Main Interface

The main Attune window is what is presented when Attune is first opened. It is also the only window that typically remains open throughout the correlation process. The main interface has five distinct functional areas, which are highlighted below.

_images/Main_Annotated.png — Main Attune Form

Configuration Region

The configuration region, in the upper left corner, is where the user creates and edits configurations containing the test and analysis data. To create a brand new configuration, press the Create button. This brings up the Configuration Form in a blank state where data can be loaded. To edit a configuration that has already been created, select the desired configuration from the list box and press Edit. The Configuration Form will be loaded and prepopulated with all the data from the selected configuration.

Once a configuration is created/loaded, the name of the configuration is placed in the listbox. There is an icon to left of the configuration’s name that indicates its status. A indicates the configuration is valid and ready to be correlated. A indicates that the configuration does not contain enough data to define the configuration. If there are any configurations that are not valid, many of the buttons and menu items are disabled until this is corrected by editing the configuration. A valid configuration has analysis shapes along with test data or a frequency only definition. The configuration can be in a pre-optimization state (SOL 103) or in an optimization state (SOL 200). This is controlled by using the Active Data Set pull down on the Configuration Form and is discussed in length in that section. As it pertains to the main form, if any of the loaded configurations are in the pre-optimization state, all features related to optimization will be disabled.

If a configuration is no longer needed, valid or not, press the Remove button and the configuration and all its data will be removed from memory.

Summary Region

The status of the correlation is summarized graphically on the axis in the upper-right corner and textually in the summary table below. If more than one configuration is present, the main form can be updated with the information for a particular configuration by selecting the configuration from the pull-down in the Summary region.

The state variables in the summary table are named automatically. The prefix of the name for frequency and shape variables corresponds to the type of variable (e.g., frequency: freq; cross-orthogonality: xort; cross-MAC: xmac). The remainder of the name indicates from which test mode the state variable is derived. For example, freq_1 refers to a frequency state variable that compares test mode 1 to some analysis mode. The analysis mode matching this test mode is identified in the second column of the table (in this case, mode 1). The mass variables are named, intuitively, mass, X_cg, Y_cg, Z_cg, I_xx, I_xy, I_xz, I_yy, I_yz, and I_zz.

Action Region

The action region contains all the buttons necessary to call up other forms. The Design Variables and State Variables buttons create (or make active, if they already exist) the Design Variable Form and State Variable Form, respectively. These forms allow the user to manage the design and state variables and are discussed in sections that follow. State variable preferences can be registered as part of the configuration creation and are discussed in the Configuration Form section. Design variable preferences are set on the Global Preferences form available under the “File” menu and are discussed at length below.

The Match Modes button opens the Mode Match Form that permits the user to redefine the mapping between test and analysis modes. This form is discussed later. The Optimization button is tied to the adjacent pull-down list. Clicking this button will create or make active the optimization window corresponding to the method selected in the pull-down. Only one type of optimization window can be open at any time. The optimization forms are discussed in sections to follow. The History button opens the Optimization History Form that allows the user to browse the design space that has been explored in the course of the current Attune session (or may include previous sessions if a checkpoint file was used). The Pushback Design button opens the Pushback Design Form which is a special optimization form used to minimize design variable changes while keeping the integrity of the correlation. The Property Status button opens the Property Status Form used to query the current value of each property and provides a contour plot on the FEM. Both of these forms are discussed in detail later.

Menu Region

The menu region at the top of the Attune form contains many important functions for Attune.

File Menu

The File menu contains commands that affect Attune as a whole and are described below.

