TurboAgent: An LLM-Driven Autonomous Multi-Agent Framework for Turbomachinery Aerodynamic Design

2026-04-08 · Source: cs.AI updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Engineering & Applied Sciences · Depth: Expert, short

Summary

TurboAgent is a new large language model (LLM)-driven autonomous multi-agent framework designed for turbomachinery aerodynamic design and optimization. This framework addresses the complexity of multi-stage design processes, which traditionally involve geometry generation, performance prediction, optimization, and high-fidelity physical validation, often relying on loosely coupled pipelines. TurboAgent uses an LLM for core task planning and coordination, while specialized agents handle generative design, rapid performance prediction, multi-objective optimization, and physics-based validation. Validated using a transonic single-rotor compressor, the system demonstrated strong agreement between target performance and generated designs, with coefficients of determination exceeding 0.91 for key metrics and normalized RMSE values below 8%. The optimization agent improved isentropic efficiency by 1.61% and total pressure ratio by 3.02%. The entire workflow completes in approximately 30 minutes with parallel computing.

Key takeaway

For turbomachinery design engineers seeking to accelerate and automate complex aerodynamic processes, TurboAgent offers a paradigm shift. Its LLM-driven multi-agent architecture enables autonomous, closed-loop design from natural language inputs to final validated designs. You should consider integrating such an autonomous framework to reduce design cycle times significantly and achieve optimized performance metrics, moving beyond traditional, fragmented design pipelines.

Key insights

TurboAgent uses an LLM-driven multi-agent system for autonomous, closed-loop turbomachinery aerodynamic design and optimization.

Principles

• LLMs can coordinate complex multi-stage engineering design.
• Specialized agents enhance specific design sub-tasks.
• Data-driven workflows improve traditional trial-and-error processes.

Method

The method involves an LLM for task planning and coordination, supported by specialized agents for generative design, rapid performance prediction, multi-objective optimization, and physics-based validation, culminating in high-fidelity CFD verification.

In practice

• Automate turbomachinery design from natural language requirements.
• Achieve rapid design iterations (approx. 30 minutes per cycle).
• Improve design metrics like isentropic efficiency and pressure ratio.

Topics

• TurboAgent
• Turbomachinery Aerodynamic Design
• LLM-Driven Multi-Agent Framework
• Multi-Objective Optimization
• CFD Simulations

Best for: AI Scientist, Machine Learning Engineer, Research Scientist

Related on AIssential

• A Lightweight Multi-Agent Framework for Automated Concrete Barrier Design — A multi-agent "generation-evaluation-optimization" framework automates concrete barrier design, achieving high accuracy with lightweight LLMs.
• Agentic Discovery of Neural Architectures: AIRA-Compose and AIRA-Design — LLM agents can autonomously discover and optimize neural architectures and training methods, surpassing human-designed baselines.
• You Probably Don’t Need an Agent Framework — Explicit LLM workflows offer more reliability and control than autonomous agents for many applications.
• EngiAI: A Multi-Agent Framework and Benchmark Suite for LLM-Driven Engineering Design — Specialized benchmarks are crucial for evaluating multi-agent LLM systems in complex engineering design tasks involving simulation, retrieval, and manufacturing.
• Exploring LLM-based Verilog Code Generation with Data-Efficient Fine-Tuning and Testbench Automation — Multi-agent LLMs can automate testbench generation, creating high-quality data for efficient Verilog code generation.
• : Structuring Your Natural Language SOPs into Tailored Ambiguity-Resolved Code Templates — A multi-agent LLM framework effectively transforms ambiguous SOPs into structured, executable code with high accuracy.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.AI updates on arXiv.org.