19767 research outputs found
Sort by
Regression-based sensitivity analysis of key building parameters influencing indoor overheating in existing residential buildings under future climates
No full textClimate change is intensifying the frequency, duration, and severity of heatwaves, increasing the risk of indoor overheating in residential buildings. Housing stocks in cold climates are particularly exposed, as they were predominantly designed to minimise winter heat losses rather than dissipate excess summer heat, limiting their adaptive capacity under future warming. This study assesses summer indoor overheating in an existing cold-climate residential dwelling in Gdynia, Poland, and quantifies the influence of key building parameters using regression-based sensitivity analysis. Building performance simulations were conducted for six future climate scenarios (2030–2090) under RCP 4.5 and RCP 8.5 and compared with a current baseline climate. Indoor overheating was evaluated using three thermal resilience indicators capturing duration, frequency, and severity: Time of Overheating (To), Percentage of Overheating (Po), and Indoor Overheating Degree (IOD). Results demonstrate systematic increases in overheating across all future scenarios, with a disproportionate escalation in cumulative severity relative to duration and frequency, particularly under late-century RCP 8.5 conditions. Regression-based sensitivity analysis identifies solar heat gains through glazing as the dominant driver of overheating across all scenarios. Ventilation acts as a secondary but climate-dependent lever, while infiltration and envelope conductive properties exhibit limited influence within the tested parameter ranges. Results also indicate declining mitigation leverage of individual passive measures under progressively warmer climates, underscoring the need for prioritised and integrated retrofit strategies. The study provides quantitative evidence to support the assessment of overheating resilience and retrofit decision-making under future climate scenarios
An analysis for decarbonized multi-vector energy systems
No full textMany governments consider new nuclear power plants to promote decarbonization. On the one hand, dispatchable nuclear plants can complement fluctuating generation from wind and PV. On the other hand, escalating construction costs and times raise economic concerns. This paper investigates the economic threshold at which nuclear plants are an efficient decarbonization option. Building on an extensive review of construction costs and times, we apply a detailed model of the European energy system to analyze the cost-efficient share of nuclear power in fully decarbonized energy systems in 2040. Our analysis finds that even if, reversing the historical trend, overnight construction costs of nuclear half to 4,000 US-$2018 per kW and construction times remain below ten years, the cost-efficient share of nuclear power in European electricity generation is only around 10%. Nuclear plants must operate inflexibly and at capacity factors close to 90% to recover their investment costs, implying that operational flexibility – even if technically possible – is not economically viable. As a result, grid infrastructure, flexible demand in multi-energy systems, and storage are more efficient options for integrating fluctuating wind and photovoltaic generation. The findings suggest that nuclear power should not be relied on for flexibility in future power systems.TU Berlin, Open-Access-Mittel – 202
Arm-current-limiting control for modular multilevel converters with combined AC and DC grid-forming control
No full textGrid faults pose a particular challenge to grid-forming (GFM) converters: The voltage-source behavior of a GFM converter requires current-limiting control to protect the semiconductors. In modular multilevel converters (MMCs), current limits must be imposed on arm currents rather than output currents to reliably limit current through the semiconductors. In addition, currents during transients may deplete the energy of one or more arms. Arm energy depletion may impose voltage limits that differ between phases. Voltage limitation has adverse effects on output voltages and controller function. This work proposes a strategy for arm current and arm voltage limitation for a grid-forming MMC. The current limitation strategy allows for open-loop voltage control; the current controller is active only during limitation. Current and voltage limitations are designed to prioritize terminal behavior, reducing distortion of the AC and DC terminal voltages and improving grid support during faults. The effectiveness of the approach is demonstrated through simulations of phase-angle jumps and an unbalanced voltage sag.TU Berlin, Open-Access-Mittel – 202
Carbazole‐based thin microporous polymer films for photocatalytic hydrogen evolution
No full textPhotocatalytic hydrogen evolution is a direct pathway to store solar energy in chemicals. Conjugated microporous polymers (CMPs) are porous organic photocatalysts, that are typically applied in powder form in heterogenous catalytic reactions. However, the use of powder photocatalysts in dispersion poses some major challenges when it comes to practical applications in larger scales. In this manuscript, the photocatalytic performance of a carbazole-based porous organic polymer (C-POP) film produced by electro polymerizing 1,2,3,5-Tetrakis(carbazol-9-yl)-4,6-dicyanobenzene (4CzIPN) is investigated, well-known for its intriguing photocatalytic properties. The thickness of the intrinsic microporous film is tuneable by the amount of cyclic voltammetry cycles but it is shown that the hydrogen production is not dependent on film thickness. It can therefore be concluded that catalysis is mainly occuring on the outer surface of the films, questioning whether high surface areas are always required for efficient photocatalysis. A microstructured film offers the advantage that, with a reduced amount of polymer material, a constant or even increased external surface area of the film can be achieved. The approach presented here is therefore advantageous for achieving high hydrogen production per unit area with minimal amounts of polymer, as very thin layers are already sufficient for high activity.TU Berlin, Open-Access-Mittel – 202
