New and Updated Operators, Functions, and Definitions
•
A CFD Data (CGNS) function is available as a new global function. You can use it to import CFD analysis data as CGNS files (*.cgns). This function requires the Acoustics Module or the CFD Module. Evaluation is only supported on Windows®.
•
For Interpolation functions with a single argument, a new option for specifying the sampling points called Adaptive is available. With this option selected, the interval between two adjacent sampling points depends on the slope of the interpolation curve.
•
A Plot Parameters section has been added to the settings for the Gaussian Pulse, Ramp, Rectangle, Step, Triangle, and Waveform functions.
•
A new built-in function, expinti(x), evaluates the exponential integral function Ei(x).
•
A new built-in global boolean variable, issolving, indicates whether a problem is being solved.
•
The built-in operators elemint and elemavg now accept a suffix _frame, where frame is spatial (gives the default operator), mesh, material, or geometry to explicitly specify the frame in which to perform the integration. For example, the operator elemavg_mesh(gporder,expr) evaluates the average of the expression expr in the mesh frame.
•
For DNN functions, the following functionality has been added:
-
There is now an Export button that lets you export the deep neural network to an ONNX (Open Neural Network Exchange) file.
-
More scaling, loss function, and activation function options have been added.
-
The Settings window now includes a Post-training test data section.
•
Weak Contribution and Constraint equation contributions now allow using a new built-in Gauss–Lobatto quadrature method or a user-defined integration point and weight pattern instead of the default Gaussian quadrature. Also, the integration order can now be controlled by a user-defined parameter expression, for example in a Parametric Sweep.
•
The new Integration Rule feature allows defining custom quadrature rules for use with Weak Contribution and Constraint equation contribution features.
•
The new P-Norm feature computes the p-norm of an expression over a selection for a chosen value of p. The p-norm provides a differentiable approximation of the maximum value.
•
The new Standard Deviation feature computes the standard deviation of an expression over a selection.