The problem on Gravity of underground structure

[shuaicai]

I created a underground structure model in plane strain, and the soil thickness (normal) is 8 m. The quadUP element is used to soil. In this elements, what's value of $b2, gravity acceleration components in vertical directions? Is it $g*$rho_soil, or $g*$rho_soil*$thickness? and displacement-Based Beam-Column Elements were used for underground structure, how to consider the gravity of underground structure? I created elemet mass, and also applied a element force in vertical directions, but the value is ? $g*$rho_underground structures? or $g*$rho_underground structures*$thickness?


the other question: Can I make the element of soil and underground structure use a comman node? They are all three degrees of freedom, if it is done , then the three degree of freedom on behalf of which degrees of freedom, pore pressure or bending moment ?

[skamalzare]

Hi Shuaicai,

I think the proper way to do it, is to input rho to you material, and gravity (g=9.807) to the element. Therefore, in case of quadUP elements b2 = -9.807.
For the dispBeamColumn, you need to only input mass per unit length (=rho*area/UnitLength).

I am not quite sure about sharing a node with two different definition for 3rd DOF. I am guessing that the node will have one output for 3rd DOF, and it will probably mess up your model. Maybe, Frank can help you better.

Bests,
Soheil

[shuaicai]

skamalzare wrote:
> Hi Shuaicai,
>
> I think the proper way to do it, is to input rho to you material, and
> gravity (g=9.807) to the element. Therefore, in case of quadUP elements b2
> = -9.807.
> For the dispBeamColumn, you need to only input mass per unit length
> (=rho*area/UnitLength).
>
> I am not quite sure about sharing a node with two different definition for
> 3rd DOF. I am guessing that the node will have one output for 3rd DOF, and
> it will probably mess up your model. Maybe, Frank can help you better.
>
> Bests,
> Soheil




Hi Soheil, Thank you very much for your kind answers. In case of element for dispBeamColumn, <-mass SmassDens > from which a lumped mass matrix is formed (optianal, default=0.0) in <OpenSees Command Language Manual> can be used to calculate formation? or the gravity induced ? but, both all?.
In my model, In addition to setting density on the element, I also apply a eleLoad on the disBeamColumn with reference to the ground structure gravity method, as follows:
#Load Case = DEAD
set gravY_concrete 2.5
pattern Plain 1 Linear {
eleLoad -ele 3001 -type -beamUnifrom 0 $gravY_concrete
……
}
Is it ok?

[skamalzare]

Shuaicai,

I think, only adding the -mass $massDens will do the job for you. Keep in mind that, FE model tries to solve the equation of motion: Mass * acceleration + C_Damping * velocity + K_stiffness * displacement = F_external
In solution system, the mass matrix will be multiplied to internal accelerations and find a form of force. Therefore, you don't need to add extra set of gravitational accelerations to the problem.

When we add gravitational accelerations to quadUP elements, we basically want to identify that these elements are/are not in a slope. Simply because these element don't have a direction!

The load pattern that you have defined is equivalent to putting 2.5 kN load in upward direction to the model! You don't need it.

Bests,
Soheil

[shuaicai]

skamalzare wrote:
> Shuaicai,
>
> I think, only adding the -mass $massDens will do the job for you. Keep in
> mind that, FE model tries to solve the equation of motion: Mass *
> acceleration + C_Damping * velocity + K_stiffness * displacement =
> F_external
> In solution system, the mass matrix will be multiplied to internal
> accelerations and find a form of force. Therefore, you don't need to add
> extra set of gravitational accelerations to the problem.
>
> When we add gravitational accelerations to quadUP elements, we basically
> want to identify that these elements are/are not in a slope. Simply because
> these element don't have a direction!
>
> The load pattern that you have defined is equivalent to putting 2.5 kN load
> in upward direction to the model! You don't need it.
>
> Bests,
> Soheil


Hi Soheil, this is a numerical example related to the gravity. Link below.

http://opensees.berkeley.edu/WebSVN/com ... e[]=/@2344

And this are Some key code, as follows:
#
……
#_____________________tag_____8 nodes____matID bf1___bf2___bf3 poro alpha rho_s rho_f kx ky kz s_bulk f_bulk pressure
element Brick8N_u_p_U 1 5 6 7 8 1 2 3 4 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 2 9 10 11 12 5 6 7 8 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 3 13 14 15 16 9 10 11 12 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 4 17 18 19 20 13 14 15 16 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 5 21 22 23 24 17 18 19 20 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 6 25 26 27 28 21 22 23 24 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 7 29 30 31 32 25 26 27 28 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 8 33 34 35 36 29 30 31 32 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 9 37 38 39 40 33 34 35 36 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0
element Brick8N_u_p_U 10 41 42 43 44 37 38 39 40 1 0.0 0.0 $Gr $poro $alpha $rho_s $rho_f $kx $ky $kz $bulk_s $bulk_f 0.0

pattern Plain 1 "Constant" {
eleLoad -ele 1 -type -BrickW
eleLoad -ele 2 -type -BrickW
eleLoad -ele 3 -type -BrickW
eleLoad -ele 4 -type -BrickW
eleLoad -ele 5 -type -BrickW
eleLoad -ele 6 -type -BrickW
eleLoad -ele 7 -type -BrickW
eleLoad -ele 8 -type -BrickW
eleLoad -ele 9 -type -BrickW
eleLoad -ele 10 -type -BrickW
}
……

[fmk]

to answer the part about soil and structural elements .. they cannot share the same node as ndf requirements are different . create duplicate nodes at points you do want them to share the node and then use equalDOF command to constrain the translational degrees-of-freedom together.