Starcross
Aerospace
- Oct 1, 2015
- 1
Hello (and HELP!),
So I am effectively trying to model an assembly and spring back behaviour of a lattice structure using the *IMPORT command. The basic jist is to perform a quasi-static analysis in Abaqus/Explicit to model to the assembly stage, this is then imported into Abaqus/Standard where by the individual parts are attached together with hinges and then the structure is "released" to determine its resting minimum energy position. Please see below for more details.
First I replicated the assembly process of a lattice structure, made of initially straight rods (modelled with b32 elements), the the rods lie tangentially on a circle with one end touching the circle, there are 8 of these even spaced around the circle, pointing in a clockwise fashion and the same again in an anticlockwise fashion (see attached image Straight_rods). The rods are then displaced (using displacement boundary condition at 5 points on each rod) to lie on the surface of a sphere with the circle at the equator and all the end points not originally lieing on the circle meeting at the meridian, This results in a spiral like pattern with points of intersection between the clockwise and anticlockwise rods (see attached image Spherical_Rods). This simulation is performed as a quasi-static analysis in Abaqus/Explicit, I have checked all the energies and the stresses and displacement and am happy that this has reach a static equilibrium.
Second, the output of the Abaqus/Explicit analysis is imported into Abaqus/Standard using *IMPORT, upon which connector elements (hinges) are added to the points of intersection, I have also imported the deformed mesh so that its easier to visualise in CAE. I specify that I want the orientation of their single rotational degree of freedom to be that around the radial direction of a spherical coordinate system defined at the centre of the sphere which the rods lie on (see attached image hinge orientation). The spring back is allowed by constraining the points of intersection on the equator of the sphere to remain in plane and only allowed to move radially, all other points previously constrained to lie on the sphere are removed.
All good so far, now it gets interesting!
When UPDATE=YES the orientation of the hinges are correct i.e. the radial direction of both attached nodes are coincident, pointing towards the centre of the sphere (see attached image Hinge_Orientation_UPDATE=YES) and they behave as expected when the structure is released but the spring back within the rods themselves is not, the initially straight rods should want to return to a configuration as straight as possible but the do not. I have deduced this down to the fact the beam normals are altered when the reference configuration is updated even though the correct stresses are mapped across.
OK no problem set UPDATE=NO (which is better as I get total strains and displacements relative to the straight rods) as well as the correct spring back behaviour, but in doing so the definition of the connector orientations are taken when the nodes are in their original position, when all the rods lie in the circle at the beginning of the assemble stage (instead of the end of the assembly stage). The problem arises when the rods and hence their nodes are deformed to the surface of the sphere, the vector defining the hinge orientation for that node also rotates so that when the come together they are no longer coincident (see attached image Hinge_orientation_Update=NO) and the hinge doesn't work and the model wont run!
So what I need to do is to be able to define the hinge orientations based on the deformed mesh i.e. when the structure is assemble in the spherical configuration and with the UPDATE=NO so that the spring back is correct. Does anyone know anyway to achieve this?
Thanks in advance
So I am effectively trying to model an assembly and spring back behaviour of a lattice structure using the *IMPORT command. The basic jist is to perform a quasi-static analysis in Abaqus/Explicit to model to the assembly stage, this is then imported into Abaqus/Standard where by the individual parts are attached together with hinges and then the structure is "released" to determine its resting minimum energy position. Please see below for more details.
First I replicated the assembly process of a lattice structure, made of initially straight rods (modelled with b32 elements), the the rods lie tangentially on a circle with one end touching the circle, there are 8 of these even spaced around the circle, pointing in a clockwise fashion and the same again in an anticlockwise fashion (see attached image Straight_rods). The rods are then displaced (using displacement boundary condition at 5 points on each rod) to lie on the surface of a sphere with the circle at the equator and all the end points not originally lieing on the circle meeting at the meridian, This results in a spiral like pattern with points of intersection between the clockwise and anticlockwise rods (see attached image Spherical_Rods). This simulation is performed as a quasi-static analysis in Abaqus/Explicit, I have checked all the energies and the stresses and displacement and am happy that this has reach a static equilibrium.
Second, the output of the Abaqus/Explicit analysis is imported into Abaqus/Standard using *IMPORT, upon which connector elements (hinges) are added to the points of intersection, I have also imported the deformed mesh so that its easier to visualise in CAE. I specify that I want the orientation of their single rotational degree of freedom to be that around the radial direction of a spherical coordinate system defined at the centre of the sphere which the rods lie on (see attached image hinge orientation). The spring back is allowed by constraining the points of intersection on the equator of the sphere to remain in plane and only allowed to move radially, all other points previously constrained to lie on the sphere are removed.
All good so far, now it gets interesting!
When UPDATE=YES the orientation of the hinges are correct i.e. the radial direction of both attached nodes are coincident, pointing towards the centre of the sphere (see attached image Hinge_Orientation_UPDATE=YES) and they behave as expected when the structure is released but the spring back within the rods themselves is not, the initially straight rods should want to return to a configuration as straight as possible but the do not. I have deduced this down to the fact the beam normals are altered when the reference configuration is updated even though the correct stresses are mapped across.
OK no problem set UPDATE=NO (which is better as I get total strains and displacements relative to the straight rods) as well as the correct spring back behaviour, but in doing so the definition of the connector orientations are taken when the nodes are in their original position, when all the rods lie in the circle at the beginning of the assemble stage (instead of the end of the assembly stage). The problem arises when the rods and hence their nodes are deformed to the surface of the sphere, the vector defining the hinge orientation for that node also rotates so that when the come together they are no longer coincident (see attached image Hinge_orientation_Update=NO) and the hinge doesn't work and the model wont run!
So what I need to do is to be able to define the hinge orientations based on the deformed mesh i.e. when the structure is assemble in the spherical configuration and with the UPDATE=NO so that the spring back is correct. Does anyone know anyway to achieve this?
Thanks in advance