RE; additional information on SSPbrick up element.

[ryancwl95]

Hi,

I am currently utilizing an SSPbrickUP element with a linear elastic isotropic model to simulate a soil column. The objective is to model the consolidation and pore pressure dissipation process of the soil column under a specific loading condition.

In the ElasticIsotropic model, an optional input is the mass density (rho). I have a query regarding whether there is an inherent mass damping included in the analysis when the mass is provided. The input syntax looks like this:
nDMaterial ElasticIsotropic $matTag $E $v <$rho>

Moving on to the SSPbrickUP element, there is an optional input for constant body forces in the global x-, y-, and z-directions, denoted as <$b1 $b2 $b3>, respectively. The documentation mentions that these values represent components of the gravitational vector. I would appreciate further clarification on how these relate to the scenario of modeling a soil column under gravity. The element syntax is as follows:
element SSPbrickUP $eleTag $iNode $jNode $kNode $lNode $mNode $nNode $pNode $qNode $matTag $fBulk $fDen $k1 $k2 $k3 $void $alpha <$b1 $b2 $b3>

Specifically, my question is about the interpretation of these body forces. Do they correspond directly to gravitational acceleration, e.g., $b3=-g ? Or, do they represent the unit weight of the soil, such as $b3 = mass density*-9.81 in the context of the soil column example I mentioned? Furthermore, should the weight of the water be considered in these body forces ($b3)?

Sincerely,

Ryan

[STKO Team]

I have a query regarding whether there is an inherent mass damping included in the analysis when the mass is provided.

No damping is automatically used by OpenSees. You have do explicitly include damping

Specifically, my question is about the interpretation of these body forces. Do they correspond directly to gravitational acceleration, e.g., $b3=-g ? Or, do they represent the unit weight of the soil, such as $b3 = mass density*-9.81 in the context of the soil column example I mentioned? Furthermore, should the weight of the water be considered in these body forces ($b3)?

Unfortunately this is sometimes confusing.
Standard (displacement-only) elements define the body forces as rho*g.
Instead, in the UP (displacement-pressure) elements, the body force actually is only the acceleration vector. I think they did it so that the element can compute the mixed rho (from the material and from the fluid) and then multiply it by the acceleration vector to compute the full body force.
Give it a try, define a rho for the material, another rho for the fluid density in the element, use b3 = -g, and check the sum of vertical reaction forces to see if they match (rho_material+rho_fluid)*total_volume*g

[ryancwl95]

No damping is automatically used by OpenSees. You have do explicitly include damping.

I am currently conducting a gravity analysis using transient calculations with numerous steps (500) and large time steps (dt=500s). In the graph depicting pore pressure at the nodal point of -10m within a 1m x 1m x 10m soil column, the pore pressure converges to a valid value of 98.1 kPa by the end of the analysis.

Screenshot 2023-11-21 173815.png (21.16 KiB) Viewed 5040 times

However, during the convergence process, there is noticeable oscillation between 102 kPa and 95 kPa. Interestingly, when I experimented with altering the density (ρ) value of the soil column, the amplitude of the oscillations increased with increased density (ρ) before ultimately converging to 98.1 kPa. No explicit damping (Rayleigh) is applied in the analysis.

I am curious about the potential reasons behind this oscillatory behavior and its eventual convergence, especially given the absence of damping inputs.

[ryancwl95]

Unfortunately this is sometimes confusing.
Standard (displacement-only) elements define the body forces as rho*g.
Instead, in the UP (displacement-pressure) elements, the body force actually is only the acceleration vector. I think they did it so that the element can compute the mixed rho (from the material and from the fluid) and then multiply it by the acceleration vector to compute the full body force.
Give it a try, define a rho for the material, another rho for the fluid density in the element, use b3 = -g, and check the sum of vertical reaction forces to see if they match (rho_material+rho_fluid)*total_volume*g

In continuation of the second point mentioned, I employed the results of the Reaction Force, including inertia, specifically focusing on the RFz component. Unfortunately, it did not produce the expected outcome. However, when examining the stresses at the bottom of the soil column, which aligned with the anticipated stresses when b3 = -g.

I would also like to confirm that in the elastic isotropic model, the input syntax appears as follows:
nDMaterial ElasticIsotropic $matTag $E $v <$rho>

Could you confirm that in this context, ρ ($rho) refers to the saturated density of the soil and not the dry density of the soil?

Additionally, concerning the SSPbrickUP element, the syntax is outlined below:
element SSPbrickUP $eleTag $iNode $jNode $kNode $lNode $mNode $nNode $pNode $qNode $matTag $fBulk $fDen $k1 $k2 $k3 $void $alpha <$b1 $b2 $b3>

In this context, does $fBulk specifically refers to the fluid bulk modulus, as opposed to the Combined Undrained Bulk Modulus $Bc (an input required in the Twenty_Eight_Node_BrickUP element)?

[STKO Team]

Could you share your file to have a look?

[ryancwl95]

Hi,

I attached my scripts, the recorders file and the SPD post processing file of this analysis. Thank you in advance for taking a look at it.

[ryancwl95]

Hi,

I was wondering if there's any follow up for the questions in this post?
Thank you in advance.

Sincerely,
Ryan Chia