Disertační práce

Nejnovější přírůstky:
2024-07-27
00:02
Impact of water deficit stress on leaf quality, predator and herbivore communities in Ficuses
XIAO, Xue
This thesis addresses the mechanism by which drought affects tropical plant-insect interactions via leaf traits, focusing on Ficus species. I discovered that both drought intensity and duration indirectly influenced the leaf palatability of Ficus through changes in leaf mechanical traits such as leaf water content and specific leaf area. Furthermore, my research revealed that drought increased leaf herbivory by altering leaf latex outflow, without affecting insect communities. Additionally, I observed that drought enhanced Ficus indirect defense through changes in herbivore-induced plant volatiles. My findings suggest that the impact of drought on Ficus-insect interactions may vary between greenhouse and tropical rainforest environments and may be species-specific. This work underscores the importance and urgency of researching tropical plant-insect interactions under climate change.

Úplný záznam
2024-07-27
00:02
The role of ecological gradients and host-parasitoid interaction in tropical insect communities
FINNIE, Sam Cameron
The thesis explores the various biotic and abiotic factors that drive change in insect communities with particular emphasis on host-parasitoid interactions. These factors range from ecological gradients to defensive traits and host specialisation. The aim of the first chapter was to determine how the assemblage of an ecologically and anthropocentrically significant group of insects changed in response to altitude. This study highlighted the elevational limit of fruit flies and discovered that a new and promising attractant lure used in agriculture was particularly useful at higher altitudes, potentially opening new avenues for pest management in mountainous regions. The aim of the second chapter was to ascertain the physiological and phylogenetic traits of galling insects that alter the diversity and host specificity of their parasitioids. This research not only uncovered a remarkably specialist community of gall parasitoids, and found significant physiological drivers of parasitoid assemblage, it also broached the possibility of a latitudinal gradient in gall parasitoid host specificity, that warrants further exploration due to its potential implications for biodiversity patterns. The final chapter, aimed to uncover patterns in the vertical stratification of a caterpillar-parasitoid community, and whether any potential patterns and interactions are driven by the defensive traits of the caterpillar hosts. This study provided further evidence that the defensive traits of caterpillars influence their vulnerability to parasitism, and discovered distinct patterns in the vertical stratification of caterpillar communities. Overall, this holistic research expands upon the existing literature on tropical insect ecology and advances our understanding on the complex interactions between insects and their environment.

Úplný záznam
2024-07-21
00:01
2D materials for electrocatalysis and hydrogen generation as clean energy source
Sanna, Michela ; Kim, Daewoo (oponent) ; Urso, Mario (oponent) ; Pumera, Martin (vedoucí práce)
The electrochemical production of hydrogen from water is gaining more attention as a clean and renewable energy source in response to the alarming environmental issues caused by the exploitation of fossil fuels during the last centuries. However, the process can be considered an environmentally friendly alternative only if it is fuelled using renewable sources of energy, like solar energy, the largest carbon-free resource available on our planet. Solar energy can be converted to electricity via solar panels and electrical energy be used for water splitting via electrocatalysts, such as platinum. Alternatively, the water splitting to hydrogen can be carried out directly via solar light energy. However, the yields of direct photochemical water splitting are low. The combination of both approaches, also called photoelectrochemical water splitting, combines the best of both worlds – electrocatalytic water splitting with the aid of photons. For these reason, the study of novel materials based on earth-abundant elements that can be applied as photoelectrocatalysts for hydrogen generation is fundamental to guiding society toward more sustainable energy production. This thesis explores the potential of the emerging two-dimensional (2D) materials and related layered compounds, alongside investigations into the utility of 3D printing for fabricating functional electrodes in the field of photoelectrochemistry. The study of several transition metal selenophosphites confirmed their potential as photoelectrocatalysts for hydrogen generation, in particular under the influence of visible light. MAX phases were modified through exposure to fluorine gas and the properties of the obtained fluorinated MAX were investigated, starting from their morphology to their potential as photoelectrocatalysts for the hydrogen evolution reaction. The fluorinated phases showed better performances compared to the untreated MAX phases. The improved catalytic activity was attributed to photoactive oxyfluorides that formed as a consequence of the fluorination process. The photoactivity of the MAX phases was further investigated both by theoretical and experimental approaches, to understand the origin of the photocatalytic behaviour. The results showed that the presence of oxide impurities on the phases plays a crucial role in the photoelectrochemical production of hydrogen. The role of the oxides in the photocatalytic activity of these compounds inspired the fabrication and investigation of 3D printed electrodes and their modification with atomic layer deposited oxides, like TiO2, SiO2, and Al2O3. Also in this case, the presence of a thin layer of oxide on the surface of the electrode contributed to significantly better performances under the influence of visible light. The obtained results demonstrated the importance of the fundamental study of novel 2D materials for application in the photoelectrochemical production of hydrogen and open new insights into the fabrication of innovative 3D printed conductive devices that can be modified with functional materials for energy conversion.

