Structural Optimization in Finite Element Analysis (FEA)

Duration:	1.5 days
Delivery:	E-learning Onsite Classroom
Language:	English
Level:	Mid-level
Availability:	Worldwide
Tutor(s):	Tony Abbey
	Request Full Details

Achieve meaningful structural optimization.

Finite Element Analysis has emerged has a tool that can play a vital part in the drive towards the ultimate goal of any manufacturing process; to produce the most effective products in the most efficient manner. This simple statement embraces all of the ‘right first time’, ‘minimum design to test cycles’ and other practices that have evolved.

The introduction of a formal structural optimization strategy into this process has met with great success in many industries. It makes the creation of the most effective product that much more attainable.

Traditionally one might think of the Aerospace Industry as the classic example with the goal of keeping weight to a minimum. Indeed the structural efficiencies of modern aircraft owe a lot to optimization methods. However it would be wrong to think of this as always a strength and stiffness against weight minimization task. The interaction of Aerodynamics, Aeroelasticity, Structures, Performance, Operating Cost and many other disciplines all have to play a role in the overall vehicle design.

This gives the clue as to the broader nature of structural optimization across all industries. The objective does not need to be weight minimization. It could be, for example driving down the overall vibration amplitude of a hairdryer, whilst keeping away from unpleasant harmonic frequencies. Weight has still to be monitored, and we can place an upper limit on this – but the other factors are more important and will feature directly in the optimization analysis.

Similarly other disciplines can play a role in structural optimization. In the case of pump housing, we want this to be stiff and strong enough to do the job, with minimum weight. However the cost of manufacture is important so a parametric penalty function can be introduced which ‘steers’ the weight reduction to a compromise solution which is cheaper to machine.

How do we define the penalty function in the above case? Well, that’s where the ingenuity of the analyst comes in! Knowing how to set up the optimization task and how to obtain innovative solutions with the tools provided is a key to success in FEA Structural Optimization.

The objective of this course is show you a broad overview of the range of FEA based tools available and what the methods and specializations of each encompass. Plentiful hints and tips will demonstrate powerful ways to use these methods. The goal is to achieve meaningful structural optimization in support of the most effective products.

Course Program

Part 1

Finite Element Analysis Overview
Background and History of Structural Optimization
Putting Optimization in perspective
The Goals of Optimization
Terminology, Definition and Classification
The upside and the downside of Optimization
Overview of Optimization Categories applied to FEA
Sizing
Shape
Topology
Discussion of internal FEA optimizers and external optimizers
Difference in Approach
Advantages and Disadvantages
Overview of Optimization Strategies
Optimality Criteria
Gradient based methods
Design Sensitivity and approximate solutions
Homogeneous Stress or Energy solutions
Design Of Experiments, Genetic Algorithm and similar methods
Some simple Case Studies to illustrate the concepts

Part 2

Theoretical background to Optimization
Implications for Practical FEA implementation
A closer look at Sizing Optimization
Background theory
Case Studies in Sizing optimization
A more sophisticated approach to objectives, variables and constraints
Linking Design Variables
Practical Gauge Constraints
Complex Responses
Response functions as Objectives
Compound Objectives
Practical Hints and Tips
Case Studies of the methods

Part 3

Shape optimization in detail
Parametric and Nonparametric issues
Traditional gradient based approaches
Homogeneous methods
DOE, GA and similar methods
Improving practicality of results
Practical hints and tips
Case studies in shape optimization
Topology Optimization in detail
Parametric and Nonparametric issues
Interface with CAD and production – concept study or practical design?
Review of methods available
Practical hints and tips
Case studies in topology optimization

Part 4

Multi Objective Methods
Background Theory
Multi-Disciplinary Optimization (MDO)
Case Studies in MDO
Optimization of Nonlinear and Dynamic Response systems
Case Studies in Nonlinear and Dynamic Response
Robust Optimization – moving away from the one point solution
Background theory and case studies for Robust Optimization

Who Should Attend?

Practicing engineers who wish to learn more about how to apply the various optimization methods available to FEA structural analysis in the most effective manner.

Interested?

Get in touch to discuss your next steps with our experienced training team. We can work closely with you to understand your specific requirements, cater for your specific industry sector or analysis type, and produce a truly personalised training solution for your organisation.

All NAFEMS training courses are entirely code independent, meaning they are suitable for users of any software package.

Courses are available to both members and non-members of NAFEMS, although member organisations will enjoy a significant discount on all fees.

