Generative Design: Your AI Partner in Product Development Date: Thursday, November 29, 2018
Webcast Logistics No Sound This is a streaming audio event. Make sure your speakers are on. Questions Type your Questions in the Ask A Question text box at anytime. Feedback Please complete the survey at the conclusion of the event Resources Download a PDF of the presentation and valuable content from the Resources box Technical Problems First Press your F5 key to refresh your browser. Click the Help icon below.
Participants Keith Meintjes Executive Consultant Simulation & Analysis CIMdata Tony Abbey Columnist FEA trainer Consultant Carina Maurer Product designer Moderator: Kenneth Wong Senior Editor Digital Engineering
CARINA S OPENING REMARKS Please submit your QUESTIONS now. Please complete our FEEDBACK SURVEY Carina Maurer Product designer Don t forget to check out our RESOURCES
RETAIL EXHIBITION Reuse of waste
Tool 40 % decrease in weight
Parametric swing Organic shape made of surfaces
Ski goggles Reduction of waste Easy to recycle
TONY S OPENING REMARKS Please submit your QUESTIONS now. Please complete our FEEDBACK SURVEY Tony Abbey Columnist FEA trainer Consultant Don t forget to check out our RESOURCES
TONY S OPENING REMARKS Traditional FEA structural simulation Target sign-off of existing design Requirements high accuracy Occurs downstream Generative Design using FEA Target explore radically new designs Requirements product improvement/innovation Occurs upstream
KEITH S OPENING REMARKS Please submit your QUESTIONS now. Please complete our FEEDBACK SURVEY Keith Meintjes Executive Consultant, Simulation and Analysis CIMdata Don t forget to check out our RESOURCES
CIMdata s Mission CIMdata is the leading independent global strategic management consulting and research authority focused exclusively on PLM and the digital transformation it enables. We are dedicated to maximizing our clients' ability to design, deliver, and support innovative products and services through the application of PLM.
Keith s Opening Remarks Ideas for generative design were developed in the late 1980s Adoption was low, except in a few cases like gauge BIW optimization and essentially 2D structures like sheet metal and brackets Maturing of additive manufacturing has led to a resurgence of interest It is NOT simply that we can now manufacture parts we could not make before
Generative Design Integrated performance-driven generative design systems are aimed at creating new design processes that produce spatially novel yet efficient and buildable designs through exploitation of current computing and manufacturing capabilities. Prof. Dr. Kristina Shea (ETH Zurich) K. Shea et al, Towards integrated performance-driven generative design tools, Automation in Construction 14 (2005) 253 264.
Traditional Design Process Previous Designs (History) Design Space (Context) Design Constraints Create Design (CAD) Design Iterations Issues Performance Requirements Evaluate Design (CAE) Valid Design Build
Generative Design Process Previous Designs (History) Design Space (Context) Design Constraints Topology Optimization via CAE Performance Requirements Document Design (CAD) Valid Design Build
Generative Lives in a Sea of Others Generative Design Advanced Materials Additive Manufacturing alone or together will enable incredible innovation in years to come aided by other process and IT technologies 1 AI and Machine Learning Generative Design 4 2 3 Industry 4.0
Integrated in the Lifecycle Digital Thread Portfolio Management Requirements Planning Conceptual Design Product Engineering Manufacturing Engineering Information: Re-use Re-purpose Re-mfg. Re-cover Re-cycle Re-tire Disposal & Recycling Maintenance & Repair In-service Operation Sales & Distribution Test & Quality Simulation & Validation Build and Produce
Alpha Go: AI and Machine Learning Go is an oriental game where tokens are placed on a grid It is more complicated than chess The best Go computer program could beat most players most of the time It learned from a data base of 100,000 games played by humans Alphabet (Google) then built Alpha Go It learned by playing against itself It was given only the rules of the game In a few days it could beat almost all players (and other computer programs almost all of the time By not using human data - by not using human expertise in any fashion - we've actually removed the constraints of human knowledge. https://www.theverge.com/2017/10/18/16495548/deepmind-ai-go-alphago-zero-self-taught
Defining Generative Design The Elbow Chair, designed in Autodesk Fusion 360. Image courtesy of Autodesk.
Generative Design Terms - 1 Parameter optimization Dimensions as design variables Shape changes Formal optimization methods Topology optimization Design space defined Topology changes Heuristic optimization methods
Generative Design Terms - 2 Lightweighting A wonderful all-embracing term! Avoids implication of the optimum Suggests a more robust engineering solution But an understrength initial design will add weight!!
WORKING WITH GENERATIVE DESIGN Generative design with lattice structure in Frustum Generate. Image courtesy of Frustum.
Designers and optimization - 1 Is design space too tight? Base could be wider Is design space too wide? Parametric too much search area Min Nom Max 0.1 1.0 10.0 Topology redundant mesh to eliminate
Designers and optimization - 2 Are these realistic loads and boundary conditions? design space axial load bending load torsional load combined loads Missing load cases really bad news
Designers and optimization - 3 Are we asking the right questions? The answer to life the universe and all that? 42 Are we asking too many questions? Avoid becoming a one-person committee With too many variables to handle
Designers and optimization - 4 Second guess the solution why this design? Aim for a robust engineering solution Not an unrealistic optimum
WHO IS IT FOR? HOW DOES IT WITH MANUFACTURING Generative design with manufacturing constraint in Siemens NX. Image courtesy of Siemens PLM.
Who should use generative design? Generative Design for all! Play to your strengths in: Design Analysis Manufacturing
Q&A Keith Meintjes Executive Consultant, Simulation and Analysis CIMdata Please submit your QUESTIONS now. Tony Abbey Columnist FEA trainer Please complete our FEEDBACK SURVEY Don t forget to check out our RESOURCES Carina Maurer Product designer
TONY S CLOSING REMARKS Please submit your QUESTIONS now. Please complete our FEEDBACK SURVEY Tony Abbey Columnist FEA trainer Don t forget to check out our RESOURCES
Closing remarks Structural analysis optimization in a renaissance at present Topology Optimization - Generative Design rapid prototyping is the catalyst exciting. After languishing as a specialist analyst tool structural optimization is now back in the mainstream Traditional methods sizing and shape tedious, difficult.
Closing remarks On my Christmas list Realistic manufacturing simulation Realistic stress constraints Optimizing structural supports Nonlinear response Linking Topology and Parametric Optimization
Closing remarks Tony Abbey contact details: Email tony@fetraining.com FEA Resource website www.fetraining.net LinkedIn www.linkedin.com/in/tonyabbey/
CARINA S CLOSING REMARKS Please submit your QUESTIONS now. Please complete our FEEDBACK SURVEY Carina Maurer Product designer Don t forget to check out our RESOURCES
KEITH S CLOSING REMARKS Please submit your QUESTIONS now. Please complete our FEEDBACK SURVEY Keith Meintjes Executive Consultant, Simulation and Analysis CIMdata Don t forget to check out our RESOURCES
Thank you! Keith Meintjes Executive Consultant, Simulation and Analysis CIMdata k.meintjes@cimdata.com Tony Abbey Columnist FEA trainer tony@fetraining.com Carina Maurer Product designer carina@maurer-missen.de Moderator: Kenneth Wong Senior Editor Digital Engineering