to the conventional p-y analysis approach. 2. p-y analysis methods for laterally loaded piles in clays The beam-on-elastic fondation (BEF) approach, often called as Winkler fondation, is a common approach in practice for the displacement analysis of laterally loaded piles. In this approach, as shown in Fig. 1, soils are assmed as a series of elastic springs with load responses defined by the p-y crves that are either linear or non-linear. Varios p-y analysis methods have been proposed (Matlock 1970, Reese et al. 1975, Mrchison and O Neill 1984, Franke and Rollins 2013). The difference of the methods mainly come from the shape of p-y crve and definition of ltimate lateral resistance p. For offshorepiles embedded in clays, the method proposed by Matlock (1970) is often sed and was adopted in varios specifications (DNV 2004, API 2000). Matlock (1970) proposed a p-y crve shown in Fig. 2(a) applicable in soft clays. The fnction for the p-y crve of Matlock (1970) in Fig. 2(a) is given in a normalized form as follows

3 CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays Fig. 1 p-y analysis model for laterally loaded piles p 0.5p y y. p 0.055p y A y. A y E p y 1/3 1/ E kz A s y 50 6A s y 50 18A s y 50 (a) Fig. 2 p-y crve models for different clay conditions of: (a) soft clay (Matlock 1970); and (b) stiff clay (Reese et al. 1975). (b) p p 0.5 y y 50 1/ 3 (1) where p = lateral load per nit length; p = ltimate lateral soil resistance; y = indced lateral displacement; y 50 = limit lateral displacement = D; 50 = limit strain corresponding to 50% of failre stress in triaxial tests; and D = pile diameter. Beyond the lateral displacement of 8y 50, the vale of p was set as a constant. In Eq. (1), p represents the ltimate resistance exerted by srronding soils given as the following relationship p N s D (2) c where s = ndrained shear strength; N c = bearing capacity factor; and D = pile diameter. N c varies

4 Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee from 3 to 9 depending on depth range given as follows N c z Jz 3 r s D ( z z ) (3) N c ( z z ) (4) 9 r where = effective nit weight of soil; z = depth from grond srface; z r = limit depth below grond srface; and J = empirical parameter that can be taken as 0.5 and 0.25 for soft and stiff clays, respectively. Eqs. (3)-(4) indicate that the vale of N c increases down to the limit depth z r and then becomes constant eqal to 9. According to Matlock (1970), the limit depth z r can be estimated sing the following relationship z r 6D D J s For the cases in stiff over consolidated clays, Reese et al. (1975) proposed a p-y crve shown in Fig. 2(b) considering the stress softening behavior. The p-y crve of Reese et al. (1975) in Fig. 2(b) consists of 5 piecewise sections with fnctions given, respectively, as follows (5) p ( k z) y (6) ) p 0.5 p ( y / y (7) ( y A y )/( A ) 0.055p s 50 s 50) p 0.5 p ( y / y y (8) s 50 s 50 p.5p (6A ) 0.411p (0.0625/ y ) p ( y 6A y ) (9) p 0.5 p (6A ) 0.411p 0. 75p A (10) s 0.5 s where k = sbgrade reaction modls; A s = depth correction factor; and y 50 = limit displacement = 50 D. The depth correction factor A s varies from 0.2 at the srface to 0.6 for depths greater than 3 times pile diameter (3D). p by Reese et al. (1975) is taken as the smaller one among the followings p 2sD Dz 2. 83s z (11) p 11s D (12) where s = average ndrained shear strength over the depth z; and = effective nit weight of clays. For the displacement analysis sing the p-y method in clays, as reviewed herein, key components are the magnitde of p and the fnction that defines the p-y crve, both of which are controlled by the effective stress and ndrained shear strength. Althogh no particlarly specific consideration has been addressed for the p-y analysis for offshore piles, the proper identification of

