Abstract
Keywords
Introduction
Problems related to the vertical steady-state vibration of floating pile in the soil layer have attracted increasing attention due to their significance in the field of civil engineering and solid mechanics.1–4 For the vibration problem of the pile–soil system, it is generally important to determine the dynamic impedance and the velocity response subjected to harmonic load at the pile head which can provide theoretical reference to the integrity detection of pile in practice.
In the dynamic system of pile–soil interaction, various mathematical models have been developed by investigators. 5 The Winkler model is extensively employed due to its simplicity in which soil layers are represented by a series of springs and dashpots.6–9 However, the Winker model is limited in describing the mechanism of wave propagation within the pile–soil system.10,11 Novak et al.12–14 presented the plane-strain model for the pile–soil interaction system to simplify the surrounding soil as a linear viscoelastic layer with the hysteretic-type damping. Manna and Baidya 15 investigated the possible reasons for the unsatisfactory performance of the Novak model. Furthermore, various axisymmetric continuum models were also proposed by researchers, in which the three-dimensional (3D) wave effect of energy propagation can be considered.16–18
The surrounding soil of pile in the aforementioned studies of the pile–soil system is assumed as a single-phase medium. However, soil is generally a multiphase medium that can be modeled as a liquid-filled porous medium. Biot19,20 developed a theory to describe the deformation and wave propagation in a fluid-saturated porous elastic medium. Based on Biot’s theory for porous elastic solid, Rajapakse and Senjuntichai 21 explicitly derived rigorous analytical solutions for the vertical vibration of pile which describes the relationship between the generalized displacement and the force of multilayered porous media in the Fourier frequency space. Zhou et al. 22 investigated the dynamic response of a pile embedded in a saturated half-space subjected to transient vertical loading by adopting Biot’s porous elastodynamic equations. Li et al. 23 used Biot’s elastodynamic theory as the basis to present an analytical solution for the vertical vibration of an end-bearing pile in a saturated soil layer. Subsequently, Zheng et al. 24 presented an analytical method to study the vertical vibration of a floating pile embedded in porous elastic soil using Biot’s theory, in which the saturated elastic soil is divided into two independent parts: a half-space underlying the pile base and a series of Novak’s thin layers along the pile shaft. However, some soil parameters defined in Biot’s theory18,19 are not clear enough and difficult to determine, which leads to inconvenience for its application in engineering, to some extent.
Moreover, Bowen 25 proposed the theory of porous media (TPM) by integrating the continuum theory of mixtures with the concept of volume fractions. In contrast to Biot’s theory, the TPM is also proven to provide a comprehensive and extensive modeling framework.26,27 Substantial developments with respect to the TPM were extended to geomechanical problems and were contributed by the pioneering studies of de Boer et al.28–30 In the subject of pile–soil dynamic interaction, Liu and Yang 31 extended the governing equations of the TPM to obtain analytical solutions that describes the vertical vibration of an end-bearing pile in saturated poroelastic soil. In addition, Yang and Pan 32 and Zhang et al. 33 deduced the frequency solutions for the vertical vibration of an end-bearing pile in the saturated porous viscoelastic soil using the variable separation method.
Based on an extensive review of the literature, it can be seen that little work has been conducted on the dynamic response of a floating pile in a viscoelastic saturated soil using the TPM. The main purpose of this article is to propose a new mathematical model for both velocity response and dynamic impedance of a floating pile embedded in a viscoelastic saturated soil which can provide theoretical reference to the low-strain integrity detection of pile in engineering practice. The saturated surrounding soil is divided into Novak’s thin layers along the pile length and a homogeneous half-space underlying the pile base to consider the effect of excitation frequency on the dynamic stiffness of the soil beneath the pile base. And the corresponding analytical solutions for the dynamic impedance and the velocity response at the pile head are derived and verified by comparing its reduced solution with the existing solutions for the end-bearing pile. In addition, an extensive parametric analysis further needs to be conducted to investigate the effects of the slenderness ratio of pile, the modulus ratio of the pile to the surrounding soil, and the permeability coefficient of saturated soil on the vertical vibration characteristics of the pile–soil interaction system.
Statement of problem and basic assumption
The problem under consideration is shown in Figure 1. An elastic solid pile is embedded in a liquid-saturated porous viscoelastic half-space subjected to a vertical time-harmonic load

