Midhun Xavier

1. Agentproof: Static Verification of Agent Workflow Graphs

   Melwin Xavier; Vaisakh M A; Melveena Jolly; Midhun Xavier

   arXiv · 2026 · doi:10.48550/arXiv.2603.20356

2. ReACT — Gen AI Agents for Reasoning, Planning, and Testing in IEC 61499-Based Control Systems

   Xavier, M.; Patil, S.; Yang, C.-W.; Vyatkin, V.

   IECON 2025, 51st Annual Conference of the IEEE Industrial Electronics Society, 1–6 · 2025 · doi:10.1109/iecon58223.2025.11221719

3. LLM-Powered Multi-Actor System for Intelligent Analysis and Visualization of IEC 61499 Control Systems

   Xavier, M.; Laikh, T.; Patil, S.; Vyatkin, V.

   IECON 2024, 50th Annual Conference of the IEEE Industrial Electronics Society, 1–8 · 2024 · doi:10.1109/iecon55916.2024.10905502

Conference contributions

1. It Runs — but Is It Right? Ontology-Constrained Synthesis of Physics-Valid Simulation Tools for AI Agents

   Xavier, M.; Berglund, D.; Babu, B.; Berglund, E.

   MSE · 2026

   +−Abstract

   A simulation tool can run flawlessly and still be wrong. For AI agents that synthesize tools for other AI agents [4], this is a particularly dangerous failure: inconsistent units, incompatible material states, or use of a solver outside its validated domain can produce convincing outputs that are physically meaningless. We present an ontology-constrained approach for PHASES [1], aligned with EMMO and OSMO [2,3], in which AI agents synthesize new application-level simulation tools from existing validated solvers and approved external physics components. Given a request such as “build a laser-hardening tool,” the agent first constructs a semantic blueprint that specifies the required quantities, units, material states, physical capabilities, solver dependencies, and operating envelopes. It then assembles a bounded workflow, namely laser energy deposition, transient heat transfer, PHASES transformation kinetics, and hardness and case-depth post-processing. The agent generates the interfaces, coupling logic, parameterization, and post-processing, but does not invent new governing physics. Before registration, each candidate undergoes dimensional, semantic, state-transfer, and applicability checks, followed by sandboxed execution and physics-based validation. If a required capability is missing or incompatible, synthesis stops safely. An accepted tool carries a machine-readable validity manifest that records its operating limits, verification evidence, and ontology, material, and solver versions. The tool can then be reused by PHASES AI or other LLM agents through a tool registry and MCP. We will evaluate the approach on custom metallurgical simulation requests that require different combinations of PHASES and approved external solvers, measuring task success, correct capability selection, safe rejection, envelope compliance, and silent physical errors. The aim is to create AI-generated simulation tools that do more than run: they know and expose the limits within which they are right.

   Abstract PDF↗

2. PHASES Agent: an AI agent for physics-based digital metallurgy

   Berglund, D.; Babu, B.; Xavier, M.; Berglund, E.

   ESTEP Annual Event · 2026

   +−Abstract

   Predictive metallurgy is often promised and rarely delivered. Most tools stop at a property lookup or a single calibrated model, and the few that reach further tend to stay inside a specialist’s script. Phases (https://phases.aerobase.se), built by Aerobase Innovations AB, works differently. It is an AI agent that runs physics-based phase-transformation models from a plain-language request and returns a finished simulation with plots, and it already drives a validated automotive process chain from hot forming through welding to crash. This abstract describes the platform and the physics behind it, then shows the result that makes the case: a single composition-dependent model, applied across a full press-hardening chain, predicts how the chemistry of recycled steel changes crash performance.

   Abstract PDF↗

3. AeroCRAFT: AI-Driven Closed-Loop Toolpath Planning and Thermal Simulation for DED

   Xavier, M.; Berglund, D.

   VMAP User Meeting · 2026

   +−Abstract

   Directed Energy Deposition (DED) enables near-net-shape fabrication of complex metallic geometries; however, the process is highly sensitive to thermally induced defects, including porosity, delamination, residual stress, and microstructural inhomogeneity. These defects arise from the interaction between heat-source dynamics and the scan strategy, yet conventional toolpath planners generate trajectories based solely on geometric considerations, without accounting for the cumulative thermal history. This work presents AeroCRAFT, an end-to-end AI-driven toolpath planning and thermal simulation platform that closes the loop between scan strategy, thermal response, and defect prediction in DED. The framework integrates deterministic geometric toolpath generation with a voxel-based thermal finite element (FEM) module to enable automatic preprocessing and transient temperature-field evaluation. AeroCRAFT addresses the limitations of traditional scanning methods - including unidirectional, bidirectional, serpentine, and concentric strategies - by introducing a transformer-based attention model that is trained using reinforcement learning to optimize scan sequencing. The AI agent determines toolpaths by considering both geometric factors and user-defined goals such as hotspot reduction, defect minimization, or achieving specific microstructures. Real-time thermal evaluation is performed using process- and temperature-based quality metrics, including Energy per Unit Length (EUL) and HAI₁₀ thermal exposure indicators, enabling quantitative assessment of build integrity during process planning. Toolpaths, process parameters (laser power, scan speed, dwell time, surface normal), orientation data, and full temporal temperature fields are exported in the VMAP HDF5 standard format, enabling seamless interoperability with downstream FEA solvers and robotic systems such as LS-DYNA (*BOUNDARY_THERMAL_WELD_TRAJECTORY), Abaqus, CalculiX, and KUKA KRL programs. The integrated deterministic and AI-driven planning framework, combined with VMAP-compliant digital thread export, demonstrates a scalable pathway toward closed-loop, defect-aware DED process planning, reducing empirical trial-and-error and enabling systematic optimisation of complex geometries prior to physical deposition.