human-in-the-loop finite element modeling

This study addresses the low automation and inefficiency in finite element modeling of safety-critical infrastructure such as bridge barriers, which remains heavily reliant on manual effort. To overcome these limitations, the authors propose HELM, a human–AI collaborative framework that decomposes the modeling process into discrete, visually verifiable steps—including geometry generation, boundary condition definition, and material assignment—and integrates ANSYS with LS-PrePost. By introducing a novel structured collaboration protocol, the framework strategically incorporates human intervention to compensate for AI agents’ deficiencies in spatial reasoning and algebraic logic, substantially enhancing modeling reliability. Experimental results demonstrate that the success rate of complete model generation improves from 20% to 75%, with near-doubling in agent pass rates for geometry and boundary condition tasks. The project’s code and prompt library have been made publicly available.

This work proposes an end-to-end automated framework that generates compliant engineering reports directly from a single image of a mechanical component. The approach employs a solver-agnostic multi-agent system operating within a shared contextual space, leveraging a quality-gated conditional iteration mechanism to collaboratively perform geometric reconstruction, material inference, adaptive mesh generation, multi-case finite element analysis, and code compliance assessment. A unified uncertainty quantification framework is innovatively formulated by integrating interval analysis, probability density functions, and fuzzy logic, complemented by task-dependent conservativeness criteria to reconcile conflicting multi-limit-state requirements. Demonstrated on a single photograph of an L-shaped steel bracket, the system autonomously produced a 171,504-node mesh, executed seven analyses, and delivered a complete report—including failure diagnosis and redesign recommendations—without any human intervention.

This work addresses the complexity and expert dependency of traditional finite element analysis by proposing the first end-to-end automation framework capable of processing both image and text inputs. The approach introduces a multi-agent system grounded in ReAct-style reasoning, integrating vision-language understanding, collaborative task planning, and a verification-first code generation mechanism. To ensure physical validity, the framework incorporates self-debugging and fallback strategies. Evaluated across diverse engineering mechanics scenarios, the method substantially outperforms existing large language model baselines, demonstrating high success rates and robustness in generating complete, correct, and physically consistent simulation models.

Existing tools automate finite element analysis (FEA) only for single components and lack support for end-to-end engineering verification. To address this, we propose the first natural language–driven, full-stack Geometric-Mesh-Simulation-Analysis (GMSA) agent framework. Our method integrates large language models, engineering semantic parsing, physics-constrained modeling, and adaptive mesh generation to enable a fully automated, human-in-the-loop–free pipeline—from natural language specifications to validated simulation outcomes—featuring boundary condition inference and physics-aware mesh optimization. The framework embeds the CalculiX solver for industrial-grade verification. Evaluated on a turbocharger case and 432 NACA airfoil configurations, the system consistently produces physically plausible, reproducible, multi-objective simulation results. These experiments demonstrate the framework’s scalability and engineering practicality in complex rotating machinery and parametric design scenarios.

This work proposes the first autonomous simulation system that integrates an agent-based architecture with domain-finetuned large language models (LLMs) to enable end-to-end modeling and solution of solid mechanics, fluid dynamics, and multiphysics problems. Addressing the limitations of conventional LLMs—which often hallucinate, lack awareness of variational structures, and fail to close the loop from problem description to verified solutions—the system incorporates retrieval-augmented multi-LLM code generation and filtering, finetuned models spanning 3B to 120B parameters, multi-agent collaboration, and runtime feedback mechanisms. A high-quality corpus of over a thousand FEniCS codes was curated to support training and evaluation. On a benchmark suite of 39 nonlinear elasticity, plasticity, and non-Newtonian fluid problems, the GPT OSS 120B model achieved a code generation success rate of 71.79%, substantially outperforming non-agent-based approaches.

Although topology optimization has matured, its reliance on manual intervention—such as modeling, meshing, and boundary condition specification—hinders accessibility for non-experts. This work proposes the first conversational framework based on a large language model (LLM) agent that enables end-to-end topology optimization through natural language instructions and optional inputs (e.g., images, geometry, or meshes), automatically invoking finite element solvers and optimization tools. The approach integrates multi-load structural and thermal optimization, handles stress constraints, and employs few-shot prompting strategies, successfully reproducing benchmark cases while solving complex engineering problems and autonomously generating optimized structures, field distributions, and convergence curves. Ablation studies confirm that prompt design critically enhances system robustness, substantially lowering the usability barrier without compromising numerical reliability.

This work addresses the challenges of high-fidelity simulation in fusion energy systems—specifically, geometric modeling, multiphysics coupling, and the integration of particle and continuum methods—by proposing a unified framework that seamlessly combines commercial CAE software with existing fusion codes. The framework enables accurate representation of complex geometries, automatic generation of unstructured meshes, and efficient coupling between particle transport and continuum solvers. The resulting simulation workflow significantly enhances geometric fidelity for critical components and strengthens capabilities in multiscale, multiphysics co-simulation, while maintaining strong scalability and computational efficiency.

This work proposes a measurement-driven, constrained natural language interface architecture to reduce manual configuration overhead in finite element simulations while mitigating the risk of unreliable code generation by large language models (LLMs) in critical solver stages. The approach confines the LLM to front-end tasks—such as prompt parsing and Gmsh script generation for non-standard geometries—while a deterministic scheduler orchestrates verified FEniCS/UFL templates for core computations across five multiphysics problem classes: linear elasticity, hyperelasticity, elastoplasticity, thermomechanical coupling, and phase-field fracture. Experimental results demonstrate 100% prompt parsing success, 97.1% field extraction accuracy, and 90% success rate in custom geometry generation. Simulation accuracy reaches sub-percent levels for smooth problems, with errors in nonlinear cases maintained within 2–5%.

This work proposes AbaqusAgent, the first end-to-end, natural language–driven multi-agent framework for finite element analysis (FEA) in solid mechanics, designed to lower the barrier to entry and reduce reliance on expert knowledge. The framework integrates six collaborative modules—Interpreter, Architect, Input Generator, Runner, Reviewer, and Visualizer—to automatically translate user-provided natural language instructions into a complete Abaqus workflow encompassing pre-processing, solving, and post-processing. Evaluated on 50 solid mechanics problems, the system achieves an 86% success rate, significantly enhancing simulation efficiency. Furthermore, it advances human–computer interaction paradigms and enables seamless integration with AI-driven optimization and material characterization pipelines.

This study addresses the high computational cost of traditional CAE simulations in pedestrian protection crash safety design, which hinders efficient exploration of high-dimensional, nonlinear design spaces. The authors propose an AI-driven engineering workflow coordinated by a foundation model, integrating large language models and vision-language models as a unified interface to combine data-driven surrogate modeling, NSGA-II multi-objective optimization, and topology-preserving geometric generation—all while supporting natural language interaction. Demonstrated on an automotive front bumper case study, the framework generates 35 diverse, safety-compliant designs within seconds, achieving a dramatic efficiency gain over conventional simulation workflows that require several hours. The approach maintains strong interpretability and enhances design space exploration capabilities.