SimScale führt Engineering AI Agent in Onshape ein

Key Takeaways

• SimScale’s Engineering AI Agent ist jetzt in Onshape eingebettet und ermöglicht Designern, eine vollständige Physiksimulation mit einem einzigen natürlichsprachlichen Prompt zu starten.
• Der Agent übernimmt die Bereinigung der Geometrie, die Netzgenerierung, die Auswahl der Physik, das Einrichten von Randbedingungen, die Ausführung und die Ergebnisinterpretation ohne den CAD-Arbeitsbereich zu verlassen.
• Unterstützte Physik: CFD, thermisch, elektromagnetisch (EM) und strukturelle FEA, alles läuft auf SimScales cloud-native Solver-Flotte (bis zu 200 k CPU-hours pro Tag).
• Im Vergleich zum klassischen Export-zu-Solver-Workflow verkürzt der KI-gesteuerte Weg die „simulationsbereite“ Zeit um 70 % und reduziert gesamte Design-zu-Prototyp-Zyklen um bis zu 7× (wie von frühen Anwendern berichtet).
• Die Lösung wird über den Onshape App Store bereitgestellt, erfordert nur eine Ein-Klick-Aktivierung und erfüllt Unternehmens-Audit- und Daten-Governance-Richtlinien über die SimScale Agent API.

SimScale Engineering AI Agent: What It Does

Natural-Language-Driven Simulation

Ingenieure geben Eingaben wie „run a steady-state thermal analysis on this housing with a 150 °C heat source“ ein und der Agent:

1. Identifies the relevant physics (thermal, CFD, etc.).
2. Prepares the CAD geometry – removes small features, creates watertight bodies, and applies defeature rules.
3. Generates a mesh (default element size 0.5 mm for solid parts, 2 mm for fluid domains).
4. Assigns boundary conditions based on the prompt and best-practice libraries.
5. Launches the solver on SimScale’s elastic cloud (auto-scales from 2 to 64 cores).
6. Returns results directly in Onshape, with visual overlays, plots, and plain-English explanations.

Integrated App Experience

• Installation: One-click from the Onshape App Store → “Enable”.
• Workflow: The AI panel appears as a docked tab inside any Onshape document.
• Control: Users may edit the auto-generated setup before execution, ensuring engineering judgment remains central.

How It Differs From Traditional Simulation

Quantitative Impact

• Design-to-prototype: Companies such as Withings reported a 7× reduction after moving simulation earlier in the design loop.
• Simulation-ready time: Average geometry preparation dropped from 45 min to 13 min (≈ 70 % faster).
• Resource utilization: The AI agent automatically provisions compute nodes, achieving 95 % average solver CPU utilization versus ~60 % in manual batch runs.

Architecture & Governance

• SimScale Agent API: Exposes job metadata, resource allocation, and execution logs.
• Auditability: Every AI decision (physics choice, mesh parameters, BCs) is recorded, enabling traceability for regulated industries.
• Security: Data remains encrypted in-transit (TLS 1.3) and at rest (AES-256); on-premise customers can route traffic through private links.

Real-World Adoption

• User base: Over 900 k engineers worldwide already run CFD, FEA, thermal, and EM analyses on SimScale.
• Case study – Withings: By embedding the AI agent in their product development pipeline, they cut the average iteration cycle from 6 weeks to ≈ 1 week.
• Industry spread: Automotive, consumer electronics, and medical device firms are piloting the solution to accelerate early-stage validation.

Bottom Line

SimScale’s Engineering AI Agent transforms the traditionally fragmented simulation workflow into a seamless, conversational experience inside Onshape. By automating geometry cleanup, meshing, physics selection, and result interpretation, it slashes set-up time by up to 70 % and enables design teams to validate concepts days instead of weeks. The cloud-native, API-driven architecture ensures scalability, auditability, and enterprise-grade security, making it a compelling option for any organization looking to embed simulation deeper into the design decision loop.