What's New in HEEDS MDO
New Features in Version 6.0
HEEDS MDO 6.0 was released in November 2011 with the following features and enhancements.
HEEDS COMPOSE
When you have a large system that you want to optimize, but the part of the model that is changing during optimization is localized to a small area (sub-system) of the full model, you can save time and find better designs with HEEDS COMPOSE.
► Learn more about COMPOSE
Collaborative Optimization
Collaborative optimization allows you to inject specific designs to the search process during a parameter optimization study. Using this feature, you can provide design ideas for HEEDS MDO to use during its search. This allows you to combine your domain knowledge and experience with the intelligent search technology in HEEDS MDO.
Benefits of Collaborative Optimization
Reliability Analysis
Reliability analysis allows you to assess the probability of failure of a design when certain features of the design will have some variability. For example, in an assembly of stamped parts, each part will have a variation in its thickness. Even though the parts were designed for a nominal thickness, the actual value of the thickness will vary with some distribution. In this case, it is important to know how many designs will fail to meet the criteria, since there will be variability from design to design.
Benefits of Reliability Analysis
New Optimization Methods: Nelder Mead Simplex and Particle Swarm
Although almost all HEEDS MDO users prefer SHERPA or MO-SHERPA for their optimization studies, we continue to add new optimization techniques to HEEDS MDO to increase the optimization methods available in HEEDS MDO for two main reasons. First, as HEEDS MDO is being used more and more in the academic world, there is a demand for additional methods so HEEDS MDO can be used as a teaching tool in the classroom, as well as in the research settings. Second, some users may prefer a certain specific optimization technique that they feel works best on their specific problem.
Augmented LHS Set
LHC (Latin Hypercube sampling) is a sampling technique that tries to distribute designs evenly in the design space. The distribution depends on the number of designs that need to be sampled in the design space. If points need to be added to an existing LHC sample, the base LHC code does not allow adding these points evenly around the existing designs. It creates a new LHC sample based on the new total number of points, and then randomly adds the number of additional points to the existing LHC sample. This can result in a bad distribution of designs in the design space along with clustering of points.
The new augmented LHC set uses a special technique to add the new designs around the existing design in the LHC sample to maintain the uniform spread of the design points. The process of adding new points evaluates the minimum distance, as well as the overall quality of the design samples, when determining locations at which to add new points.
Benefits of the Augmented LHS set
New Features in Version 5.4.1
New Features in Version 5.4
HEEDS® MDO 5.4, which was released in February 2011, introduced important new features that streamline the design optimization process for engineers.
New product name and branding
More efficient parallel design evaluations and queuing
Forming Limit Diagram (FLD) Criteria
Additional new features
HEEDS MDO 6.0 was released in November 2011 with the following features and enhancements.
HEEDS COMPOSE
When you have a large system that you want to optimize, but the part of the model that is changing during optimization is localized to a small area (sub-system) of the full model, you can save time and find better designs with HEEDS COMPOSE.
► Learn more about COMPOSE
Collaborative Optimization
Collaborative optimization allows you to inject specific designs to the search process during a parameter optimization study. Using this feature, you can provide design ideas for HEEDS MDO to use during its search. This allows you to combine your domain knowledge and experience with the intelligent search technology in HEEDS MDO.
Benefits of Collaborative Optimization
- Greater knowledge discovery and transfer
- Faster development of innovative solutions
- Combined math-based and intuition-based search intelligence
- Greater control over the search process
Reliability Analysis
Reliability analysis allows you to assess the probability of failure of a design when certain features of the design will have some variability. For example, in an assembly of stamped parts, each part will have a variation in its thickness. Even though the parts were designed for a nominal thickness, the actual value of the thickness will vary with some distribution. In this case, it is important to know how many designs will fail to meet the criteria, since there will be variability from design to design.
Benefits of Reliability Analysis
- It allows you to calculate the reliability of a design using HEEDS MDO.
- Within MDO, you can very easily switch from an optimization study to a reliability study, once a design has been identified for which the reliability needs to be calculated.
