While
analysis helps you simulate a product development cycle on the computer quickly and
inexpensively, you still need to create several studies and simulate many scenarios. Each time
you make a change, you need to run the analysis and examine the results.
Even in a relatively simple design, you can change several dimensions.
Deciding on what combinations to try and the associated bookkeeping and result viewing can
become cumbersome.
Design Study exploits the parametric, feature-based modeling, and the
automatic regeneration capabilities of the software to automate the optimization process. The
software is equipped with a technology that quickly detects trends and identifies the optimum
solution in the least number of runs. The program uses a method based on the Design of
Experiments.
The program offers two different qualities in the properties of the design
study. The software runs a number of trials based on the quality level and the number of
variables. For each trial, the program runs all the associated simulation studies with a
strategically determined set of variable values. The following table lists the number of
iterations for the high quality and fast results methods for continuous variables (Range option). The program uses the Box-Behnken quadratic plan
for High Quality setting and the Rechtschafner
quadratic plan for the Fast Results setting. Although
the Rechtschafner plan performs certain precalculations that Box-Behnken design does not
require, it needs fewer experiments to form the response function and optimize.
| Number of Design
Variables (for continuous variables) |
High
Quality |
Fast
Results |
| 1 |
3 |
N/A |
| 2 |
9 |
N/A |
| 3 |
13 |
N/A |
| 4 |
25 |
15 |
| 5 |
41 |
21 |
| 6 |
49 |
28 |
| 7 |
57 |
36 |
| 8 |
N/A |
45 |
| 9 |
121 |
55 |
| 10 |
161 |
66 |
| 11 |
177 |
78 |
| 12 |
193 |
91 |
| 13 |
N/A |
105 |
| 14 |
N/A |
120 |
| 15 |
N/A |
136 |
| 16 |
385 |
153 |
| 17 |
N/A |
171 |
| 18 |
N/A |
190 |
| 19 |
N/A |
210 |
| 20 |
N/A |
231 |
After running the experiments, the program calculates the
optimal design variables by forming a response function to satisfy the optimization goal. As a
goal, you can select to minimize, maximize, or define specific values for simulation data
tracked by sensors. The program then runs the associated simulation studies to evaluate the
results for the optimal design.