Preferences: This option will bring up the global preferences form described below. These are preferences that can be set that persist between each Attune sessions.
Change Working Directory: This command changes the current directory that Attune is pointing to. By default, this is the directory that Attune was launched from, but it can be changed with this option. This directory is the one that is the default when any open or save dialog is created by Attune.
Load Checkpoint: This menu item reads data that has been previously saved in a checkpoint file. Loading a checkpoint file is different than loading a configuration file. When a checkpoint file is loaded, Attune is returned to the same state as when the checkpoint file was saved. Changes that have been made locally to files that are references by the checkpoint will not have been updated.
Save Checkpoint: This item saves the current state of attune into a checkpoint file.
Load Configuration: This item is used to add a new configuration to Attune via a previously created configuration file (.cfg). This will reload the files using their current state, and apply any mapping, settings, etc., that were originally used when the configuration was created. Only new configurations can be loaded with this command. To reload the data for configurations that are already a part of Attune, Reload All Configurations is used.
Reload All Configurations: This command reloads the data from all active configurations so the data loaded into Attune matches that in the referenced files. This command is used after both the bulk data has been updated with new design variable values, and those analysis files have been resolved. This will reinitialize Attune with those updated values so the next iteration of the correlation process can begin.
Restart: This command will close and reopen Attune.
Exit: This item will close Attune.

Export Menu

The Export Menu contains functions that export information about the correlation to outside formats. This includes saving the correlation status to an Excel file, as well as updating the bulk data between iterations and to its final state after the correlation is complete.

Current Design to Plot: This item creates a visual representation of the current design variable status. This plot is similar to the plot that can be created by selecting “Best Design” on the Genetic Algorithm form.
Current Design Variables To Bulk: This item brings up the Input Data Update Form, which is used to update the design variables in the input files so the next iteration of the correlation can be run. It will change name based on the type of analysis data being used.
Current Design to Spreadsheet: This item brings up the Excel Output Form. From this window, the user can save the current status of the correlation to an XML or CSV file.
Current XORTHO/XMAC Plot To File: This will export the cross-orthogonality/cross-MAC plot on the main form to an image file. Upon selecting the menu item, an input dialog will appear where the filename and image format can be set. The exported plot will be oriented as it is on the main form when the menu item is selected.
Current XORTHO/XMAC Matrix to XML: This is similar to exporting the XORTHO/XMAC plot to a file but will store it in an XML file instead. This XML file is readable by Microsoft Excel, where it can be further manipulated.
Commond DOF Set to USET: This command will create a bulk data file with USET1 cards for each of the DOF in the set common to the analysis and test shapes. This can be used, for example, to define the USET cards necessary to produce a reduced mass matrix.
Final Property Values to Bulk: This item brings up the Input Data Update Form, which is used to update the properties in the input file to the final values after the correlation is completed. It will change name based on the type of analysis data being used.

Preferences

The preferences form has four tabs: Design, State, Configuration, and Optimization.

Design Tab

_images/Preferences_Design.png — Design Tab of Preferences Form

The design tab contains the following options:

Color Design Variable Changes: This option will color the design variable changes based on their current delta value. An increase in design variable value will be colored green; a decrease will be colored red. Gradations of these colors will used between delta values of -100% to 100%. After these extremes, the color will not change.
Design Cost Relative to 1.0: This capability gives the user the flexibility to control the cost associated with changing a design variable. Selecting this checkbox means that when the objective function is evaluated, the cost is based on the difference between the current design variable value and 1.0. Otherwise, the cost is based on the difference between the current design variable value and the initial design value for the current session (or the last time sensitivity data was read).
Persist Design Variable Weights: Persisting design variable weights will cause the design variable weights to be stored. Despite the fact that design variables may span configurations, the design variable weights will be stored in a configuration file. If more than one configuration is loaded, it is possible for a conflict to exist in the weights stored in different configurations for the same design variable. If a conflict is detected, the weight will be set to the default value of 1.0.
Persist Design Variable Active Status: This is similar to persisting the design variable weights. It will cause the design variable active status to be stored. Despite the fact that design variables may span configurations, the design variables’ active status will be stored in a configuration file. If more than one configuration is loaded, it is possible for a conflict to exist in the active status stored in different configurations for the same design variable. If a conflict is detected, the active status will be set to the default value of “active.”
Pushback Tolerance: The amount of root-mean-square (RMS) error increase allowed over all of the loaded configurations while the design is pushed back.
Pushback Step Size: The step size used when pushing back design variables. A smaller step size here will allow greater refinement available during pushback, but the operation will take longer.
Overall Design Weight: The overall design weight allows the user to emphasize or deemphasize the cost of changes to design variables as a whole with respect to improvements in the fidelity of the model. A larger overall design weight means that the user is willing to trade some improvement in model fidelity to ensure smaller changes in design variables. A smaller overall design weight indicates the converse. If the overall design weight is set to zero, changes in design variables have no cost.