5 CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays these parameters and their depth profiles is crcial, as it is often sbjected to varios ncertainties nder offshore environment with larger variability than for inland cases. 3. Lateral displacement analysis sing CPT reslts 3.1 Modified p-y fnction based on effective cone resistance The soil reactions from the soil springs for the BEF approach vary with depth as a reslt of increasing stiffness and strength properties of the p-y crves. The ndrained shear strength s is the governing soil variable for clays, which characterizes the main featres of the p-y crve and the ltimate lateral soil resistance p. Varios experimental methods have been proposed to estimate s (Ladd et al. 1977, Teh and Holsby 1991, Stewart and Randolph 1994, Lnne et al. 2005). The typical examples of sampling-free, in-sit testing methods poplar in practice inclde the field vane test (FVT) and cone penetration test (CPT). The cone penetration test can be particlarly effective for offshore soils as the test is condcted by a single penetration process providing continos and detailed depth profiles of seabed soil conditions. The cone factor method is a common approach to estimate s sing the CPT cone resistance given as the following relationship qt v0 s (13) N where s = ndrained shear strength; q t = cone resistance; v0 = overbrden total stress at cone tip level; and N k = cone factor. As the cone factor method of Eq. (13) involves additional nknown parameter v0, additional experimental procedre is reqired, which redces the effectiveness of CPT application. As a recent development, the effective cone factor method was proposed, where s is given as a sole fnction of the cone resistance (Lee et al. 2010). The effective cone factor method is given by qt 0 qe s (14) N N e where q e = effective cone resistance = q t 0 ; 0 = hydrostatic pore pressre; and N e = effective cone factor 16. It is noted that no additional testing procedre is reqired for Eq. (14) as 0 can be directly obtained from CPT reslts. Introdcing the effective cone factor into the ltimate lateral soil resistance p of Eq. (2), the following CPT-based p eqation can be obtained p N k e c D qe (15) Ne N c and D in Eq. (15) are the bearing capacity factor and pile diameter, respectively. As N c given by Eq. (3) varies with depth, the vale of N c /N e also varies with depth. Below the limit depth z r, however, N c becomes constant as specified by Eq. (4). For depths greater than z r, the vales of N c and N e can be taken eqal to 9 and 16, which prodces the vale of N c /N e eqal to Using the p correlation of Eq. (15), the p-y fnction by Matlock (1970) can be modified in terms of the effective cone resistance q e as follows

6 Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee Fitting c ε CPT = 0.02 Proposed 거듭제곱 Fitting crve (Proposed) CPT (a) ε_cpt0.1848q_e/p_a ^ (b) q e /p A Fig. 3 Correlation of limit strain vales for p-y method: (a) ε 50 -s correlation; and (b) ε CPT -s correlation N p 0.5 N c e q e D y y 50 1/ 3 (16) Note that, as Eq. (16) does not involve the overbrden stress v0, the continos cone resistance profiles from CPT can be flly tilized into the analysis. According to Matlock (1970), the limit displacement y 50 can be evalated by sing the limit strain vale 50 based on the stress-strain crve of the clay at the site. As the original method specifies the discretized vales of 50 for a given range of s as indicated in Fig. 3(a), a fitting crve was obtained and inclded in Fig. 3(a) to describe the 50 -s correlation. Following the effective cone resistance of Eq. (14), the 50 -s correlation in Fig. 3(a) can be frther modified in terms of the effective cone resistance as plotted in Fig. 3(b). CPT in Fig. 3(b) can be directly estimated from CPT cone resistances. Note that all correlations given in Fig. 3 were fit to be compatible to the existing reslts adopted in crrent practice. Skempton (1951) and Matlock (1970) sggested the vales of 50 in the range from 0.02 to for most clayey soils. The maximm vale of 50 and CPT eqal to 0.02 given in Fig. 3 was also set to maintain the consistency with the vale specified in the original method. The CPT -q e correlation given in Fig. 3(b) is qe CPT (17) p A where q e = effective cone resistance; p A = reference stress = 100 kpa. Using Eqs. (16)-(17), the modified CPT-based p-y fnction is then obtained as follows N p N c e q e yd 2 CPT 1/ 3 (18)

