In this paper, we present a novel framework that combines large language models (LLMs), digital twins and industrial automation system to enable intelligent planning and control of production processes. Our approach involves developing a digital twin system that contains descriptive information about the production and retrofitting the automation system to offer unified interfaces of fine-granular functionalities or skills executable by automation components or modules. Subsequently, LLM-Agents are designed to interpret descriptive information in the digital twins and control the physical system through RESTful interfaces. These LLM-Agents serve as intelligent agents within an automation system, enabling autonomous planning and control of flexible production. Given a task instruction as input, the LLM-agents orchestrate a sequence of atomic functionalities and skills to accomplish the task. We demonstrate how our implemented prototype can handle un-predefined tasks, plan a production process, and execute the operations. This research highlights the potential of integrating LLMs into industrial automation systems for more agile, flexible, and adaptive production processes, while also underscoring the critical insights and limitations for future work.

翻译：本文提出了一种新型框架，该框架融合大语言模型（LLMs）、数字孪生与工业自动化系统，以实现生产过程的智能规划与控制。我们的方法包括构建包含生产过程描述性信息的数字孪生系统，并对自动化系统进行改造，使其提供统一接口，支持由自动化组件或模块执行的细粒度功能或技能。随后，设计大语言模型代理（LLM-Agents）以解读数字孪生中的描述性信息，并通过RESTful接口控制物理系统。这些LLM代理作为自动化系统的智能体，能够实现柔性生产的自主规划与控制。给定任务指令作为输入，LLM代理可通过编排一系列原子化功能与技能来完成该任务。我们展示了所实现的原型系统如何处理非预定义任务、规划生产流程并执行操作。本研究揭示了将大语言模型集成到工业自动化系统中以实现更敏捷、灵活和自适应生产过程的潜力，同时指出了未来研究的关键见解与局限性。