9.16     Transmitting Nodal Boundary Element

 

            The element is used to provide a transmitting nodal boundary for incident propagating motions, such as occur in seismic response calculation. The boundary is frequency independent, and is local in space and time (see Note at end of this chapter for more details).  It is exact for vertically propagating wave motions and linear systems only.  At the boundary:

 

 

where  time,  mass density of underlying medium,  = wave speed,  incident motion,  associated displacement motion at the node, and s = multiplier.  The incident motion  is defined by a corresponding load-time function.

 

 

NODAL_TRANSMITTING

 

 

            Element_name = NODAL_TRANSMITTING  etc...

                  < material data >

                  < output requests >

                  < connectivity data >

 

 

 


 

9.16.1     Element Group Control Information

 

            Must follow the element name (same data record), and define the following:

 

Note    Variable Name                       Type    Default    Description

 

            Number_of_material_sets    integer      [1]        Number of material sets, Numat

 

            File_name                              string    [none]     File name (optional).  Name must be

                                                                                            enclosed in quotation marks.

 

            Input_format                           list         [*]        Input format

                 keywords / list

 

 

 

 

9.16.2     Geometric / Material Properties Data (Numat sets)

 

 

Note    Variable Name                       Type    Default    Description

 

Keywords Read Method

            Material_set_number            integer      [0]        Material set number  Numat

 

            Incident_motion_definition    list         [*]        Incident motion definition

                 displacement

                 velocity

                 acceleration

 

            Incident_motion_load_time integer      [0]        Incident motion load-time function number

 

            Motion_multiplier_i                real       [0.0]       Motion multiplier for degree of freedom
                                                                                            ().

 

            Rhoc_i                                     real       [0.0]       for degree of freedom  ().

 

List Read Method

            Geometric / material data must follow in the form:

            <   >

                 <  >

            < terminate with a blank record >.

 

 


 

EXAMPLE

            Define_Element_Group  /

                  name = "Group_3"  /

                  element_type = nodal  /

                  element-shape = one_node  /

                  number_of_material_sets = 1

 

                  material_set_number = 1  /

                        incident_motion_definition = displacement  /

                        incident_motion_load_time = 1  /

                        motion_multiplier_1 = 1.00  /

                        rhoc_1  = 105.175e6

 

            NODAL_CONNECTIVITY  etc…

 

 

9.16.3     Element Nodal Connectivity Data

 

            Consult Chapter 11 for details; for this element  NEN = 1.

 

 


Note/

 

Transmitting Boundary

 

1.         Introduction

 

            The application of finite elements to the solution of problems involving the propagation of waves requires the development of special boundary conditions referred to as transmitting, non-reflecting, silent or energy-absorbing boundaries.  These boundary conditions are required to use at the boundary of the necessarily finite mesh to simulate the infinite extent of the domain.  For instance, when an infinite domain in the vertical direction is modeled by a finite mesh, there is danger that waves reflected from the free-surface will be reflected back off the artificial bottom boundary and cause errors in the response calculations, unless special boundary conditions can be imposed at the base of the column.  In the following, a rigorous formulation of an appropriate boundary condition is presented.  The proposed boundary condition is frequency independent, and is local in space and time.  It is exact for linear systems only, and therefore requires that the boundary be placed at a sufficiently large distance such that the response be linear at that distance.

 

            Seismic site response calculations are usually performed for a given seismic input prescribed in the form of an acceleration time history to be applied at the base of the soil column.  As discussed hereafter, the implementation of an appropriate boundary condition at the base of the soil column requires detailed knowledge of the nature of the prescribed seismic input, viz. whether it corresponds to an incident vertically propagating motion or is the sum of an incident and a reflected motion.

 

            The features of one-dimensional wave propagation in a semi-infinite system are first reviewed before the boundary condition is developed.

 

 

2.         One-Dimensional Vertical Wave Propagation

 

            For the purpose of illustrating the features of the boundary formulation, the vertical propagation of shear waves is considered.  The equation of motion may be expressed as:

 

                                                                                                                       (1)

 

where a comma is used to indicate partial differentiation;  mass density;  shear modulus;  horizontal displacement;  time; and  depth coordinate, with the -coordinate assumed oriented upwards positively.  The fundamental solution of Eq. 1 can be expressed as:

 

                                                                                       (2)

where

                                                                                                                             (3)

 

and I and R are two arbitrary functions of their arguments:  represents a wave motion propagating upwards in the positive -direction with the velocity C, and is referred to as the
incident motion;  presents a wave motion propagating downwards in the negative ‑direction with the velocity C, and is referred to as the reflected motion.  The following two identities apply:

                                                                                                                  (4a)

 

                                                                                                                (4b)

 

and therefore, if one differentiates Eq. 2 with respect to  and  in turn:

 

                                                                                                           (5)

 

                                                                                                                      (6)

 

The shear stress can therefore be expressed as

 

                                                                                         (7)

 

and upon elimination of  the following relation is obtained:

 

                                                                                                       (8)

 

            At this stage it is instructive to study the total wave pattern when an incident wave motion  encounters an artificial boundary at .  Three extreme cases can be considered as follows:

 

 

2.1       The boundary at  is fixed. 

 

Setting   in Eq. 2 leads to:

 

                                                                               (9)

 

resulting in the total wave motion:

 

                                                                     (10)

 

Therefore, at a fixed boundary, the incident wave is reflected back with the same shape but opposite sign.