7 CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays The vale of N c /N e varies down to the depth of z r, below which it is eqal to As the proposed p-y fnction of Eq. (18) tilizes the continos CPT profile, any changes in depth profile of soil characteristics can be readily considered into the analysis withot a need of assmed simplification or idealization of in-sit soil profiles. 3.2 Load transfer analysis and calclation The load-transfer analysis is the main calclation step in the BEF approach sing the p-y crve for the displacement analysis of laterally loaded piles. It simlates the load responses of assmedpile segments and soil springs pon lateral loading. The schematic illstration of load-transfer mechanism is shown in Fig. 4. A series of pile segments with the length of h are connected at nodes where the shear stress (V), axial stress (Q) and bending moment (M) are transferred. At each node, the bending stiffness is introdced in terms of the elastic modls (E) and the second moment of inertia (I). The governing differential eqation for the eqilibrim condition of a pile segment shown in Fig. 4(b) is given by 4 2 d y d y EI Q p W 0 (19) 4 2 dz dz where EI = flexral rigidity of pile; Q = axial load; p = soil reaction per nit length; and W = distribted load along pile. Applying the finite difference scheme, the governing differential (a) (b) Fig. 4 Configration of laterally loaded piles: (a) lateral deflection of pile segments and (b) forces acting on elementary pile segment

8 Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee eqation of Eq. (19) can be rewritten as the following discretized formlation a y i b y c y d y e y i2 i i1 i i i i1 i i2 f i (20) where a, b, c, d, e, and f are model coefficients. The sbscript i represents the node nmber of discretized pile segments. Each model coefficient in Eq. (20) is given by a (21) i R i1 b i 2R R Qh i 1 2 i1 2 (22) c R i i 1 Ri Ri Qh E h 2 pyi 4 (23) d i Ri 2Ri 1 2 Qh 2 (24) e (25) i R i1 4 f i Wh (26) where R i = flexral rigidity at node i = (EI) i and E pyi = secant modls of soil springs at node i from the proposed p-y crve of Eq. (18). If lateral load imposed on pile head is the only external load, Q and W are set eqal to zero. The model coefficients of Eqs. (21)-(26) can be determined from the system eqations established for the assigned nodes between pile segments and two additional imaginary nodes assigned at the top and bottom ends of pile. If the pile is divided into n segments, n + 5 nodes are generated and ths n + 5 eqations are established. If bondary conditions are given at the top and bottom of pile, the set of algebraic eqations for the assigned nodes can be solved. At the bottom of pile, the bondary condition is set as zero-bending moment and zero-shear given as follows y y y 0 (27) R 2h Q 2h 0 ( y 3 2 2y1 2y1 y2) ( y1 y1) V0 (28) where, y -2, y -1, y 0, y 1, y 2 = lateral deflections defined at the bottom segment and imaginary nodes of pile shown in Fig. 5(a); R 0 = flexral rigidity at the bottom node of pile = (EI) 0, h = segment length, Q = axial load, and V 0 = shear force at the bottom node of pile. Note that the assmption of zero-bending moment and zero-shear is applicable for long flexible piles and short rigid piles that show rotation behavior pon lateral load as assmed in most cases. However, if a pile is short rigid and yet fixed at base, the bending moment wold be generated at pile base and the assmption of zero-bending moment at pile base wold be not be valid (Reese et al. 1970). For the top bondary condition for free head illstrated in Fig. 5(b), the following eqations for moment and force eqilibrims are imposed

9 CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays (a) (b) Fig. 5 Node configration for bondary conditions: (a) bottom bondary of pile; and (b) top bondary of pile R 2h R h t ( yt 2yt yt ) M t (29) Q 2h t ( yt 2yt 2yt yt ) ( yt yt ) P t (30) where, M t and P t = bending moment and shear force at the top node of pile, y t-2, y t-1, y t, y t+1, y t+2 = lateral deflections at the top segment and imaginary nodes of pile shown in Fig. 5(b), R t = flexral rigidity at the top node of pile = (EI) t. In a matrix form, the governing eqation of Eq. (20) can be given as follows A y f (31) where [A] = stiffness matrix; (y) = lateral displacement vector matrix; and (f) = load vector matrix. The stiffness matrix [A] contains the system eqations of the model parameters and bondary conditions given by Eqs. (21)-(30). The lateral displacement vector of pile can be obtained taking [A] -1 for each side. 3.3 Calclation algorithm The load transfer mechanism and the proposed CPT-based p-y analysis method described previosly were programmed sing the commercial programing software MATLAB. Fig. 6 shows the schematic flow and steps of the developed program for the calclation algorithm sing the

