This thesis presents a CFD and experimental analysis associated with the parameter compressor fouling and a CFD analysis associated with the parameter on-line compressor washing of industrial gas turbines. On-line compressor washing is very popular and quite effective in the industrial gas turbine operational scheme. Many companies apply on-line washing with the engine running at normal speed so as to avoid downtime periods for off-line cleanings that could cause significant economic drawbacks. At this thesis vital parameters affecting compressor cleaning of an industrial gas turbine were examined and combined in such a way so as to provide adequate coverage of the frontal inlet guide vane area which is critical for effective cleaning. The parameters investigated were water injection nozzle position, inclination with respect to the engine centerline, injection velocity and nozzle spray angles. However, before applying compressor washing, compressor fouling comes into consideration. For this purpose a compressor cascade tunnel (test rig) was designed and come into operation in order to examine different levels of fouling. The cascade test rig involved as well a washing kit for future cleaning of the cascade blades. This work related to the cascade design released a lot of information about designing suction type compressor cascade test rigs by analysing the flow inside the cascade rig computationally and three-dimensionally via CFD tools. The results in terms of the quality of flow obtained for the current test rig were also compared with modified versions of the test rig, one which involved a bigger plenum area behind the cascade test section and one involving the current rig running in a blowing type mode. The CFD results coming out from the compressor cascade tunnel for the different fouling levels, were analysed in terms of mass flow capacity and polytropic efficiency reduction due to fouling by using Howell’s theory (1945) and they were used as inputs for running performance simulation in terms of an industrial gas turbine engine using the performance simulation code Turbomatch. Therefore, a correlation between cascade fouling and real engine uniformly stage spread fouling was achieved. At high levels fouling where the 254 microns roughness height takes place, the nondimensional air mass flow reduction can reach levels of 1.6% and the drop in compressor efficiency can touch the value of 5%. The CFD results obtained after running all the simulation scheme for the different roughness levels, were compared to the actual experimental results coming from running the compressor cascade rig with the same fouling scheme of roughness. Applying Howell’s theory (1945), the fouled cascade was correlated to a uniformly fouled stage and a real industrial gas turbine. This time input in the Turbomatch code was the percentage deterioration in compressor efficiency calculated from correlated cascade data. This deterioration reaches a high level of 11 % when the fouling particle size is 254 microns.

No description available.

No description available.

Uncertainty analysis in computer models has seen a rise in interest in recent years as a result of the increased complexity of (and dependence on) computer models in the design process. A major problem however, is that the computational cost of propagating uncertainty through large nonlinear models can be prohibitive using conventional methods (such as Monte Carlo methods). A powerful solution to this problem is to use an emulator, which is a mathematical representation of the model built from a small set of model runs at specified points in input space. Such emulators are massively cheaper to run and can be used to mimic the "true" model, with the result that uncertainty analysis and sensitivity analysis can be performed for a greatly reduced computational cost. The work here investigates the use of an emulator known as a Gaussian process (GP), which is an advanced probabilistic form of regression, hitherto relatively unknown in engineering. The GP is used to perform uncertainty and sensitivity analysis on nonlinear finite element models of a human heart valve and a novel airship design. Aside from results specific to these models, it is evident that a limitation of the GP is that non-smooth model responses cannot be accurately represented. Consequently, an extension to the GP is investigated, which uses a classification and regression tree to partition the input space, such that non-smooth responses, including bifurcations, can be modelled at boundaries. This new emulator is applied to a simple nonlinear problem, then a bifurcating finite element model. The method is found to be successful, as well as actually reducing computational cost, although it is noted that bifurcations that are not axis-aligned cannot realistically be dealt with.

No description available.

No description available.

