Intelligent Control Platform

Compared to conventional energy systems, future energy systems will have fluctuating generation subject to uncertainties and a higher number of components due to the many units. In the future, "consumers" will take on the role of "prosumers", who can actively contribute to the integration capability of renewable energies by acting as local energy producers. Accordingly, intelligent control of the energy system is necessary in order to make the best possible use of available resources. Overall, the complexity of the system increases significantly, so that new control approaches are needed.

Within the framework of the LLEC project network, such future district energy systems are being investigated, which are primarily supplied by renewable energies and waste heat. For this purpose, central parts of the energy infrastructure as well as buildings of the FZJ campus are supplemented by further technologies for the generation, conversion and storage of energy. These form the basis for the innovative district energy system. Currently, the "LLEC infrastructure" consists of the following components:

• Various photovoltaic systems (ground-mounted systems, roof-mounted PV systems and façade-integrated systems)
• Two lithium-ion batteries
• Hydrogen infrastructure (including electrolyzer, AFC fuel cell, LOHC storage, H2 admixture in combined heat and power plant (full heat supply center))
• ow-temperature network with waste heat recovery of a water-cooled mainframe computer
• 15 cognitive existing buildings, which are equipped with sensors and actuators on a room-by-room basis for testing advanced room control concepts, among other things
• Electric vehicles / uni-/bi-directional charging stations

Both existing and newly installed components of the system are digitally networked via a cloud-based information and communication platform (ICT platform). This means that current operating data for the entire energy system is available at all times. With the help of this data, detailed modeling of the energy system can be carried out using "digital" twins. Furthermore, the available data can be evaluated by machine and characteristics of the operation can be determined, which allow statements about improvements of the operation strategies. In addition, the digitization of the energy system makes it possible to control all components from a central location, taking all operating data into account. This makes it possible to test innovative control concepts, such as predictive (model-predictive) control concepts, in addition to conventional, mostly rule-based operating strategies. A model predictive control uses a mathematical model that describes the operating dynamics of the real system to determine an optimal driving mode for the components. For this purpose, the system behavior is mapped in mathematical form and converted into a so-called optimization problem. By using forecasts regarding future energy consumption and generation, the behavior of the energy system can be calculated and optimal manipulated variables (driving mode) can be determined by solving the optimization problem. These optimal control variables are sent by the ICT platform in real time to the real components and implemented there.

The LLEC offers a variety of different energy systems and thus allows the investigation of many different control approaches and operating strategies with regard to, for example, optimal integration and use of renewable energies and waste heat and relief of the power grid through the active use of flexibility potentials on the consumer side. After successful testing, the approaches developed here should also be transferable to other applications outside the research center.

Through the "Dashboard Suite" developed within the LLEC, users can gain better insight into consumption data at the building level as well as consumption data and comfort parameters at the room level. By coupling a virtual business game for the FZJ property with offices in the "real" world, research is being conducted to determine whether game-based approaches can help to sustainably improve user behavior with regard to an energy-saving scope.