Machine learning approach to fast thermal equilibration
No full textWe present a method to design driving protocols that achieve fast thermal equilibration of a system of interest using techniques inspired by machine learning training algorithms. For example, consider a Brownian particle manipulated by optical tweezers. The force on the particle can be controlled and adjusted over time, resulting in a driving protocol that transitions the particle from an initial state to a final state. Once the driving protocol has been completed, the system requires additional time to relax to thermal equilibrium. Designing driving protocols that bypass the relaxation period is of interest so that, at the end of the protocol, the system is either in thermal equilibrium or very close to it. Several studies have addressed this problem through reverse engineering methods, which involve prescribing a specific evolution for the probability density function of the system and then deducing the corresponding form of the driving protocol potential. Here we propose a new method that can be applied to more complex systems where reverse engineering is not feasible. We simulate the evolution of a large ensemble of trajectories while tracking the gradients with respect to a parametrization of the driving protocol. The final probability density function is compared to the target equilibrium one. Using machine learning libraries, the gradients are computed via backpropagation and the protocol is iteratively adjusted until the optimal protocol is achieved. We demonstrate the effectiveness of our approach with several examples.TU Berlin, Open-Access-Mittel – 202
Synthesis, characterization, and catalytic testing of η‐carbide‐type phases in the system Fe–Co–Mo–N
No full textFive phases of Fe3–xCoxMo3N were synthezised and characterized by energy-dispersive X-ray spectroscopy, hot gas extraction analyses and X-ray diffraction. The catalytic activity for the decomposition of ammonia was investigated. The data was compared with the results for Fe3Mo3N. Co3Mo3N and Fe3Mo3N show a higher activity than the solid solutions.TU Berlin, Open-Access-Mittel – 2025BMFTR, 03HY203C, Verbundvorhaben TransHyDE_FP3: Reformierung von Ammoniak - Transport von H2 über Derivate - Teilvorhaben der TU Berlin: Multinäre (Oxid-)Nitrid-Katalysatoren und nachhaltige Reaktorkonzepte für Niederdruck- und Hochdruckrouten der Ammoniakreformierun
Agent-based platform development in the context of infrastructure-enhanced mobility
No full textIn der jüngeren Vergangenheit haben sich Plattformen zunehmend als Beschleuniger für Innovation und Interaktion etabliert. Einerseits sind hier technische Plattformen wie Firebase zu nennen, welche den Prozess von der konkreten Idee hin zur Lösung beschleunigen. Zum anderen sind dies Plattformen, die die Kommunikation und Interaktion zwischen unterschiedlichen Einheiten ermöglichen, wie beispielsweise soziale Medien.
Im Folgenden wird das Thema der Plattformentwicklung im Rahmen der infrastrukturgestützten Mobilität erörtert. Im Rahmen dessen werden Konzepte entwickelt, die die Realisierung solcher Plattformen ermöglichen. Zunächst wird elaboriert, inwiefern datenschutzkonform und föderal Modelle der Künstlichen Intelligenz auf einer solchen Plattform trainiert werden können. Es wird eine neue Architektur vorgestellt, die es ermöglicht, Modelle im Rahmen des föderalen Lernens zuverlässig zu trainieren, ohne das Risiko eines Ausfalls des zentralen Aggregators einzugehen. Die Ergebnisse zeigen, dass die Architektur verfügbar bleibt, auch im Falle eines Ausfalls der zentralen Instanz. Zusätzlich wird die Frage beantwortet, wie die Kommunikation und Interaktion zwischen der Plattform und den darauf befindlichen Diensten standardisiert werden kann. Dazu wird ein Agenten-basiertes Framework präsentiert, welches die Realisierung dieser Standardisierung ermöglicht. Hierbei wird nachgewiesen, dass Agenten welche in dem Framework entwickelt werden, sich nahtlos in bestehende Systeme integrieren lassen, mit Komponenten außerhalb des Frameworks kommunizieren können, inhärent wiederverwendbar sind und den Entwicklungsprozess neuer Anwendungen erheblich vereinfachen. Abschließend wird das Framework anhand von zwei Anwendungsbeispielen im Kontext der Mobilität validiert. Eines der Anwendungsfälle umfasst eine Datenerfassungspipeline, mit der Bilddaten auf der Edge erfasst und anonymisiert werden können. Die entwickelte Pipeline zeigt bei verschiedenen Konfigurationen eine gleichbleibende Leistung, unabhängig von der zeitlichen Dauer und der Dichte des Datensatzes. Abschließend wird ein System zwecks adaptiven Fahrstrategien bei wechselnden Wetterbedingungen vorgestellt, mit dem Vehikel mittels einer Pub-and-Subscribe Methode und einem leichtgewichtigen Modell der Künstlichen Intelligenz über verändernde Wetterbedingungen informiert werden können, sodass sie ihr Verhalten anpassen. Das entwickelte Modell weist eine hohe Genauigkeit bei der Vorhersage der Wetterbedingungen vor. Die Architektur selbst arbeitet effizient und informiert das Fahrzeug schnell über Wetterveränderungen.In the recent past, platforms have increasingly established themselves as accelerators for innovation and interaction. There are technical platforms, such as Firebase, that accelerate the process from idea to solution, and there are platforms that enable communication and interaction between different entities, such as social media.
