An Analyst s View STEP-Enabled CAD-CAE Integration Stephen Gordon

An Analyst’s View: STEP-Enabled CAD-CAE Integration Stephen Gordon NASA’s STEP for Aerospace Workshop JPL, Pasadena, CA January 17, 2001 SFG-GD/EB-1/17/01 - 1

Outline of Topics CAD-CAE Integration Domains Purpose, Scope, and Scale of Analysis CAD-Centric vs. CAE-Centric Processes Categories of Integration (What is truly ‘Seamless’? ) De-Featuring Geometry (Content, Amount of Detail) Geometry ‘Gender Changing’ Needs and Tools “Simulation-Specific Geometry” vs. CAD Geometry Collaboration (Inter- and Intra-Company) ISO STEP 10303 is not just about Data Exchange AP 209 is an Enabler for CAD-CAE Integration AP 209 = CAD + CAE + FEM + FEA + PDM Backup Slides SFG-GD/EB-1/17/01 - 2 (Mnemonic)

CAE Functionality Computer-Aided Engineering (CAE) is made up of many varied and diverse functions, depending upon one’s organization and it’s structure. My focus will be on the exchange and sharing of design and engineering data (largely geometry and material information) for use in computational analysis tools, specifically finite element method codes. My purpose today is to define where I see, as an analyst, the potential and growing use for ISO STEP AP 209. There is no attempt to slight many other important (nonanalytical) engineering functions in the CAD-CAE domain related to the overall design, approval, and life-cycle of the products we create. SFG-GD/EB-1/17/01 - 3

CAD and CAE Integration The key to understanding CAD-CAE Integration, is related to the scale, scope and purpose of the required engineering analysis - e. g. Finite Element Analysis (FEA). It is not simply related to the existence of captured CAD geometry, a perception unwittingly left during product model ‘walk-throughs’. The closer the scale, scope and purpose of an engineering analysis is to the type and detail of the existing CAD product model geometry, the greater the likelihood that a closely-coupled, automated, or even seamless integrated CAD-CAE process can be implemented. SFG-GD/EB-1/17/01 - 4

Scope, Purpose, and Scale Small-scale, single part optimization “Widgets & Gadgets” Large-scale simulation, e. g. vehicle crashworthiness These may both employ the same software, but with significantly different models! SFG-GD/EB-1/17/01 - 5

Large-Scale Simulation, Full Vehicle Models Part of the CAE Domain Such analyses and simulations likely include multi-physics and mediastructure interaction SFG-GD/EB-1/17/01 - 6

CAD and CAE Integration When the scale, scope and purpose of an engineering analysis are not consistent with the type and detail of the existing CAD product model geometry, a computer-assisted “man-in-theloop” or semi-automated process may be more feasible and appropriate than a fully-automated process. (Still a place for engineering judgement. ) We have found that the time and cost to change, re-work, and de-feature CAD geometry can sometimes be greater than that for creating analysis models from readily-generated “idealized” geometry. (Not a popular view in today’s world!) The more abstract, idealized geometry used for analysis is also referred to as “Simulation-Specific” or “FEA-Specific Geometry”. SFG-GD/EB-1/17/01 - 7

Analysis Model Creation (e. g. FEA) Geometry is not always the same! Change Type or “Gender” Captured CAD Geometry Simplify Idealize De-Feature Derived Idealized Geometry Pave Mesh Discretize Engr. Anal. Model (FEM) A mechanical engineer, a structural engineer, and a piping engineer may each require different forms of geometry capture. SFG-GD/EB-1/17/01 - 8

Geometry Representation 1 D Line (Curve) Same Object. . . Multiple/Different Forms of Geometry Capture SFG-GD/EB-1/17/01 - 9 2 D Surface (Shell) 3 D Solid (Volume)

Exploded View - Same Geometry, Different Data Capture 1 D Line (Curve) 3 D Solid (Volume) 2 D Surface (Shell) g” d in hang er C “Gen SFG-GD/EB-1/17/01 - 10

Integrated CAD-CAE Process The value of engineering analysis and optimization early (‘up front’) in the design process is now readily accepted and is generally unassailable. Unfortunately, there is often the perception (sometimes from MCAD vendor hype) that engineering analysis is a totally seamless process within CAD. This view can be a disservice to many sectors of business, where solid product models have become the CAD approach of choice, but the wrong geometry for analysis. Fortunately, articles in the literature have recently begun to reflect some of these different views on delivering analysis. SFG-GD/EB-1/17/01 - 11