This thesis focuses on the design and development of a rolling-ball mechatronic system for on-line testing and measurement of surface contact fatigue of dental composites, and is based on a technique initially developed at the Newcastle Dental School. The mechatronic system synergistically combines the mechanical/electronic hardware with a low-cost embedded digital signal controller (DSC microcontroller) hardware and software to monitor and measure in real-time surface wear due to contact fatigue. ISO/TS 14569-2.2001 standard specification for testing of dental materials was used for selecting appropriate test variables. The mechatronic system attempts to simulate the human oral environment with temperature and moisture being controlled. A closed-loop PI control algorithm combining both optical encoder pulse timing and counting methods is used to drive a dc brushless motor at speeds of 240 and 2040 rpm. A small (2mm diameter) ruby ball is mounted in a V-grooved mandrel which over time creates a circular orbital wear path in the dental composite material. One algorithm has been designed to acquire and process the on-line measurement of wear using a linear voltage differential transformer (LVDT), with another monitoring the fatigue cycling process. A graphical user interface (GUI) has also been designed and implemented on a laptop which is connected to the rig embedded controller. A kinematic model of the rolling ball constrained in a V-groove has been developed along with a finite element analysis of the surface deformation. This has been augmented by a comprehensive test programme, in dry, moisturized and elevated temperature (i.e. 37°C), using Synergy D6 specimens. Using ANOVA test, 70% reproducibility of fatigue track measurements was attained. A comparison of LVDT transducer and profilometer measurements indicated 5% consistence with each other. The insight gained from the testing programme sets a basis for an extensive programme to qualify and validate the measurement system basing on ISO/TS 14569-2.2001 specifications.

Particle accelerators are devices that are capable of forcing charged particles to very high energy levels. Modern particle accelerators are required to produce conditions of extremely high electromagnetic fields in the Radio Frequency (RF) range and produce extremely high accelerating gradients. This has led to various practical issues, one of the most significant being the phenomenon of RF breakdown in the accelerating cavities. When exposed to an intense electromagnetic field, a conducting surface can emit electrons. In the case of an accelerating cavity, these electrons are further accelerated by the RF field. Such emissions are capable of inflicting irreversible damage on the cavity surface and need to be avoided. Among the factors responsible for initiating such emission, the quality of the cavity surface in combination with the operating conditions has been identified as the main one. The mechanisms involved in the initiation of RF breakdown need to be better understood and related to the cavity design criteria, such that they can lead to correctly specified and reproducible designs. The cavity surface quality may be characterised in terms of surface finish and its chemical composition, both of which are strongly affected by the manufacturing processes. Unlike previous studies, this research has focused on the analysis of the effects of fabrication procedures on surface quality. The work involved manufacture and surface characterisation of button shaped samples, which were produced using relevant metal forming and polishing techniques, in preparation for future experimentation in a test system at Fermilab, USA. Although RF breakdown may be initiated locally at the metallic surface, its effects propagate globally across the entire cavity. Surface defects and impurities act as emission sites by inducing local field enhancements. Simulation methods were developed in order to simultaneously study electron emission due to local field enhancement and electron propagation across the cavity, with the benefit of being able to perform tracking in 3D and to verify enhancement factors obtained by theoretical measurements using the surface properties observed on the buttons. This research was conducted in close collaboration with Daresbury Laboratory and Lancaster University in UK as well as the MTA group Fermilab USA.

