This cumulative dissertation investigates the operation of storage-augmented energy systems and their interaction with the overall energy system. A storage-augmented energy system, in this con-text, is defined as an electric energy storage system in close proximity to consumers and distributed generation units under joint control. This work consists of four papers published in scientific, peer-reviewed journals and conference proceedings that aim to answer the following Re-search Questions (RQs): (RQ1): What is the status of research of mathematical decision support models for operating storage-augmented energy systems? (RQ2): How do thermal and electrical energy storage systems in hybrid energy systems influence each other, and how does their interaction influence the way the superordinate system should be operated? (RQ3): Which models are suitable to include battery aging costs into the operation problem, and how does this cost-factor change the way the storage-augmented energy system should be operated? (RQ4): To what extent does including an EESS into an industrial production facility enhance the flexibility offering to the overall energy system?

All four papers focus on various combinations of the above RQs, and apply different research methodologies to address them. Paper 1 begins with a systematic and comprehensive literature review on the current status of research of energy management for storage-augmented systems in stationary applications. The paper first develops a conceptual framework, which is then used to structure and discuss the relevant literature. Paper 1 concludes with a set of propositions for future research based on the identified research gaps, and hence prepares Papers 2 to 4. Paper 2 und Paper 4 develop mathematical models for operating storage-augmented energy systems in residential and industrial applications, respectively, and discuss the results of computational studies on exemplary configurations. Paper 3, in contrast, formulates a conceptual framework on demand-side flexibility measures in industrial production facilities as a preliminary work for Paper 4. In the following, the different research areas of Papers 1 to 4 are outlined in more detail, and the specific research gaps addressed by the four papers are explained.

The systematic and comprehensive literature review presented in Paper 1 develops an overall view on the current status of research in the field (RQ1). Paper 1 provides an introduction to energy management of electric energy storage systems in general, and the multifarious aspects to be considered when operating stationary systems in particular. Research in this field has received more and more attention in recent years. The vast amount of publications on the management of electric energy storage systems, especially those that appeared in the last ten years, has created a need for a structured review and classification of existing research. Although several papers re-viewing the matter have been published, the review in Paper 1 differs from existing research in terms of its focus on mathematical models and its systematic review approach. In the synthesis of the reviewed publications, Paper 1 outlines propositions for future research, which were partially addressed in Papers 2 to 4.

Paper 2 analyzes operations of a storage-augmented, hybrid residential microgrid. The paper con-tributes to research by investigating the case of a local energy supplier. The local energy supplier is responsible for meeting local hybrid, i.e. electrical and thermal, energy demands while interacting with the grid at real-time pricing. The major benefit for the energy supplier comes from efficiently using non-renewable decentralized generation units by leveraging thermal energy storage systems and electric energy storage systems. Compared to classical, thermal power plants, distributed generation units utilize primary energy resources more efficiently as they offer the opportunity to use excess heat to serve local thermal demand. Gas-fired combined heat and power plants can operate at combined efficiencies ranging between 70 % and 80 %. This is well above the efficiency levels of conventional power plants without waste heat utilization that usually do not exceed 30 %. Thermal and electrical energy demand in hybrid systems are for the most part uncorrelated, whereas combined generation units generate thermal and electrical energy simultaneously in a fixed ratio. Therefore, in practice, combined generation units follow either electrical or thermal loads when operated heuristically. Two approaches have been applied in Paper 2 to respond to these challenges. On the on hand, optimization methods support economic and reliable operations of microgrids and have already attracted much attention among researchers and practitioners in recent years. On the other hand, hybrid energy storage systems, a combination of electric and thermal energy storage systems, can be applied to decouple both types of demands. Paper 2 first contributes to research by revisiting current work on optimization models for microgrids that include battery energy storage systems and take battery aging into account (RQ3). Most of current research has focused on using batteries to optimize energy systems for economic, ecological, and technical objectives, but barely considered battery aging in the optimization models. Especially battery aging models that consider specific usage conditions have been underrepresented. Paper 2 addresses this research gap by deriving a weighted cost model, considering both cyclical and calendrical aging components, from the domain-specific literature on battery lifetime prediction. The paper further integrates the piecewise-linearized battery aging model into a mixed-integer linear programming formulation for a hybrid microgrid application. The influence of the battery aging model formulation on microgrid operations in a cost-optimal schedule is illustrated in a computational study for a real-world example. Secondly, Paper 2 contributes to research by investigating the interdependencies of the thermal and electrical systems in a parameter study on component sizing. Sensitivities are investigated through selected key parameters and show that both storage types can significantly reduce the grid-provided energy without losing economic viability.

Paper 3 and Paper 4 put the spotlight on the industrial consumer. By size, the industrial sector was responsible for around 42.5% of world-wide electricity consumption in 2014. This entails a large potential for generating flexibility by demand-side management. Paper 3 addresses research efforts undertaken to tap this potential and to enable industrial consumers to offer short-term flexibility. Paper 3 fosters the idea that production facilities incorporate a versatile set of flexibility measures that enable them to modulate their electricity consumption time- and volume-wise and, as a result, to participate in respective flexibility markets. Paper 3 develops a conceptual framework for an energy-aware view on production facilities to identify the various resources of flexibility. Besides the production system, whose energy consumption is adjustable by changing the production schedule, there are many examples for additional resources of flexibility such as local generation, energy conversion systems, and other auxiliary systems, of which many show a storage-equivalent behavior. As a final note, the paper proposes a control architecture to coordinate the different sources of flexibility.

Paper 4 concludes this dissertation. The paper elaborates on the ideas outlined in Paper 3 and presents an in-depth analysis of a storage-augmented industrial production facility participating in demand response. For simplicity, this work concentrates on the production system and a co-located battery. Paper 4 outlines challenges for industrial consumers participating in demand response and provides an overview of the corresponding literature. For residential and commercial consumers, interdependencies between the scheduling of different applications (e.g., refrigerators or air conditioning equipment) are negligible and scheduling can be performed independently. Prior research has investigated various residential applications and to what extent they are compatible with demand response, e.g. for air conditioning, cloth dryers, or dishwashers. For industrial consumers, however, participating in demand response is more difficult as the scheduling of processes within facilities is often subject to many interdependencies. While in many traditional demand response programs, system operators require direct control of single consumers for short-term flexibility, the aforementioned complexity within industrial consumers falsifies the appropriateness of such approaches and reveals the need for other solutions. In industrial applications, demand response thus requires sophisticated models that account for the influence of demand response on production processes and vice versa. Firstly, Paper 4 contributes to this research by proposing an incentive-based program according to which the facility operator determines alternative electricity consumption scenarios and communicates discrete load reduction potentials to the system operator without disclosing internal processes. Secondly, Paper 4 develops a flexible flow shop formulation for a discrete manufacturing process. A reference model is extended to account for the operating-mode-specific energy consumption of machines with specific consumption trajectories per product-machine-combination. A mixed-integer linear programming formulation is suggested to model and solve the problem in three stages. First, a base-line solution is developed by minimizing total weighted completion time. Then, based on the baseline solution, additional solutions with different responses to the demand response are calculated and a load reduction curve as a potential means of communication is established. Finally, the effects of using a battery to allow easy-to-apply and economically better responses are studied. A numerical example is provided and analyzed to give a zest of the suggested solution.