TIB Open Publishing
Not a member yet
1994 research outputs found
Sort by
Predictive Maintenance in Tree Care - TreeAngel
Full text linkDie Gewährleistung der Verkehrssicherheit von Bäumen stellt für städtische Behörden eine große Herausforderung dar.Herkömmliche manuelle Inspektionsmethoden sind sowohl zeitaufwändig als auch ressourcenintensiv und unterliegen menschlichen Fehlern. Der folgende Artikel stellt ein innovatives System zur automatisierten Beurteilung des Baumzustands unter Verwendung moderner Kameratechnologien und Künstlicher Intelligenz (KI) vor. Im Rahmen einer Machbarkeitsstudie wurden Bilddaten analysiert, die von verschiedenen Kamerasystemen generiert wurden. Auf der Grundlage dieser Daten wurde ein YOLOv8-Modell trainiert, das eine präzise Erkennung von Bäumen und Schäden, wie z. B. Totholz, ermöglicht. Die Ergebnisse desvorgestellten Prototyp-Systems sind hinsichtlich Genauigkeit und Effizienz vielversprechend und lassen das Potenzial erkennen, manuelle Inspektionen durch automatisierte Verfahren zu ergänzen oder zu ersetzen. Die Ergebnisse dieser Studie legen den Grundstein für nachhaltige und skalierbare Ansätze in der Baumpflege und können zu einer Erhöhung der öffentlichen Sicherheit und Effizienz in der Stadtverwaltung beitragen.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 treecare and can contribute to increasing public safety and efficiency in urban management
Agentic Recommender System Concept for Sustainable Knowledge Management: AI-Enabled Knowledge Management in Companies
No full textUnternehmen stehen vor erheblichen Herausforderungen beim Erhalt, bei der Nutzung und Weitergabe zentralen Wissens. Demografischer Wandel, zunehmender Mitarbeiterfluktuation und steigende organisatorische Komplexit ¨at erschweren klassische Dokumentations- und Austauschprozesse, welche auf erhebliche Mitarbeitendenmitwirkung angewiesen sind. Mithilfe generativer KI und agentischer Recommender-Systeme lassen sich diese Hindernisse ¨uberwinden, 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 unterst ¨utzen und Wissen so kontinuierlich aktualisieren. So entsteht eine effektive Wissenskultur mit niedrigen Barrieren für die Nutzung dank minimalem Mehraufwand f ¨ur die menschlichen Nutzer. 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
Encasement of Pre-Placed Reinforcement in Injection 3D Concrete Printing: the Effect of Rheology and Process Parameters
Full text linkInjection 3D Concrete Printing (I3DCP), where material is robotically injected into a carrier liquid and remains stable has been successful at producing complex concrete structures. I3DCP is capable of overcoming the directional limitations faced by other additive fabrication methods. However, this technology has been limited to producing compression-only structures, as its thin concrete strands are incapable of withstanding significant tensile loads. A potential solution is the introduction of reinforcement into concrete structures. This study focuses on the injection of a fine grain concrete into a carrier liquid with defined rheological properties and the capability to encase reinforcement bars, which are spatially fixed inside the carrier liquid, during the printing process. The effect of material- und process related parameters on the encasement quality are studied. The rheological parameters of the carrier liquid are varied by solid volume fraction and the addition of viscosity modifying admixtures. The shape of the nozzle (flat/U-Shape), the nozzle traverse speed (ranging from 20 mm/s to 60 mm/s) and the distance from nozzle to rebar (ranging from 5 mm to 15 mm) are systematically studied. The quality of the encasement is evaluated by image analysis. Selected probes are mechanically tested in pull-out-tests. It is observed that with increasing yield stress of the carrier liquid the reinforcement is less encapsulated. This effect can be counteracted by changing the nozzle shape and/or print speed. Finally, the potential and limitations of using reinforcement bars in I3DCP are discussed
Robotic Fibre Winding Reinforcement: Fundamental Structural Engineering Principles
Full text linkThe Digital Fabrication with Concrete (DFC) enables freedom of form. In order to fully leverage this freedom, the reinforcement should be rethought, as well. Such an opportunity is provided by the Robotic Fibre Winding (RFW). This in-situ and on-demand produced fibre reinforced polymer reinforcement can provide endless strands, which can be either digitally deposited on concrete (concrete defines the form) or on a frame with subsequent application of concrete (reinforcement defines the form), and combined particularly successful with the digital shotcrete or Shotcrete 3D Printing (SC3DP).