This research concerns the application of the Probability Density Function (PDF) on Large Eddy Simulations (LES) of turbulent reacting flows in a wide range of open flame configurations spanning between the premixed and non-premixed regime. The aim is to validate the applicability of the PDF model on a wide range of flames without any special treatment. Additionally, the \textit{a-posteriori} Chemical Exposive Mode Analysis (CEMA) has been applied to the results in order to examine the flame structure and identify locations of extinction, re-ignition, etc. Four different series of flames are studied, each one of them belonging to a completely different combustion regime. The F1-F3 premixed turbulent flames is the first family of flames where the PDF method is applied. The LES-PDF model is shown to accurately predict the flow field and the scalar field even on a very coarse grid. The simulations were performed on a personal computer, so the computational power was severely restricted. Nevertheless, the PDF model was able to give accurate predictions, so one of the flames was chosen for a further sensitivity analysis. A large number of modelling parameters were studied and the results show little sensitivity to them in contrast to RANS-PDF approaches in premixed flames. Finally, the model is able to capture large scale quenching at qualitatively the correct extinction speed. The Cambridge-Sandia series of swirling stratified flames was also examined. It encompasses a wide range of flames with various combinations of swirl and stratification ratio levels. Four distinct cases were selected and tested. For the most simple flames (SwB1 and SwB5), the model gives excellent prediction for both the flow field and the scalar distribution. The introduction of the additional fields improves slightly the results, especially at locations further away from the nozzle exit. For the flames which exhibit more complex flow fields and complex characteristics (SwB6 and SwB11), the model gives reasonable results, given the complexity of the flow field. The introduction of differential diffusion and heat losses towards the ceramic cap was studied independently on the SwB11 flame and was found to have counteracting effects. Therefore, their combination was tested and was found to give a significant improvement. The next series of flames is the Sydney Swirl flames. The SM1 and SM2 flames are two complex swirling flames with a difficult flow field to capture. The field is composed of recirculating zones and vortex break-down bubble areas. The SM2 has not been tested in the literature and this work is the first modelling approach. The flow field simulation results are reasonable, given the complexity of the flame. The biggest discrepancies are observed close to the nozzle exit. The Chemical Explosive Mode Analysis is also performed to give information about the flame structure. The flame is divided into three distinct zones with the second one being a very large quenching region. The CEMA analysis explains why the flame does not quench, but re-ignites further down. Finally, the Delft III premixed flame is studied, a difficult flame to model as it shows quenching with large extinction pockets despite the moderately low Reynolds number. The major flow characteristics were accurately captured by the simulation and the introduction of the additional stochastic fields improves the results close to the nozzle exit. Contrary to most researchers that model the pilot flow as a single heat source close to the nozzle exit, in this work the pilot flow is modelled as a separate flow stream, something that increased the complexity of the simulations due to the extremely thin pilot rim which was comparable to the cell size. Nevertheless, the model was able to accurately capture the localized extinction throughout the flame and the application of the Chemical Explosive Mode Analysis gave further insight into the structure of the flame.

Fusion blankets are required to operate in a harsh environment under the influence of a number of synergistic physical phenomena, working across several length scales. The ability to model the thermal-hydraulics of a blanket effectively is key for analysis and design purposes. For magnetic confinement reactor blanket designs using a conducting fluid as a coolant and tritium breeder, the difficulties in flow modelling are particularly challenging due to interactions with the large magnetic field. Blanket analysis is an ideal candidate for the application of a code coupling methodology, with a thermal-hydraulic systems code modelling portions of the blanket amenable to 1D analysis, and a CFD or rather CMHD (Computational Magnetohydrodynamics) solver providing detail where necessary. It is the aim of this study to develop a coupled systems code - CMHD based approach to the modelling of fusion blanket thermal-hydraulics. In particular, it addresses some of the problems associated with the flow of electrically conducting fluids in a high magnetic field. This will enable extensive thermal-hydraulic simulations of the blanket and associated systems to be performed, accounting for MHD effects in a computationally efficient manner that lends itself to the design process. Novel analytical solutions have been developed to address the problem of the electromagnetic coupling of flows between adjacent conducting walled ducts, and for the related heat transfer problem. The resulting correlations have been used in the development of a one dimensional thermal-hydraulic systems code, MHD-SYS. The code has been coupled via TCP socket communications to a CMHD solver (mhdFoam) and the resulting coupled solver has been validated for several test cases. Studies have been performed on simple blanket relevant geometries, comprised of a manifold and several ducts, in order to demonstrate the potential of the coupled solver to capture MHD effects such as modified velocity profiles, increased pressure drops and flow redistribution.