Library

Description: In this paper we revisit the a priori turbulent flame speed tabulation (TFST) technique for a given parameter space within the region of flamelet combustion-regimes. It can be used as a subgrid-scale (SGS) model in Large Eddy Simulation (LES). In a first step, stationary laminar flamelets are computed and stored over the progress variable following the ideas of flamelet generated manifolds (FGM). In a second step, the incompressible one-dimensional Navier-Stokes equations supplemented by the equation for the progress variable are solved on a grid that resolves all turbulent scales. Additionally, turbulent transport is implemented via the linear eddy model (LEM). The turbulent flame structures are solved until a statistically stationary mean value of the turbulent flame speed has been reached. The results are stored in a table that could be used by large scale premixed combustion models, e.g. front tracking schemes. First results are compared quantitatively with direct numerical simulations (DNS) taken from literature. Then it is illustrated in one example how the new method could help to fix constants in algebraic models for turbulent flame speeds. Further it is shown how the technique can be extended to incorporate turbulent strain effects. Finally we investigate the effect of the use of detailed and tabulated chemistry under unsteady conditions.

Description: A world-wide used program for the simulation of fire-induced flows is the Fire Dynamics Simulator (FDS) which originally was developed for a purely serial execution on single-processor computing systems. Due to steadily increasing problem sizes and accuracy requirements as well as restrictions in storage capacity and computing power on single-processor systems, the efficient simulation of the considered fire scenarios can only be achieved on modern high-performance systems based on multi-processor architectures. The transition to those systems requires the elaborate parallelization of the underlying numerical methods which must guarantee the same result for a given problem as the corresponding serial execution. Unfortunately, one fundamental serial serial solver of FDS, the pressure solver, only possesses a low degree of inherent parallelizm. Its current parallelization may cause additional numerical errors, casually leading to significant losses of accuracy or even numerical instabilities. In order to ensure that the parallelization errors are limited by the leading error of the numerical scheme such that second order convergence for the whole method can be acchieved, optimized parallelization concepts must be designed. With respect to these considerations this articles gives an overview of the current parallel pressure solver as well as the problems related to it and presents an alternative method, SCARC, to overcome the existing complicacies. Part I explains the theory, concept and implementation of this new strategy, whereas Part II describes a series of validation and verification tests to proof its correctness.

Description: Because CFD programs, like FDS, generally consist of a large number of different components representing the variety of participating numerical algorithms and chemical / physical processes, it is nearly impossible to verify such codes in their entirety, for example with comparisons of fire tests. Instead, a careful verification and validation with respect to the underlying mathematical conditions and applied numerical schemes is indispensable. In particular, error cancelations between single program components can only be detected by such detailed component-level tests. In part I of this article series a conceptual deficiency of the FDS program package with regard to multi-mesh computations was illustrated and an alternative domain decomposition strategy FDS-ScaRC was introduced. In this second part we will present the structure of a comprehensive test concept and the needs for a more mathematically and numerically orientated test procedure that is much more suited for a reliable evaluation than only a simple visual comparison of the numerical results with experimental fire tests. After a general introduction of our test concept we will demonstrate the high potential of the new FDS-\scarc{} technique compared to the FDS-FFT technique which is used in the FDS program package as yet. Based on this concept, we will present a comprehensive set of analytical and numerical test results.

Description: In this paper we propose a technique for a priori turbulent flame speed tabulation (TFST) for a given parameter space in standard combustion-regime diagrams. It can be used as a subgrid-scale (SGS) model in Large Eddy Simulation (LES). In a first step, stationary laminar flamelets are computed and stored over the progress variable following the ideas of flamelet generated manifolds (FGM). In a second step, the incompressible one-dimensional Navier-Stokes equations supplemented by the equation for the progress variable are solved on a grid that resolves all turbulent scales. Additionally, turbulent transport is implemented via the linear eddy model (LEM). The turbulent flame structures are solved until a statistically stationary mean value of the turbulent flame speed has been reached. The results are stored in a table that could be used by large scale premixed combustion models, e.g. front tracking schemes. Results are compared to an algebraic model and to direct numerical simulations (DNS).