10 Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee Fig. 6 Calclation flow and algorithm proposed method. The developed program is composed of three main parts. The first part incldes the inpt process of reqired soil and pile parameters, division of pile segments, node assignment, and constrction of the p-y crve. As the cone resistance and its depth profile are sed as inpt soil parameters, CPT reslts obtained from field are directly introdced into the program. CPT and p are then calclated to constrct the p-y crves at each node, which is incorporated into the main calclation step of the load-transfer analysis. The second part consists of assembling the stiffness matrix for the system eqations and solving the matrix eqation by iteration. The five-diagonal banded matrix is created with the coefficients given by Eqs. (21)-(26) at nodes. The set of algebraic eqations are then solved for lateral pile deflections sing the inverse matrix and imposed bondary conditions at the top and bottom of pile. The third, final step is to calclate the bending moments, shear forces, and soil reactions along pile from calclated lateral deflections. 4. Comparison and validation To check the validity of the proposed CPT-based p-y analysis method, a case example was

11 CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays (a) (b) (c) Fig. 7 Soil parameter at test site: (a) soil profile; (b) ndrained shear strength profile; and (c) CPT profile selected from the literatre and sed to compare measred and calclated lateral load responses. The selected example is a field lateral pile load test by Rollins et al. (1998), condcted at Salt Lake City, Utah in USA. Varios in-sit and laboratory tests were condcted to characterize the soils at the test site, inclding the standard penetration test (SPT), cone penetrometer test (CPT), dilatometer test (DMT), pressremeter test (PMT), field vane test (FVT), and other fndamental laboratory tests. The soil profile at the test site showed a composite soil layering condition with clay, silt, and some sand layers. Fig. 7(a) shows the depth profiles of soil type at the test site. There was a gravel fill layer near srface, which was excavated before the test pile installation. Low-plasticity silts and clays existed within the depths of 1.7 m to 4.5 m. The depth profiles of ndrained shear strength s and CPT cone resistance are given in Figs. 7(b)-(c), respectively. It is seen that the vales of s were typically in the range between 25 and 60 kpa while some exceptionally higher vales arond 100 kpa were observed near srface de to the over consolidated stress condition. The cone resistances were lower than 3 MPa for the clay layer down to the depth of 4 m and higher cone resistances were observed for the sand layer between the depths of 4.5 to 6.2 m. The test pile was made of close-ended steel pipe with an inside diameter eqal to mm. The embedded pile depth was 7.9 m with the vertical load eccentricity above the grond of 0.4 m. The p-y analysis was performed sing both reslts of s and q t given in Figs. 7(b)-(c), respectively. For the CPT-based p-y analysis, the proposed method was adopted while the original Matlock s method was sed for the s -based p-y analysis with the well-known commercial program LPILE (Reese et al. 2004). For the inpt s profiles for LPILE, two differently assmed profiles were prepared and adopted into the p-y analysis. One is a simplified profile sing a single

12 Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee Fig. 8 Undrained shear strength profiles Fig. 9 Measred and predicted pile head deflections average vale of s for a given layer and the other is a detailed profile reflecting most significant variation of s within the layers. These two assmed profiles are shown in Fig. 8. For the proposed method, the CPT profile given in Fig. 7(c) was adopted. For the calclation within the sand layers, the API method (API 2000) was sed. Fig. 9 shows measred and calclated lateral load responses. It is seen that the calclated reslts sing the proposed CPT-based method and detailed s -profile are both in similarly close agreement with the measred crve. The simplified s -profile on the other hand prodced nderestimated lateral load response. While the reslts in Fig. 9 validates the proposed CPT-based method, it is also indicated that the resoltion of depth profile for the inpt parameters plays an important role in the p-y analysis and affect significantly calclated load responses. 5. Effect of property profile and soil layer condition 5.1 Effect of profile resoltion The main advantage of the proposed CPT-based p-y analysis is that detailed soil profiles, often with complexity, can be readily considered in the analysis withot additional sampling and testing procedre. It is distingished from the conventional way that tilizes individal property vales with a certain depth interval. The effectiveness of the proposed method wold differ depending on the soil profile and layering conditions. In order to investigate the effect of property profile and its depth resoltion on the p-y analysis, a series of clay soil deposits were assmed and parametric stdy was performed. The assmed soil profiles were defined sing the following s profile relationship

