Unterschiede

Hier werden die Unterschiede zwischen zwei Versionen angezeigt.

--- en:versuchsdatenverwaltung:versuchsplanung:versuchsplanung_-_d.o.e [2025/05/13 15:23] – angelegt deppe2
+++ en:versuchsdatenverwaltung:versuchsplanung:versuchsplanung_-_d.o.e [2025/12/06 18:01] (aktuell) – deppe2
@@ Zeile 1: / Zeile 1: @@
 ====== Design of Experiments ======
+===== Design of Experiments =====
+**Path:** Main menu > Experimental data > Design of experiments...\\
+{{:versuchsdatenverwaltung:versuchsplanung:icondoe-versuchsdaten.ico?nolink&60x60 |}}
+\\
+\\
+\\
+{{ :en:versuchsdatenverwaltung:versuchsplanung:en_sigma150_dlg_versuchsdatenverwaltung_009.png?nolink |}}
+**Figure:** Design of experiment 1: Selection of the target values
+Besides the manual set up of measuring data tables described above, SIGMA enables the user to generate a measuring data table according to a design of experiment. In that case the defined target variables will be transformed to measuring points in a measuring data table and control variables are represented by special columns. By choosing Experimental data, Design of experiments in the menu a wizard dialog opens, which assists the user by designing an experiment. The wizard is build up of three steps at which end a measuring data table is generated by SIGMA according to the chosen target variables, control variables and the selected design.
+==== Control Variables ====
+The next step is to choose independent variables, which are expected to have an effect on the defined target variables. These variables are known as control variables.
+This page (see figure) is similar to the previous one. Control variables can be chosen from independent process variables on the left hand side of the dialog. By clicking ”>>” a control variable is created on basis of the selected independent process variable and listed in the upper right list box. Depending on the selected process variables SIGMA distinguishes four different types of control variables to be created, which have to be further defined in the lower right panel. All in common is a name, the independent process variable, denoted as path, and the current value of that process variable.
+=== Continuous Control Variables ===
+A continuous control variable is created if the value of the process variable is a real number (is element of R).
+The design space of the continuous control variable has to be defined by its minimal and maximal values. The number of variations for continuous factors can be chosen freely. Optionally the value may be set to ”r;auto”, meaning that the user let the design determine an appropriate number of variations. If the number of variations is given here, the design has to deal with it.
+=== Discrete Control Variables ===
+A discrete control variable is created if the value of the process variable is an integer (is element of R).
+In contradiction to continuous control variables the values of discrete control factors can only be varied in whole numbers. The figure  shows a control factor for the number of flights of a conveying element. The design space has to be defined by its minimal and maximal values. The default value for the number of variations is two, because a discrete factor can always be varied at two levels. If the number of variations is given by the user, the user has to make sure that the resulting values for each level are also whole-numbered.
+=== Discrete Control Variables (Exchange Screw Elements) ===
+To study the effect of different screw configurations one can define control variables that exchange a given screw element by another. In the figure a kneading block is exchanged by an eccentrical kneading block.
+Such a discrete control variable can only be varied at tow levels, so that the user can’t define the number of variations for such a variable. The elements to be exchanged are defined by their Ids. The first ID is the current value of the chosen process variable and the second can be chosen from the available elements for the current machine.
+=== Discrete Control Variables (Exchange Files) ===
+Besides the above described control variables that exchange a single element, there are control variables in SIGMA that may exchange a whole screw configuration or used materials. Everything that is managed in single files by SIGMA e.g. process parameters, screw configurations and materials may be exchanged by such a control factor.
+In Fig. the definition of a control variable is depicted that exchanges the used materials in the first material stage.
+Such control factors can be varied at two levels. The first value is the current value of the process variable and the second value can be chosen by a file selection dialog.
+==== Target Variables ====
+The first step is to choose variables to study, namely the target variables.
+In the figure this first step is depicted. On the left hand side a type for a target variable has to be chosen from calculated results in SIGMA. By clicking the ”>>”-button (at the centre of the dialog) a new target variable is created. All defined target variables are displayed in the upper right corner. By selecting a target variable in this list box the target variable is displayed with further information in the lower right panel. Here the user has to define some more input like a name for the target variable and the position of the probe along the screw. The position of the probe is optionally and can be set to ”total”. The value ”total” means that there is no probe along the screw and the target variable is measured elsewhere.
+Having all target variables defined, the user can proceed with the next step by clicking the ”>>”-button in the lower right corner.
+==== Design: ====
+If all control variables are defined, the next step will be to choose an appropriate design. This step is depicted in the figure.
+In the upper left corner there is a summary of defined target and control variables. In the upper right corner an appropriate can be chosen. The calculated number of experiments for the selected design and the defined control variables is displayed below. In the lower right some properties of the selected design can be defined. This panel changes with the selected design. In the lower left a graphic of the selected design in R3 is displayed.
+If the target variables, the control variables and the design are defined properly a measuring data table can be generated by SIGMA according to the given design. To do so, the generate button has to be clicked on. The target variables are transformed to measuring points in the measuring data table and control variables are represented by special columns, whose cells in the grid are filled according to the chosen design space.
+The user may select one of four different designs.
+=== Factorial Design ===
+In SIGMA there are three types of factorial designs with different properties and restrictions to generate a measuring data table accordingly.
+  * Full factorial $2^k$ designs: Any type of control variable can be used. The values of the control variables are varied at two levels.
+  * Half fractional factorial $2^{k-1}$ designs: There has to be at least three control variables defined. The control variables are varied at two levels.
+  * Full factorial $3^k$ designs: Each control variable is varied at three levels.
+=== Central Composite Design ===
+A central composite design is composed of a full factorial 2k design plus star points and centre points. The number of centre points is given by the user. The user can choose between orthogonal, rotable and face-centred designs. The factor alpha is computed accordingly. For this design each control variable needs to be a continuous one, because the location of the star points is computed according to the factor alpha.
+=== Box-Behnken Design ===
+Only continuous control factors can be used for this design. The user can define the number of centre points.
+=== Multi-level Design ===
+For the multi-level design any type of control factor with any number of variations can be used. Each control variable is varied with the defined number of variations. For continuous control factors the number of variations may not be ”auto”, but has to be defined by the user.