Adaptive Meshing and XFEM in Different Parts - Error

[aurum]

Hello,

I am modeling a dam-reservoir interaction using XFEM enriched elements in the dam structure and acoustic domain to model water elements in the reservoir region. Both dam and reservoir domains are tied at their interface. However, the job has an error showing Adaptive Meshing can't be used with enriched elements even if they are different domains. Is it because they are tied, and there is an interaction between dam and reservoir domains. How can I solve this error?

 

[FEA way]

Why do you need adaptive remeshing in this analysis ? Is it really necessary ? Maybe the study could be performed without changing the mesh during the analysis. After all, there shouldn't be large mesh distortions, apart from cracking.

 

[aurum]

Thanks! I forgot to mention that this is a dynamic analysis (earthquake). I included the adaptive meshing because if ignored, the tie constraints are ignored as well. The following is the **warning message**

"Unless the *adaptive mesh option is used in a *static analysis: acoustic elements and their *tie constraints are ignored; large solid medium motions may impose distortions on acoustic elements on the structural/acoustic boundary; gaps between the acoustic and the solid media may form; coupling effects are treated using the original, undeformed geometry. Use the *adaptive mesh option to adjust the acoustic mesh if this behavior is of concern. "

Is there anything I am missing? I will be good if I can do away with adaptive meshing, but there is a separation( gap) formation between the dam and the reservoir domain if ignored.

 

[FEA way]

Can you also paste the whole message that you get when submitting the analysis with XFEM and adaptive remeshing ?

 

[aurum]

*Error*

The *adaptive mesh option is not supported with enriched elements

*Warnings*

For two-dimensional models, if a non-unity thickness is specified for two-dimensional solid elements and these elements are involved in an interaction such as contact, the same thickness should be specified for the out-of-plane thickness of the corresponding surface under *surface interaction.

Slave surface assembly__pickedset46_cns_ is a node-based surface used with surface to surface approach. In order to obtain improved stress accuracy with this approach, an element-based surface should be used instead.

For *tie pair (assembly__pickedset46_cns_-assembly__pickedset45_cns_), adjustment was specified but no node was adjusted more than the adjustment distance = 2.22000e-16.

For *tie pair (assembly_resevoirupstreamedge-assembly_damupstreamedge), adjusted nodes with very small adjustments were not printed. Specify *preprint,model=yes for complete printout.

Some surfaces used with the structural-acoustic *tie are two-dimensional. In order to obtain correct results, please make sure that the out-of-plane thickness of the acoustic elements is the same as the out-of-plane thickness of the solid elements.

The energy balance contributions for acoustic elements are only present in *steady state dynamics steps. For other procedures, the acoustic element contributions to the energy balance are ignored.

Unless the *adaptive mesh option is used in a *static analysis: acoustic elements and their *tie constraints are ignored; large solid medium motions may impose distortions on acoustic elements on the structural/acoustic boundary; gaps between the acoustic and the solid media may form; coupling effects are treated using the original, undeformed geometry. Use the *adaptive mesh option to adjust the acoustic mesh if this behavior is of concern.

Acoustic-structural interface is present in geometrically nonlinear analysis step. Nodal coordinates at some acoustic elements on acoustic-structural interface can thus be updated. If such acoustic elements become very distorted during analysis, the solution validity can be questionable, especially in acoustic domain.

 

[FEA way]

Maybe you could split this study into two separate analyses - calculate the variations of the pressure acting on the dam during an earthquake in one analysis involving acoustics and then import this pressure distribution to fracture mechanics analysis with XFEM and no acoustic elements. This would be a sort of one-way coupling but could let you bypass the limitations of Abaqus.