Úplný záznam
2024-07-21
00:01
Field demagnetised artificial square ice magnets: long-range interactions, origin of stochasticity and effective thermodynamics
Brunn, Ondřej ; Lacour, Daniel (oponent) ; Lassailly,, Yves (oponent) ; Kolařík, Vladimír (vedoucí práce)
In the past two decades, artificial spin ice systems have become a powerful experimental platform to investigate cooperative magnetic phenomena often associated with highly frustrated magnets. Compared to their natural counterparts, artificial spin ice systems made of interacting magnetic nanostructures offer several key advantages. Being engineered through nanofabrication processes, an extensive palette of geometries can be designed. In addition, their magnetic configuration can be visualised directly, at the scale of the spin degree of freedom, using magnetic imaging techniques. Local and global quantities can then be measured conveniently, in real space and time, at almost any desired temperature. This PhD work focuses on such artificial spin systems, and more specifically on the square geometry, which was initially proposed as a two-dimensional (2D) counterpart of the three-dimensional (3D) pyrochlore crystal structure. However, this 2D approach removes the magnetic frustration present in 3D, and the system orders in a conventional antiferromagnetic fashion rather than exhibiting a highly degenerate, liquid-like ground state. Following a strategy proposed in the literature, arrays of nanostructures consisting of two vertically offset sub-lattices were fabricated to restore frustration, enabling to reach a spin liquid regime experimentally. Imaging the magnetic configurations obtained after a field demagnetisation protocol, the analysis of the spin-spin correlations reveals deviations from what is predicted by the (short-range) square ice model. Comparing the experimental findings to Monte Carlo simulations, our results indicate that long-range magnetostatic interactions are not washed out in our arrays, contrary to what was initially thought. Then, these artificial square ice structures were used to understand to what extent the field demagnetisation protocol we apply is a stochastic process. To do so, we studied the magnetic configurations obtained after successive field protocols. Our results show that each captured magnetic micro-state differs substantially from the previous one, but not entirely. Analysing the corresponding spin and vertex configurations, we demonstrate that our field protocol is a stochastic process, although we also observe unambiguous signatures of magnetic determinism that we attribute to the presence of quenched disorder. The possible sources of randomness in our experiment are discussed. Finally, we explore the behaviour of a series of field-demagnetised conventional (non-offset) square arrays, in which the lattice parameter is gradually varied to tune the interaction strengths. Comparing the experimental vertex populations and spin-spin correlations to Monte Carlo predictions, we show that the lattice series is well approximated by a unique short-range spin Hamiltonian probed at different effective temperatures. In other words, the lattice parameter can serve as a knob to probe the thermodynamics of a given spin model.

Úplný záznam
2024-07-21
00:01
Rozhodování andělského investora o exitu v podmínkách trhu rizikového kapitálu v České republice
Vejmělková, Lada ; Doskočil, Radek (oponent) ; Horvátová, Eva (oponent) ; Łapińska, Justyna (oponent) ; Zinecker, Marek (vedoucí práce)
Tato disertační práce předkládá výzkumné výsledky v oblasti identifikace interních a externích faktorů ovlivňujících rozhodování andělských investorů o exitu na trhu rizikového kapitálu v podmínkách České republiky. Výzkum vychází z polostrukturovaných rozhovorů a dotazníkových šetření s 25 andělskými investory. Hlavním nástrojem zpracování získaných dat je obsahová analýza kvalitativních a kvantitativních dat. Data obsažená ve výsledné datové matici jsou podrobena statistické analýze s využitím softwaru IBM SPSS Statistics. Výsledky naznačují, že hlavní formou exitu andělských investorů je obchodní prodej, který je vzhledem k vývoji interních a externích faktorů načasován tak, aby maximalizoval výtěžnost investice. Navíc je exit zpravidla výsledkem oportunistického přístupu k plánování výstupu z cílového podniku. Z výsledků výzkumu dále vyplývá, že se mezi nejdůležitější interní faktory ovlivňující exit andělských investorů řadí důvěra, kvalitativní kritéria hodnocení investice ze strany andělského investora, úroveň technologické inovace a smluvní úprava exitu. V případě externích faktorů jsou pro andělského investora nejdůležitější změna a úroveň budoucí poptávky po cílovém podniku, geopolitický vývoj nebo daňový systém. Výzkumné výsledky jsou využity k formulaci vícekriteriálního rozhodovacího modelu, který slouží jako nástroj podpory rozhodování andělských investorů o exitu.

