When performing a nonlinear static structural analysis in Ansys Mechanical, users often encounter convergence challenges. These issues are particularly common in large-scale models, where identifying the root cause of non-convergence can be complex. This post outlines practical strategies for diagnosing and resolving convergence problems.
In an static analysis, every node must be constrained in some way; otherwise, the solver cannot converge to a unique solution. A quick way to check for missing constraints is to run a modal analysis. In the modal results, unconstrained motion appears as 0 Hz modes (rigid body modes). Plot each 0 Hz mode to identify which parts of the model are moving freely, then apply the necessary constraints or connections.
In the details of Solution Information there is an option to create result objects for Newton-Raphson Residuals and Element Violations. The number in the field is how many results to generate. When that number is greater than one, Mechanical will generate results object when the analysis fails to converge.
The resulting plots for Newton-Raphson Residuals can highlight the location in the model that has a high residual that could be preventing the model from converging. The element violation results can show elements that violate predetermined quality, and could isolate areas of excessive distortion that can also contribute to non convergece.
Nonlinear equilibrium needs gradual loading, especially across contact changes or as the material starts yielding. Changing Auto Time Stepping from Program Controlled to On will allow for more controll over the time stepping behavior. To help improve convergence, reduce the initial step, and set tight min/max bounds. The algorithm will predict step sizes and bisect on divergence, retreating to the last converged state.
Contact is a common source of non-convergence. Contact trackers are a very helpful tool to help identify which contact definitions are contributing to non-convergence. The use of contact trackers are described in detail in a previous post.
Once a troublesome contact is identified there are several settings that can be changed to help with convergence. Reducing the Normal Stiffness is sometimes necessary to acheive convergence, but it does come at the cost of additional penetration in the contact, however this additional penetration may be very small. It is recommended to start with a normal stiffness factor of 0.1 and see if it helps with convergence. If it does, then gradually increase the stiffness factor until you arrive at the highest value that still allows for convergece.
Changing the Update Stiffness parameter can also help. Changing it to "Each Iteration, or "Each iteration, aggressive" can give your model more of a chance to converge, at the cost of a small increase in compute time.
For frictional contact, the Interface Treatment is an important behavior to control. If your model has contact interfaces that supposed to be coincident at the start of a simulation, then changing the interface treatment to "Adjust to Touch" can help to clean up any incidental gap in between surfaces that could cause convergence problems in a static analysis.
Finally if all else fails, many times the best way to find the causes of non-convergence is scale down the problem complexity as much as possible. This can take the form of removing geometry, changing nonlinear contacts to bonded contacts etc. If you can create a simplified model that convergence, then gradually adding back in complexity is a valid way of finding the cause of non convergence.
There are many different potential sources of error that can affect a models convergence and it can be very difficult to pinpoint a cause especially in large models. This is by no means an exhaustive list, nor will it find every issue, but it can provide direction and bring attention to some features of Ansys Mechanical that are helpful.
We offer support, mentoring, and consulting services to enhance the performance and reliability of your systems. Trust our proven track record to accelerate projects, optimize performance, and deliver high-quality, cost-effective results for both new and existing water control systems. For more information, please visit https://ozeninc.com.