This thesis describes the work completed to create long distance directional discharge using the exploding wire technique. A historical literature review gives the previous works completed on exploding wires dating back to 1780. The review also gives the current understanding of the exploding wire phenomena. The design and construction of a test set to create long distance exploding wires is outlined. A capacitor bank, high voltage switch, charging circuit, hand earths design and construction methods are given. The measurement of the impulse and charging voltage and current is explored and solutions found. Design and construction of a Rogowski Coil and associated circuitry is outlined. Development of software made for reconstruction of Rogowski signals is outlined. Results of testing of different wire explosions is given. It is found that if the charge voltage of the capacitors is to high the wire explodes to fast to remove all charge from the capacitors. If the charge is too low the wire may not have enough energy to explode. Conditions for maximum energy discharged by a plasma shrouded exploding wire are found for varying wire diameters and lengths up to 10m.

Electrical arcs are usually created with the breakdown of air, requiring an average electric field (AEF) of at least 100 kV/m in long spark gaps. This thesis explores a novel method of creating long electrical arcs using exploding wires (EWs). Arcs up to 60 m long have been produced with AEFs of just 4.5 kV/m. Extensive observations of the EW process are presented, which demonstrate that the arcs, which are a type of ‘restrike', form via the seldom-reported ‘plasma bead' restrike mechanism. Beads of plasma appear to form at sites of wire fragmentation, and can expand and coalesce into a continuous plasma column. There are strict conditions under which the plasma beads, and hence restrike channels, are produced. A restrike prediction model has been developed to provide a reliable method of producing restrike. The model was derived from the improved understanding of the restrike mechanism, and uses the wire's length and the energy supply voltage and characteristics as inputs. Capability diagrams are then constructed, which allow researchers to easily design experiments that will produce restrike. Extensive descriptions are provided of the experimental environments that were designed and constructed to facilitate long distance EW experiments. Experiments have been performed inside a high voltage laboratory, in the laboratory's outdoor compound, off the laboratory's earth grid and completely off-site. The off-site location allowed vertical experiments, suspended by a weather balloon, to be performed. This led to a theory on artificially triggered lightning, which is one of many exciting future applications that are suggested. It is also predicted that other research groups will be able to create arcs of several hundred metres long.

Thesis (M.E.)--University of Canterbury, 2008. / Typescript (photocopy). Includes bibliographical references (p. [153]-155). Also available via the World Wide Web.

Increasing insensitivity demands on designing and producing munitions necessitates utilizing primarily insensitive initiation trains specifically in missile systems. Exploding Foil Initiator (EFI) is a high voltage detonator that is used as the initiation elementof rocket motor and warhead initiation trains of modern insensitive missile systems. In this thesis, EFI prototypes are designed and manufactured with the knowledge gained from detailed literature studies. An experimental setup is constructed including firing and testing means for EFI prototypes. That experimental setup is capable of firing EFI prototypes from 500 volts to 3000 volts voltage range. Besides, it allows measuring electrical characteristics like current and voltage traces and average velocity of the flyer plates of these prototypes.Using EFI prototypes,detonation tests of HNS &ndash / IV and PBXN &ndash / 5 explosive pellets are carried out.Function times and detonation outputs of the prototypesare measured with the same experimental setup. A numerical study which predicts electrical performance of EFI prototypes and impact characteristics of flyer plates are carried out. Numerical code is validated with the experimental results.

The development of both experimental and diagnostic equipment to assist with simulating the mechanical effects of cold X-ray deposition is covered by this work. This thesis reviews the various experimental techniques suitable for conducting the electromagnetic launch of flyer plates and the chosen technique is developed into a fully functional experimental facility. The development of a bespoke 1-dimensional computer model is also described in the text. A novel current distribution measurement technique is also fully described. This new diagnostic approach will allow the variation of the current across the width of a large conductor to be easily determined which is something not previously demonstrated.

Andy Warhol's Exploding Plastic Inevitable (EPI) was a drug-fueled rock concert-cum-multimedia art event where layers of mediation mixed with immediate experience: The Velvet Underground performed their innovative music in front of films of themselves performing, Factory Superstars danced and performed poetry, various Warhol films projected on the walls, flashing lights flickered on mirrored surfaces, and a crowd of spectators – both famous and unknown – packed in to see and be seen, to dance and trip into the early hours of the morning. The experience of the EPI was a potent combination of alienation, mediation, and commercialization. The EPI was a promotional vehicle for Warhol, Warhol's Factory crew, and the Velvet Underground, but is also a complex example of spectacle that has been under-analyzed in recent scholarship. The EPI's unabashed emphasis on marketing, packaging, consumer goods, and empty celebrity are all manifestations of fears of late capitalist excess, but beneath the veneer of vapidity was an undercurrent of counterculture political activism and social awareness. Original contributions include cultural analysis, interpretation of contemporary reviews and reports, examination of the event's lack of art historical presence, and incorporation of music scholarship into the Warhol historical canon.