The current paper critically analyses the results of the three previously published experimental campaigns with glass fibre wound reinforcement: 1) one set of pull-out and direct tensile tests with various RFW reinforcement types embedded in cast concrete, and 2) two sets of four-point bending tests with two different RFW reinforcement types. In all cases the experimental results are re-analysed from the structural engineering point of view, with the special consideration of anchorage length required for the full activation of reinforcement.
On basis of the obtained results it is concluded, that despite theoretically good bond, resulting in required anchorage comparable to this of classical steel bars, many bending tests unexpectedly ended in pull-out failure. Hence, the mode of pull-out failure should be carefully observed in the upcoming experiments. Likewise, the focus should be put in the future on collection of all relevant data required for such detailed investigation of the bonding zone
DREAM-LLMs at LLMs4OL 2025 Task B: A Deliberation-Based Reasoning Ensemble Approach With Multiple Large Language Models for Term Typing in Low-Resource Domains
Full text linkThe LLMs4OL Challenge at ISWC 2025 aims to advance the integration of Large Language Models (LLMs) and Ontology Learning (OL) across four key tasks: (1) Text2Onto, (2) Term Typing, (3) Taxonomy Discovery, and (4) Non-Taxonomic Relation Extraction. Our work focuses on the Term Typing Prediction task, where prompting LLMs has shown strong potential. However, in low-resource domains, relying on a single LLM is often insufficient due to domain-specific knowledge gaps and limited exposure to specialized terminology, which can lead to inconsistent and biased predictions. To address this challenge, we propose DREAM-LLMs: a Deliberation-based Reasoning Ensemble Approach with Multiple Large Language Models. Our method begins by crafting few-shot prompts using training examples and querying four advanced LLMs independently: ChatGPT-4o, Claude Sonnet 4, DeepSeek-V3, and Gemini 2.5 Pro. Each model outputs a predicted label along with a brief justification. To reduce model-specific bias, we introduce a deliberation step, in which one LLM reviews the predictions and justifications from the other three to produce a final decision. We evaluate DREAM-LLMs on three low-resource domain datasets: OBI, MatOnto, and SWEET using F1-score as the evaluation metric. The results, 0.908 for OBI, 0.568 for MatOnto, and 0.593 for SWEET, demonstrate that our ensemble strategy significantly improves performance, highlighting the promise of collaborative LLM reasoning in low-resource environments
LLMs4OL 2025 Overview: The 2nd Large Language Models for Ontology Learning Challenge
Full text linkWe present the results of the 2nd LLMs4OL 2025 Challenge, a shared task designed to evaluate the effectiveness of large language models (LLMs) for ontology learning. The challenge attracted a diverse set of participants who leveraged a broad spectrum of models, including general-purpose LLMs, domain-specific models, and embedding-based systems. Submissions covered multiple subtasks such as Text2Onto, term typing, taxonomy discovery, and non-taxonomic relationship extractions. The results highlight that hybrid pipelines integrating commercial LLMs with domain-tuned embeddings and fine-tuning approaches achieved the strongest overall performance, while specialized domain models improved results in biomedical and technical datasets. Key insights include the importance of prompt engineering, retrieval-augmented generation (RAG), and ensemble learning. This paper presents the second benchmark of LLM-driven ontology learning, serving as an overview of the participants’ contributions to the challenge. Building on this, this overview presents findings, highlights emerging strategies, and offers practical insights for researchers and practitioners seeking to align unstructured language with structured knowledge
Optimization of Hybrid Renewable Energy System Incorporating Heliostats with Thermal Energy Storage, PV and Battery Storage for Enhanced Energy Flexibility and Reliability
Full text linkThe urgent need for sustainable energy solutions to combat climate change and the growing global energy demand has stimulated the development of renewable energy technologies. Among these, hybrid renewable energy systems (HRES) that combine multiple energy sources promise enhanced efficiency, reliability and flexibility compared to single-source systems. This research focuses on a novel HRES configuration that integrates a concentrated solar power named solar tower with thermal energy storage (TES), photovoltaic modules and battery energy storage systems (BESS) in North Togo. The study aimed to maximize the energy produced, while reducing costs and finally evaluated the environmental aspect saved. We use parametric optimization through SAM to determine the optimum levelized cost of energy and net present value. SAM software was used for heliostat modelling. The single-owner model was used for financial analysis. The results demonstrated the project\u27s financial viability and environmental impact, with an LCOE of 9,307,001 indicated some good interest in the proposed HRES. Furthermore, the significant reduction of 25,815.7513 tons of CO2 emissions highlights the project\u27s substantial contribution to sustainability and carbon footprint reduction
