ODYSSEE Lunar - GUI
- In Nastran parser, the pressure can be now extracted from the HDF5 file.
- Visual and intuitive improvements:
- In Data management tab, when a column was selected, and the user expanded the column width to display a long label, unchecking the column led to column width reset to default value. From now on, the width set is fixed.
- In the Sensitivity tab, in case of multiple datasets, a title to the colour scale of the "colormap" graphic has been added. This title allows to identify what the colour means: "multi datasets parameter’s effect coefficient".
- In the interpolation window, the term “multiquadric” has been changed to “multiquadratic”.
- After naming a new project, the interface switches automatically on the data management window. The tree on the left presenting the X and Y is now expended to have direct access at the import CSV and DOE tool buttons.
- The validation cases used to validate the solver were decomposed in the data management tree with the parameters: “XN – validation cases” and the exact responses: “YN – validations cases”. Now, the name will be “Xval – validation cases” and “Yval – validations cases” to limit the confusion with the standard predictions named YN.
- Bug Fixes
- In the DOE tool, even though user saved its file under .csv with comma separator, semi-colon separator was used the file generated.
- The “enable sort” in Sensitivity tab was reset each time the user switched to another dataset.
- When the user was on the animation tab and was running an optimization, ODYSSEE CAE crashed.
- In animation, if the DOE path was incorrect, a crash occurred after clicking on “Run animation prediction” button. Now a warning message is displayed.
- In certain situations where the same optimization was executed twice in a row, ODYSSEE CAE crashed.
- Exact expected curves were not correctly displayed when hiding base curves and predicted curves, due to scaling issue.
Known Bugs
- Kriging interpolation (direct or with POD, FFT, Clustering) with cubic basis function is returning the error
“LU decomposition failed. at “ - In the “data management window”, if we import one CSV file containing X, Y, Xval and Yval, depending on the way to define the validation cases, there is an issue to fill the Yval part: if we start with a right click on rows from the original data to “use selected rows as validation cases”, and next a right click on columns from the original data to “Create Y-dataset from selected columns”, then there are no rows added in the Yval – validation cases.
- In Animations configuration dialog box, if the paths for the input animation files and the database directory are different, then a UI crash can occur.
ODYSSEE Quasar – Solver
- Bug Fixes
- The INVD solver gave different results if the XN file had a different number of cases due to the standardization.
- ext("mining","entropy",matGrey,nintervals,nstart,windowsize, print,lastline,flag,stand,0, mb,OUTPUT) function has an 11th additional argument set by default to 0.
- When calling valueReference = [0] from FMU v3.0 generated by ODYSSEE, an incorrect result was given. From now on, a default 0 value is retrieved.
- With the MLP solver since 2023.2, if the data contained columns full of zeros, normalization was not possible, which blocked interpolation. This is fixed, if there are columns full of zeros, these are skipped, and normalization is not applied.
- The Quasar Help document has been updated.
ODYSSEE Nova – Optimizer
Known Bug
- The summarize printed at the end of the log file is not correct if constraints H or G are defined. If you use ODYSSEE Solver directly, to obtain the optimal point open the CSV file nova_results.csv or optimization.csv to find the optimal point among the tested points. This issue does not affect the optimal point presented in ODYSSEE CAE.