New Optimization Methods: Nelder Mead Simplex and Particle Swarm
Although almost all HEEDS MDO users prefer SHERPA or MO-SHERPA for their optimization studies, we continue to add new optimization techniques to HEEDS MDO to increase the optimization methods available in HEEDS MDO for two main reasons. First, as HEEDS MDO is being used more and more in the academic world, there is a demand for additional methods so HEEDS MDO can be used as a teaching tool in the classroom, as well as in the research settings. Second, some users may prefer a certain specific optimization technique that they feel works best on their specific problem.
Augmented LHS Set
LHC (Latin Hypercube sampling) is a sampling technique that tries to distribute designs evenly in the design space. The distribution depends on the number of designs that need to be sampled in the design space. If points need to be added to an existing LHC sample, the base LHC code does not allow adding these points evenly around the existing designs. It creates a new LHC sample based on the new total number of points, and then randomly adds the number of additional points to the existing LHC sample. This can result in a bad distribution of designs in the design space along with clustering of points.
The new augmented LHC set uses a special technique to add the new designs around the existing design in the LHC sample to maintain the uniform spread of the design points. The process of adding new points evaluates the minimum distance, as well as the overall quality of the design samples, when determining locations at which to add new points.
Benefits of the Augmented LHS set
- LHC-based robustness and reliability studies can be extended without degradation of the sample quality
- LHC-based sampling used for fitting response surfaces can be augmented with additional designs to improve the quality of the fit, if needed
New Features in Version 5.4.1
HEEDS MDO 5.4.1 was released in June 2011 with the following features and enhancements.
New portals
New portals
- Input/output portals for SolidWorks, SolidWorks Simulation and Adams
New Features in Version 5.4
HEEDS® MDO 5.4, which was released in February 2011, introduced important new features that streamline the design optimization process for engineers.
New product name and branding
- HEEDS Professional is now HEEDS MDO (Multi-Disciplinary Optimization).
New portals and portal enhancements
- Output portals for ANSYS Workbench and Nastran
- Input portals for Abaqus, ANSYS Workbench, LS-DYNA and Nastran
- Enhanced MoldFlow portal
- Time is now available in the output portals, which makes material calibration problems easy to set up.
- With HEEDS PARALLEL (formerly MPD), you can run HEEDS MDO optimization studies faster and make more efficient use of your hardware and software resources.
- HEEDS PARALLEL also offers support for interfacing with external queuing systems.
- HEEDS internal queuing (now called HEEDS Q) offers the ability to copy files across machines.
New Evaluation agent
This new agent type allows you to use the process automation capabilities in HEEDS MDO to evaluate specific designs without performing a complete study.Significant enhancements to the Analysis Manager
- Analysis success check — Eliminates the need to define additional responses to determine analysis success.
►Watch the Video: Analysis success check and FLD criteria - Analysis completion criteria — Eliminates the need to create special scripts to check for completion of an analysis.
- Unique folder for each analysis — Reduces clutter, improves file organization, prevents overwriting of output files with the same names, and allows simultaneous execution of multiple analyses.
Forming Limit Diagram (FLD) Criteria
- Allows you to define an FLD in HEEDS MDO for use as a failure criterion in manufacturing and structural performance design studies.
- Eliminates the need for complex post-processing scripts to use an FLD.
Additional new features
- Enhanced objective/constraint definition, including more control over weights and normalizing coefficients
- Bounded constraints
- Ability to define maximum analysis time for cases in which a process may "hang" indefinitely
- Response surface generation from existing designs
- Text parameter definition
- Detailed time stamp for each design
- New marking design
- Fast Save option
- Significantly improved project mobility
- User control over the format used to write variable values to the input file
Red Cedar Technology | 4572 S. Hagadorn Road, Suite 3A | East Lansing, Michigan 48823 | Phone: 517.664.1137 | Fax: 517.664.1175
Copyright © 2012 Red Cedar Technology, Inc. All rights reserved.