State Tab

_images/Preferences_State.png — State Tab of Preferences Form

The State Tab controls the Goals for the three types of state variables (mass state variables are not included here). The Goal and Near Goal settings control the coloring of both the state variable table and the XML. If the Apply Targets to Objective Function box is checked, the goal settings will be used to weight the objective function. State variables are considered to have reached their goals if they are within the stated percentage of the target value. A state variable that has reached its goal will be treated as having the same contribution to the objective no matter what its actual value. This will have the effect of concentrating the optimization on state variables that have not yet reached their goals.

The Ortho, MAC, and Frequency panels at the top of the form are used to control what the Goal and Near Goal errors are. These three state variable types can be controlled independently. The fourth state variable type, Mass is not currently supported in this manner. The Colors panel in the middle of the form is used to control the colors that will used to highlight the state variable table and XML based on the Goal and Near Goal values. Away From Goal is any value larger than the assigned Near Goal value. To change the color, either change the R, G, or B values in the corresponding edit box, or press the ? button to use the color picker to choose a new color. The assigned color will appear in the box to the right of the ? button.

The Color State Variable Error checkbox at the bottom of the tab controls whether the colors will be applied to the state variable tables. The XML output will be colored with the settings on this tab regardless of the checkbox status.

Configuration Tab

_images/Preferences_Configuration.png — Configuration Tab of Preferences Form

The Configuration tab controls the default values applied to newly defined configurations. Each of the settings on this tab corresponds to an identical setting on the preferences tab of the Configuration form when creating or editing a preference, but the latter tab only controls the behavior of that configuration rather than the defaults.

Important

The configuration form only changes the defaults for new configurations. Any currently loaded configurations will not change when items on this tab are changed.

Optimization Tab

_images/Preferences_Optimization.png — Optimization Tab of Preferences Form

The Optimization tab controls the defaults for the different optimization routines in Attune. The General Options tab controls the settings that are common to all of the routines. This includes the Population Size, Max Iterations, and the Plot Interval in which plots will be created. The GA Options and Gradient Options panels control the defaults for those specific routines.

If the checkbox for Enforce the match threshold during optimization is selected, mode matches during optimization will be subject to the match threshold set by the user on the Mode Match Form. This implies that target modes which initially are assigned a match may not have a match in all designs. Likewise, a target mode which is initially unmatched may acquire a match during optimization. Note that this preference has no effect on optimization if automatic mode matching is not being used. To allow mode matches to appear and disappear, it is necessary to have some way of expressing the cost of having no match in the objective function. If this preference is selected, any target mode without a match will be assigned 100% error for its frequency and 100% error for its shape (cross-orthogonality or cross-MAC). The RMS error will be affected in the same way. Thus, the error (and objective value) associated with a design with this selection may be significantly higher than the same design with this preference deselected since unmatched modes do not otherwise contribute to the RMS error or the objective function. Note that if this preference is not selected, Attune will not allow new mode matches to be identified during optimization, nor will it allow existing modes to fail to find a match no matter how poor the match is. So, if the user is only interested in the set of modes which initially have matches defined, it may be a good choice to leave this preference deselected. However, if the user is interested in finding matches for modes which do not initially have mode matches identified, the user will benefit from selecting this preference.