Archive: 2023 Telediscussions

[arthurrge]

Kick-off Call (1-JUN-2023)

Focus / Scope Discussion

• AIAA DEIC Digital Thread paper pre-print to be shared to better understand landscape (Melanie?) - "word match rubric", reference model?
• AIAA DEIC Ecosystem - HPC-related use case a candidate crossover (Graves)
• Overall DEIC goal use case "Certification by Analysis" will rely upon advanced / advancing computational methods and HPC-driven scale (realism, fidelity, UQ, etc.)
• NASA Vision2040 end state - how do the listed tools/methods (as nuts & bolts) map as enablers of the envisioned ecosystem? where are the gaps? how is the linkage accomplished? what toolsets/capabilities need to better interoperate?
• Consider: fusion of measured + synthetic data
• Examples: NOAA schema for unified forecasting, medical/cancer research sharing, others?
• Workforce: alongside tool/techniques - what is the workforce readiness / gap to fully leverage?

On the call:

• Steve Arnold (NASA)
• Rick Arthur (GE)
• Michael Belisle (NG)
• Gavan Burke (RR)
• Kuan Collins (Axient)
• Jay Ganguli (P&W)
• Christopher Juannet (Saab)
• Don Kinard (Lockheed)
• Melanie Lorang (Boeing)
• Laura Mainini (Collins)
• Ian Marks (Virgin Galactic)
• David Minor (GE)
• Abhishek Ram (Draper)
• Greg Roth (AFRL)
• Mitch Wolff (Wright State U)

Outreach needs:

• OUSD / Pentagon ~ DTO
• ACE / HPCMP ~ CREATE-AV (Scott Morton? Ryan Bond? Owen Eslinger?) - via Fran Hill?
• DOE / NNSA ~ Sandia

AIAA AVIATION Conference

• Ian Marks and Melanie Lorang - potential face-to-face discussion

Other points to add anyone?

[arthurrge]

Fall Call (8-Nov-2023)

Focus: formation & alignment

• Prioritization/scope, definitions
• Linkage, overlaps, gaps (AIAA CFD2030, USSF SDA, DOE/NNSA, DAF DTO/AFIT, HPCMP, ASME, AIAA DEIC, ...?)
• Conferences (SC|23 ~ November 13-17), (SciTech ~ January 8-12)

On the call

• Rick Arthur (GE Aerospace)
• Kuan Collins (Axient)
• Mitch Wolff (Wright State)
• Melanie Lorang (Boeing)

Actions

• (Rick) Continue to converge consensus on scope, report to AIAA DEIC (4Q23 / SciTech 2024)
• (Melanie) Outreach to CFD2030 working group (synergy in data pipelines for analytics?)
• (Kuan) Sandia contact re: USSF/SDA activities

Notes

Group participation has been sparse (both in scheduled calls and on github discussions), despite strong stated interests at SciTech 2023. Two hypotheses on potential causes:

• immaturity of technical topics (compared to more deeply-established thread/twin topics)?
• interest in "the results from others" but short on "bringing content to the discussion"?
• Strategies: move forward with survey (stimulate via inclusion/absence), leverage CFD2030, USSF/NNSA activities

General sense there are important topics on the table for AIAA, that are insufficiently covered by other working groups.

• need for aerospace industry lens on general advances in practice (computational engineering)
• need for coordination / harmonization between military and commercial roadmaps, nomenclature, standards/regulation
• potential urgencies as related to the AI Executive Order and the DOD Data, Analytics, and AI Adoption Strategy
• Strategy: survey to document (early adopter use cases/learnings) + (gaps, risks, & grand challenges)

Topping the list of prioritized topics (effectively the present scoping down-select):

Advancing Computational Methods for Aerospace Science & Engineering

(Architectures and frameworks for application of novel methods and capabilities)

• Enabling Multi-X (multiscale, multi-fidelity, co-simulation/coupled multiphysics, multi-disciplinary/transdisciplinary design + manufacturing + sustainment) modeling and analyses
• Hybrid Methods (AI/ML + Mod/Sim), (fusing synthetic + measured data), (formulaic-models + data-derived models/surrogates)
• System Co-design (of hardware platform + software implementation + domain expertise), potentially leveraging emerging capabilities in AI/LLMs (ChatGPT/Copilot) toward composability and optimization in such multidimensional trade-off spaces.
• Generative & Inverse Methods (spec desired effect, derive candidate causes), (trade-off-space mapping), etc.

Framework for Adoption

(of these Advancing Computational Methods for Aerospace Science & Engineering )

• Modeling Maturity Assessment (physics-basis, data-derivation quality, predictive accuracy, bounded confidence, numerical robustness, software engineering quality, etc.)
• Uncertainty Quantification (for data-derived models/probabilistic methods, bounding region of model competence, data requirements)
• Verification & Validation (Traceability, Reproducibility, Explainability, & Provenance) - including tracking and retiring Intellectual Debt
• Certification (gaps, risks, & grand challenges) + (impact, value, & opportunities)

Examples of important factors to ensure are included:

• Implications for both data scientists and engineers for lifecycle management (of data, models, tools) to succeed, comply, innovate upon
• Consistency for commercial and military domains
• Digital Thread as conduit of data to/from/between tools & methods
• Enterprise architecture implications (ERP, PLM, MES, MRO, etc.)
• Full lifecycle (design + manufacture + operation + sustainment)
• Modes of data (unstructured text, structured data, boundary conditions, geometry, tolerances, visual)

[arthurrge]

SciTech 2024 briefing at DEIC face-to-face:

• We have heard computational methods, ML, AI in many of the panels at SciTech 2024 already
• Karen Wilcox observation: We may moderate our perspective on AI/ML as more of an incremental computational method in our industry (as compared to its disruption in other industries) given the degree to which aerospace engineering has already adopted and matured computational methods, including intelligent/cognitive tools.

2024 Subcommittee plans

• reboot (note @aramdraper will be taking the co-lead on the subcommittee)
• interface with other committees
• generate notional table of contents for position paper
• begin to gather use cases to highlight
• gather baseline references for guidelines / assessments already matured/maturing (NAFEMS, INCOSE, DAFDTO, ASME, NNSA, ...)
• coordinate with DEIC Digital Ecosystem team to build library reference of open and commercial tools
• begin plans for a workshop at SciTech 2025
• set up rhythm for virtual meetings (at least quarterly for broad group)