On Integrating Resilience and Human Oversight into LLM-Assisted Modeling Workflows for Digital Twins

TL;DR

This paper proposes design principles for integrating resilience and human oversight into LLM-assisted workflows for creating digital twins, emphasizing interpretability, error-resilience, and the use of Python as a density-preserving intermediate representation.

Contribution

It introduces three key design principles derived from the FactoryFlow framework to improve resilience and oversight in LLM-assisted digital twin modeling workflows.

Findings

Characterization of LLM-induced errors across different model descriptions

Demonstration of Python as an effective density-preserving intermediate representation

Guidelines for building resilient, interpretable, and error-resilient workflows

Abstract

LLM-assisted modeling holds the potential to rapidly build executable Digital Twins of complex systems from only coarse descriptions and sensor data. However, resilience to LLM hallucination, human oversight, and real-time model adaptability remain challenging and often mutually conflicting requirements. We present three critical design principles for integrating resilience and oversight into such workflows, derived from insights gained through our work on FactoryFlow - an open-source LLM-assisted framework for building simulation-based Digital Twins of manufacturing systems. First, orthogonalize structural modeling and parameter fitting. Structural descriptions (components, interconnections) are LLM-translated from coarse natural language to an intermediate representation (IR) with human visualization and validation, which is algorithmically converted to the final model. Parameter inference, in contrast, operates continuously on sensor data streams with expert-tunable controls. Second, restrict the model IR to interconnections of parameterized, pre-validated library components rather than monolithic simulation code, enabling interpretability and error-resilience. Third, and most important, is to use a density-preserving IR. When IR descriptions expand dramatically from compact inputs hallucination errors accumulate proportionally. We present the case for Python as a density-preserving IR : loops express regularity compactly, classes capture hierarchy and composition, and the result remains highly readable while exploiting LLMs strong code generation capabilities. A key contribution is detailed characterization of LLM-induced errors across model descriptions of varying detail and complexity, revealing how IR choice critically impacts error rates. These insights provide actionable guidance for building resilient and transparent LLM-assisted simulation automation workflows.