Description: The influence of thermal stratification on autoignition at constant volume and high pressure is investigated under turbulent conditions using the one-dimensional Linear-Eddy Model (LEM) and detailed hydrogen/air chemistry. Results are presented for the influence of initial temperature inhomogeneities on the heat release rate and the relative importance of diffusion and chemical reactions. The predicted heat release rates are compared with heat release rates of Chen et al. and Hawkes et al. obtained by two-dimensional Direct Numerical Simulations (DNS). Using the definition of Chen et al. for the displacement speed of the H2 mass fraction tracked at the location of maximum heat release, and a comparison of budget terms, different combustion modes including ignition front propagation and deflagration waves are identified and the results are compared to the DNS data. The LEM approach shows qualitatively and quantitatively reasonable agreement with the DNS data over the whole range of investigated temperature fluctuations. The results presented in this work suggest that LEM is a potential candidate as a sub-model for CFD calculations of HCCI engines.

Description: This diploma thesis deals with the restoration problem in telecommunication networks. The goal is to find a cost minimal capacity capacity assignment on the edges and nodes of a network such that given demands can be satisfied even in case of the failure of an edge or node in the network. Moreover, restrictions on the routing paths (like length restrictions) and hardware constraints have to be satisfied. A Mixed Integer Programming model is presented which takes into account restoration requirements as well as hardware constraints and which abstracts from a particular restoration protocol and failure situation. This abstraction provides new insight into the structure of the network restoration problem and shows that from a mathematical point of view, the commonly used restoration techniques Link Restoration, Path Restoration and Reservation are not as different as they seem to be from a practical point of view. In addition, our model allows (but is not limited to) optimizing working capacity, intended for normal use, and spare capacity, intended for rerouting purposes in case of a failure, in one step. Furthermore, our formulation of capacity cost allows taking into account the effects of discrete, non-linear cost structures which are common in practice. Up to our knowledge, no publication in the existing literature covers all these aspects, let alone in one model, although they are of major practical interest. The model has been implemented in a Branch and Cut framework. The theoretical background of the algorithmic procedure is presented in detail, including computational complexity investigations on the pricing problem. The abstraction from a particular restoration protocol turns out to be useful both from a theoretical and computational point of view. In fact, our investigations suggest a distinction into Local Restoration and Global Restoration rather than into Link Restoration,Path Restoration, Reservation and mixtures of these concepts. In addition to the theoretical aspects of the algorithmic procedure, some implementational details are briefly discussed. Our implementation has been tested on 14 real world instances, which is described in detail. One part of the computational results consists of a comparison of optimal network cost values using diffeent restoration mechanisms, applied to securing either all single node failures, all single edge failures or both. In addition, the effects of a discrete cost structure are investigated, which has rarely been considered yet in literature. Furthermore, the cost ifference between joint and successive working and spare capacity optimization is investigated. In the second part of the computational results, several heuristics for the network restoration problem are compared with respect to both solution quality and time. This diploma thesis deals with the restoration problem in telecommunication networks. The goal is to find a cost minimal capacity capacity assignment on the edges and nodes of a network such that given demands can be satisfied even in case of the failure of an edge or node in the network. Moreover, restrictions on the routing paths (like length restrictions) and hardware constraints have to be satisfied. A Mixed Integer Programming model is presented which takes into account restoration requirements as well as hardware constraints and which abstracts from a particular restoration protocol and failure situation. This abstraction provides new insight into the structure of the network restoration problem and shows that from a mathematical point of view, the commonly used restoration techniques Link Restoration, Path Restoration and Reservation are not as different as they seem to be from a practical point of view. In addition, our model allows (but is not limited to) optimizing working capacity, intended for normal use, and spare capacity, intended for rerouting purposes in case of a failure, in one step. Furthermore, our formulation of capacity cost allows taking into account the effects of discrete, non-linear cost structures which are common in practice. Up to our knowledge, no publication in the existing literature covers all these aspects, let alone in one model, although they are of major practical interest. The model has been implemented in a Branch and Cut framework. The theoretical background of the algorithmic procedure is presented in detail, including computational complexity investigations on the pricing problem. The abstraction from a particular restoration protocol turns out to be useful both from a theoretical and computational point of view. In fact, our investigations suggest a distinction into Local Restoration and Global Restoration rather than into Link Restoration, Path Restoration, Reservation and mixtures of these concepts. In addition to the theoretical aspects of the algorithmic procedure, some implementational details are briefly discussed. Our implementation has been tested on 14 real world instances, which is described in detail. One part of the computational results consists of a comparison of optimal network cost values using different restoration mechanisms, applied to securing either all single node failures, all single edge failures or both. In addition, the effects of a discrete cost structure are investigated, which has rarely been considered yet in literature. Furthermore, the cost difference between joint and successive working and spare capacity optimization is investigated. In the second part of the computational results, several heuristics for the network restoration problem are compared with respect to both solution quality and time.