NAFEMS course tutors enjoy a world-class reputation in the engineering analysis community, and with decades of experience between them, will deliver tangible benefits to you, your analysis team, and your wider organisation.

Find out more

PSE Competencies addressed by this training course

ID	Competence Statement
OPTpr4	Familiarity with at least two of the traditional problem definition methods such as Simplex methods, Linear Programming, Geometric Programming, Quadratic Programming
OPTpr5	Familiarity with gradient search methods such as steepest descent
OPTpr6	Understanding of unconstrained and constrained strategies
OPTpr7	Understanding of at least one of the -modern-methods of search strategy such as Neural Networks, Genetic Algorithms, etc.
OPTpr8	Ability to carry out Linear Static Analysis or similar level of analyses in other core disciplines and produce validated results
OPTpr9	Thorough awareness of effects of bad modelling practice and need for adequate checking
OPTpr10	Awareness of difference between global and local minima
OPTpr11	Awareness of parametric controls such as CAD geometry dimensions.
OPTkn1	List the various steps in a general optimisation study.
OPTkn2	List the various types of optimisation search algorithms available in the system(s) you use.
OPTkn3	State whether the optimisation system(s) you use are controlling CAD geometry or finite element parameters (or both).
OPTkn4	State the maximum problem size recommended for your optimization tool in terms of design variables and constraints
OPTkn5	Define the convergence criteria used in your optimization tool for establishing an optimum
OPTkn6	List some direct and indirect methods used for the optimum solution of a constrained nonlinear programming problem.
OPTkn7	State whether your system can handle multiobjective functions
OPTkn8	Outline via a sketch a typical 2 variable optimization problem using variables as x and y axes and show objective function and constraints on the sketch.
OPTkn9	State if linearization of local design space can be used during an optimization with your system
OPTkn10	List which Artificial Intelligence based approaches that your system uses
OPTkn11	State whether your system can deal with discrete variables as well as continuous variables
OPTkn12	State whether your system can define objective functions of more than one term, such as weight AND cost
OPTkn13	List the various methods of establishing feasible search directions
OPTkn14	List methods which transform constrained problems to unconstrained problems
OPTkn15	Define a discrete design variable
OPTco1	Explain the terms goal (objective function), variable and constraint.
OPTco2	Explain why an optimum solution is not always a robust solution.
OPTco3	Describe the basic methodology used to achieve shape modification in any system(s) you use.
OPTco4	Describe the basic methodology used to create structural holes in any system(s) you use.
OPTco5	Explain the concept and usage of a Pareto Set.
OPTco6	Explain the concept of Objective Space and Design Space.
OPTco7	Explain the terms local minima, global minima and saddle point.
OPTco8	Describe the advantages and disadvantages of the search algorithms available in the software tools you use.
OPTco9	Describe Basis Vector methods to reduce the number of design variables
OPTco10	Describe Design Variable Linking
OPTco11	Describe the Kuhn Tucker conditions
OPTco12	Explain how you would investigate the design evaluation trends shown by your software using GUI based graphs, tables tec.
OPTco13	Describe how you would confirm that the optima found is not a local minima
OPTco14	Explain the importance of the definition of the applied loading case set to be used in the optimization
OPTco15	Describe the process to take the optimum solution found and map it into a practical CAD design
OPTco16	Describe how you would review the final design to understand what the main driving parameters are
OPTco17	Describe what steps you may take to understand why an optimization problem will not converge to a solution and how to improve the strategy
OPTco18	Discuss how important it is to find the absolute minima relative to practical limits on design, manufacturing etc.
OPTco19	Define the difference between sizing, shape and topology optimization
OPTco20	Describe how linearization of design space is used, with pros and cons
OPTco21	Explain the difference between parameter based and non-parameter based optimization and where each is most effective
OPTco22	Discuss why mutation in a gene pool is important in a Genetic Algorithm
OPTco23	Describe the difference in approach to an objective function between topology optimization and sizing optimization
OPTco24	Describe what is meant by a stochastic approach to optimization
OPTco25	Describe what is meant by an optimality criterion based method and give an example
OPTco26	Describe typical ways of dealing with discrete variables and their pros and cons
OPTco27	Describe a typical multi-term objective function and mention any drawbacks with this approach
OPTco27b	Discuss the importance of an accurate baseline FE Analysis with validated results as the starting point for optimization
OPTco28	Describe parameter linking in a design variable with pros and cons
OPTco29	Describe synthetic type constraints created from multiple responses with pros and cons
OPTco30	Explain why an optimum solution may actually violate one or more constraints
