Open training course 'Design optimisation'

Date: From 28/10/2013 to 01/11/2013

Location : University of Leeds

The Design Optimisation course is split into two sections that can be attended independently or together

Design Optimization, Introduction: 28 – 30 October 2013

Design Optimization, Advanced Concepts: 31 October – 1 November 2013

 

For a print friendly version of the programme, please click here. Timetable and room directions are given at the bottom of this page.

 

Speakers

Organising Lecturer: Professor Vassili Toropov

 

Guest lecturers:

  • Professor Karen Willcox (MIT)
  • Dr Jaroslaw Sobieski (NASA)

 

AMEDEO ITN lecturers:

  • Professor Harvey Thompson (Leeds). Review of metamodelling techniques, results of testing.
  • Professor Zafer Gürdal (University of South Carolina). Composite optimization.
  • Professor Fabian Duddeck (TU Munich). Optimization for crashworthiness.
  • Dr Tom Verstraete (VKI, Brussels). Multidisciplinary optimization of turbomachinery components using EAs.
  • Prof. Shahrokh Shahpar (Rolls-Royce). Turbomachinery applications of design optimization.
  • Prof. Royston Jones (Altair). Use of structural topology, shape and sizing optimization tools in aerospace industry.
  • Mr Gerald Carrier (ONERA). Some experiments with high fidelity aero-structural optimization approaches.

 

External lecturer:

  • Mr Gordon May (Rolls-Royce). Robust design optimization.

 

Design Optimization – Introduction

 

Monday 28 October 2013 Room: Electrical Engineering Agilent LT (1.52)

  • Introduction to the course, motivation for the systematic design improvement. Examples of structural optimization: sizing, shape, topology and material optimization. Examples of multidisciplinary optimization. Lessons from Nature. 
  • Criteria of design quality. Formulation of an optimization problem as a nonlinear mathematical programming problem. Choice of design variables. Continuous and discrete problems.  Choice of the objective function. Formulation of typical constraints on the system’s behaviour. 
  • Classification of design optimization problems. Constrained and unconstrained problems. Graphical solution. 
  • Optimality conditions for unconstrained optimization. Kuhn-Tucker optimality conditions for constrained optimization. Global and local optima.
  • Multi-objective problems. Pareto optimum solutions. Basic approaches to the formulation of a combined criterion. Bound formulation for minmax problems. 
  • Numerical optimization techniques. Methods for one-dimensional optimization:  bisection, Golden search, use of approximations.
  • Unconstrained multi-parameter optimization techniques. Simplex method, Nelder-Mead sequential simplex method. Hooke-Jeeves pattern search method. 
  • Gradient-based optimization. Difficulties in typical problems. Effect of scaling. Overview of: steepest descent, Fletcher-Reeves' conjugate gradient method, second order methods: Newton's Method, Quasi-Newton’s Methods. 

 

Tuesday 29 October 2013 Room: Mechanical Engineering Board Room (1.31)

  • Constrained optimization. Penalty functions. Exterior and interior penalty functions. General constrained optimization techniques: Sequential Quadratic Programming, Method of Feasible Directions. 
  • Evolutionary and heuristic optimization. Genetic Algorithms, Particle Swarm Optimization, Ant Colony Oprimization. Review of evolutionary techniques. Applications to design optimization problems (F1 front wing, structural optimization).  Permutation GA and applications to composite optimization. GA: pros and cons. 
  • Multidisciplinary optimization of turbomachinery components using Evolutionary Algorithms (Dr T. Verstraete)
  • Structural optimization and fully-stressed design. structural topology optimization. Examples of topology optimization. Fully-stressed design vs minimum weight structures. Main approaches to structural topology optimization: Homogenisation, Solid Isotropic Material with Penalization (SIMP), Ground Structure, level set, evolutionary structural optimization approaches.
  • Examples of topology optimization in fluid dynamics problems. 
  • Design sensitivity analysis. Direct and adjoint methods. 
  • Approximations for design optimization. Local, global (metamodels) and mid-range approximations .
  • Review of designs of experiments (DOEs). Uniform DOEs. Optimum Latin hypercube, nested, constrained, generalised optimum Latin hypercube DOEs. Examples of use of DOE for empirical model building (Prof. H.M. Thompson).

