Nonlinear Ansys APDL Contact Model for Gear Teeth – Forum

Gear Contact Analysis in ANSYS APDL: Addressing a Rotational DOF Coupling Dilemma

Nonlinear contact simulation of spur gear teeth under load is a recurring challenge in mechanical design. A recent discussion among simulation specialists on the Italian engineering forum cad3d.it dissected a specific modeling approach that, while delivering results within 1% of Hertzian theory, raised a subtle but important question about the coupling of elements with different degrees of freedom (DOF). The case offers valuable insights for anyone working with multipoint constraints in ANSYS APDL for 2D contact problems.

Modeling Approach: Torque Transfer via CERIG and a Central Node

The analyst built a 2D model of two spur gear teeth in contact. To apply the load realistically, the sector of the gear tooth was connected to a central node using CERIG elements. These rigid constraints linked the PLANE183 elements (the 8-node structural solids representing the tooth) to a single pilot node located at the gear center.

By default, this central node lacks a rotational degree of freedom. For applying a torque directly at the center and having it transmitted through the CERIG elements to the tooth flank in contact, the user needed to activate the rotational DOF at that node. The solution adopted was to place a MASS21 element at the center point, assigning a very small mass (10⁻⁹) to avoid altering the inertial characteristics of the system. This allowed the moment to be imposed, and the rigid elements transferred the load to the tooth, bringing it into contact with the second gear tooth.

Validation Against Hertzian Theory

The simulation produced a contact pressure distribution with deviations limited to 1% compared to classical Hertzian theory. This level of agreement suggests that the kinematic and load transfer mechanism, as implemented, is fundamentally sound. The user, however, remained cautious about a specific aspect of the coupling methodology.

The Core Doubt: Coupling Elements with Mismatched DOFs

The critical question raised by the analyst centers on the compatibility between the MASS21 element and the CERIG constraint applied with the UXYZ option. MASS21 is a point element that, depending on the keyoption settings, can have up to six DOFs (UX, UY, UZ, ROTX, ROTY, ROTZ). The CERIG element, when configured with the UXYZ option, rigidly connects the translational DOFs of the slave nodes (the PLANE183 solid elements) to the master node. The PLANE183 element for 2D analysis has only UX, UY, and optionally ROTZ as DOFs.

The concern is whether coupling a MASS21 node (which has been given rotational capability) to PLANE183 elements (which operate with translational DOFs only in the 2D formulation) via CERIG with UXYZ creates a mathematically consistent or potentially overconstrained system. The different intrinsic DOF sets between the structural solid and the point mass can introduce artificial stiffness or constraints if not managed correctly, even when the results appear physically plausible.

Practical Considerations for the Workflow

From the comparison among professionals on cad3d.it it emerges that the combination of MASS21 and CERIG is a known workaround for activating rotational DOFs at a pilot node in a 2D solid model. The approach is widely used, but the validity depends on the specific element formulations and the keyoption settings. The user’s caution is well-placed: a mismatch in DOF types between the master and slave nodes in a constraint equation can sometimes lead to unexpected stiffness or spurious modes, especially in nonlinear contact scenarios.

For practitioners, this case reinforces the importance of checking not only the global results (contact pressure, stress distribution) but also the local behavior at the constraint boundaries. The fact that the contact footprint matches Hertzian theory so closely is a strong indicator that the coupling methodology is effective, but a sensitivity study on the MASS21 mass value and CERIG keyoption settings would further solidify confidence in the approach.

The dialogue around this model reflects a mature understanding of the nuances of APDL and the need to carefully validate each step of a simulation when mixing element types with different kinematic capabilities.

To explore all the technical details, read the full responses, and actively participate in the debate, we invite you to visit the original discussion on the cad3d.it forum (in Italian).