13 CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays s z (32) where s = ndrained shear strength; = strength increase factor; and z = depth. While the range of vales for NC clays is typically between 1 and 2, the vales eqal to 0.5, 1, 2, and 3 were considered in the analyses to cover fll range of possible clay conditions. The assmed soil conditions are shown in Fig. 10(a). The assmed s profiles shown in Fig. 10(a) were divided into several sb-layers and average s vales were assigned for each sb-layer as inpt vales for the p-y analysis. Fig. 10(b) shows examples of the considered sb-layer conditions adopted in the p-y analysis. Note that, for all the cases shown in Fig. 10(b), the vales of average s along the pile embedded depth are the same. In order to consider the effect of pile rigidity, different EI valesfor the pile, eqal to , and kn m, were considered. The total pile length was 35 m with the pile embedded depth of 30 m. Fig. 11 shows lateral deflections (y) at pile head [Fig. 11(a)] and the maximm bending moments (M max ) [Fig. 11(b)] with the nmber of sb-layer for different s profile conditions. The lateral load and EI of the pile in Fig. 11 was 100 kn and kn m, respectively. In Fig. 11, the vales of y and M max were normalized with those obtained for the original continos s profile (i.e., y 0 and M max,0 ). The single sb-layer case prodced the vales of y and M max nderestimated by arond 40% and 72% compared to those for the continos s profile. As the nmber of sb-layer increases, the difference became smaller. For the cases with more than 5 sb-layers, calclated reslts showed the degrees of match higher than 90% to y 0 and M max,0. Similar reslt were obtained for the other pile cases with EI = and kn m. These confirm that the calclated reslts of the Fig. 10 Soil profiles for parameter stdy: (a) assmed s -profiles; and (b) assmed sb-layer conditions

14 Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee y h / y Nmber of sb-layers α=0.5 α=1 α=2 α=3 (a) (b) Fig. 11 Normalized deflections and bending moments: (a) normalized head deflections; and (b) normalized maximm bending moments M max / M max, Nmber of sb-layers α=0.5 α=1 α=2 α=3 z c s,1 L EI = kn m 2 s,2 D = 1 m Fig. 12 Assmed soil sb-layer and pile conditions for parameter stdy p-y analysis are significantly affected by assmed sb-layer condition and resoltion of inpt soil profile, jstifying the effectiveness of the proposed CPT-based method. 5.2 Significant inflence layer depth The depths or thicknesses of sb-layers affect the lateral pile load responses as the p-y behavior changes with depth. If there is a depth range within which the soil condition dominates the overall lateral load response of the pile, inpt soil variables within this significant inflence depth range shold be identified in more detail. To analyze the significant inflence depth range, additional parametric stdy was performed assming different soil layer conditions. The assmed soil conditions consisted of two sb-layers with different layer thicknesses as illstrated in Fig. 12. The ndrained shear strengths (s ) for the pper and lower sb-layers were designated as s,1 and s,2, respectively. Zc in Fig. 12 represents the thickness of the pper sb-layer. For the parametric stdy, two grops of soil conditions were considered. s,1 of the first grop is