2.2       The boundary at  is free.

 

Setting  in Eq. 7 leads to:

 

                                                                             (11)

 

resulting in the total wave motion:

 

                                                               (12)

 

Therefore, at a free boundary, the incident wave is reflected back with the same shape and the same sign.

 

 

2.3       The boundary at  is silent.

 

Selecting Eq. 4a which is identically satisfied for I as the boundary condition for

 

                                                                                       (13)

 

results in .  Eq. 13 is called the radiation condition.  It is obtained by selecting:

 

                                                                                         (14)

 

When the incident wave I encounters that boundary, it passes through it without modification and continues propagating towards .  No reflected wave R, which would propagate back in the negative -direction can arise.

 

 

3.         Semi-Infinite Column

 

            Consider the situation shown in Fig. 9.13.1.  An incident vertically propagating wave I (coming from infinity) arrives at the site, and it is sought to compute the site response for this incident motion.  The finite element mesh has been selected to extend down to the depth , and an appropriate boundary condition at the base of the soil column is sought to simulate the infinite extend of the soil domain in the vertical downward direction.  For the purpose of illustration, it is assumed that the site consists in general, of two homogeneous deposits with material properties as follows:

 

                above the base of the soil column: 

 

                below the base of the soil column: 

 

 

 

In order to separate the influence of the incident wave from the reflected wave on the site response, it is assumed that the incident motion disturbance spans over a duration  with:

 

                                                               

 

and that it reaches the location  at time .  Several cases are considered as follows:

 

 

 

Figure 9.16.1 Semi-Infinite Layered Soil Profile


 

3.1       Case 1: Homogeneous semi-infinite deposit (viz., ):

 

In that case the incident vertically propagating wave arriving at  at time , will reach the free surface  at time , will be reflected back from the free surface with the same shape and sign, and must cross the boundary at  (at times ) without any further modification and continue propagating back towards infinity.  The resulting motions are as follows:

 

     at

 

                                                       (15)

 

     at

 

                                                                 (16)

 

 

where H is the Heaviside function.  this is illustrated in Fig. 9.13.1.  The desired response in the finite soil column can be achieved by prescribing at the base of the soil column either the total motion or the incident part of the motion only, as follows:

 

>    Prescribed motion (fixed base case).  In that case the base input motion must be made up of the incident and reflected motions to reproduce the specified site response as

 

                                                       (17)

 

The first part of the input corresponding to  in Eq. 17 will propagate towards the surface and reproduce the prescribed surface motion.  It will then be reflected back off the free surface towards the fixed base where it will be reflected again with a negative amplitude:

 

                            

 

This reflected wave is canceled exactly by the second part of the input motion in Eq. 17 thereby preventing any further propagation of waves towards the surface.  In other words, the incident wave  produces the surface motion and the reflected wave cancels the reflection from the rigid base.

 

Remark:   The total motion is the one computed in standard deconvolution procedures implemented in computer programs such as SHAKE (1972).


 

>    Prescribed traction (non-reflecting case).  From Eq. 8 the stress in the semi-infinite soil deposit at location  can be expressed directly in terms of the motion at the location and the incident wave motion.  Therefore, it suffices to apply at the artificial boundary  the traction:

 

                                                                      (18)

 

In that case, the incident input motion is absorbed exactly at the base after reflection from the surface.  Eq. 18 is the most general boundary condition since it only requires knowledge of the incident motion.

 

 

3.2       Case 2: Non-homogeneous semi-infinite deposit. 

 

In that case only the incident motion is known as it arrives at location .  In order to compute the site response for this incident motion, accounting for the effects of ensuing reflections (or no reflections if  and ) at the boundary , one must prescribe the input at the base of the finite soil column in terms of prescribed tractions as:

 

                                                                 (19)

 

This will ensure proper simulation of the infinite extend of the soil domain in the downward direction.

 

 

 

References / Bibliography

 

1.      Schnabel, P.B., J. Lysmer and H.B. Seed, "SHAKE: A Computer Program for Earthquake Response Analysis of Horizontally Layered Sites", Report No. EERC 72-12, University of California, Berkeley, (1972).

 


 

 

Notes . .