No description available.

A dual-piston type two-stroke spark-ignition free-piston engine generator prototype has been developed. A comprehensive review on recent published researches and patent documents from academia and industrial organisations on free-piston engine generator, especially on the applications for series hybrid electric vehicles, was conducted. Relevant parameters affecting the operating performance and a number of challenges had been identified as the common denominator for this technology. Modelling and simulations using one-dimensional tools were conducted in parallel with the development activities. Three main simulation models for the crankshaft engines were developed, validated and optimised before converted into the free-piston engine model. This was done by using imposed-piston motion sub-model. The two-stroke free-piston engine model had undergone parametric study for valve timing optimisation. This model was validated by using motoring experimental results using the developed free-piston engine generator prototype. From the experimental results, the free-piston engine generator motoring performance was able to meet the targeted cyclic speed and compression pressure for starting. However, the free-piston engine generator operating speed was limited to 5Hz and below due to valve delay inherent in the pneumatic actuators. The motoring results were used to validate the free-piston engine model which showed a good agreement at various starting speeds. Finally, performance and parametric investigations were conducted using the final validated and refined free-piston engine model. From the simulation, it was found that the free-piston engine had similar response to air-fuel ratio and ignition position variations compare to crankshaft engine with the free-piston engine performance was slightly reduced. Further, the reduced frictional losses contributed little to its performance gain. However, the high influence of piston motion around TDC on the engine performance, observed in free-piston engine, could be manipulated to increase its performance significantly.

Industrial gas turbines generally operate at a bit stable power levels and the hot section critical components, especially high pressure turbine blades are prone to failure due to creep. In some cases, plants are frequently shut down, thus, in addition to creep low cycle fatigue failure equally sets in. Avoiding failure calls for proper monitoring of how the lives of these components are being consumed. Efforts are thus being made to estimate the life of the critical components of the gas turbine, but, the accuracy of the life prediction methods employed has been an issue. In view of the above observations, in this research, a platform has been developed to simultaneously examine engine life consumption due to creep, fatigue and creep-fatigue interaction exploiting relative life analysis where the engine life calculated is compared to a reference life in each failure mode. The results obtained are life analysis factors which indicate how well the engine is being operated. The Larson-Miller Parameter method is used for the creep life consumption analysis, the modified universal slopes method is applied in the low cycle fatigue life estimation while Taira's linear accumulation method is adopted for creep-fatigue interaction life calculation. Fatigue cycles counting model is developed to estimate the fatigue cycles accumulated in any period of engine operation. Blade thermal and stress models are developed together with a data acquisition and pre-processing module to make the life calculations possible. The developed models and the life analysis algorithms are implemented in PYTHIA, Cranfield University's in-house gas turbine performance and diagnostics software to ensure that reliable simulation results are obtained for life analysis. The developed life analysis techniques are applied to several months of real engine operation data, using LM2500+ engine operated by Manx Utilities at the Isle of Man to test the applicability and the feasibility of the methods. The developed algorithms provide quick evaluation and tracking of engine life. The lifing algorithms developed in this research could be applied to different engines. The relative influences of different factors affecting engine life consumption were investigated by considering each effect on engine life consumtion at different engine operation conditions and it was observed that shaft power level has significant impact on engine life consumption while compressor degradation has more impact on engine life consumption than high pressure turbine degradation. The lifing methodologies developed in this work will help engine operators in their engine conditions monitoring and condition-based maintenance.

No description available.

The present study aims at conducting a numerical investigation of the classic film cooling scheme of transpiration film cooling and effusion film cooling for validation through computational methods. Steady-state simulations were performed and the flow was considered incompressible with low turbulence. The CFD package FLUENT 6.2.16 was used to solve the Navier-Stokes equations numerically, and the pre-processor, Gambit 2.2.30, was used to generate the required grid. The research aims at perfonning computational predictions on the film cooling performance and the aerodynamics aspect of flat plate film cooling ·on the transpiration and effusion film cooling. It was determined that the proposed scheme and type of geometry, coupled with the hybrid mesh generation, can validate the classic experimental results. with reasonable agreement. Computational predictions on the transpiration film cooling have shown that different boundary conditions used for the porous media may lead to different results, whether it is over-prediction or under-prediction results in comparison with the experimental data. It has been observed for the effusion film cooling, on the case of co-flow coolant , . ejection into the mainstream, that the adiabatic film cooling effectiveness continuously increases with the axial distance towards the leading edge where the flow of the coolant is fully-developed. Furthermore, the streamwise cooling uniformity was better than in the upstream region at the middle region of the test wall. In contrast, the adiabatic film cooling effectiveness for the opposed flow coolant ejection into the mainstream flow was gradually decreasing with the axial distance. Coflow coolant ejection into the mainstream has provide better cooling effectiveness but the oppose flow coolant ejection from the cooling holes has proved to be good aerodynamics in protecting the adjacent wall due to the large area of the film cooling coverage of the combustor wall. The present study was concerned only with the downstream effectiveness aspect on the performance of the coolant mass flow on the geometrical parameters effects; for transpiration film cooling - the pore size, and effusion film cooling - hole diameter, film cooling hole arrangement, number of holes, inclination and orientation of cooling hole with respect to the mainstream flow. The performance related to the heat transfer coefficient and conjugate heat transfer is a prospective topic for future studies. Advanced and innovative cooling techniques are essential in order to improve the efficiency and output power of the gas turbines. The CFD predictions performed have utilised a scalar tracer gas in the coolant flow and has been very effective at visualizing the coolant to the mainstream mixing phenomenon, determining the boundary layer development and directly predicting the adiabatic film cooling effectiveness. Current methods in determining the film cooling effectiveness using the scalar tracer gas concentration facilitate the future study on the conjugate heat transfer pred,iction where the temperature profiles cannot be used because conjugate heat transfer is highly affected by the effect ofthe temperature in the system. The technique· of providing an alternative method using the heat and mass transfer analogy in quantify the cooling effectiveness combines the advantages of using a scalar tracer gas in determining the cooling effectiveness and also provide clear insight into the film cooling structure in the cooling hole and coolant interaction in the mainstream when the experimental method is at 'off-limit'. The results of the present investigations performed were used to validate the computation model. Therefore, this study is of value for those interested in gas turbine cooling.