Recent Articles Show Enlightened Views “Three-Dimensional CAD Design and Analyzing with Shell Elements - A Soluble Contradiction? ”, by M. W. Zehn, H. M. Baumgarten, & P. Wehner, NAFEMS 7 th Int’l. Conf. , Newport, RI, April 1999 “Don’t Change the Model Till the Simulation Finishes”, by Paul Kurowski, Machine Design, August 19, 1999 “Rookie Mistakes - Over Reliance on CAD Geometry”, by Vince Adams, NAFEMS Benchmark, October 1999 “Common Misconceptions About FEA”, by Vince Adams, ANSYS Solutions, Fall 2000 “Eight Tips for Improving Integration Between CAD and CFD”, by Scott Gilmore, Desktop Engineering, May 2000 SFG-GD/EB-1/17/01 - 12

“Don’t Change the Model Till the Simulation Finishes” by Paul Kurowski Machine Design August 19, 1999 When analysis geometry is not the same as design geometry! “Simulation-specific geometry”or “FEA-specific geometry” SFG-GD/EB-1/17/01 - 13

CAD-CAE Integration Whether CAD-CAE applications can be closely-integrated and automated depends upon: • The scale, scope, and purpose of the CAE analysis. • The nature and type (order, or “gender”) of the captured CAD geometry. • The amount of detail required for the CAE application. Today’s bottleneck in CAD-CAE integration is not automated mesh (grid) generation, it lies with efficient creation of appropriate simulation-specific geometry. SFG-GD/EB-1/17/01 - 14

The First “Problem” - Geometry Type In general, the most automated CAD-to-CAE processes are for MCAE “same gender” geometry classes, eg. widgets and gadgets, or 3 D volumes filled with 3 D elasticity tet or hex solid elements; or piping analysis employing 1 D geometry. In the structures discipline, our ship product lines most often involve large scale, simplified geometry and analysis models, e. g. large assemblages of stiffened 2 D plate/shell surfaces and framework structures. However, enterprise CAD product model geometry capture is fundamentally 3 D solid modeling supplemented with 1 D “structures” entities. Thin-walled structure is more accurately and efficiently analyzed as plates and shells. SFG-GD/EB-1/17/01 - 15

Same “Gender” Geometry Creating Engineering Analysis Models Geometry Type Analysis Model Type 1 D Lines, Curves Beams, Trusses Axisymmetric Shells Stiffened Plates & Shells 2 D Surfaces 2 D ‘Cut” Surfaces (or Sections) Plane Stress / Strain Elasticity Axisymmetric Solids (“Quasi-3 D”) 3 D Volumes SFG-GD/EB-1/17/01 - 16 Plates & Shells 3 D Solid Elasticity

CAD-Centric Approaches Creating Engineering Analysis Models Captured CAD Geometry CAD Structures (1 D) Idealized Geometry Analysis Model Type (Eg. FEM) 1 D Lines, Curves Beams, Trusses Axisymmetric Shells Stiffened Plates & Shells CAD Surfaces (2 D) Plates & Shells 2 D ‘Cut” Surfaces (or Sections) CAD Solids (3 D) 2 D Surfaces Plane Stress/ Strain Elasticity Axisymmetric Solids (“Quasi-3 D”) 3 D Volumes 3 D Solid Elasticity Can be ‘Seamless’ ‘Gender-Changing’ Required SFG-GD/EB-1/17/01 - 17

CAD-Centric Process with a 3 D Solid Product Model Creating Engineering Analysis Models Captured CAD Geometry Idealized Geometry 1 D Lines, Curves Requires “Gender Changing” Analysis Model Type (Eg. FEM) Beams, Trusses Axisymmetric Shells Stiffened Plates & Shells 2 D Surfaces Plates & Shells 2 D ‘Cut” Surfaces (or Sections) 3 D Solid Product Model SFG-GD/EB-1/17/01 - 18 Plane Stress/ Strain Elasticity Axisymmetric Solids (“Quasi-3 D”) 3 D Volumes 3 D Solid Elasticity

The Second “Problem” - Model Content and Amount of Detail In general, the captured CAD geometry contains a great deal of detail, necessary for creating drawings and for manufacturing support, but too much detail for most idealized FEA models. Therefore, the idealization portion of FEA requires simplifying the geometry, removing unwanted details which are not commensurate with the scale of the idealized FEA model. Examples include removing small holes, adding or removing fillets, even eliminating whole portions which may be idealized as a rigid mass, or may not be in the analysis at all! This process of simplification is sometimes referred to as “suppressing the details” or “de-featuring” the geometry. For welded structure adding features (weld fillets) to the CAD product model may be required for detailed stress analysis. SFG-GD/EB-1/17/01 - 19

