OPTICast UNIT 12: Reviewing Optimization Results
While an optimization run is in progress, no operator involvement is necessary. When the run is complete, the system will bring you back to the SOLIDCast Main Window, just as if you had completed a regular simulation.
At this point the Project Tree will have two sets of entries. The first will be your original model and mesh, along with the Optimization Icon. The second will be the final model, mesh and simulation of the optimized casting/process. This data can be examined and plotted just like any other simulation. Those procedures are covered in detail in the SOLIDCast training workbook.
In addition to the normal simulation results, you can view a summary of the iteration history, which is contained in an Excel spreadsheet, and also a series of graphs, which plot the variables, constraints and objective function over the range of simulation runs.
To view these results, load the project, if it isn’t already, and click on the Optimization Icon. Then click on the Optimization Menu and choose Report.
To look at the summary of each iteration, click on View Iteration Data( .CSV file)… This will automatically start Excel and load the data into the spreadsheet. Note: This requires that you have a copy of Excel installed on your computer. If you do not have Excel, the .CSV file can be loaded manually into any other spreadsheet. A typical example of what the file looks like is shown here:
There will be an entry for each iteration , which will list the values of each variable, the constraints, and the objective function. This is a quick way to see how the optimization run progressed. As you can see from the sample file, the first simulations are single variable changes, to find the slope of the response surface of each variable. Once each contribution is found, the optimization runs begin in earnest.
The other way to view the data is in graphical format. To do this, go to the Optimization Menu and select Report… View Graphs. The first graph to be displayed will be the Objective Function, as shown here:
In this example, the casting yield was improved from about 48% to over 78%, all without operator involvement!
Other data that can be plotted include the design variables, model variables and constraints. Several examples of these are shown on the following pages.
The figure above shows the progression of the vertical scale, or height, of a riser. Note that the final height is one half( 0.500) of the original height.
This chart shows the progression of a riser horizontal scale, or diameter, through the 100 runs. The final diameter was reduced to about 78% of the original diameter.
This chart shows the constraint of Material Density. The final value was 0.9954, which was acceptable in this particular run. The minimum value specified was 0.994.
A final comparison can be made by visual inspection. The screen below shows the first and last model, side by side, so you can compare the rigging systems.
All changes to the model, meshing and simulation work was done completely without operator help, once the optimization run was set up.