Analysis Settings in Ansys Software

In the Ansys tool, the Analysis settings contain the parameters determining how the simulation should run. These include vital factors like the size / number of the load steps, size / number of the sub-steps, type of solver, tolerance criterion etc. 

These parameters ultimately determine if the solution will converge or not. Hence, the Analysis settings play a crucial role in the setup of an FEA problem.

Before exploring the Analysis settings, we must first understand the difference between the explicit and implicit methods. This is explained in our previous blog, which we highly recommend that you read.

Here’s a short summary of the differences between the Implicit and Explicit methods:

In the implicit method, the intervals are called Load steps, because the applied load is divided into the given number of load steps. 

For example, if a body is to be rotated 360 degrees and the number of load steps chosen by the user are 12, the 360 degrees rotation is applied in 12 steps. Also the user can decide the amount of rotation which can happen in each of the load steps. Either constant rotation of 30 degrees in each load step can be given (static loading) or different rotation angles in different load steps can also be given (transient loading).

In the Ansys tool, the number of load steps can be specified using the “number of steps” option in Analysis settings as shown below. In the following image, the number of load steps is 5.

“Step end time” option denotes the size of the load step. In the image above, it is 1 second.

Now, to reduce the pressure on the solver, Ansys again divides a particular load step into sub-steps of smaller size. This means the load to be applied in that particular load step is again divided and applied in steps. 

Consider the example above in which a rotation of 360 degrees is applied to a body. Suppose a constant loading of 30 degrees is applied in each load step. If the number of sub-steps within the load step is specified as 10, then the 30 degrees rotation is again divided into 10 steps, say of 3 degrees each. 

These sub-steps can be specified in 2 ways: either by specifying their number or by specifying their size (time duration) as shown in the below image:

When the sub-step is defined using its time duration, Ansys uses the term “time step” for such sub-steps. However, do keep in mind that the word “time-step” is technically associated with the explicit method. 

Here, in the Analysis settings, the word “time step” is only a nomenclature used by Ansys, although the method used here is the implicit method. It is important to note that the Ansys tool uses the implicit method by default. 

Now, there are 3 ways to bring variation in the size of the sub-step, as shown below in the “Auto time stepping” option:

If the “Auto time stepping” option is “Off” - the size of the sub-steps is constant. In the image below it is 0.1 seconds:

If the “Auto time stepping” option is “On”, the size of the sub-steps varies automatically, based on the difficulty level of the problem. 

If, at a particular time, the problem becomes difficult to solve, the size of that sub-step is automatically reduced. As we saw in the previous blog, this is because the time interval of measurement reduces, and therefore the error also reduces. As the error is smaller, it takes less number of iterations to converge the solution and it becomes easier to solve.

As the size of the sub-step is variable, it varies within a particular range. This means that the lower and upper limits of this range have to be specified. The lower limit is specified as the “minimum time step” and the upper limit is specified as the “maximum time step” as shown below:

If the “Auto time stepping” option is “Program controlled” - the solver itself decides whether to use the constant time step or variable time step based on the difficulty level of the problem.

For the 1st load step, the “Initial time step” option specifies the size of the first sub-step of that load step as shown below:

From the 2nd load step onwards, if the “Carry over time step” option is “Off” - the “Initial time step” option specifies the size of the first sub-step of that load step as shown below:

If the “Carry over time step” option is “On” - the size of the first sub-step of a load step is the same as the size of the last sub-step of the previous sub-step. There is no need to specify the value of the initial time step in this case as shown below:

For example, if the last sub-step of the 1st load step is of 0.1 seconds and the “Carry over time step” option is “On” - then the first sub-step of the 2nd load step will also be of 0.1 seconds duration.

The Analysis settings also allow the user to specify the “tolerance” value to be used in the Newton-Raphson method. As discussed in the previous blog on the Explicit vs. Implicit methods, the Newton-Raphson method tries to reduce the error below the value of the tolerance.

Therefore, the value of tolerance determines if the solution will converge or not. If the tolerance value is very strict, more iterations will be needed to converge the solution. The solution may not even converge if after multiple iterations, the error is not reduced below the strict tolerance value.

In the Ansys tool, one of the ways to decide the tolerance is to specify the “stabilization” settings under the “Non-linear controls” section of the Analysis settings.

As shown in the image above the “method” used is “Energy” - meaning the tolerance value for energy is to be used. 

The “stabilization” option is constant - meaning a constant value of tolerance is to be used and that value is 0.1 as specified by the “Energy dissipation ratio” option.

By default Ansys uses the “Energy dissipation ratio” value of 0.0001, which is very strict compared to 0.1. Hence the value of 0.1 will help to converge the solution easily, although by compromising a little on the accuracy of the results. 

In the above image taken from Ansys, the purple line represents the error values while the cyan blue colored line represents the tolerance. The dots on the purple line represent the iterations. The error is reduced through multiple iterations to a value below the tolerance. When this happens, the solution converges.

Do apply these tips when you work with the Ansys tool. We hope these methods help you solve your problems better!