15 CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays y h / D s,1 - s, y h / D s,1 -s, z c / D -2E z c / D (a) (b) Fig. 13 Calclated head deflection for different sb-layer thickness condition for: (a) long flexible; and (b) short rigid and piles greater than s,2 and the other grop represents the opposite condition. For each grop, different vales of s,1 and s,2 were considered. Two different pile lengths of 15 and 30 m were considered to check the reslts for both short (rigid) and long (flexible) pile conditions. Short and long piles in this stdy were classified based on the pile characteristic length (L) proposed by Broms (1964). Short piles correspond to L smaller than 2.3. The vales of pile characteristic length parameter [ = (E s /E p I p ) 1/4 ] were calclated for each given soil and pile conditions. The vales of the characteristic pile length (L) for the 15-m pile were between 1.8 and 2.2 indicating short piles. The 30-m pile was classified into a long pile as L ranged from 3.7 to 4.5. Figs. 13(a)-(b) show calclated pile head deflections (y) from the p-y analysis as a fnction of z c for the long and short pile cases, respectively. Both z c and y in Fig. 13 were normalized with the pile diameter (D) of 1 m. The reslts obtained for different s conditions were all inclded in the figre. As shown in Fig. 13(a) for the long pile case, no significant changes in the calclated pile deflections are observed for z c greater than 10D. This means that inpt soil profile down to the depth of 10D dominates calclated lateral load responses and the effect of soil profile below 10D is minor. However, for the short pile case with L = 15D in Fig. 13(b), no converged vales of z c are observed, indicating the soil profile throghot the entire pile embedded depth contribtes the overall load response. This represents that the soil profile identification becomes more important for short pile cases and the proposed method with detailed soil profiling wold be more beneficial. 6. Conclsions In this stdy, a CPT-based p-y analysis method was proposed for the displacement analysis of laterally loaded piles. Key consideration was the continos soil profiling capability of CPT and cone resistance profiles that do not reqire artificial assmption or simplification for inpt parameter selection. The proposed method is focsed on the application into mono-piles embedded in offshore clayey soils where the soil characterization is limited. The correlations of p-y fnction

16 Garam Kim, Donggy Park, Doohyn Kyng and Jnhwan Lee components to the effective cone resistance were proposed, which can flly tilize CPT measrements. The proposed CPT-based p-y model and the load transfer calclation procedre were programmed and sed to obtain load responses of laterally loaded piles. A case example was selected from the literatre and sed to compare measred and calclated reslts. Close match was observed from the reslts measred and sing the proposed method. It was indicated that the resoltion of depth profile for inpt parameters in the p-y analysis is important and affect significantly calclated load responses. Varios parametric stdies were performed to examine the effectiveness of the proposed CPT-based method and investigate the effect of property profile and its depth resoltion on the p-y analysis. A series of assmed clay soil deposits were prepared and introdced into the calclation and comparison of load responses of laterally loaded piles. It was fond that the calclation cold be largely misleading if wrongly interpreted sb-layer condition or inappropriate resoltion of inpt soil profile was involved in the analysis. From the analysis of the significant inflence depth range, it was fond that there is a depth range where the soil profile and properties affect the calclated load responses, conseqently dominating the overall load response. It was also fond that the soil profile identification becomes more important for short piles indicating that the proposed method with detailed soil profiling wold be more beneficial. Acknowledgments This research was spported by Basic Science Research Program throgh the National Research Fondation of Korea(NRF) fnded by the Ministry of Science, ICT and Ftre Planning(No ) and the Ministry of Edcation(No.2013R1A1A ). References API (2000), Recommended Practice for Planning, Designing and Constrction Fixed Offshore Platforms - Working Stress Design, American Petrolem Institte, API Recommended Practice 2A-WSD (RP2A- WSD), (21st Edition), Dallas, TX, USA. Broms, B.B. (1964), Lateral resistance of piles in cohesive soils, J. Soil Mech. Fond. Div. ASCE, 90(2), Briad, J.L., Tcker, L., Lytton, R.L. and Coyle, H.M. (1985), Behavior of piles and pile grops in cohesionless soils, Final Report, Report No. FHWA/RD-83/038, NTIS PB /AS. DNV (2004), Det Norske Veritas, Design of Offshore Wind Trbine Strctres, Offshore Standard, Norway. Dncan, J., Evans, L. Jr. and Ooi, P. (1994), Lateral load analysis of single piles and drilled shafts, J. Geotech. Eng., ASCE, 120(6), Franke, K. and Rollins, K. (2013), Simplified hybrid p-y spring model for liqefied soils, J. Geotech. Geoenviron. Eng., ASCE, 139(4), Gabr, M., Lnne, T. and Powell, J. (1994), P-y analysis of laterally loaded piles in clay sing DMT, J. Geotech. Eng., ASCE, 120(5), Haldar, S. and Bab, G.S. (2009), Design of laterally loaded piles in clays based on cone penetration test data: A reliability-based approach, Géotechniqe, 59(7), IEA (2008), World Energy Otlook 2008, International Energy Agency, Paris, France. Ladd, C.C. (1977), Stress-deformation and strength characteristics, state of the art report, Proceedings of the 9th ISFMFE, Vene, Month, Vol. 4, pp