Le mécanisme de la transition préarc-arc dans les éléments de coupure du type fusible est encore mal connu à ce jour. La compréhension du phénomène requiert encore de nombreuses données fondamentales tels que la température ou la densité des vapeurs métalliques créées. Des hypothèses sont avancées pour expliquer la différence du temps de préarc prévue par les modélisations effectuées au sein du LAEPT avec celui constaté lors des expérimentations menées dans ce même laboratoire. Cet ouvrage tente de les vérifier par une approche expérimentale menées sur des fils explosés et des rubans fusibles. L’étude expérimentale est complétée par une recherche bibliographique sur les fils explosés. Ce complément propose de nouvelles pistes d’investigation pour la compréhension de l’amorçage de l’arc électrique sur les rubans fusibles. Enfin, toutes les méthodes de diagnostic et les grandeurs obtenues dans les tests les plus représentatifs sont données en fin d’ouvrage. / Transition between prearcing and arcing stage remains not well known. Several fundamental data are needed as temperature or density of metalic vapour created to better understand the phenomenon. Hypothesis are mentioned to explain the difference of prearc time between modelisations and experiences that have been made in LAEPT. This study try to verificate them among experiences made on exploding wire and fuses ribbon. Experimental study is completed by a bibliographic review on exploding wires. This review permits to show other lines of investigation to understand the transition between prearcing and arcing stage on fuses ribbon. Finally, all the methods of diagnostic and the results obtained with the most representative tests are given at the end of the thesis.

Recurrent Neural Networks (RNNs) are pivotal in deep learning for time series prediction, but they suffer from 'exploding values' and 'gradient decay,' particularly when learning temporally distant interactions. Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU) have addressed these issues to an extent, but the precise mitigating mechanisms remain unclear. Moreover, the success of feedforward neural networks in time series tasks using an 'attention mechanism' raises questions about the solutions offered by LSTMs and GRUs. This study explores an alternative explanation for the challenges faced by RNNs in learning long-range correlations in the input data. Could the issue lie in the movement of the representations - how hidden nodes store and process information - across nodes instead of localization? Evidence presented suggests that RNNs can indeed possess "moving representations," with certain training conditions reducing this movement. These findings point towards the necessity of further research on localizing representations.

No description available.

The work in this thesis is separated into two parts. The first part deals with the load planning and sequencing problem for double-stack intermodal railcars, an operational problem found at many rail container terminals. In this problem, containers must be assigned to a platform on which the container will be loaded, and the loading order must be determined. These decisions are made with the objective of minimizing the costs associated with handling the containers, as well as minimizing the cost of containers left behind. The deterministic version of the problem can be cast as a shortest path problem on an ordered graph. This problem is challenging to solve because of the large size of the graph. We propose a two-stage heuristic based on the Iterative Deepening A* algorithm to compute solutions to the load planning and sequencing problem within a five-minute time budget. Next, we also illustrate how a Deep Q-learning algorithm can be used to heuristically solve the same problem.The second part of this thesis considers sequential models in deep learning. A recent strategy to circumvent the exploding and vanishing gradient problem in recurrent neural networks (RNNs) is to enforce recurrent weight matrices to be orthogonal or unitary. While this ensures stable dynamics during training, it comes at the cost of reduced expressivity due to the limited variety of orthogonal transformations. We propose a parameterization of RNNs, based on the Schur decomposition, that mitigates the exploding and vanishing gradient problem, while allowing for non-orthogonal recurrent weight matrices in the model. / Le travail de cette thèse est divisé en deux parties. La première partie traite du problème de planification et de séquencement des chargements de conteneurs sur des wagons, un problème opérationnel rencontré dans de nombreux terminaux ferroviaires intermodaux. Dans ce problème, les conteneurs doivent être affectés à une plate-forme sur laquelle un ou deux conteneurs seront chargés et l'ordre de chargement doit être déterminé. Ces décisions sont prises dans le but de minimiser les coûts associés à la manutention des conteneurs, ainsi que de minimiser le coût des conteneurs non chargés. La version déterministe du problème peut être formulé comme un problème de plus court chemin sur un graphe ordonné. Ce problème est difficile à résoudre en raison de la grande taille du graphe. Nous proposons une heuristique en deux étapes basée sur l'algorithme Iterative Deepening A* pour calculer des solutions au problème de planification et de séquencement de la charge dans un budget de cinq minutes. Ensuite, nous illustrons également comment un algorithme d'apprentissage Deep Q peut être utilisé pour résoudre heuristiquement le même problème. La deuxième partie de cette thèse examine les modèles séquentiels en apprentissage profond. Une stratégie récente pour contourner le problème de gradient qui explose et disparaît dans les réseaux de neurones récurrents (RNN) consiste à imposer des matrices de poids récurrentes orthogonales ou unitaires. Bien que cela assure une dynamique stable pendant l'entraînement, cela se fait au prix d'une expressivité réduite en raison de la variété limitée des transformations orthogonales. Nous proposons une paramétrisation des RNN, basée sur la décomposition de Schur, qui atténue les problèmes de gradient, tout en permettant des matrices de poids récurrentes non orthogonales dans le modèle.