optiSLang and ANSYS Maxwell

This training unit will introduce you to optiSLang's basic functionality allowing the process integration, sensitivity analysis, and optimization of Maxwell-Models.

Any Maxwell simulation model can be coupled to optiSLang via standard functions for text-based integration. On the output side, the handling of scalar parameters and signals in both Maxwell and optiSLang's calculator environment represents a central aspect. On the output side, the handling of scalar parameters and signals in both Maxwell and optiSLang's calculator environment represents a central aspect. The definitions made here determine the database upon which the visualizations of the statistics and metamodel postprocessing build up. You will learn how to conduct a sensitivity analysis in optiSLang and to interpret the graphically displayed results, e.g. to identify the most influential model parameters and to bring design conflicts into clear view. Next, you will use the gained knowledge for efficient design optimization. The procedure will be exemplified through a case study of a 2D electric motor model.

Agenda

1. Overview of the graphical user interface of optiSLang

2. Process integration

  • Direct integration of a Maxwell model via text files
  • Extraction of result variables and signal processing
  • Integration via ANSYS-Workbench 

3. Application of a sensitivity analysis

  • Definition of parameter spaces and Design of Experiments
  • Statistics postprocessing: visualization of correlations (e.g. torque-efficiency conflict)
  • Metamodel postprocessing: interpretation of visualizations and discussion of physical causes

4. Solution of an exemplary optimization task

  • Definition of objective and constraint functions taking into account knowledge gained during the sensitivity analysis
  • Exploiting the metamodel for optimization
  • Direct optimization of the Maxwell model

5. Outlook: parallelization via Large-Scale DSO

  • GUI-based DSO versus Large-Scale-DSO
  • Basic concept of using Large-Scale-DSO with optiSLang
  • Function with DoE algorithms and iterative optimizers
  • Speed-up behavior of different optimizers in comparison

As a prerequisite, basic knowledge about statistics and optiSLang is required which can be obtained by attending the training "optiSLang basics".

Details

Date:
Dec 6, 2017 in Weimar

Map

Please register not later than 14 days prior to the event date.

Time:
9am - 5pm

Fee: 
600 Euro (excluding VAT) per participant in Weimar or Vienna
2000 Euro (excluding VAT) per on-site training including maximum 12 participants

Discount:
50% for students
30% for university members / Ph.D. students