FEM - Changes in rigid links - since SCIA Engineer 24.0

Introduction

This chapter is to provide a closer information about change originated from the FEM solver implemented in SCIA Engineer version 24.0. There are several changes concerning rigid links and line rigid links.

Hinges only on one end

First of all, there is no longer possible to define hinge on both ends of these entities. The reason of this is that such hinge is considered to release all 3 rotations, and if applied on both ends (master and slave node), the rigid link could rotate along its axis. Before version 24.0, there was certain numerical noise so this was not causing any problems. Since version 24.0, this is not possible any more. The user will encounter a warning upon opening of older project from previous version where rigid links with hinges on both ends (on both master and slave nodes) were present. In such case, the hinges are deactivated and the user needs to manually check the corresponding locations in order to justify this setting.

FEM formulations

The internal FEM formulation of these entities has also changed. Before version 24.0, the rigid links were internally considered by beams of large rigidity. Since version 24.0, these entities are treated as rigid body constraints, and their behaviour will be also influenced by the " FEM - Multi Freedom Constraints (MFC) - Lagrange Multipliers" property (in most cases).

Certain numerical differences might occur if the rigid links are present in the project, however these differences are expected to be rather negligible.

An example of formulation difference is shown in the figure below, where there is a schema of a line rigid link between master node and slave nodes of the line.

Before version 24.0, in case there are no hinges at the ends of these rigid links, the formulation is expressed in the top-left corner, as connection of each slave node with a master node by an internal beam of large stiffness (beam stiffness matrix based on the physical properties of this very stiff beam). In case there were hinges (bottom-left corner), the nodes were duplicated internally at both ends of these stiff beams. This multiple duplication of nodes at some positions might have brought numerical noise, hence not resulting in singularity (by allowing rotations along x-axis of these very stiff-beams).

Since version 24.0, the internal connections of these nodes is done in a bit different way (top-right corner) - the connections between nodes are forming the smallest path (blue continuous line), what is formally the same as before, but due to the fact "very stiff" beam is not the same as "ideally rigid", and also mathematical formulations of these connections (influenced also by the " FEM - Multi Freedom Constraints (MFC) - Lagrange Multipliers" settings), there might be little differences. In case there are hinges (bottom-right schema), the nodes are again internally duplicated, and constraints between the coincident nodes are handled.