Introduction to Ansys LS-OPT
Overview
This course overviews using the optimization code, LS-OPT, for design. It covers both theoretical concepts and practical aspects of design optimization. An emphasis is placed on interfacing LS-OPT with LS-DYNA. The course includes work¬hop sessions in which the covered theoretical topics are applied. The LS-OPT graphical user interface is used to teach input preparation and post-processing.
Over the duration of the class, you will work individually (some¬times in groups of 2) to solve the exercises. The exercises are simple, so that the run times are short, but contain enough complexity to give insight into the optimization process. Most of the problems are non¬linear dynamic and will be solved using LS-DYNA.
Learning Outcome
Following the completion of this course, you will be able to:
- Apply fundamentals and theoretical concepts of design optimization and metamodeling for industrial applications.
- Setup and solve nonlinear design optimization problems using LS-DYNA.
- Apply sensitivity analysis to screen out insignificant design parameters.
- Apply design optimization techniques for multi-disciplinary and multi-objective problems.
- Calibrate unknown material model parameters by matching computed data to target experimental data.
Prerequisites
- An introductory class in LS-DYNA is recommended but not necessary.
Target Audience: CAE Engineers, Mechanical Engineers including Automotive Engineers, Aerospace Engineers, Biomechanical Engineers, and other subdisciplines, Students, and academic researchers.
Teaching Method: Lectures and workshop sessions to apply theoretical knowledge to practical examples. A major emphasis is placed on teaching by software demonstration and on the development of a solution to a design challenge.
Learning Options: Training materials for this course are available with a Ansys Learning Hub Subscription. If there is no active public schedule available, private training can be arranged. Please contact us.
Agenda
SUBSCRIBE TODAY
Topics Covered :
- LS-OPT overview
- Optimization fundamentals
- Direct simulation-based optimization
- Metamodeling theory
- Polynomial response surface methodology
- Experimental design
- Advanced metamodels
- Metamodel accuracy and error analysis
- Simple optimization with LS-DYNA stage
- Setting up a simple optimization with LS-DYNA stage from start
- Sampling, metamodeling and stage options
- LS-DYNA interface features, such as ASCII database, binary database, filtering, time history functions, injury criteria
- Composite functions
- Simple design optimization formulation
- Program execution
- Database and output
- Post-processing using the viewer, such as simulation & approximation results, optimization history, etc.
- Repair options, discrete optimization, importing user-results
- Sensitivity analysis
- Metamodel-based optimization strategies
- Optimization with user-defined stage/solver
- Modal analysis and Multidisciplinary design optimization (MDO)
- Shape optimization
- Material parameter estimation
- Theory- Curve similarity measures
- Setting up, running, and post-processing material parameter estimation examples
Agenda :
This is a 2-day classroom course covering both lectures and workshops. For virtual training, this course is covered over 4 x 2-hour sessions lectures only.
Virtual Classroom Session 1 / Live Classroom Day 1
Module 1 – LS-OPT overview
- LS-OPT capabilities
- Design process setup and flexibility
- Optimization fundamentals
- Gradient-free optimization using genetic algorithm
- User interface, setup, and results
- Workshop 1.1 – Direct optimization
Virtual Classroom Session 2 / Live Classroom Day 1
Module 2 – Metamodeling
- Polynomial response surface methodology
- Advanced metamodels
- Metamodel accuracy and error analysis
- Workshop 2.1 – Simple metamodel-based optimization and results
- Workshop 2.2 – Run from scratch
- Workshop 2.3 – Repair optimization
- Workshop 2.4 – Discrete optimization
- Workshop 2.5 – Import user-results
Virtual Classroom Session 3 / Live Classroom Day 2
Module 3 – Sensitivity analysis
- Linear ANOVA
- Global sensitivity analysis
Module 4 – Metamodel-based optimization strategies
- Sequential optimization
- Sequential optimization with domain reduction
- Workshop 4.1 – Sequential optimization with domain reduction
- Workshop 4.2 – Working with non-LS-DYNA solvers
- Workshop 4.3 – Using dependent variables
Module 5 – Model analysis and multidisciplinary optimization
- Mode tracking
- Variable deactivation
- Workshop 5.1 – Modal analysis and tracking- DOE (optional)
- Workshop 5.2 – Modal analysis and tracking- optimization (optional)
- Workshop 5.3 – Multidisciplinary design optimization
Virtual Classroom Session 4 / Live Classroom Day 2
Module 6 – Shape optimization examples
Module 7 – Material parameter estimation
- Mean square error, partial curve mapping
- Dynamic time warping (DTW) for handling noise
- Workshop 7.1 – Ordinate-based mean square error (MSE)
- Workshop 7.2 – Ordinate-based (MSE) for multiple cases
- Workshop 7.2 – Point-based MSE (optional)
- Workshop 7.2 – Hysteretic response- Multiple cases
- Workshop 7.2 – GISSMO failure model (shear load case)