Description: Mobile cellular communcication is a key technology in today's information age. Despite the continuing improvements in equipment design, interference is and will remain a limiting factor for the use of radio communication. This Ph. D. thesis investigates how to prevent interference to the largest possible extent when assigning the available frequencies to the base stations of a GSM cellular network. The topic is addressed from two directions: first, new algorithms are presented to compute "good" frequency assignments fast; second, a novel approach, based on semidef inite programming, is employed to provide lower bounds for the amount of unavoidable interference. The new methods proposed for automatic frequency planning are compared in terms of running times and effectiveness in computational experiments, where the planning instances are taken from practice. For most of the heuristics the running time behavior is adequate for inter active planning; at the same time, they provide reasonable assignments from a practical point of view (compared to the currently best known, but substantially slower planning methods). In fact, several of these methods are successfully applied by the German GSM network operator E-Plus. The currently best lower bounds on the amount of unavoidable (co-channel) interference are obtained from solving semidefinite programs These programs arise as nonpolyhedral relaxation of a minimum /c-parti tion problem on complete graphs. The success of this approach is made plausible by revealing structural relations between the feasible set of the semidefinite program and a polytope associated with an integer linear programming formulation of the minimum ^-partition problem. Comparable relations are not known to hold for any polynomial time solvable polyhedral relaxation of the minimum ^-partition problem. The appli cation described is one of the first of semidefinite programming for large industrial problems in combinatorial optimization.

Description: In der vorliegenden Dissertation untersuchen wir die Optimierung von ausfallsicheren Telekommunikationsnetzwerken. Wir präsentieren unterschiedliche gemischt-ganzzahlige Modelle für die diskrete Kapazitätsttruktu,, sowie für die Sicherung des Netzes gegen den Ausfall einzelner Komponenten. Die Modelle wurden in einer Kooperation mit der E-Plus Mobilfunk GmbH verwendet. Die theoretischen Resultate wurden in Algorithmen umgesetzt und in das von uns entiickllte Netzwerksoptimierungswerkzeug Discnet (Dimensioning Survivable Capaiitated NETworks) integriert, welches seit mehreren Jahren in der Planung bei E-Plus eingesetzt wird. Wir betrachten das Transportnetzllanungsproblem eines Telekommunikationsanbieters. Dieses Problem setzt auf logischen Kommunikattonsanforerrungen zwischen den Standorten (Knoten) des zu planenden Netzes und potentiell inslallirrbaren Verbindungen (Kanten) zwischen derselben Knotenmenge auf. Ein Kapazitätsmodell stellt die Information bereit, welche Kapazitäten auf den potentiellen Kanten verfügbar sind. Wir betrachten zwei Modelle. Entweder ist eine explizite Liste der verfügbaren Kapazittten gegeben oder eine Menge von sogenannten Basiskapazitäten, die auf jeder Kante indiviuelll kombiniert werden können. Die Basiskapazitäten müßen paarweise ganzzahlige Vielfache voneinander sein. Man beachte, daß diese Eigenschaft von den internationalen Standards PDH und SDH erfüllt wiid. Ein Ausfallsicherheitsmodell stellt die Information bereit, wie das zu planende Netz gegen den Ausfall einzelner Netzkomponenten geschützt werden soll. Wir betrachten sinnvolle Kombinationen der Modelle Diversification, Reservation und Path Restoration. Das erste Modell garantiert Ausfallsicherheit durch kommunikationsbedarfsabhängige Beschränkung des Prozentsatzes, der durch einzelne Netzkomponenten geroutet werden darf. Bei den beiden anderen Modelle können Kommunikationsbedarfe bei Ausfall einer Netzkomponente auf unterschiedliche Weise neu geroutet werden. Ziel der Planung ist eine ktstenminimlle Kapatitätsentscheidung, die eine Routenllanung aller Kommunikationsbedarfe gemäß den Ausfallsicherheitsanforderungen ermöglicht. Wir entwickeln ein Schnittebenenverfahren zur Lösung der betrachteten Optimiergngsrrobleme. Zu diesem Zweck untersuchen wir Polyeder, die mit den verschiedenen Problemen assoziiert sind. Wir präsentieren neue Klassen von Ungleichungen, entwickeln Separationsalgorithmen und Heuristiken. Mit dem Schnittebenenverfahren werden untere und obere Schranken für den Wert von Oitimallösungen berechnet, und daher ist es möglich, Qualitätsgarantien für die berechneten Löungen anzugeben. Parallel zur Beschreibung der implementierten Algorithmen präsentieren wir umfangreiche Tests mit praktisch relevanten Daten, die zu Problemen mit mehr als 2 Billionen Variablen führen.