Welded Thin-Walled Structure with a 3 D Solid Product Model Joint Surface Index JXXX Portion of foundation for resilient mounts and shock snubbers Weld material is annotated (e. g. weld symbols), but not explicitly captured as fillets in the product model (as it would be for a machined part) Typical de-featuring might include eliminating the small holes but keeping the larger ones. Whereas, a featuring change could be adding weld fillets to avoid stress concentrations or singularities at sharp corners. SFG-GD/EB-1/17/01 - 20

CAE-Centric Process CAD-Centric Process Categories of CAD-CAE Integration Category I - The CAD Geometry and the Simulation-Specific Geometry are the same (identical). This is the truly “seamless” case; there is no change in detail, no de-featuring, and no geometry gender changing required. Analysts and designers use the same (or duplicate copies of the same) geometry. Category II - Existing (available) CAD geometry has the wrong content; it is too detailed and/or of the wrong type to support the scale, scope, and purpose of the required or most appropriate type of analysis. Changes are required to add features or remove unnecessary detail from, and/or modify the gender of, the CAD geometry to create Simulation-Specific Geometry amenable to analysis. Automated and semi-automated procedures are required. Category III - Engineering analyses are performed first to define and refine a design concept using idealized geometry prior to establishment of the enterprise (CAD) product model. Simulation-Specific Geometry employed for analysis models will require modification and the addition of details and features to support drawings and manufacturing. Automated and semiautomated procedures are desirable. SFG-GD/EB-1/17/01 - 21

CAD-Centric Approaches Category I CAD Geometry = Simulation. Specific Geometry Start Change Type or “Gender” Captured CAD Geometry SFG-GD/EB-1/17/01 - 22 Simplify Idealize De-Feature Pave Mesh Discretize Engr. Anal. Model (FEM) Category II Simulation. Specific Geometry Engr. Anal. Model (FEM) Pave Mesh Discretize

A CAE-Centric Approach Start Modify Type or “Gender” Create CAD Geometry Add Details & Features Simulation. Specific Geometry Category III Pave Mesh Discretize Engr. Anal. Model (FEM) More mature or optimized concept prior to CAD geometry capture; designers add detail later for drawing creation, design disclosure, and manufacturing SFG-GD/EB-1/17/01 - 23

Category I Solids Examples Mechanical parts and components SFG-GD/EB-1/17/01 - 24

3 D Solids & 3 D Elasticity Analysis TET Meshing SFG-GD/EB-1/17/01 - 25 Automated model building options are readily available in almost all CAD and CAE tools HEX Paving

Categories II & III Thin-Walled Structures (Where the product model is solids) SFG-GD/EB-1/17/01 - 26

EB Example Automated Mid -Surfacing CAD-Centric Category II (Solids -to-Shells) Welded Plate Tank Structure - Multiple Brep Manifold Solids SFG-GD/EB-1/17/01 - 27

Solids 1 2 Mid-Surfaces Automated Mid -Surfacing Category II (Solids -to-Shells) SFG-GD/EB-1/17/01 - 28 Trimmed and Adjusted Mid -Surfaces

Category II 1. Thin-walled solid part 2. Mid-surface geometry 3. Meshed FEA shell model SFG-GD/EB-1/17/01 - 29 COTS capabilities now exist for automatic creation of mid-surface geometry and shell FEA mesh.

Section of Stiffened Deck Plate “DECK_ASSY” Assembly with deck plate and replicated (dittoed) tee stiffeners (Example used in Nov. ‘ 98 PDES demo using PATRAN and COMMANDS) SFG-GD/EB-1/17/01 - 30

Solid geometry Automatically created mid-surface geometry SFG-GD/EB-1/17/01 - 31

FEA Idealization #1 Explicitly Modeled Stiffeners (8 -noded shell elements shown) FEA Idealization #2 Eccentric Beam Stiffeners (4 -noded shell elements with 2 -noded eccentric beam elements) SFG-GD/EB-1/17/01 - 32

Collaboration You’ve heard a lot at this workshop about inter-company collaboration, multi-tiered supply chains, even world-wide collaboration. Large companies, such as those many of us work for, often have separate groups and departments which requires intracompany cooperation and collaboration. Integration with standards (such as ISO STEP) is a logical way to build an intra-company architecture of sharing between separate design and analysis organizations. Such an in-house, multi-department business process built on standards is easily transitioned into an external collaborative teaming endeavor, when and if that would be prudent. SFG-GD/EB-1/17/01 - 33

