HEEDS Technical Tips
Our engineers have written these tips to share their strategies for accomplishing specialized tasks using HEEDS MDO software. If you have a question you would like our engineers to answer with a technical tip, please feel free to contact us.
Controlling Minimum Variable Change Using Variable Resolution
During optimization, the values of variables are automatically changed to create new designs. In many studies, it is important to control the minimum amount by which a variable can change during optimization. For example, with design variables that cannot be fractional (e.g., the number of teeth on a gear), can only be controlled to a coarse precision level, or are only available in certain sizes, you need to be able to control step size during optimization. In HEEDS MDO, there are two methods you can use to control the minimum amount by which a variable value can change to produce a new design. This tip describes one of those methods.
Using Constraints to Improve the Resolution of the Pareto Front
To capture the trade-offs between objectives, it is important to have good resolution in the region of interest on the Pareto front curve. Although the area of interest is typically the “knee” region, in some studies, you may wish to focus on another area of the curve. In HEEDS, the resolution of the Pareto front is determined by the archive size (the number of points on the plot, per cycle), which you specify. The larger the archive size, the higher the resolution. However, a larger archive size also increases the total number of design evaluations required to generate the optimal Pareto front. In this tip, we demonstrate how to effectively improve the resolution of the Pareto front without increasing the archive size.
Quickly Identifying Runtime Errors
Before launching a HEEDS study, it is important to make sure that there are no errors in the model setup. Any unresolved errors can result in the loss of valuable time. For example, a HEEDS study with undetected runtime errors, executed to run unmonitored over the weekend, could result in the loss of up to three days of computational time. This tip will describe a few different techniques you can use to quickly identify potential runtime errors without performing complete design evaluations.
Eliminating Setup Errors before a Shape Optimization Run
In shape optimization studies, it is important to have a robust parametric model for creating new designs. This tip demonstrates how to quickly identify potential errors in the parametric model prior to running the full HEEDS study.
Imposing Constraints on Design Variables
This tip demonstrates how to redefine design variables such that HEEDS only creates designs that meet the imposed constraints.
Searching the Log Space of a Design Variable
In cases where a design variable ranges over several orders of magnitude, it often makes sense to define the variable range in log space instead of linear space. Searching the log space allows you to find the value of the design variable that is within a certain percentage of the actual best value, as opposed to an absolute precision. Instances when it is appropriate to use this technique may include curve-fitting and determining chemical rate constants. In this HEEDS tip, search of the log space is explained and instructions are given for how to set this up in HEEDS.