AgentAM: An AI Agent for Spatter- and Defect-Aware Process Optimization in Laser Powder Bed Fusion

| Source: arXiv AI

Tags: AgentAM, Model Context Protocol, laser powder bed fusion, additive manufacturing, SS316L, CFD surrogates

AgentAM gives metal 3D-printing engineers plain-language access to CFD-trained spatter and melt-pool models through MCP, passing the right inputs in 98.5% of parsing prompts and landing within a median 1.0% of brute-force optimum.

Details

Predictive models for laser powder bed fusion rarely reach the shop floor because using them takes programming skill. AgentAM is an LLM agent that exposes CFD-trained surrogates for SS316L spatter and melt-pool geometry through the Model Context Protocol. Polynomial surrogates trained on 279 single-track simulations predict melt-pool dimensions with R-squared of at least 0.95 and spatter volume at 0.83. Geometric rules classify defect regimes, matching simulated geometry in 93.4% of tracks. Minimum-spatter optima often sat within 10 micrometers of the lack-of-fusion boundary; a margin-aware optimizer moved them away and raised estimated robustness under process drift. On a 145-prompt benchmark over 725 runs, the agent passed intended inputs in 98.5% of parsing prompts and met every constraint in 65 feasible optimization runs.