Real-Time Image Enhanced Data-Driven Digital Twin (Real-TImE 3DT) for Flux Density Measurements : A Novel Non-Disruptive Universal Approach
Full text linkConcentrated solar power (CSP) plants are considered one of the most attractive renewable energy producers (used for green fuel and electric power). Consequently, it is essential to enhance the different elements of their power cycles, especially the central receiver and the heliostat field. To achieve this, flux density measurement (FDM) is highly recommended. In this work, a novel methodology for FDM is presented, based on the usage of real-time data-driven simulation models in parallel with the traditional camera methods for the real-time enhancement of the results. In order to improve the output, a graph neural network is used, giving as a result a Real-Time Image-Enhanced Data-Driven Digital Twin (Real-TImE 3DT). All the data are obtained from a pre-existing data platform, where the signals from the sensors of the power plant are logged and stored. This way, it is possible to operate the model manually, or to let it work automatically with these sensors\u27 outputs. Latencies smaller than 10 seconds are achieved and the results from the digital twin showed coherence, easy handling and great inter-operability with the neural network enhancement. On the AI-enhancement’s side, the suppression of up-to 80% of the inaccuracies can be expected under parametrically semi-controlled conditions. Further work is being performed in Solar Tower Jülich (STJ) in order to contrast these results against real experimental measurements
Factor Estimation for Accounting Dynamic Start-Up Effects in Solar Heat Plants Using Quasi-Static Modeling
Full text linkSolar heat plants represent a promising solution to decarbonize the industry. However, the industry is reluctant to adopt this technology, largely due to concerns regarding the variability of solar resources. Early-stage projects rely heavily on annual simulation models to inform decision-making processes, underscoring the critical need for accurate predictions of solar plant energy yield. To address this, a methodology has been developed to enhance the precision of heat production estimates in annual steady-state modelling by integrating dynamic effects through the introduction of a heat-up factor. This heat-up factor will be dependent on the initial temperature, the DNI conditions, the heat capacitance of the installation, and the chosen control strategy. The findings demonstrate a significant enhancement in the accuracy of steady-state model simulations with the inclusion of the heat-up factor, effectively capturing the dynamic influence on energy consumption during start-up
Comparison of Shell and Tube Heat Exchangers for CO2 and CO2+SiCl4 Mixtures Transcritical Cycles
Full text linkConcentrated solar power (CSP) plants have great potential for clean energy production, but their electricity cost is higher than that of noncontrollable renewable energy sources. The main ways to lower the electricity cost are equipment cost reduction and plant efficiency increase. Looking at the power unit, the closed supercritical CO2 (sCO2) cycles offer the efficiency advantage over both the steam Rankine and helium Brayton cycles at high turbine inlet temperatures. Further improvement could be achieved by increasing the critical temperatures using mixtures, allowing condensation at temperatures typical for air-cooled condensers. The sCO2 and CO2 mixture cycles are significantly affected by the performance of the recuperative system and require high pressures (comparable to steam cycles). An increase in efficiency compensates for higher complexity in the design and construction of the plant, according to the authors. High thermal power together with high pressures and temperatures demand customized CO2 heat exchanger designs, which makes them a major part of power cycle specific cost. This paper provides a robust technological solution for the implementation of the above cycles in an industrial setup based on shell-and-tube heat exchangers. Based on the thermohydraulic and mechanical design of EMbaffle® Technology, heat exchangers weight reduction can be quantified in the range of 30 to 60% depending on the application, with additional advantages in terms of logistics and installation (footprint, foundations, etc…). Using the CO2+SiCl4 mixture instead of pure sCO2 leads to a lower weight of primary and high-temperature heat exchangers, while the weight of low-temperature heat exchanger increases