OPTco31	Discuss and sketch what is implied by a "best infeasible" solution
OPTco32	Describe the terms mean and standard deviation
OPTco33	Describe the Normal Probability distribution
OPTco34	Describe the method of Genetic Algorithms
OPTco35	Explain the process of Neural Network based optimization
OPTco36	Describe the training phase of a Neural Network
OPTco37	Describe the Quasi-Newton root finding method
OPTco38	Describe the Secant root finding method
OPTco38b	Describe Convex and Non-Convex sets
OPTco39	Explain the difference between a gradient based and non-gradient based search method
OPTco40	Describe how various optimization strategies such as Shape, Sizing, Topology may be combined in a single project
OPTco41	Describe DOE usage in optimization and how the resulting surface model may be used
OPTco42	Describe the various DOE search strategies
OPTco43	Describe Topometry optimization and its relationship to shape optimization
OPTco44	Describe Topography optimization and how it is used in the overall optimization process
OPTco45	Describe the implications of non-linear Optimization
OPTco46	Explain the design variable and Objective function options available to Composite Structural Analysis as opposed to Isotropic Structural Analysis
OPTco47	What special care is needed when carrying out optimization of Composite Structures
OPTap1	Employ available software tools to carry out parameter, shape and topology optimisation studies.
OPTap2	Use appropriate software tools to carry out multidisciplinary optimisation studies, if relevant.
OPTap3	Conduct sensitivity studies to inform optimisation studies.
OPTap4	Utilise appropriate and efficient optimisation algorithms, where a choice is given.
OPTap5	Demonstrate the definition and execution of an optimization task, starting with a baseline FE Analysis
OPTap6	Conduct an optimization analysis of a composite based structure
OPTan1	Analyse the results from sensitivity studies and draw conclusions from trends.
OPTan2	Determine whether the results from an optimisation study represent a robust solution.
OPTan3	Determine whether an optimization study should use discrete variables and the practical benefits gained from this approach
OPTan4	Determine the best design variables and optimization technique to use for composite structures
OPTsy1	Plan effective analysis strategies for optimisation studies.
OPTsy2	Formulate a series of simple benchmarks in support of a complex optimisation study.
OPTsy3	Plan an evaluation study for a new optimization tool to brought into your operation
OPTsy4	Create a process to take an FE based optimum design and evolve into a practical CAD design
OPTsy5	Formulate a check list of do's and dont's for setting up a realistic optimization problem, include practical, logistic and FE solver and optimizer specific issues
OPTsy6	Prepare an overview of your complete optimization process from concept to product
OPTsy7	Describe how your company uses optimization and recommend areas for improvement
OPTsy8	Describe a range of ideas for objective functions, other than weight minimization, with practical examples
OPTsy9	Describe the technical and resource management issues associated with Multidisciplinary Optimization
OPTsy10	Create a presentation to give at Management Level to justify a major purchase and implementation of Optimization in the design and manufacturing process
OPTev1	Justify the appropriateness of goals, constraints and variables used in an optimisation study.
OPTev2	Select suitable idealisations for optimisation studies.
OPTev3	Provide effective specialist advice on optimisation to colleagues.
OPTev4	Assess appropriate hardware and software solutions to meet the needs of planned optimisation studies.
OPTev5	Justify an optimum design configuration by comparing with initial solution and simple variations or information from the optimization tool.
OPTev6	Justify an optimum design based on its applicability to manufacture and assembly
OPTev7	Assess the application and effectiveness of using EXCEL Solver, MATLAB, open source or programmatic in-house solutions to an optimization problem as an alternative to COTS

Special Note(s):

Telephony surcharges may apply for attendees who are located outside of North America, South America and Europe. These surcharges are related to individuals who join the audio portion of the web-meeting by calling in to the provided toll/toll-free teleconferencing lines. We have made a VoIP option available so anyone attending the class can join using a headset (headphones w/ microphone) connected to the computer. There is no associated surcharge to utilize the VoIP option, and is actually encouraged to ensure NAFEMS is able to keep the e-Learning course fees as low as possible. Please send an email to the e-Learning coordinator (e-learning @ nafems.org ) to determine if these surcharges may apply to your specific case.

Just as with a live face-to-face training course, each registration only covers one person. If you plan to register a large group (5+), please send an email to e-learning @ nafems.org in advance for group discounts.

For more information, please email e-learning @ nafems.org .

Not Available to Attend this Time?

Would you like us to notify you when the next course on Fluid Dynamics Review for CFD is open for enrollment? If so, add yourself to the Waitlist!