No description available.

Variable Geometry Turbochargers (VGT) are widely used to improve engine-turbocharger matching and currently common in diesel engines. VGT has proven to provide air boost for wide engine speed range as well as reduce turbo-lag. This thesis presents the design and experimental evaluation of a variable geometry mixed flow turbocharger turbine. The mixed flow rotor used in this study consists of 12 blades with a constant inlet blade angle of +20°, a cone angle of 50° and a tip diameter of 95.2mm. A variable geometry stator has been designed within this work, consists of 15 vanes fitted into a ring mechanism with a pivoting range between 40° and 80°. A novel nozzle vane was designed to have 40° lean stacking (from the axial direction). This geometrically achieves 3-dimensional match with the mixed flow rotor and aims to improve the turbine stage performance. A conventional straight nozzle vane was also constructed in order to have a comparative design to assess the benefits of the new lean vane. The steady flow performance results are presented for vane angle settings of 40°, 50°, 60°, 65° and 70° over a non-dimensional speed range of 0.833-1.667. The tests have been carried out with a permanent magnet eddy current dynamometer within a velocity ratio range of 0.47 to 1.09. The optimum efficiency of the variable geometry turbine was found to be approximately 5 percentage points higher than the baseline nozzleless unit. The peak efficiency of the variable geometry turbine corresponds to vane angle settings between 60° and 65°, for both the lean and straight vanes. The maximum total-to-static efficiency of the turbine with lean vanes configuration was measured to be 79.8% at a velocity ratio of 0.675. The equivalent value with straight vanes configuration is 80.4% at a velocity ratio of 0.673. The swallowing capacity of the turbine was shown to increase with the lean vanes, as much as 17% at 70° vane angle and pressure ratio of 1.7. The turbine pulsating flow performance is presented for 50% and 80% equivalent speed conditions and a pulse frequency range of 20-80 Hz, these frequencies correspond to an engine speed range of 800-3200 RPM respectively. The turbine was observed to go through a period of choking within a pulse for vane angle settings between 60°-70°. The unsteady efficiency of a nozzled turbine was found to exhibit larger deviation from the quasi-steady curve compared to a nozzlesless turbine, by as much as -19.4 percentage points. This behaviour was found to be more pronounced towards the close nozzle settings, where the blockage effect is dominant. The nozzle ring was also shown to act as a 'restrictor' which shields the turbine rotor from being completely exposed to the unsteadiness of the flow. This coupled with the phase shifting ambiguity was shown to result in the inaccuracy of the point-by-point instantaneous efficiency; where as much as 25% of a cycle exhibits instantaneous efficiency above unity. Finally the turbine was tested by adapting to the pulsating flow (20-60 Hz) by cyclic variation in the opening and closing of the nozzle vanes, called Active Control Turbocharger (A.C.T.). The nozzle vane operating schedules for each pulse period were evaluated experimentally in two general modes; natural oscillating opening/closing of the nozzle vanes due to the pulsating flow and the forced sinusoidal oscillation of the vanes to match the incoming pulsating flow. The spring stiffness was found to be a dominant factor in the effectiveness of the natural oscillation mode. In the best setting, the turbine energy extraction was shown to improve by 6.1% over a cycle for the 20 Hz flow condition. In overall it was demonstrated an optimum A.C.T. operating condition could be achieved by allowing the nozzle ring to oscillate naturally in pulsating flow, against an external spring pre-load, which eliminates the use of complex mechanism and external drive. However, the current result suggest the benefits of A.C.T. are best realised in large low speed engines.

No description available.