The topic of platform development in the context of infrastructure-enhanced mobility is discussed in the following thesis. As part of this, concepts are developed that enable the realization of such platforms. First, the extent to which artificial intelligence models can be trained on such a platform in compliance with data protection regulations and in a federated manner is elaborated. A new architecture is presented that makes it possible to reliably train models in the context of federated learning without the risk of a failure of the central aggregator. The results show that the architecture remains available even in the event of a failure of the central instance. In addition, the question of how communication and interaction between the platform and the services on it can be standardized is answered. For this purpose, an agent-based framework is presented that enables the realization of this standardization. It is shown that agents developed in the framework can be seamlessly integrated into existing systems, can communicate with components outside the framework, are inherently reusable and considerably simplify the development process of new applications. Finally, the framework is validated using two use cases in the context of mobility. One of the use cases involves a data acquisition pipeline that captures and anonymizes image data on the edge. The developed pipeline shows consistent performance with different setups, both in terms of temporal length and density of the dataset. Further, a weather reporting system is presented that can be used to inform vehicles about changing weather conditions using a pub-and-subscribe method and a lightweight artificial intelligence model so that the vehicles can adapt their behavior. The developed model shows a high precision regarding the predicted weather conditions. The architecture itself works efficiently and informs the vehicle quickly about weather changes
Mixture effect in dimethyl ether auto‐ignition
No full textUndesired detonation development is an obstacle in the development of modern combustion systems based on auto-ignition. Excitation time is one of two time scales that affect detonation development. It describes the time interval, during which heat is released. Extending excitation time decreases the propensity to detonation development by inhibiting the coupling between heat release and pressure waves emerging from reactivity gradients, which are often present in technical systems. As excitation time is mixture-dependent, mitigation of detonation development is possible through mixture tailoring. This work investigates the underlying physico-chemical processes that are responsible for the effect of dilution and equivalence ratio on excitation time. The numerical investigation is performed for dimethyl ether/air mixtures at 15 bar, which feature multistage ignition depending on initial temperature. The resulting nonmonotonous evolution of the heat release rate requires to adapt the analysis methods and utilize a novel excitation time definition. Diluted and off-stoichiometric mixtures feature longer excitation times compared to undiluted stoichiometric mixtures, which is favorable for decreasing the detonation propensity of a mixture. The results demonstrate that excitation time is mainly controlled by reactions that affect reactivity and the production of important intermediate species, which are related to the underlying heat release chemistry. Dilution impacts excitation time by thermal effects, related to the diluent's heat capacity, and chemical effects, such as scavenging of important radicals by third-body collision of the diluent. The current work illuminates which physico-chemical processes extend the excitation time when mixture composition changes, which supports future work on mixture tailoring for mitigation of detonation development.TU Berlin, Open-Access-Mittel – 2025DFG, 200291049, Optimization of combustion mixtures for shockless explosion combustion (A08
Diastatic activity of German hop cultivars with respect to variety, crop year, and separated hop cone parts
No full textDry hopping of beer can result in unintended refermentation, also known as hop creep, because of intrinsic hop diastatic activity. The objective of the work described herein was to determine the enzymatic activity across 16 different hop cultivars grown in Germany in crop years 2019, 2020, and 2021. Optimized enzyme kit protocols were used to quantitate hop α-, β-amylase, amyloglucosidase, and limit dextrinase activities, while a recently published method measured hop diastatic activity. Clear varietal distinctions exist, and hops of harvests 2019, 2020, and 2021 were subsequently classified into three groups depending on their enzymatic activity. With respect to different harvest years, the results imply an annual influence on the amylolytic activity of hops in principle, but more monitoring is needed. Processing methods such as pelletization and storage under different conditions showed a minimal impact on enzymatic activity. Based on further sampling from hops of the harvest 2022, it was observed that differences among hop fractions are pronounced, with the vegetative material and strig exhibiting higher enzymatic activity compared to the lupulin fraction.TU Berlin, Open-Access-Mittel – 202
Distinguishing graph states by the properties of their marginals
No full textGraph states are a class of multipartite entangled quantum states that are ubiquitous in quantum information. We study equivalence relations between graph states under local unitaries (LU) to obtain distinguishing methods both in local and in networked settings. Based on the marginal structure of graph states, we introduce a family of easy-to-compute LU invariants. We show that these invariants uniquely identify the entanglement classes of every graph state up to eight qubits and discuss their reliability for larger numbers of qubits. To handle larger graphs, we generalize tools to test for local Clifford (LC) equivalence of graph states that work by condensing large graphs into smaller graphs. In turn, we show that statements on the equivalence of these smaller graphs (which are easier to compute) can be used to infer statements on the equivalence of the original, larger graphs. We analyze LU equivalence in two key settings, with and without allowing for the permutation of qubits. We identify entanglement classes, whose marginal structure does not allow us to distinguish them. As a result, we increase the bound on the number of qubits where the LU-LC conjecture holds from 8 to 10 qubits in the setting where qubit permutations are allowed.TU Berlin, Open-Access-Mittel – 202