Intuition plays a critical role in all stages of a design exploration study, from defining the problem statement to building the simulation model to interpreting the results. But what about the search process itself? Should we make design improvements based on intuition, or should we allow a mathematical search engine to explore the design space for better designs? The answer is both. We call this shared process collaborative design exploration.
The SHERPA search strategy allows you to inject your design ideas before and during an exploration study. Before you start a study, you can seed it with multiple ideas (in the form of actual designs) that might help SHERPA to locate productive regions of the design space more quickly, thus speeding up the overall search. For example, in addition to the baseline design, you might consider seeding the study with other potentially good designs that:
- you have investigated or produced in the past
- your competitors have used
- are feasible, but perhaps not optimal
- are high performing relative to one or more criteria, but not all of them
- have some desirable features, but don’t necessarily perform well
- you have a hunch may work well
- are from a previous HEEDS MDO study
One or more of these injected ideas might contribute to a more efficient search, while the cost of doing this is only the time to enter the variable values that define each of the designs. SHERPA will evaluate the injected designs when the search process is launched, so there is no need to simulate them before injection. Continue reading
We all feel the “need for speed” when trying to find better design solutions through simulation. What can we do to speed up our design exploration studies? Let’s discuss all the options here, including one that may not be obvious.
In a typical study, the total CPU time needed to perform a design exploration study is determined using this simple formula:
Highlighting a Few New Features that Help You Discover Better Designs, Faster
Often, improvements to the simplest things can have a big impact on your daily tasks. There are many tasks we perform repeatedly when working with HEEDS, and streamlining those saves time and reduces effort. HEEDS 2015.11 contains many enhancements focused on simplifying workflows and I want to highlight a few that help in exploring design performance relationships.
To explore relationships between variables and responses in detail, you typically require multiple plots of the same type, but with different variables to gain a clearer understanding of dependency or influence. However, there are many plot features that are tailored to suit the particular way you want to view the results such as axis scales, data symbols, curves styles, title fonts, and so on.
To avoid having to create a new plot from scratch and redefine all these settings, you can now right click and select the Copy Plot option. This makes an exact copy of the existing plot, with all the customization. You then just need to alter the variables or responses being displayed saving a lot of setup time.
Figure 1. Make a copy of an existing plot with a single right click option
There are a lot of great tools in HEEDS to help you gain insight into finding the best design. One area of enhancements in HEEDS 2015.11 focused on parallel plots. In this article, we’ll highlight some ways to use new features of parallel plots in HEEDS to discover better designs, faster.
Parallel plot background
To help show the new capabilities in the context of an engineering problem, let’s look at exploring shape options for a human powered vehicle. There are obviously many dimensions that can be adjusted to improve the design.
Figure 1. Possible parameters to change for a Human Powered Vehicle
We often need for a design or a model to perform in a specified way. For example, the parameters in a nonlinear material model should be selected to best match the experimental stress-strain response. The geometrical parameters of a rubber bushing should be designed so that its force-deflection response matches the desired nonlinear stiffness behavior.
Optimization problems like these arise frequently. We refer to them as curve-fitting problems, because the goal is to minimize the difference between the specified target curve and the actual response curve
of our design or model.
Figure 1. The difference between a target curve and a design curve is minimized in a curve fitting optimization problem.
One of the challenges to finding optimal solutions is the coupling of variables. If we could change one variable at a time, the search process would be so much easier. But in most problems the variables are strongly coupled, so the best value for one variable depends on the values of many other variables.
Often, we have no control over this variable coupling, since it is inherent in the physics model that defines our objectives and constraints. In these cases, we need a powerful optimization search strategy like SHERPA to figure out the complexities of the design landscape and to produce an optimized solution.
But in some cases we make this task harder than it needs to be because of how we represent the problem. That is, sometimes the way we define the problem creates unnecessary coupling or increases the complexity of existing coupling among variables. This makes the optimization search harder, and may decrease the chance of finding the optimal solution within our limited optimization search budget. The good news is that we can often alleviate this situation with a different representation. Continue reading
Quite often, design optimization problems involve semi-independent design variables. That is, some of the design variables may have to satisfy a certain relationship, but they vary independently. This would be true, for example, if you had three variables that were independent, but you wanted their sum to equal a certain value. In general, there are two ways to deal with these types of problems:
- You can impose a constraint on the design variable values using a formula-based response.
- You can redefine the design variables such that only designs that meet the imposed constraints can be created.