Volume 58, Issue 1 - Energy Informatics / Sebastian Lehnhoff, Astrid Nieße

The operation of smart grids is based on different energy management and control systems. Despite their variety, all of them offer similar functionalities to achieve a stable, reliable, sustainable and economic supply of energy. Similarity offers advantages for the construction of systems that are not well reflected in current system architectures. This work introduces foundations for specifications and development of fractal architectures that are based on self-similarity. We aim to apply those principles on two smart grid use cases: virtual power plants and optimal power flow control applications.

The increasing share of volatile renewable energy sources leads to new challenges for the power system. Flexible operation of power consumers is getting more important to keep power generation in balance with power consumption. Large potential loads are available in the industrial sector. Flexibility of these large loads is hard to utilize and requires consideration of the specific operational constraints. In the automotive industry test rigs for the development of cars or components, like engine or gearbox, exist, which are characterized by a high energy demand. In this work the idea of an energy-aware scheduling concept for test rigs in the automotive industry is presented. The goal is to provide an optimized schedule based on internal (e. g., prediction of local power generation) or external (e. g., electricity prices) signals. To this end, a methodological approach for the classification, identification and selection of suitable test rigs is presented. Based on this, a model of the test rigs has to be developed which can be used in a simulation to evaluate the optimization algorithm. Finally, the need for design guidelines for the implementation of an energy-aware test rig scheduling in a real-world evaluation environment are discussed.

Battery electric vehicles (EVs) have a more limited range and longer refueling, i. e. charging times accompanied by battery aging as conventional vehicles. Initial field studies about the behavior of EV users also have identified the presence of range anxiety, which is expressed by frequent recharging. Battery aging is a non-linear process with numerous interactions, but generally pronounced around higher state of charge (SoC) levels. Therefore, range anxiety potentially provokes elevated battery degradation due to the resulting operating conditions. Based on 2966 empirical driving profiles, I analyze different user groups as well as the effect of different charging strategies in a simulation-based analysis. In comparison to charging strategies, for instance as fast as possible (AFAP) charging or critical SoC charging (such that the next trip is feasible), I calculate the degradation optimal charging pattern based on a realistic battery cell aging model. The model is derived from accelerated aging tests of Li-NMC 18650 cells, but can be adjusted to individual degradation parameters, in order to account for different battery chemistries. Preliminary results show that degradation can be reduced by up to 46% by the optimized charging strategy as compared to AFAP or critical SoC charging.

Due to the integration of renewable energy ressources in the electricity grids, many new challenges have to be coped with. Therefore, new architectures are developed. To validate if a new architecture meets the requirements of transforming electricity grids into Smart Grids, a general evaluation-metric should be used. Not all requirements and their satisfaction can be mapped on monetary values, so a differentiated view on different aspects of an Smart Grid architecture is preferable. To meet this purpose, the idea is to develop separate evaluation-metrics on each layer of the Smart Grid Architecture Model and to build one comprising evaluation scheme for Smart Grid Architectures.

Buildings acting as flexible loads have been often proposed to mitigate the volatility of renewable energy sources. Thereby, an accurate short-term demand forecast is indispensable for effective demand side management. At the same time, standardized load profiles, commonly used in the distribution grid, are inadequate for load forecasting within building domain. For this PhD, project a novel short-term forecasting model is proposed for that domain. It considers not only residual load, but also scheduled demand response as well as the PV-generation of the building. Moreover, it is not building specific and is, therefore, suitable for area-wide application within building domain.

Cognitive impairments are a rising challenge in society. Everybody has experienced events of forgetting what one was going to do when entering a certain room, or where one has put the keys, the purse or the smartphone. Such problems mostly occur under stress or fatigue. And they may increase with age up to diseases like senile dementia or Alzheimer which affect mental tasks like memory or reasoning, and thus often lead to the need of comprehensive assistance. Due to the societal change, the number of people suffering from such impairments is continuously growing. Cognitive modeling may provide human-centered solutions for this challenge since a cognitive model of a person's behavior, regarding activities of daily living, can serve as a knowledge base for support actions. This paper presents a Domain Specific Modeling Language for Ambient Assistance: The Human Cognitive Modeling Language (HCM-L). It was developed to preserve the episodic memory of a person in the form of conceptual behavior models including relevant context. The work is part of the Human Behavior Monitoring and Support (HBMS) project, a research project in the field of Ambient Assisted Living funded by the Klaus Tschira Stiftung gGmbH. HBMS aims at monitoring a person's behavior using activity recognition techniques, generating models from their output, and providing focused and timely support by the use of intelligent reasoning mechanisms.

We present the architecture of the new ABR-KFUE decision-support system for nuclear emergency management used in Germany. Such systems assist decision makers in taking countermeasures in case of releases of radioactive materials into the environment. The specificity of these systems is that they use simulation software in a safety-critical application context. The new architecture of the system thus aims at fulfilling non-functional requirements for improved reliability, performance, availability, and maintainability. The proposed solutions are evaluated using a stimulus/response analysis.