Úplný záznam
2024-07-21
00:01
Ceramic materials and composites for advanced applications
Mařák, Vojtěch ; Částková, Klára (oponent) ; Michálková,, Monika (oponent) ; Drdlík, Daniel (vedoucí práce)
This thesis focuses on advanced oxide ceramics and particle and laminate composites. The materials studied include bioceramic hydroxyapatite, thermomiotic aluminium tungstate, and lead-free barium titanate-based piezoelectrics. These ceramics suffer from several limitations in their processing and properties, including the use of toxic stabilisers, low densification, susceptibility to cracking, undesirable chemical reactions, and poor mechanical and electrical performance. This work addresses these issues by tailoring individual processing steps and employing novel fabrication and treatment methods to optimise the desired performance of the materials. Plasma treatment of hydroxyapatite powder allowed electrophoretic deposition without toxic stabilisers, resulting in denser and crack-free coatings, as was also showcased on orthopaedic screws. Thermomiotic aluminium tungstate, known for its near-zero coefficient of thermal expansion, was synthesised by optimised co-precipitation to produce a powder with improved sinterability. Rapid pressure-less sintering and spark plasma sintering were used to densify the material to the highest density at the lowest temperature to date. The microstructural evolution of barium titanate piezoceramics was studied using rapid pressure-less sintering and radiation assisted sintering. The gained knowledge was used to sinter barium titanate-based particle composites with reinforcing tougher oxide ceramics. High reactivity between selected materials had a negative influence on the properties of the composites, and one of the reaction products was characterised in greater detail. Laminated composites were prepared by alternating barium titanate-based and dielectric zirconia layers. The high reactivity of piezoelectric materials was inhibited by the sintering strategy. In summary, this dissertation thesis aims to improve the understanding of ceramic material properties and processing techniques, provide insight into performance optimisation, and address critical limitations across diverse applications.

Úplný záznam
2024-07-21
00:01
From Modular to End-to-End Speaker Diarization
Landini, Federico Nicolás ; Ganapathy, Sriram (oponent) ; Bredin, Herve (oponent) ; Burget, Lukáš (vedoucí práce)
Speaker diarization is usually referred to as the task that determines ``who spoke when'' in a recording. Until a few years ago, all competitive approaches were modular, i.e. voice activity detection, segmentation, embedding extraction, clustering and overlapped speech detection and handling were tackled by different sub-systems and applied in a cascaded fashion. Systems based on this framework reached state-of-the-art performance in most scenarios but had major difficulties dealing with overlapped speech. More recently, the advent of end-to-end models, capable of dealing with all aspects of speaker diarization with a single model and better performing regarding overlapped speech, has brought high levels of attention.   This thesis is framed during a period of co-existence of these two trends. We describe a system based on a Bayesian hidden Markov model used to cluster x-vectors (speaker embeddings obtained with a neural network), known as VBx, which has shown remarkable performance on different datasets and challenges. We comment on its advantages and limitations and evaluate results on different relevant corpora. Then, we move towards end-to-end neural diarization (EEND) methods. Due to the need for large training sets for training these models and the lack of manually annotated diarization data in sufficient quantities, the compromise solution consists in generating training data artificially. We describe an approach for generating synthetic data which resembles real conversations in terms of speaker turns and overlaps. We show how this method generating "simulated conversations" allows for better performance than using a previously proposed method for creating "simulated mixtures" when training the popular EEND with encoder-decoder attractors (EEND-EDA). We also propose a new EEND-based model, which we call DiaPer, and show that it can perform better than EEND-EDA, especially when dealing with many speakers and handling overlapped speech. Finally, we compare both VBx-based and DiaPer systems on a wide variety of corpora and comment on the advantages of each technique.