Conceptual model of a floating pile embedded in the saturated porous viscoelastic half-space.
Variables and symbols of the material parameters.
It is assumed that the equilibrium of shear stress and the continuity of displacement are both satisfied at the interfaces; the deformation of the soil–pile system is infinitesimal; the soil is divided into two parts along the pile length: the upper part consists of a series of independent Novak’s thin layers, while the lower part is simplified as a porous viscoelastic half-space.
Formulation of the governing equations
On the basis of the TPM proposed by Bowen, 25 the 3D dynamic governing equations for the saturated porous viscoelastic soil layer in axisymmetric cylindrical coordinates can be formulated as
where
As shown in Figure 1, the interaction system of a floating pile embedded in the saturated porous viscoelastic half-space is subjected to an axial time-harmonic load
where
Analytical solution for the dynamic impedance of a floating pile
Combining the cylindrical symmetry conditions of the interaction system and Novak’s plain-strain model, equations (1a)–(1c) can be expressed in the frequency domain by equations (3a) and (3b)
Substituting equation (3a) into equation (3b), equation (4) can be obtained
where
The general solution of equation (4) is given by
where
Taking into account the boundary conditions expressed in equations (6a) and (6b)
it is obtained that
The solution of equation (4) can be further expressed as
Therefore, the shear stresses at the pile–soil interface in the saturated soil layer can be written as
Integrating
where
Furthermore, the dynamic equilibrium equation of the floating pile can be expressed as
where
Inserting
where
Thus, the solution of equation (11) can be given by
where
The boundary conditions at the head and base of the pile can be expressed using equations (13) and (14), respectively
where
The soil reaction at the pile base
where
The undetermined function
The kernel function
where
Substituting equation (12) into equations (13) and (14), equations (19) and (20) can be obtained in the following forms, respectively
where
Therefore, the vertical dynamic impedance at the pile head can be expressed as
The dimensionless vertical dynamic impedance at the pile head can be further written as
where the real part
Furthermore, the frequency response function of velocity at the pile head can be obtained as
where
Using the convolution theorem and inverse Fourier transformation, the velocity response in the time domain at the pile head can be written as
where
For the integrity detection of pile foundation, the hammer excitation
where

A half-sine pulse excitation.
Combining equations (24) and (25), the dimensionless frequency response function of velocity at the pile head
where
Results and discussion
This section presents the numerical results to demonstrate the validity of the obtained analytical solutions and to investigate the vertical dynamic response of the floating pile in the saturated porous viscoelastic soil. Unless otherwise specified, the parameter values in Table 2 are used. For the convenience of comparison, the dynamic impedance in the following figures is normalized as
Parameter values of the surrounding soil and the solid pile.
The impedance solution expressed in equation (22) for a floating pile can be reduced to describe the vertical vibration of an end-bearing pile in the saturated viscoelastic soil by setting the complex impedance at the pile base

Comparison of the reduced vertical impedance with the solution of Yu et al. for an end-bearing pile.
Dynamic impedance at the pile head
The dynamic impedance at the pile head to time-harmonic loads is generally used to investigate the vibration characteristics of the pile–soil system. Figures 4–6 show the effects of the slenderness ratio of pile, the modulus ratio of pile to soil, and the permeability coefficient of soil on the dynamic impedance at the pile head, respectively.

Effect of the pile slenderness ratio on the vertical impedance.

Effect of the modulus ratio of pile to soil on the vertical impedance

Effect of the permeability coefficient on the vertical impedance
Figure 4 shows the dimensionless vertical dynamic impedance
Figure 5 shows the dimensionless vertical dynamic impedance
Figure 6 shows the dimensionless vertical dynamic impedance
Velocity response at the pile head
Due to the impact at the pile head, it generates the downward stress wave along the pile shaft. When the downward stress wave encounters a change in cross section or in concrete quality, the reflected wave of velocity response can be observed at the pile head. The analytical solution for the velocity response at the pile head can depict the typical patterns of the reflected signal corresponding to different pile defects, which is able to provide theoretical reference to the low-strain integrity detection of pile in engineering practice.
Figure 7 shows the dimensionless reflected wave signal of vertical velocity in the time domain at the pile head

Effect of the pile slenderness ratio on the reflected signal.
Figure 8 shows the dimensionless reflected wave signal of vertical velocity in the time domain at the pile head

Effect of the modulus ratio of pile to soil on the reflected signal

Effect of the permeability coefficient on the reflected signal
Conclusion
Based on the TPM, a new mathematical model of a floating pile embedded in a viscoelastic saturated soil under vertical time-harmonic load is proposed for the evaluation of low-strain integrity detection. The corresponding analytical solutions for the dynamic impedance and velocity response at the pile head are derived and then verified. The proposed analytical solution for velocity response at the pile head can also be reduced to investigate the vertical vibration problem of end-bearing piles in saturated soil and solid piles embedded in single-phase half-space. Moreover, it should be noted for the limitations of the obtained solution that the transverse isotropy of the surrounding soil, the vertical inhomogeneity of soil layers around the pile, and the discontinuous effect of the pile–soil interface are not considered in the proposed model.
An extensive parametric analysis is further conducted to investigate the effects of the pile slenderness ratio, the relative modulus of pile to soil, and permeability of the saturated soil on the dynamic impedance and the reflected wave signal of vertical velocity. The parametric analysis shows that (1) with the increase of the pile slenderness ratio and the decrease of the modulus of pile to soil the resonance frequencies decrease and the oscillation amplitudes of the dynamic impedance at resonance frequencies decrease; (2) the effect of the permeability coefficient on the velocity response at the pile head can be practically negligible; (3) only within the high-frequency range the resonance frequencies increase and the oscillation amplitudes of the dynamic impedance at resonance frequencies slightly decrease with the increase of the permeability coefficient.
