This article proposes a novel framework for the rapid implementation of automation programs in industrial heating and cooling systems. The global push for sustainability necessitates significant infrastructural transformations within these systems, which currently rely heavily on fossil fuels and are responsible for 75% of industrial final energy consumption. Our research highlights the critical role of design patterns and object-oriented programming principles to address the complex integration of additional energy converters and storage into automation programs. By leveraging design patterns, our framework encapsulates the intricacies of various components, such as actuators, sensors, and storage, within a comprehensive object-oriented model that also allows the integration of different control strategies. Qualitatively, this approach enhances the reusability, scalability, and adaptability of automation programs. Therefore, quantitatively, our framework enables a more resilient and efficient energy system. The framework is validated through its application to a complex, cross-linked industrial heating and cooling system at the ETA Research Factory of the Technical University of Darmstadt. Using the developed framework reduces implementation effort significantly due to its consistent and modular structure resulting from the reusable design patterns.
Deep Reinforcement Learning (DRL) can optimize the operating strategies of industrial energy supply systems (IESS), enhancing energy efficiency and flexibility while reducing costs. This work conducts and evaluates a real-world application of Proximal Policy Optimization (PPO), a widely used DRL algorithm, on the complex ETA Research Factory’s supply system. Our approach involves five steps: system boundary definition and data accumulation, modelling and validation, DRL implementation, simulation-based evaluation and finally, application of the DRL-based controller on the real system. In simulation experiments, the controller demonstrates a 43% reduction in operational costs with respect to the conventional controller while maintaining operational stability within safety limits, thus proving to be a cost-effective and reliable solution. During real-world application, the DRL-based controller upheld safety and robustness. Future research should aim at establishing comparability under various environmental conditions and closing the gap between simulation and reality.
Industrial energy systems face increasing complexity driven by decarbonization requirements, renewable integration, and price volatility, making conventional rule-based control insufficient to exploit energy efficiency and flexibility potentials. Advanced control strategies can improve the operation strategy, yet their real-world performance and transferability remain largely unexplored due to inconsistent system representations and the absence of quantitative benchmarking frameworks. This work addresses these gaps by presenting ETA Factory Thermal System Operation (EFTSO), a framework for the development, benchmarking, and simulation-to-reality transfer of optimized control strategies for industrial energy systems. The application of the framework is demonstrated on a representative industrial heating network. A rule-based controller is used as baseline and compared with reinforcement learning, model predictive control, and a market-based controller. The evaluation metrics incorporate key performance indicators including energy efficiency, operating costs, security of supply, system wear, and CO$_2$ emissions. Real-world experiments over three representative days reveal that each advanced control strategy exhibits distinct trade-offs between competing objectives, while none achieves optimal performance across all metrics. Moreover, simulation results alone prove unreliable predictors of real-world behavior. Therefore, this study demonstrates the essential role of real-world validation in control strategy assessment and provides a framework to systematically benchmark different control approaches.