Úplný záznam
2024-07-21
00:01
Investigating the evolution of laser induced plasmas by complementary optical and spectroscopic methods
Buday, Jakub ; Novotný, Karel (oponent) ; Pouzar, Miroslav (oponent) ; Kaiser, Jozef (vedoucí práce)
Laser-Induced Breakdown Spectroscopy (LIBS) is a method of analytical chemistry that provides a qualitative and quantitative sample analysis. As this method has been implemented in more and more applications, certain limitations and drawbacks have been raised from this progress. Some of these limitations are connected to the fact, that the ablation process of a~sample is fast and dynamic and the changes of the plasma plume in time and space are a question of a few microseconds. As this can cause certain limitations in the analysis process, there has been a question of how to minimize or eliminate these problems. For this purpose, supplementary methods have been more and more implemented in the LIBS experiments. The goal is simple, acquire more information from the complex process of the ablation and improve the analytical performance of LIBS, be it in the scope of qualitative or quantitative analysis. Here lies the main motivation of this dissertation thesis. Goal of this thesis was to research, implement and utilize supplementary methods, that could be used in combination with the existing LIBS experimental setup. As a result, direct plasma imaging, shadowgraphy and sound analysis systems have been implemented and streamlined into the typical analytical process. Each of these methods brings certain additional information. For direct plasma imaging, it can be the spacious-temporal morphology of the plasma plume, or the distribution of individual elements within its volume. The Shadowgraphy method can be used to calculate the energy spent on the ablation itself, pointing at the thermal properties of the analyzed sample. Lastly, the sound information from the generated shock wave is connected to the amount of ablated mass. Within this thesis, all of the mentioned supplementary methods were examined, and utilized. Insight into these methods together with the LIBS analysis was done individually for each method as well as from the point of their mutual combination. The theoretical information as well as their background is described and discussed. Moreover, the technical information is discussed and analyzed, be it the implementation of the LIBS experiment, or the actual data analysis process. This work will serve as a foundation for further research in this area, where a selection of these methods will be used in several applications that are already using LIBS but suffer from certain limitations. These methods will be used to mitigate the limitation and to elevate the potential use of the LIBS in certain areas and applications.

Úplný záznam
2024-07-21
00:01
Metodika jízdních testů motocyklu pro identifikace parametrů modelu pneumatik
Tobiáš, Martin ; Ph,D,, Pavel Ščerba, (oponent) ; Magdolen, Luboš (oponent) ; Porteš, Petr (vedoucí práce)
Tématem práce je analýza dynamických jízdních stavů motocyklu v rámci dlouhodobého cíle tvorby nových nástrojů a metod pro účely výzkumu vozidel na Odboru motorových vozidel ÚADI FSI VUT. Cílem práce je vytvoření metodiky umožňující získat z měřené jízdy motocyklu data, využitelná pro tvorbu a ladění modelů pneumatik. Metoda spočívá v realizaci inversních multibody simulací pomocí interně vyvinutých softwarových nástrojů. Ty vyžadují vytvoření věrohodného multibody modelu motocyklu a měřicího systému, který během jízdy s dostatečnou přesností zaznamená veličiny odpovídající stupňům volnosti multibody modelu motocyklu. Součástí vytvářené metodiky je i definice vhodných jízdních manévrů. V práci jsou představeny realizace těchto požadavků, včetně výsledků inversních multibody simulací. Ty jsou v práci dále zpracovány, za účelem ukázky jejich použitelnosti k zamýšlenému účelu. Následně je nabídnuta diskuze nad dosaženými výsledky a možnému pokračování výzkumu.

Úplný záznam
2024-07-21
00:01
Modelování dynamických vlastností převodových ústrojí pomocí virtuálních prototypů
Otipka, Václav ; Trochta,, Miroslav (oponent) ; Žlábek,, Pavel (oponent) ; Píštěk, Václav (vedoucí práce)
Převodová ústrojí s ozubenými koly jsou velmi masivně rozšířena napříč všemi kategoriemi technických výrobků. Jejich funkce je nezastupitelná, nicméně jejich provozování je vždy spojeno s generováním nezanedbatelné úrovně vibrací a hluku, což je také považováno za jednu z hlavních nevýhod, která se projevuje především u vysoce zatěžovaných převodů a představuje nevýhodu hlavně z pohledu konkurenceschopnosti výrobků a případně také možných legislativních problémů. Zmíněný problém lze částečně vyřešit prostřednictvím důmyslné optimalizace převodového ústrojí, což však vyžaduje vynaložení značného úsilí, času a finančních prostředků. Toto je hlavní důvod, proč je vyvíjen neustálý nátlak na nalezení efektivních cest pro optimalizaci a proč je této oblasti stále věnována značná pozornost vědecké komunity. V současnosti se pro celkovou optimalizaci v této oblasti využívá kombinace výpočtového a experimentálního přístupu. Tato disertační práce je zaměřena na predikci a porozumění dynamického chování převodového ústrojí, což si žádá sestavení pokročilého, komplexního výpočtového modelu. Výpočtové modelování je podpořeno řadou dílčích technických experimentů, které slouží k validaci výsledků.

Úplný záznam