IljaK
Mechanical
- Jul 18, 2019
- 4
Hi Forum,
I am attempting to make a pinned connection (with 1 rotational DOF) between 5 rigid beams in the following configuration -o-o-o-o-, where 'o' are cylindrical hinges and '-' are beams. The objective of the simulation is to lift and stretch a piece of flat rubber that goes over the top, the rubber is fixed at both ends. The rubber is stretched by lifting the center beam upwards in Step-1 (so beams look like this _/-\_ )and then in Step-2 pushing the pushing side beam axially to further expand the rubber (the other side is fixed in Step-2). As you can see, side beams are only allowed to translate axially, center beam is allowed to move upwards (no canting allowed), and middle beams are allowed to rotate.
The beams are made rigid to reduce the computational cost, which eventually brought up an issue with penetration between rigid bodies. To avoid this I decided to create a rigid pinned connection between the beams.
Since there are many ways to model such a connection, I will only list two that worked for me(to some degree) and state the issues that I faced.
(1) A reference point attached with MPC BEAM to a set of nodes on each hinge feature on the beams. A wire connects two reference points between the neighboring hinges and HINGE type connector is assigned with a rigid-like elastic behavior.
Diagram of one hinge: RP (MPC) – WIRE (HINGE) – RP (MPC).
After inspecting the animation of the analysis I found that the hinges shift away from their rotational axis towards the of the first step, making some beams penetrate into each other to a small degree. At the start of the second step, the petals on one side of the assembly rapidly come out of their sockets striking the rubber and distorting the elements, which essentially aborts the analysis. This motion is highly unrealistic and releases a high amount of kinetic energy. We attempted to damp the inertia accumulated in the petals by applying pressure to counter that motion, but no luck. Alternatively, the petals were fixed in all DOF in an intermediate step to lose their inertia, no luck too.
(2) A reference points attached with Kinematic Coupling to a set of nodes on each hinge feature on the beams. One reference point is allowed to have one rotational DOF between the neighboring hinges. An MPC type BEAM constraint connects two reference points between the neighboring hinges.
Diagram: RP (Kinematic Coupling w/ one rotational DOF) – MPC (BEAM) – RP (Kinematic Coupling, all DOFs constrained).
The issue is that the reference points that have the rotational DOF are slowly drifting away from their initial position, hence shifting the rotational axis. This is a bit odd as each hinge has its own local coordinate system and both reference points constrained from translating in any direction within that coordinate system.
I am attempting to make a pinned connection (with 1 rotational DOF) between 5 rigid beams in the following configuration -o-o-o-o-, where 'o' are cylindrical hinges and '-' are beams. The objective of the simulation is to lift and stretch a piece of flat rubber that goes over the top, the rubber is fixed at both ends. The rubber is stretched by lifting the center beam upwards in Step-1 (so beams look like this _/-\_ )and then in Step-2 pushing the pushing side beam axially to further expand the rubber (the other side is fixed in Step-2). As you can see, side beams are only allowed to translate axially, center beam is allowed to move upwards (no canting allowed), and middle beams are allowed to rotate.
The beams are made rigid to reduce the computational cost, which eventually brought up an issue with penetration between rigid bodies. To avoid this I decided to create a rigid pinned connection between the beams.
Since there are many ways to model such a connection, I will only list two that worked for me(to some degree) and state the issues that I faced.
(1) A reference point attached with MPC BEAM to a set of nodes on each hinge feature on the beams. A wire connects two reference points between the neighboring hinges and HINGE type connector is assigned with a rigid-like elastic behavior.
Diagram of one hinge: RP (MPC) – WIRE (HINGE) – RP (MPC).
After inspecting the animation of the analysis I found that the hinges shift away from their rotational axis towards the of the first step, making some beams penetrate into each other to a small degree. At the start of the second step, the petals on one side of the assembly rapidly come out of their sockets striking the rubber and distorting the elements, which essentially aborts the analysis. This motion is highly unrealistic and releases a high amount of kinetic energy. We attempted to damp the inertia accumulated in the petals by applying pressure to counter that motion, but no luck. Alternatively, the petals were fixed in all DOF in an intermediate step to lose their inertia, no luck too.
(2) A reference points attached with Kinematic Coupling to a set of nodes on each hinge feature on the beams. One reference point is allowed to have one rotational DOF between the neighboring hinges. An MPC type BEAM constraint connects two reference points between the neighboring hinges.
Diagram: RP (Kinematic Coupling w/ one rotational DOF) – MPC (BEAM) – RP (Kinematic Coupling, all DOFs constrained).
The issue is that the reference points that have the rotational DOF are slowly drifting away from their initial position, hence shifting the rotational axis. This is a bit odd as each hinge has its own local coordinate system and both reference points constrained from translating in any direction within that coordinate system.