The cruise ship industry is under increasing pressure to reduce greenhouse gas emissions, as international regulations define ambitious requirements and goals for modern cruise ships. One of the most significant consumers of energy onboard cruise ships are their Heating Ventilation and Air Conditioning (HVAC) systems. However, the energy optimization of HVAC systems is challenging, as they are impacted by a number of uncontrolled variables, such as changing weather conditions, passenger behavior, and the demands of other significant energy consumers, such as propulsion systems. Reinforcement Learning (RL) is often used to tackle such complex optimization tasks, however concerns over ensuring the safety of RL optimized systems hinders its adoption in industry, especially in the context of safety-critical systems. This paper presents the initial findings of applying a novel approach to ensure safety in RL: a safety shield developed utilizing a novel hazard analysis method, System-Theoretic Process Analysis. In this work the safety shield is used to both train the RL agent as well as block unsafe behavior in operation. Preliminary findings suggest that blocking unsafe behavior during training hinders the ability to learn a safe RL policy, however, when used in testing the approach is capable of significantly reducing the number of safety violations.

Large Language Models (LLMs) are increasingly being used in software development and in applications like code generation. While LLMs can provide significant value in the form of time savings in common programming languages like Python, their usability in generating automation software has yet to be studied extensively. In the context of generating control software in the form of IEC 61131-3 compliant code, initial studies suggest LLMs provide a promising avenue for increasing control engineer productivity. However, similar code generation for IEC 61499-based control applications is still scarce. While tools are being developed for this purpose, their capabilities are not yet fully understood, and they often require significant human input to generate the intended outcomes. This paper explores LLM-based code generation for IEC 61499-based applications through iterative prompting. The prompts for the experiments are derived from requirements generated by System-Theoretic Process Analysis (STPA), which provides a systematic approach to creating prompts that also connect to the larger systems engineering workflow. The results indicate that while the approach may be successful in some instances, more work is required to mitigate the issues arising from its application.

Cruise ships are large greenhouse gas emitters. Given the net-zero emission goal by 2050 set by the International Maritime Organization, it is crucial to focus on different strategies for reducing emissions. One promising strategy is to work on reducing energy losses due to non-optimal management of a cruise ship’s complex energy system. Effectively managing a ship’s energy consumption is a difficult task. Modern reinforcement learning approaches have been employed to the task at hand. Results have shown improvements over traditional optimization methods that fail to scale to large stochastic problems such as ship energy optimization. However, reinforcement learning approaches lack a direct focus on ensuring the safety of the system, which is critical for a real-world application.This paper showcases work in progress related to developing a safe reinforcement learning (RL) framework for ship energy optimization. We utilize a formal safety shield to block unsafe control actions and a large language model (LLM) to generate a reward function (RF) that incorporates safety constraints. We construct a white-box physical system of a ship’s energy management system and a faster black-box surrogate model as environments to train and test our RL methods. Preliminary results show that the integration of the safety shield and the LLM-generated reward function is promising in reducing safety violations.

Industry 4.0 demands intelligent and customized production, driving the upgrade of traditional systems. However, many industrial devices lack source code or documentation, making control program migration highly challenging. This issue is critical for PLC replacement, system modernization, and integration of heterogeneous devices. Recently, large language models (LLMs) have demonstrated strong capabilities in code generation and logical reasoning. When combined with the IEC 61499 standard—known for its modular and event-driven design—LLMs offer new possibilities for automatic control software generation. This paper proposes a method that uses LLMs to transform industrial control logic into IEC 61499-compliant programs. It extracts finite state machines representing software logic, converts them into formal requirements, and generates the IEC 61499 function block with the assistance of LLMs. The method enables efficient, data-driven migration of control software.

Modernization of currently operational nuclear power plants is becoming increasingly important to maintain their performance and safety. Ensuring the safety of newer Instrumentation and Control (I&C) systems used in modernization efforts requires hazard analysis techniques suitable for the analysis of complex and software-heavy systems. System-Theoretic Process Analysis (STPA) has proven to be a suitable hazard analysis method for these complex I&C systems, however, its practical use is still often limited by its labor-intensive and time-consuming nature, partially due to the limitations of the tools used to perform the analysis: common Office tools such as Microsoft Excel or Visio. Conducting an STPA analysis could be simpler and more attractive with software tools specific to the method. This work introduces the requirements for these software tools and lays the foundation for further work, in which software tools will be evaluated against these requirements.