The ISO STEP 10303 Standards are Enablers for Improved Design-Analysis. Construction Processes Today’s business enterprises must have access to enterprise -wide PDM information which integrates design, analysis, construction and life-cycle support. AP 209 provides a means to more closely integrate design and analysis, by including nominal (CAD) geometry, various idealized CAE geometries, and associated FEM analysis models and results, along with PDM and separate version control. Mnemonic: AP 209 = CAD + CAE + FEM + FEA + PDM SFG-GD/EB-1/17/01 - 34

What is ISO STEP 10303 AP 209? Nominal CAD Geometry Idealized CAE “Simulation-Specific” Geometry Product Data Management Info AP 209 = CAD + CAE + FEM + FEA + PDM Finite Element Models Mnemonic Finite Element Analysis Controls & Results (Engineers like equations!) One can use AP 209 with any one or more of these pieces, but the real power lies with the assemblage of all these parts. SFG-GD/EB-1/17/01 - 35

Design Analysis STEP AP 209 Repository File Archived Design/Analysis AP 209 Snapshots The ISO STEP 10303 Standards are not just about data exchange! AP 209 captures and integrates design, analysis, and CM/PDM information. Design STEP AP 209 File FEA Results FEA Controls FE Models Idealized Geometry Nominal Geometry Design Model - PDM SFG-GD/EB-1/17/01 - 36 Analysis STEP AP 209 File FEA Results FEA Controls FE Models Idealized Geometry Nominal Geometry Design Model - PDM

Detailed AP 209 PDM Concepts Part Version Assembly Design Discipline Product Definition Nominal Design Shape SFG-GD/EB-1/17/01 - 37 Allow Analysis to Revise Independently of Design Analysis Design Version Relationship Analysis Version Analysis Discipline Product Definition Idealized Analysis Shape Finite Element Analysis Shape

Nominal CAD Geometry Idealized CAE Geometry Recommended Practices for AP 209 ME 007. 01. 00 June 25, 1999 SFG-GD/EB-1/17/01 - 38 FEA Model

CAD-CAE Integration Status COTS Vendor Report Card Category I A Mature, MCAD for solids good Category II B-, C+ Improving, recent mid-surfacing attention Category III D, F Very little for CAE-centric ‘leading design’, need shell ‘thickening’ tools, or ‘solids-on-demand’ Overall: Still too CAD-Centric Continued role for traditional FEA pre- and post-processors AP 209 is ready to support / enable more CAD-CAE integration AP 209 is more appropriate for CAE than AP 203 Need more vendor support for AP 209 SFG-GD/EB-1/17/01 - 39

Back-up Slides PDES & NAFEMS Activities EB’s Prototype AP 209 SFG-GD/EB-1/17/01 - 40

Working with larger-scale test cases for AP 209 coverage 3847 Nodes 6743 Elements PATRAN COMMANDS SFG-GD/EB-1/17/01 - 41

Joint PDES & NAFEMS Activity Recasting NAFEMS FEA Benchmarks into AP 209 Format NAFEMS Benchmark LE 5 Z-Section Cantilever NAFEMS Benchmark LE 1 2 D Plane Stress NAFEMS Benchmark LE 10 Thick Plate Pressure SFG-GD/EB-1/17/01 - 42

AP 209 Part 21 File Ascii AP 209 Translator Binary Electric Boat’s AP 209 Translator is Interfaced with the COMMANDS FEA System SFG-GD/EB-1/17/01 - 43

AP 209 Translator AP 209 Part 21 File Ascii Binary COMMANDS Data Base (CDB) EB’s currently implemented prototype AP 209 Translator is closely interfaced with EB’s COMMANDS Data Base using binary reads and writes. It was kept separate to enable appending multiple analysis models and results onto a single repository (Part 21 file), or selecting and extracting one model with results. Other features (next two slides) were implemented to aid developers as we learned about STEP and Express representation. SFG-GD/EB-1/17/01 - 44

SFG-GD/EB-1/17/01 - 45

SFG-GD/EB-1/17/01 - 46

Initial Checkout of AP 209 Import to COMMANDS SFG-GD/EB-1/17/01 - 47

Initial Checkout of AP 209 Import to COMMANDS SFG-GD/EB-1/17/01 - 48

EB’s AP 209 Prototype FEA Model & Results Geometry - Nominal & Idealized SFG-GD/EB-1/17/01 - 49

EB Chart 1 Common Intersection Example = HEX 20 Case Dynamic Version of the Mac. Neal -Harder Twisted Beam EB Chart 2 SFG-GD/EB-1/17/01 - 50

EB Chart 1 SFG-GD/EB-1/17/01 - 51

EB Chart 2 SFG-GD/EB-1/17/01 - 52