On average, one percent of reservoir storage capacity is lost annually to sediment deposition in reservoirs. Several methods for sediment removal do exist, but most are inefficient and costly in terms of money or water usage. One method known as siphoning has been shown to adequately remove sediment, but present knowledge is lacking to optimize this method for removal. Three relationships for sediment transport in a pipe were compared against data collected from a physical model resembling a reservoir siphoning system. None of the three accurately predicted the physical model results. However, some trends among the relationships were observed, indicating that with additional modification to the relationships, parameters could possibly be developed to design a prototype system.

Acoustical characteristics are used to classify the structural and operational conditions in underground pipes with advanced signal classification methods.

This thesis is concerned with the development and study of a pattern recognition system for siphon and sewer condition/defect analysis based on acoustic characteristics. Pattern recognition has been studied and used widely in many fields including: identification and authentication; medical diagnosis and musical modelling. Audio based classification and research has been mainly focusing on speech recognition and music retrieval, but few applications have attempted to use acoustic characteristics for underground pipe condition classification. Traditional CCTV inspection methods are relatively expensive and subjective so remote techniques have been developed to overcome this concern and increase the inspection efficiency. The acoustic environment provides a rich source of information about the internal conditions of a pipe. This thesis reports on a classification system based on measuring the direct and reflected acoustic signals and describing the energy spectrum for each condition/pipe defect. A K-nearest neighbour classifier (KNN) and Support vector machines (SVMs) classifier have been adopted to train the classification system to identify sediment and pipe surface defects by comparing the measured acoustic signals with a database containing a range of typical conditions. Laboratory generated data and field collected data were used to train the proposed system and evaluate its ability. The overall accuracy of the system recognizing blockage and structural aspects in each of the series of experiments varies between 70% and 95%.

Nos travaux portent sur l'analyse des systèmes de production automatisée à l'aide de réseaux de Petri. Le problème posé est de savoir si un système peut se bloquer complètement ou partiellement et si besoin de calculer un contrôleur garantissant son bon fonctionnement. Les systèmes de production se modélisent naturellement à l'aide d'une sous-classe des réseaux de Petri, les S3PRs. Ce modèle a été très largement étudie par le passe conduisant à des méthodes basées uniquement sur la structure du modèle. Dans ce travail, nous généralisons ces travaux aux modèles des WS3PR, une extension des S3PR ou la réalisation d'une active nécessite non par une ressource mais plusieurs ressources d'un même type et pour lesquels nous proposons des techniques originales combinant des éléments de Théorie des graphes et de théorie des nombres, améliorant même les méthodes du passe sur le modèle simple des S3PR.On présente une caractérisation fine de la vivacité d'un tel modèle basée la notion d'attente circulaire. Une attente circulaire peut être vue comme une composante connexe du sous graphe réduit aux transitions et aux places ressources du modèle. Puis nous démontrons que la non vivacité d'un WS3PR est équivalente a l'existence d' ≪ un blocage circulaire dans une attente circulaire ≫. Ce résultat généralise finement la caractérisation de la vivacité d'un S3PR. Apres avoir introduit la notion de ≪ circuits du graphe de ressources ≫ (WSDC), on construit une méthode de contrôle de ces verrous garantissant la vivacité du modèle d'autant plus efficace qu'une méthode de décomposition du réseau est proposée. Enfin, une traduction de traduit la condition de vivacité des WS3PR sous la forme d'un programme linéaire en nombres entiers est établie et des expérimentations ont démontré l'intérêt de la méthode pour contrôle de systèmes l'allocation des ressources. / As an indispensable component of contemporary advanced manufacturing systems, flexible manufacturing systems (FMSs) possess flexibility and agility that traditional manufacturing systems lack. An FMS usually consists of picking and placing robots, machining centers, logistic systems, and advanced control systems. Some of them can be recognized as its shared resources, which result in its flexibility but may lead to its deadlocks. As a classic problem in resource allocation systems, deadlocks may arise in a fully automated FMS and bring about a series of disturbing issues, from degraded and deteriorated system productivity and performance to low utilization of some critical and expensive resources and even long system downtime. Therefore, the analysis of and solution to deadlock problems are imperative for both a theoretical investigation and practical application of FMSs. Deadlock-freedom means that concurrent produc-tion processes in an FMS will never stagnate. Furthermore, liveness, another significant behavioral property, means that every production process can always be finished. Liveness implies deadlock-freedom, but not vice versa. The liveness-enforcement is a higher requirement than deadlock-freedom.From the perspective of the behavioral logic, the thesis focuses on the intrinsicallylive structures and deadlock control of generalized Petri nets modeling flexible manufacturing systems. Being different from the existing siphon-based methods, a concept of intrinsically live structures becomes the starting point to design, analyze, and optimize a series of novel deadlock control and liveness-enforcing methods in the work.The characteristics and essence of intrinsically live structures are identified and derived from subclasses of generalized Petri nets modeling FMSs with complex resource usage styles. In addition, the numerical relationship between initial markings and weights of connecting arcs is investigated and used to design restrictions that ensure the intrinsical liveness of global or local structures.With the structural theory, graph theory, and number theory, the thesis work achieves the goals of deadlock control and liveness-enforcement.The proposed methods are superior over the traditional siphon-based oneswith a lower computational complexity (or a higher computational efficiency),a lower structural complexity, and a better behavioral permissiveness of the controlled system.

Classics and Modern European Languages / M.A. (Ancient History)

The thesis contributes to the theoretical study of Petri net theory. We conduct boundedness and liveness structural analysis of Synchronized Choice nets (SNC) based on fundamental structures in Petri nets and identified as first-order structures. By studying these structures, the study proposes two ways of preserving good properties: addition of second-order structures or other asymmetric structures. Liveness of these new SNC nets is studied based on the concept of siphons and traps. We prove that SNC nets thus formed are structurally bounded and live. The thesis extends this class of nets to those with pure TP and PT first-order structures and explores its structural and marking conditions. Based on this, we introduce a new class of Synchronized Choice nets called Expanded Synchronized Choice nets.

Classics and Modern European Languages / M.A. (Ancient History)