17 CPT-based lateral displacement analysis sing p-y method for offshore mono-piles in clays Lee, J., Kim, M. and Kyng, D. (2010), Estimation of lateral load capacity of rigid short piles in sands sing CPT reslts, J. Geotech. Geoenviron. Eng., ASCE, 136(1), Lee, J. and Randolph, M. (2011), Penetrometer-based assessment of spdcan penetration resistance, J. Geotech. Geoenviron. Eng., ASCE, 137(6), Lee, J., Seo K., Kang, B., Cho, S. and Kim, C. (2010), Application of effective cone factor for strength characterization of satrated clays, Proceedings of GeoFlorida 2010 (Advances in analysis, modeling and design), West Palm Beach, FL, USA, Febrary. Lnne, T., Randolph, M.F., Chng, S.F., Anderson, K.H. and Sjrsen, M.A. (2005), Comparison of cone T-bar factors in two onshore and one offshore clay sediments, Proceedings of the 1st Intnernational Symposim on Frontiers in Offshore Geotechnics, Perth, WA, USA, September, pp Matlock, H. (1970), Correlation for design of laterally loaded piles in soft clay, The 2nd Annal Offshore Technology Conference, Hoston, TX, USA, April, pp Mrchison, J.M. and O Neill, M.W. (1984), Evalation of p-y relationships in cohesionless soils, Anal. Des. Pile Fond., ASCE, Reese, L.C., Cox, W.R. and Koop, F.D. (1975), Field testing and analysis of laterally loaded piles in stiff clay, Proceedings of Offshore Technology Conference, Hoston, TX, USA, May, pp Reese, L.C., Wang, S.T., Isenhower, W.M., Arrellaga, J.A. and Hendrix, J. (2004), Compter Program: LPILE Version 5 Technical Manal, Ensoft Inc., Astin, TX, USA. Robertson, P.K., Davis, M.P. and Campanella, R.G. (1989), Design oflaterally loaded driven piles sing the flat dilatometer, Geotech. Test. J., 12(1), Rollins, K., Peterson, K. and Weaver, T. (1998), Lateral load behavior of fll-scale pile grop in clay, J. Geotech. Geoenviron. Eng., ASCE, 124(6), Rollins, K.M., Johnson, S.R., Petersen, K.T. and Weaver, T.J. (2003), Static and dynamic lateral load behavior of pile grops based on fll-scale testing, Proceedings of the 13th International Offshore and Polar Engineering Conference, Honoll, Hawaii, USA, May. Schnaid, F. and Holsby, G.T. (1991), An assessment of chamber size effects in the calibration of in sit tests in sand, Géotechniqe, 41(3), Skempton, A.W. (1951), The bearing capacity of clays, Proceedings, Bilding Research Congress, Division 1, London, England, Month, pp Stewart, D. and Randolph, M. (1994), T bar penetration testing in soft clay, J. Geotech. Eng., ASCE, 120(12), Teh, C.I. and Holsby, G.T. (1991), An analytical stdy of the cone penetration test in clay, Géotechniqe, 41(1), Titi, H.H., Mohammad, L.N. and Tmay, M.T. (2000), Miniatre cone penetration tests in soft and stiff clays, Geotech. Test. J., ASTM, 23(4), Tmay, M.T. and Krp, P.U. (2001), Development of a continos intrsion miniatre cone penetration test system for sbsrface explorations, Soil. Fond., 41(6), Zhang, L., Silva, F. and Grismala, R. (2005), Ultimate lateral resistance to piles in cohesionless soils, J. Geotech. Geoenviron. Eng., ASCE, 131(1), GC