1. MARCUS SANDBERG, marsan@ltu.se A knowledge-based master modelling approach for whole engine design Marcus Sandberg Luleå University of Technology

2. MARCUS SANDBERG, marsan@ltu.se Some challenges of jet engine development Mass Fuel consumption Development cost Performance

3. MARCUS SANDBERG, marsan@ltu.se System-level performance SYSTEM LEVEL

4. MARCUS SANDBERG, marsan@ltu.se OEM vs. component developers Continuous updates on configuration and design at system level not readily available Need for whole engine models for component developers OEM Component developer

5. MARCUS SANDBERG, marsan@ltu.se AS-IS: Unlinked simulation models

6. MARCUS SANDBERG, marsan@ltu.se NFFP5 - METOPIA Mechanical whole engine conceptual design and analysis –METhod for OPtimization Integration and Automation Continuation of the pilot project NFFP4 Mechanical whole engine modeling Luleå University of Technology Volvo Aero Corporation

7. MARCUS SANDBERG, marsan@ltu.se METOPIA research questions How to integrate the product definition with analysis models to enable effective mechanical optimisation of different jet engine architectures? How can a common information be distributed effectively between several disciplines during whole engine system optimisation?

8. MARCUS SANDBERG, marsan@ltu.se State-of-the-practice VITAL, FP6 VIVACE, FP6 CRESCENDO, FP7

9. MARCUS SANDBERG, marsan@ltu.se TO-BE: Knowledge-based master model approach

10. MARCUS SANDBERG, marsan@ltu.se Knowledge based engineering (KBE) Fundamental concept of the Master-model approach KBE-definition by Stokes (2001): –“the use of advanced software techniques to capture and re-use product and process knowledge in an integrated way” KBE aims at making engineering design more effective by –Automating repetitive CAD/CAE tasks –Showing design change implications on downstream activities

11. MARCUS SANDBERG, marsan@ltu.se The State-of-the-art Semi configurable finite-element models –Lacking effective integration with e.g. the CAD-definition Common limitation –Either stand-alone integrated solutions or –supports only domain specific applications Opportunity for modelling methods that handle and create models for multidisciplinary applications Knowledge-based master models: an upcoming topic

12. MARCUS SANDBERG, marsan@ltu.se METOPIA technology Fully automated design and analysis process Analysis –Weight –Rotordynamics –Displacement

13. MARCUS SANDBERG, marsan@ltu.se METOPIA geometry

14. MARCUS SANDBERG, marsan@ltu.se Automatic geometry configuration Number of struts Thicknesses Radii Lengths Cone angles Mount lugs

15. MARCUS SANDBERG, marsan@ltu.se Automated design and analysis Automatic geometry generation and weight analysis Automatic finite element model generation Automatic Rotordynamics analysis 1-D cylindrical beam elements Automatic displacement due to rotordynamical loads analysis

16. MARCUS SANDBERG, marsan@ltu.se Gate way Execute (through Menu) startPartGeomMassUniteMeshConstrainStructure.vb Adv sim.fem-file Adv sim.sim-file Modeling.prt-file Change window to.fem-file (VB) Find spider mesh nodes (VB-KF-Matlab) Add bearing nodes (VB) Add spider mesh (VB) Add material (VB) Write input file.dat (VB) Change window to.sim-file (VB) Edit.dat-file for Stiffness matrix output (VB-KF-Matlab) Add max forces to mesh in Solution 1 (VB) Solve Solution 1 (VB) Prepare for Add_bearing_forces.vb (VB) Add output reqs to Solution 4 and Solve Solution 4 (VB) Change window to.fem-file (VB) Change window to.sim-file (VB) Run Nastran (VB-KF) Assemble stiffness matrix solve rotordynamics and generate VB for dynamic forces (VB-KF-Matlab) Journal chain

17. MARCUS SANDBERG, marsan@ltu.se Siemens PLM NX - Matlab Knowledge fusion (KF) –Generate geometry –3D meshing –Constraints –Start Matlab- executables –Start Nastran for stiffness matrix –Calculate weight Matlab –Write journals –Edit Nastran input file –Find spidermesh nodes –Assemble stiffness matrix Journals –Add childrule (KF) –1D meshing –Add material –Write Nastran input file –Add rotordynamical loads –Switch windows

18. MARCUS SANDBERG, marsan@ltu.se Knowledge fusion classes Geometry –17 classes Mass Meshing and constraints –8 classes

19. MARCUS SANDBERG, marsan@ltu.se Benefits Develop components to optimise whole engine Automatic design Get boundary load more frequently –Swift change of model according to OEM requirements One product definition to rule them all!

20. MARCUS SANDBERG, marsan@ltu.se Ongoing work Optimisation 3D rotordynamics Validation More models More geometry Architectures

21. MARCUS SANDBERG, marsan@ltu.se Thank you!