Inertia Relief
Inertia relief analysis is a type of stationary analysis, where the external loads are balanced by (unknown) inertial forces. Thus, even though rigid body motion constraints must be present to avoid a singular stiffness matrix, there will be no reaction forces.
In a 3D geometry, three selected control points are given by their radius-vectors rj. Using these points, a local coordinate system is defined with the origin at point 1 and base vectors given by
The transform from the global to local system is given by
The displacement vectors in the local system at point (j) are computed as
where uj are displacement vector DOF at the corresponding point.
The radius vector at the local system is given by
and the frame acceleration field in the global system is computed as
Here, a is the translational acceleration vector and o is the angular acceleration vector.
During the training study step, unit linear or angular accelerations are applied as
In the seventh and last load case, both a and o are set to zero. Thus, the external load is present in all load cases.
Apply the following six weak point constraints:
at point 1, which constrains translations in directions e1, e2, and e3. The corresponding three scalar constraints are numbered as 1,2,3. At point 2,
which constrains rotations around axes e3 and e2. The corresponding two scalar constraints are numbered as 4,5. Finally,
at point 3, which constrains rotation around axis e1. This scalar constraint is numbered as 6.
The corresponding constraint force components for the six load cases are represented by a 6x6 matrix
where the first index (m) is the number of the constraint, and the second index (n) is the number of the load case. For load case 7, the constraint force components are represented by a 6-component vector
For any other study than the training study, the frame acceleration is computed using
with the values of f and F precomputed during the training study. Applying the corresponding frame acceleration in the global system will result into zero constraint forces for all six constraints.
In the case of 2D geometry, only the constraints 1,2 and 4 are needed. The corresponding load cases are 1,2 and 6. Thus, f can be reduced to a 3x3 matrix, and F will be a 3-component vector.
In 2D axial symmetry, only constraint 3 and load cased 3 are needed.