DOES - An ANSYS Example


	*Design Optimization Expert System*



*Free*		*Free*

Documentation
Case Study
Commercial CAE Product Interfaces

DOES On-Line Documentation DOES User Manual (pdf)

ANSYS Virtual Engines

ANSYS Virtual Engines

And ANSYS Workbench Model Optimization Part 2: Optimizing with DOES Download the complete Case Study here (PDF)
To optimize any model with DOES one must first use its Knowledge Engineering user interface to create an Expert Design. The Expert Design defines the objectives and constraints and the model parameters that can be varied to create optimized designs. Using DOES' intelligent ANSYS tool partner interface, the process of loading information from the Flexible Spring Mount ANSYS Workbench Project into the Reference Run of the Expert Design is easy. The tool partner interface recognizes the type of model, automates many of the definition tasks and provides all of the input and output variables associated with the ANSYS model. Next the engineer creates the Expert Design’s Task which specifies the Objectives and Constraints of the Expert Design. DOES supports multiple Objectives; however, in this study the only Objective was to create a mount with the largest amount of deflection in the –Y direction. Remember that this Objective must be achieved subject to the Constraint that the maximum equivalent stress in the model must be greater than 1 and less than 2050 MPa. Once the design Task’s Objectives and Constraints are defined it is necessary to identify the Application Model parameters which can be varied in order to achieve the design’s Task. In this study we optimized the values of five (5) design parameters contained in the Application Model. The window below shows the Design Space Variables and the ranges for their values. Optimizing the Spring Mount
	Once all of the desired input Design Space Variables have been specified, optimizations can be launched by clicking on the Green Arrow to obtain improved designs. The Optimize dialog box controls the amount of computing that will be performed to optimize the design which in turn determines how long the optimization will run and normally how good the designs will be. The default Optimize parameters, shown in the box to the left were used to obtain the initial results for this study. Exploration Power – determines the level of the Design of Experiments (DoE) that DOES uses during the 1st stage, of the optimization process. If it is not zero, DOES uses its proprietary DoE algorithms to locate the most promising solution areas within the design space. If it is zero, NO exploration is performed. Refinement Power – is the number of levels of refinement that DOES applied to each Solution that was created during Exploration. Refinement is the second stage of the optimization process. Local Opts – determines the number of improved designs that DOES will attempt to create. The first Local Opt is the best new design. The second Local Opt is the second best design, etc.
Page 3: Results of the DOES Optimized Model ->

Global-OPT, Multi-Objective-OPT, and DOES have been developed by OPTIMUM Power Technology. For more information contact:
500 Miller's Run Road • P.O. Box 509 • Morgan, PA 15064 • Phone: 412-257-9070 • Fax: 412-257-9011 • sales@optimum-power.com