Data-driven approaches such as reinforcement learning (RL) allow a controller design without a priori system knowledge with minimal human effort as well as seamless self-adaptation to varying system characteristics. However, RL does not inherently consider input and state constraints, i.e., satisfying safety-relevant system limits during training and test. This is challenging in power electronic systems where it is necessary to avoid overcurrents and overvoltages. To overcome this issue, a standard RL algorithm is extended by a combination of constrained optimal control and online model identification to ensure safety during and after the learning process. In an exemplary three-level voltage source inverter for islanded electrical power grid application, it is shown that the approach does not only significantly improves safety but also improves the overall learning-based control performance.
Production and maintenance scheduling for total cost-effective and machine longevity manufacturing is a key aspect in dealing with the ever-increasing energy prices, competitiveness, added maintenance costs, and environmental pressures that the manufacturing sector faces nowadays. This paper addresses these issues by proposing a novel intelligent production scheduling system for joint optimization of production and maintenance for overall cost minimization and machine longevity improvement. To achieve this, it is proposed a Genetic Algorithm (GA) for production and maintenance scheduling of flexible job shop manufacturing environments. The proposed GA takes into account volatile market energy prices, Renewable Energy Resources (RERs), surplus energy selling, maintenance activities, and constraints imposed on the production plan. A case study from the literature is used as a baseline scenario to validate the proposed scheduler. It uses real-production data and considers three unique machines with 275 tasks to be scheduled among them. Using the baseline scenario, it was possible to demonstrate the robustness of the proposed scheduler in reducing total costs, by taking advantage of the volatility of energy prices, as well as utilizing RERs to cover energy expenses or for selling excess energy. Furthermore, it highlights the ability to reduce the overload of single machines. Accordingly, it was possible to achieve cost reductions of up to 11.0% and improvements of 24.4% in machine longevity when compared to the baseline scenario.
Cyber-physical energy systems with high penetration of distributed energy resources provide several grid support functions including black-start in islanded mode. In an inverter dominated power system, both grid-forming (GFM) and grid-following (GFL) inverters have a role to play in regulating system frequency/voltage and maintaining proportional power sharing through grid support functions. A self-sufficient power system can restore critical loads by forming microgrids (MGs) around black-start distributed generators (DGs). This work proposes a service restoration framework that adopts GFM inverters as black-start units and achieves load restoration in the context of dynamic MGs with automatic sectionalization and network reconfiguration. The proposed approach is structured as a two-layer framework consisting of an optimization and power systems simulation layer. Mathematical optimization models can only solve for the steady-state power equations. However, the associated dynamic transient response can cause instability in a weak power grid. Therefore, the optimization layer explicitly incorporates dynamic stability constraints developed from the power systems simulation layer. An optimal network reconfiguration sequence and power reference commands for DGs are generated during the restoration process. Moreover, a consensus based multi-agent distributed control is designed for coordination among GFM and GFL inverters. The approach is validated over a modified IEEE-123 node test feeder using real-time controller-in-the-loop simulation with OPAL-RT simulator.
Energy flexible buildings are expected to become a crucial component of our energy grid to balance the mismatch between energy demand and energy production. Therefore, it could be of great interest to develop tools that allow building designers and engineers to make the best choices regarding energy flexibility. The aim of this paper is to propose a method to generate design and control recommendations to increase the energy flexibility of buildings based on open source computer simulations. The proposed method uses OpenModelica and the building component library of the IBPSA project to generate a building model. The method has been tested using a case study. The values of specific parameters with high influence on energy flexibility have been modified and the results of the changes on the energy flexibility KPIs have been measured. The results are presented in three different possible visualisations. The lessons learned from the process of applying the method have been used to outline a possible future energy flexibility analysis tool.
The purpose of this research is to develop a theoretical method for predicting the wake region downstream of the flow from a top view of a vertical axis wind turbine (VAWT) and to examine the wake loss effects of the VAWT. Previous literature has reported that the wake region of the VAWT is not symmetric, and its downstream is similar to a circular cylinder wake street. The airflow of the VAWT is turbulent, and the inviscid incompressible flow can be applied to its range. Therefore, the current study analyzed the flow over a cylinder with circulation to investigate the velocity, pressure coefficient, and tip speed ratio and to observe the similarity between the cylinder and the VAWT. The results of this study demonstrate remarkable similarity, particularly with respect to the asymmetry of the wake region due to the distinct velocity speed in two opposite locations of the VAWT. Moreover, the pressure coefficient functionality and curvature are similar to the available measurements of the VAWT. This model can be further developed and examined under varied conditions to provide a theoretical wake loss model for VAWTs. This will lead to the expansion of VAWTs’ applications in green, sustainable, and clean energy production.