 

Wednesday 30 October 2013 Room: Baines Wing SR (4.12)

  • Metamodelling techniques. Polynomial response surface approximation and Moving Least Squares method (MLSM). Kriging, Review of metamodelling techniques, results of testing, pros and cons (Prof. H.M. Thompson). 
  • Metamodel assembly. Measures of metamodel accuracy. Gradient-assisted metamodelling. Metamodel-based optimization.
  • Nonlinear regression. Genetic Programming for explicit metamodel building.
  • Use of mid-range metamodels for large scale optimization problems. Multipoint Approximation Method (MAM). Examples of turbomachinery optimization applications. 
  • Multifidelty. Use of low fidelity models for global metamodel building. 
  • Multifidelity modelling for aircraft design and optimization (Prof. K. Willcox).
  • Dimensionality reduction, global design sensitivity indices. Use of MLSM for the variable importance ranking. 
  • Analysis and optimization under uncertainty. Deterministic vs stochastic optimization. 6-sigma design. Reliability, robustness. Monte Carlo simulation, first and second order reliability methods (FORM, SORM). Use of metamodels for stochastic analysis and optimization. 
  • Polynomial chaos expansion. Industrial applications: ExoMars example, JLR knee bolster example. 

 

Design Optimization – Advanced Concepts

Thursday 31 October 2013 Room: Electrical Engineering 1.53

  • Some experiments with high fidelity aero-structural optimization approaches (G. Carrier).
  • Uncertainty quantification and robust design optimization in industry (G. May, Rolls-Royce).
  • Multidisciplinary optimization frameworks. Examples of MDO problems (Dr J. Sobieski).
  • Use of structural topology, shape and sizing optimization tools in aerospace industry (Prof. R. Jones).

 

Friday 1 November 2013 Room: Mechanical Engineering Board Room (1.31)

  • Review of MDO frameworks, use of MDO in future aerospace vehicle design  (Dr J. Sobieski). 
  • Optimization for crashworthiness (Prof. F. Duddeck).
  • Turbomachinery applications of design optimization (Prof. S. Shahpar).
  • Composite optimization (Prof. Z. Gürdal).
  • Review of availability of commercial software. Conclusions and suggestions for further study.

 

Timetable

Timetable of training sessions for each day:
  • Session 1:  9:00 – 10:30
  • Session 2: 10:45 – 12:15
  • Session 3: 13:30 – 15:00
  • Session 4: 15:15 – 16:45

 

Directions to Rooms

For an interactive map of University of Leeds campus or a downloadable pdf, please click here

 

Electrical Engineering Agilent Lecture Theatre (1.52) and Electrical Engineering 1.53

The Electronic and Electrical Engineering Building is number 51 on the campus map. Enter the building from the street and cross over the seating area. Take the stairs up to the first level. Go straight ahead to the end of the corridor for Agilent Lecture Theatre (1.52). Electrical Engineering room 1.53 is just on the left before you reach the Agilent Lecture Theatre.

 

Mechanical Engineering Board Room (1.31)

The Mechanical Engineering Building is number 50 on the campus map, next to and connected to the Electrical Engineering Building. Enter the Mechanical Engineering building and take the stairs or lift to the first floor. Turn right at the top of the stairs (or left out of the lift) and go through the double doors. Mechanical Engineering Board Room 1.31 is on the right before the corridor turns left.

 

Baines Wing Seminar Room (4.12)

The Baines Wing is number 58 on the campus map. From the main street you can go through an archway between the Parkinson Building (60) and Michael Sadler Building (78). The Baines Wing entrance is to the right hand side as you walk away from the street. On entering the Baines Wing go straight ahead and turn left end of the corridor. Then go right next to the café and left following the corridor round. Helpfully Room 4.12 is signposted from here (although the name changes!). Go through two sets of double doors and take the stairs up to the right. At the first landing there is a lift to the right, this takes you straight up to Room 4.12. Alternatively, you can carry on up the stairs. Room 4.12 is on your left at the top. When making the return journey down the stairs, make sure to stop going down when the sign shows you are on Ground Floor West. Do not continue down to the Lower Ground Floor unless you wish to get lost!

Training Events