TY - CHAP A1 - Dietrich, Alexander A1 - van de Sand, Ron A1 - Reiff-Stephan, Jörg T1 - Implementation of a Machine Tool Retrofit System T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) N2 - Small and medium-sized companies increasingly turning their attention towards the fourth industrial revolution. In order to increase their own long-term competitiveness, there is a growing desire to make production smarter, more efficient, safer and more sustainable through new technologies. Often, however, existing plants cannot be easily replaced by modern equipment. The reasons for this can be high investment costs, excessive downtimes or the unavailability of an equivalent machine. An alternative solution to the purchase of new equipment is the modernisation or expansion of existing systems, also called retrofitting. Thus, this paper deals with the retrofit process of a machine tool, whereby the software architecture of the control unit is the primary concern of this work. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - retrofitting KW - CPS KW - machine tool KW - software architecture Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-15004 SP - 239 EP - 247 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Prell, Bastian A1 - Günther, Norman A1 - Reiff-Stephan, Jörg T1 - Cyber-Physical Production Systems in Settings with Limited Infrastructure BT - Blind Spots and Implications T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) N2 - During the last decade production innovation was mainly focused on connectivity aspects. The vision of smart factories running on software, that uses collected machine data, has become true but foremost for leading industrial companies in highly developed countries. Apart from these, production can also be found in non-industrialized craft professions as well as in less developed countries. As digitalization does not necessarily require an industrial or developed setting the latter could possibly benefit from it as well. Socio-cyber-physical production systems have been used to describe the interdependencies of linked production systems but usually focus on highly developed regions as well as for industrial applications. This paper lines out similarities and differences for each case, introduces the concept of cyber-physical production systems (CPPS) and its extension to socio-CPPS (SCPPS), which emphasizes the role of human workers in the production environment. The relation between industrial, non-industrial production and innovations is examined. Furthermore, the widening of SCCPS concepts for non-industrial production is discussed. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - socio-cyber-physical production system KW - limited infrastructure KW - developing country KW - craft KW - non-industrialized KW - digitalization Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14993 SP - 229 EP - 238 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Szymanski, Mateusz A1 - van de Sand, Ron A1 - Tauscher, Esther A1 - Rieckmann, Olaf A1 - Stolpmann, Alexander T1 - Concept Towards Segmenting Arm Areas for Robot-Based Dermatological In Vivo Measurements T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) N2 - Dermatological in vivo measurements are used for various purposes, e.g. health care, development and testing of skin care products or claim support in marketing. Especially for the last two purposes, in vivo measurements are extensive due to the quantity and repeatability of the measurement series. Furthermore, they are performed manually and therefore represent a nonnegligible time and cost factor. A solution to this is the implementation of collaborative robotics for the measurement execution. Due to various body shapes and surface conditions, common static control procedures are not applicable. To solve this problem, spatial information obtained from a stereoscopic camera can be integrated into the robot control process. However, the designated measurement area has to be detected and the spatial information processed. Therefore the authors propose a concept towards segmenting arm areas through a CNN-based object detector and their further processing to perform robot-based in vivo measurements. The paper gives an overview of the utilization of RGB-D images in 2D object detectors and describes the selection of a suitable model for the application. Furthermore the creation, annotation and augmentation of a custom dataset is presented. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - object detection KW - convolutional neural network KW - RGB-D image Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14986 SP - 217 EP - 228 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Falk, Constantin A1 - van de Sand, Ron A1 - Corasaniti, Sandra A1 - Reiff-Stephan, Jörg T1 - A Comparison Study of Data-Driven Anomaly Detection Approaches for Industrial Chillers T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) N2 - Faults in industrial chiller systems can lead to higher energy consumption, increasing wear of system components and shorten equipment life. While they gradually cause anomalous system operating conditions, modern automatic fault detection models aim to detect them at low severity by using real-time sensor data. Many scientific contributions addressed this topic in the past and presented data-driven approaches to detect faulty system states. Although many promising results were presented to date, there is lack of suitable comparison studies that show the effectiveness of the proposed models by use of data stemming from different chiller systems. Therefore this study aims at detecting a suitable data-driven approach to detect faults reliable in different domains of industrial chillers. Thus, a unified procedure is developed, to train all algorithms in an identical way with same data-basis. Since most of the reviewed papers used only one dataset for training and testing, the selected approaches are trained and validated on two different datasets from real refrigeration systems. The data-driven approaches are evaluated based on their accuracy and true negative rate, from which the most suitable approach is derived as a conclusion. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - fault detection KW - refrigeration system KW - data-driven machine-learning Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14970 SP - 165 EP - 174 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - van de Sand, Ron A1 - Corasaniti, Sandra A1 - Reiff-Stephan, Jörg T1 - Positive-Unlabelled Learning based Novelty Detection for Industrial Chillers BT - A Data-Driven Approach to Avoid Energy Wastage T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) N2 - Chiller systems are used in many different applications in both the industrial and the commercial sector. They are considered major energy consumers and thus contribute a non-negligible factor to environmental pollution as well as to the overall operating cost. In addition, chillers, especially in industrial applications, are often associated with high reliability requirements, as unplanned system downtimes are usually costly. As many studies over the past decades have shown, the presence of faults can lead to significant performance degradation and thus higher energy consumption of these systems. Thus, data-driven fault detection plays an ever-increasing role in terms of energy efficient control strategies. However, labelled data to train associated algorithms are often only available to a limited extent, which consequently inhibits the broad application of such technologies. Therefore, this paper presents an approach that exploits only a small amount of labelled and large amounts of unlabelled data in the training phase in order to detect fault related anomalies. For this, the model utilizes the residual space of the data transformed through principal component analyses in conjunction with a biased support vector machine, which can be ascribed to the concept of semi-supervised learning, or more specifically, positive-unlabelled learning. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - chiller CBM KW - machine learning KW - energy efficiency Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14965 VL - 1 SP - 155 EP - 164 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Paetow, Clara A1 - Szymanski, Mateusz A1 - Detje, Johannes Siljas A1 - Stolpmann, Alexander T1 - A Model-Based Approach to Decarbonize an Island's Energy System BT - Case Study on the Island of Föhr, Germany T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) N2 - To achieve climate goals and contain further global warming, it is inevitable to reduce CO2 emissions especially in energy consumption. A way to do so is by integrating renewable energy sources (RES) into an energy system’s power generation. However, there is no standard procedure to decarbonise a locally restricted system. Therefore, the various local conditions have to be analysed and taken into consideration. The authors propose a model-based approach to decarbonise the energy system of the island Föhr, Germany. This includes various collected data sets on local conditions such as climate data and heat and power demand. The data is used to represent the island’s energy system and design a model-based solution in a simulation software. The authors identify potentials by comparing costs and revenues by addressing the deployment of different RES technologies. One finding is that heat generation causes 91 % of CO2 emissions making it the major producer. However, with the designed solution, emissions could be reduced to a third. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - decarbonisation KW - renewable energy KW - island Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14952 SP - 145 EP - 153 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Gribova, Maria A1 - Giese, Lutz B. T1 - Sustainably Produced Hydrogen BT - Possible Variants and its Main Supply Paths T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021 N2 - The fossil fuels used to provide energy, such as coal, crude oil and natural gas, are largely responsible for carbon dioxide emissions and other so-called greenhouse gases. Hydrogen gas (H2) can make a key contribution to decarbonization. It can be produced using various processes. Several processes are available to produce hydrogen, such as (i) steam reforming, (ii) cracking process or (iii) electrolysis. Depending on the source of origin, there is a distinction made between different “colours”. Gray, blue, turquoise, yellow and green hydrogen is avail­able, the latter made using Renewable Energies. However, items such as (i) possible variants of hydrogen, (ii) leading export countries or (iii) provision paths to be preferred in the future will be discussed in this paper. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - hydrogen KW - green gas supply KW - renewable energy source Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14940 SP - 107 EP - 116 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Giese, Lutz B. A1 - Reiff-Stephan, Jörg T1 - Hydrogen and Usability of Hydrogen Storage Technologies T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) N2 - Science, technology and politics agree: hydrogen will be the energy carrier of the future. It will replace fossil fuels based on a sufficient supply from sustainable energy. Since the possibilities of storing and transporting hydrogen play a decisive role here, the so-called LOHC (Liquid Organic Hydrogen Carriers) can be used as carrier materials. LOHC carrier materials can reversibly absorb hydrogen, store it without loss and release it again when needed. Since little or no pressure is required, normal containers or tanks can be used. The volume or mass-related energy densities can reach around a quarter of liquid fossil fuels. This paper is to give an introduction to the field of hydrogen storage and usage of those LOHC, in particular. The developments of the last ten years have been related to the storage and transport of hydrogen with LOHC. These are crucial to meet the future demand for energy carriers e.g. for mobile applications. For this purpose, all transport systems are under consideration as well as the decentralized supply of rural areas with low technological penetration, e.g. regions of Western Africa which are often characterized by a lack of energy supply. Hydrogen bound in LOHC can provide a hazard-free alternative for distribution. The paper provides an overview of the conversion forms as well as the chemical carrier materials. Dibenzyltoluene as well as N-ethylcarbazole - as examples for LOHC - are discussed as well as chemical hydrogen storage materials like ammonia boranes as alternatives to LOHC. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - hydrogen storage KW - LOHC KW - climate-neutral mobility Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14937 SP - 97 EP - 106 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Schmidt, Marcus A1 - Giese, Lutz B. T1 - Bi-facial Open-Space Photovoltaic Systems versus Conventional Systems using Mono-facial Modules T2 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) N2 - As part of a scientific work within the solar company Sunfarming GmbH, the aim was to find out whether bi-facial modules on open spaces deliver better results economically than conventional mono-facial solar modules. In this context, an already installed 750 kWp PV system with mono-facial solar modules was compared directly with a structurally identical PV system with bi-facial modules, which, however, does not exist in practice but was only simulated with PV software. The second part of the investigation includes the comparison of four different assembly systems or elevation variants in order to determine the system with the best relationship between system yield and costs. The final result of the first investigation showed that the use of bi-facial modules reduced the specific costs per kWh by approximately 5 %. In order to improve this effect, the use of compact assembly systems is recommended, e.g. five rows of modules per table with horizontal alignment. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 KW - bi-facial photovoltaics KW - open-space PV system KW - renewable energy source KW - solar energy Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14922 SP - 37 EP - 42 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP ED - Reiff-Stephan, Jörg ED - Amouzou, Kou 'santa ED - Adanlete Adjanoh, Assiongbon T1 - 2nd German-West African Conference on Sustainable, Renewable Energy Systems (SusRES2021) BT - 6th April 2021 – Kara, Togo : Conference Proceedings N2 - Renewable energy sources are more important today than ever before, as they represent a key factor in reducing greenhouse emissions and fossil fuel consumption. Avoiding excessive burdens on future generations has now become the declared challenge of this century and justifies the efforts we must undertake to reduce global warming. This can only be achieved by conceiving new technical approaches, exploiting alternative resources and by raising people's awareness of the sustainable use of their own livelihoods. However, this change in awareness is a process that needs to be driven forward, especially in the field of research and education, and is therefore the declared aim of the Joint German-West African Conference on Sustainable, Renewable Energy Systems – SusRES. This annual conference provides a platform for exchange between scientists, students and many other interested parties, with technological progress at the forefront of the event. The SusRES conference is known as an accelerator for an ever-growing international network, not limited to the circumstances of a single country or even continent, which strives to take on and meet the global challenges of the present and future. This year's guiding theme of the event -Decentralized systems as a key factor in clean energy production for locations with limited infrastructure- of the event is primarily based on the consideration of how the energy supply of the future is to be designed. For some years now, a paradigm shift towards decentralised energy supply has been observed, which opens up new possibilities especially in areas with weak infrastructural supply. This is not just about providing energy from renewable resources, but also about exploiting this energy we have today efficiently and sustainably, whereby intelligent and interconnected systems play an ever-increasing role. The reader is provided with exciting insights into these topics within this conference proceedings elaborated by stakeholders from twelve time zones and four continents. In their contributions, the authors address important findings, especially from research but also from teaching, which represent an important milestone in their activities and thus make a significant contribution to our declared goals. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 1 Y1 - 2021 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-14887 PB - TIB Open Publishing ER - TY - CHAP A1 - Fiebelkorn, Richard A1 - Kugel, Rafael A1 - Günther, Norman A1 - Reiff-Stephan, Jörg ED - Reiff-Stephan, Jörg ED - Beuster, Anja T1 - Predictive Maintenance in Tree Care - TreeAngel T2 - Wildauer Konferenz für Künstliche Intelligenz 2025 (WiKKI25) N2 - Ensuring the traffic safety of trees poses a significant challenge for urban authorities. Conventional manual inspection methods are both time-consuming and resource-intensive, and they are subject to human error. The following paper presents an innovative system for automated tree condition assessment using modern camera technologies and artificial intelligence (AI). As part of a feasibility study, image data generated by various camera systems was analyzed. Based on this data, a YOLOv8 model was trained, which enables precise detection of trees and damage, such as deadwood. The results of the prototype system presented are promising in terms of accuracy and efficiency, suggesting the potential to supplement or replace manual spections with automated procedures. The results of this study lay the foundation for sustainable and scalable approaches in tree care and can contribute to increasing public safety and efficiency in urban management. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 2 KW - machine learning KW - predictive maintenance KW - tree care KW - YOLO Y1 - 2025 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-20801 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Ojo, Emil A1 - Walther, Mathias ED - Reiff-Stephan, Jörg ED - Beuster, Anja T1 - Emotion Perception in AI-Generated Facial Expressions T1 - Emotionswahrnehmung in KI-generierten Gesichtsausdrücken T2 - Wildauer Konferenz für Künstliche Intelligenz 2025 (WiKKI25) N2 - This study analyses whether facial expressions in generated images are perceived in the same way as in real images. It examines the perception of emotional intensity and sympathy, the ability to recognise emotions, and the distinction between real and generated images. To answer these questions, an online survey was conducted with more than 200 participants. The results show slight to moderate differences in perception between generated and real images, but it also becomes clear that the distinction between the two is becoming increasingly difficult. N2 - Diese Studie analysiert, ob Gesichtsausdrücke in generierten Bildern genauso wahrgenommen werden wie in realen Bildern. Dabei werden die Wahrnehmung der Emotionsintensität und Sympathie, die Fähigkeit zur Emotionserkennung sowie die Unterscheidung zwischen echten und generierten Bildern untersucht. Zur Beantwortung dieser Fragen wurde eine Online-Umfrage mit mehr als 200 Teilnehmern durchgeführt. Die Ergebnisse zeigen geringfügige bis moderate Unterschiede in der Wahrnehmung zwischen generierten und realen Bildern, allerdings wird auch deutlich, dass die Unterscheidung zwischen beiden zunehmend schwieriger wird. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 2 KW - artificial intelligence KW - emotion perception KW - generative AI KW - Künstliche Intelligenz KW - Generative KI KW - Emotionswahrnehmung Y1 - 2025 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-20799 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Kleinsorge, Alexander A1 - Fauck, Alexander A1 - Kupper, Stefan ED - Reiff-Stephan, Jörg ED - Beuster, Anja T1 - A Novel Exponential Continuous Learning Rate Adaption Gradient Descent Optimization Method T2 - Wildauer Konferenz für Künstliche Intelligenz 2025 (WiKKI25) N2 - We present two novel, fast gradient based optimizer algorithms with dynamic learning rate. The main idea is to adapt the learning rate α by situational awareness, mainly striving for orthogonal neighboring gradients. The method has a high success and fast convergence rate and relies much less on hand-tuned hyper-parameters, providing greater universality. It scales linearly (of order O(n)) with dimension and is rotation invariant, thereby overcoming known limitations. The method is presented in two variants C2Min and P2Min, with slightly different control. Their impressive performance is demonstrated by experiments on several benchmark data-sets (ranging from MNIST to Tiny ImageNet) against the state-of-the-art optimizers Adam and Lion. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 2 KW - neural network KW - optimizer KW - training Y1 - 2025 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-20785 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Reiff-Stephan, Jörg A1 - Zeise, Dominique ED - Reiff-Stephan, Jörg ED - Beuster, Anja T1 - AI-Based Cooperation Management in the Early Stages of Entrepreneurship: Matchmaking at Start-Up Events T1 - KI-gestütztes Kooperationsmanagement in der frühen Phase des Entrepreneurships BT - Matchmaking auf Gründerveranstaltungen BT - Matchmaking at Start-Up Events T2 - Wildauer Konferenz für Künstliche Intelligenz 2025 (WiKKI25) N2 - In today's dynamic industry, successful start-ups require not only innovative ideas, but also the right partners – artificial intelligence can help bring them together efficiently and in a targeted manner. This article discusses the role of artificial intelligence (AI) in cooperation management at start-up events. The authors present the AI-supported matchmaking tool ‘KIKma’, which uses profile and interest analyses to make targeted cooperation suggestions to founders, investors and mentors. The system is based on a transformer model that classifies participant profiles and generates suggestions for targeted speed dating slots. The results show a high degree of accuracy, but occasional misclassifications occur due to ambiguities. This could be improved by fine-tuning the model or optimising the questionnaires. The aim is to facilitate exchange in the start-up landscape and establish an intelligent, adaptive cooperation network in the long term. The authors see great potential for future applications in innovation management in the continuous optimisation of the matching algorithm. N2 - In der heutigen dynamischen Industrie erfordern erfolgreiche Gründungen nicht nur innovative Ideen, sondern auch die richtigen Partner – Künstliche Intelligenz kann helfen, diese effizient und gezielt zusammenzubringen. Der Beitrag thematisiert die Rolle Künstlicher Intelligenz (KI) im Kooperationsmanagement von Gründungsveranstaltungen. Die Autoren stellen das KI-gestützte Matchmaking-Tool „KIKma“ vor, das Gründern, Investoren und Mentoren anhand von Profil- und Interessenanalysen gezielte Kooperationsvorschläge macht. Das System basiert auf einem Transformer-Modell, das Teilnehmerprofile klassifiziert und für gezielte Speed-Dating-Slots Vorschläge generiert. Die Ergebnisse zeigen eine hohe Treffsicherheit, allerdings treten gelegentlich Fehlzuordnungen durch Ambiguitäten auf. Dies könnte durch ein Finetuning des Modells oder eine Optimierung der Fragebögen verbessert werden. Ziel ist es, den Austausch in der Gründerlandschaft zu erleichtern und langfristig ein intelligentes, adaptives Kooperationsnetzwerk zu etablieren. Die Autoren sehen in der kontinuierlichen Optimierung des Matching-Algorithmus großes Potenzial für zukünftige Anwendungen im Innovationsmanagement. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 2 KW - AI match making KW - cooperation management KW - entrepreneurship KW - KI-Matchmaking KW - Kooperationsmanagement KW - Entrepreneurship Y1 - 2025 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-20779 PB - TIB Open Publishing CY - Hannover ER - TY - CHAP A1 - Wilbers, Simon A1 - van de Sand, Ron A1 - Prell, Bastian A1 - Reiff-Stephan, Jörg ED - Reiff-Stephan, Jörg ED - Beuster, Anja T1 - Agentic Recommender System Konzept für nachhaltiges Wissensmanagement T1 - Agentic Recommender System Concept for Sustainable Knowledge Management BT - KI-befähigtes Wissensmanagement in Unternehmen BT - AI-Enabled Knowledge Management in Companies T2 - Wildauer Konferenz für Künstliche Intelligenz 2025 (WiKKI25) N2 - Unternehmen stehen vor erheblichen Herausforderungen beim Erhalt, bei der Nutzung und Weitergabe zentralen Wissens. Demografischer Wandel, zunehmender Mitarbeiterfluktuation und steigende organisatorische Komplexität erschweren klassische Dokumentations- und Austauschprozesse, welche auf erhebliche Mitarbeitendenmitwirkung angewiesen sind. Mithilfe generativer KI und agentischer Recommender-Systeme lassen sich diese Hindernisse überwinden, indem wichtiges Wissen ”on-the-fly“ und weitestgehend automatisiert erfasst wird. Ein zentrales Element bildet hierbei die Kombination aus semantischen Embeddings, einer Graphdatenbank und spezialisierten KI-Agenten, die Dokumente und Chatverläufe analysieren, Mitarbeiter gezielt mit Vorschlägen unterstützen und Wissen so kontinuierlich aktualisieren. So entsteht eine effektive Wissenskultur mit niedrigen Barrieren für die Nutzung dank minimalem Mehraufwand für die menschlichen Nutzer. N2 - Companies face considerable challenges in retaining, utilising and passing on key knowledge. Demographic change, increasing employee turnover and growing organisational complexity are making traditional documentation and exchange processes, which rely heavily on employee participation, more difficult. With the help of generative AI and agentic recommender systems, these obstacles can be overcome by capturing important knowledge ‘on the fly’ and in a largely automated manner. A central element here is the combination of semantic embeddings, a graph database and specialised AI agents that analyse documents and chat histories, provide employees with targeted support in the form of suggestions and thus continuously update knowledge. This creates an effective knowledge culture with low barriers to use thanks to minimal additional effort for human users. T3 - TH Wildau Engineering and Natural Sciences Proceedings - 2 KW - graph database KW - knowledge management KW - recommendation system KW - semantic search KW - Wissensmanagement KW - Graphdatenbank KW - Semantische Suche KW - Recommender-System Y1 - 2025 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:kobv:526-opus4-20764 PB - TIB Open Publishing CY - Hannover ER -