The aim of this study is to develop a new process for high precision hot forging of Ti-6Al-4V gas turbine blade. In this new process, the work-piece is hot formed and then is clamped between the dies at high pressure for a certain time in order to decrease the distortion and increase the geometric accuracy. The feasibility study of the new process has been carried out in this thesis by using experiments and finite element (FE) modelling, providing a scientific understanding of the process. From the experimental and modelling work, it has been demonstrated that the new process proposed in this thesis is an effective way to reduce distortion in gas turbine blade forging. The study can be divided into three parts: interfacial heat transfer coefficient determination, material modelling, and Ti-6Al-4V hot forming. A closed form method to determine an interfacial heat transfer coefficient (IHTC) was developed, and a one-dimensional heat transfer model was proposed and validated. Heat transfer tests were performed to study the heat transfer between Ti-6Al-4V work-pieces with an initial temperature of 920°C and H13 dies with an initial temperature of 150°C. Temperature histories measured by thermocouples were obtained, and were used as an input for the closed form method. The effects of pressure, glaze thickness and surface roughness on IHTC between Ti-6Al-4V work-pieces and H13 steel dies were studied. Thermo-mechanical properties of Ti-6Al-4V were investigated for a temperature range of 820°C to 1120°C and a strain rate range of 0.1s-1 to 10.0s-1, using a Gleeble thermo-mechanical simulator. The flow softening mechanisms of Ti-6Al-4V during hot forming were studied. A set of unified elastic-viscoplastic constitutive equations for Ti-6Al-4V during hot forming were developed. Plastic strain for alpha and beta phase, isotropic hardening, normalised dislocation density, adiabatic heating, phase transformation, and globularisation of alpha phase were described in the set of constitutive equations. The developed material constitutive model was determined by fitting with experimental strain-stress curves from uniaxial compression tests, using an Evolutionary Programming (EP)-based optimisation method. Good agreements between the experimental and computed results were obtained: the error of predicted stress is under 10%. The general trend exhibited by softening mechanism of Ti-6Al-4V during hot forging is correctly fitted. Hot forming tests were conducted to study the effects on distortion during hot forming strips used as analogous to gas turbine blades. The effect of work-piece thickness and holding time on spring-back were investigated. The unified constitutive equations for Ti-6Al-4V and IHTCs were integrated with commercial FE software DEFORM/2D to simulate material flow and heat transfer in the hot forming process. The FE model was validated by experimental results. A good agreement was obtained between experimental and FE results for temperature history and thickness distribution. The stress state, temperature field and phase volume fraction were obtained from FE simulation. The FE model can be used to predict the extent of distortion in an appropriate FE software, and thus to optimise the new hot forging process for Ti-6Al-4V high precision gas turbine blade.

Engine performance modelling is a major part of the engine design process, in which specialist solvers are employed to predict, understand and analyse the engine’s behaviour at various operating conditions. Sub-idle whole engine performance synthesis solvers are not as reliable and accurate as design point solvers. Lack of knowledge and data result in component characteristics being reverse-engineered or extrapolated from above-idle data. More stringent requirements on groundstart and relight capabilities, has prompted the need to advance the knowledge on low-speed engine performance, thereby requiring more robust sub-idle performance synthesis solvers. The objective of this study, was to improve the accuracy and reliability of a current aero gas turbine sub-idle performance solver by studying each component in isolation through numerical simulations. Areas researched were: low-speed and locked-rotor com- pressor characteristics, low-power combustion efficiency, air blast atomizer and combustor performance at sub-idle, torque-based whole engine sub-idle performance synthesis, and mixer performance at far off-design conditions. The observations and results from the numerical simulations form the contribution to knowledge of this research. Numerical simulations of compressor blades under highly negative incidence angles show the complex nature of the flow, with the results used to determine a suitable flow deviation model, a method to extract blade aerodynamic char- acteristics in highly separated flows, and measure the blockage caused by highly separated flow with operating condition and blade geometry. The study also concluded that the use of Blade Element Theory is not accurate enough to be used at such far off-design con- ditions. The linearised parameter-based whole engine performance solver was converted to used torque-based parameters, which validated against engine test data, shows that it is suitable for low-power simulations with the advantage of having the potential to start engine simulations from static conditions. A study of air-blast atomization at windmilling relight conditions has shown that current established correlations used to predict spray characteristics are not suitable for altitude relight studies, tending to overestimate the atomization quality. Also discovered is the highly influential interaction of compressor wakes with the combustor and atomizer under altitude relight conditions, resulting in more favourable lighting conditions than previous assumptions and models have shown. This is a completely new discovery which will result in a change in the way combustors are designed and sized for relight conditions, and the way combustion rig tests are conducted. The study also has valuable industrial contributions. The locked-rotor numerical data was used within a stage-stacking compressible flow code to estimate the compressor sub- idle map, of which results were used within a whole engine performance solver and results validated against actual engine test data. The atomization studies at relight were used to factor in the insensitivity of current spray correlations, which together with a newly de- veloped sub-idle combustion efficiency sub-routine, are used to determine the combustion efficiency at low-power settings. The interaction of compressor wakes with the atomizer showed that atomizer performance at relight is underestimated, resulting in oversized combustors. By using the knowledge gained within this research, combustor size can be reduced, resulting in lower NOx at take-off and a smaller and lighter core, with a com- bustor requiring less cooling air. The component research has advanced the knowledge and modelling capability of sub-idle performance solvers, increasing their reliability and encouraging their use for future aero gas turbine engines.