Advanced film-cooling systems are necessary to guarantee safe working conditions of high-pressure turbine stages. A fair prediction of the inherent unsteady interaction between the main-flow and the jet of cooling air allows for correctly describing the complex flow structures arising close to the cooled region. This proves to be crucial for the design of high-performance cooling systems. Results obtained by means of an experimental campaign performed at the University of Karlsruhe are shown along with unsteady numerical data obtained for the corresponding working conditions. The experimental rig consists of an instrumented plate where the hot flow reaches Mach = 0.6 close to the coolant jet exit section. The numerical campaign models the unsteady film cooling characteristics using a third-order accurate method. The ANSYS ® FLUENT ® software is used along with a mesh refinement procedure that allows for accurately modelling the flow field. Turbulence is modelled using the k-ω SST model. Time-averaged and time-resolved distributions of adiabatic effectiveness and Net Heat Flux Reduction are analysed to determine to what extent deterministic unsteadiness plays a role in cooling systems. It is found that coolant pulsates due to fluctuations generated by flow separation at the inlet section of the cooling channel. Visualizations of the fluctuating flow field demonstrate that coolant penetration depends on the phase of the pulsation, thus leading to periodically reduced shielding. Eventually, unsteadiness occurring at integral length scales does not provide enough mixing to match the experiments, thus hinting that the dominant phenomena occur at inertial length scales.

Modern industrial gas turbines typically employ lean-premix combustors, which can limit pollutant emissions thanks to premixed flames, while sustaining high turbine inlet temperatures that increase the single-cycle thermal efficiency. As such, gas-turbine first stage nozzles can be characterized by a highly-swirled and temperature-distorted inlet flow field. However, due to several sources of uncertainty during the design phase, wide safety margins are commonly adopted, having a direct impact on engine performance and efficiency. Therefore, aiming at increasing the knowledge on combustor-turbine interaction and improving standard design practices, a non-reactive test rig composed of real hardware was assembled at the University of Florence, Italy. The rig, accommodating three lean-premix swirlers within a combustion chamber and two first stage film-cooled nozzles of a Baker Hughes heavy-duty gas turbine, is operated in similitude conditions. The rig has been designed to reproduce the real engine periodic flow field on the central vane channel, also allowing for measurements far enough from the lateral walls. The periodicity condition on the central sector was achieved by the proper design of both the angular profile and pitch value of the tailboards with respect to the vanes, which was carried out in a preliminary phase via a Design of Experiments procedure. In addition, circular ducts needed to be installed at the injectors outlet section to preserve the non-reactive swirling flow down to the nozzles’ inlet plane. The combustor-turbine interface section has been experimentally characterized in nominal operating conditions as per the temperature, velocity and pressure fields by means of a five-hole pressure probe provided with a thermocouple, installed on an automatic traverse system. To study the evolution of the combustor outlet flow through the vanes and its interaction with the film-cooling flow, such measurements have been replicated also downstream of the vanes’ trailing edge. This work allowed for designing and providing preliminary data on a combustor simulator capable of equipping and testing real hardware film-cooled nozzles of a heavy-duty gas turbine. Ultimately, the activity sets the basis for an extensive test campaign aimed at characterizing the metal temperature, film effectiveness and heat transfer coefficient at realistic aerothermal conditions. In addition, and by leveraging experimental data, this activity paves the way for a detailed validation of current design practices as well as more advanced numerical methodologies such as Scale-Adaptive Simulations of the integrated combustor-turbine domain.

The even more stringent limitations set by the European Commission on pollutant emissions are forcing gas turbine manufacturers towards the redesign of the most important components to increase efficiency and specific power. Current trends in gas turbine design include an increased attention to the design of cooling systems and enhanced best practices for the study of components interaction. At the same time, the recent crisis suffered by the oil and gas industry reduced the interest in brand new gas turbines, thus increasing the service market. Therefore, original equipment manufacturers would rather propose the replacement of specific components within the gas turbine plant during its maintenance with compatible elements that are likely to guarantee increased performance and longer residual lifetime at a more desirable nominal working point. In the present activity the cooling system of the TG20 heavy-duty gas turbine has been redesigned to increase the turbine inlet temperature while contemporaneously reducing the total amount of coolant mass-flow. Specifically, the cooling scheme of the rotating blade of the first turbine row has been reviewed at the Department of Energy (DENERG) of Politecnico di Torino in cooperation with EthosEnergy Italia S.p.a.. The paper presents a new design, which, starting from the original solution featuring fifteen smooth pipes, adopts an improved geometry characterized by the presence of turbulators. The activity has been carried out using Computational Fluid Dynamics (CFD) for the coolant/blade interaction and one-dimensional models developed at EthosEnergy for the redistribution of the cooling flows in the cavities. The mutual effects between the coolant fluid and the blade are analyzed using a Conjugate Heat Transfer (CHT) approach with Star-CCM+. The validation of the computational approach has been performed exploiting the experimental data available for the NASA C3X test case. The TG20 rotating blade of the first turbine row has been analyzed considering the two different coolant configurations. The impact of the main flow on the thermal field has initially been included by imposing a temperature field on the blade surface. The latter field has in turn been obtained by means of a separate computation for the solid only. Full CHT simulations has hence been performed, thus quantifying the accuracy of the proposed approach. The obtained results are discussed in terms of thermo-fluid-dynamic effects.

Energy companies in the power generation field are continuously searching for green technologies to reduce pollutant emissions. In that context, small hydropower plants represent an attractive solution for distributed electricity generation. Reverse-running centrifugal pumps (also known as “pump-as-turbines”, PaT) are increasingly selected in that field. Amongst the existing type of pumps, drag-type regenerative pumps (RP) can perform similarly to radial centrifugal pumps in terms of head and efficiency for low specific speed values. For a fixed rotational speed, RPs with linear blades work as pump or turbine only depending on the flow rate. Such peculiarity makes it particularly intriguing to evaluate RPs working characteristic in the turbine operating mode. In the present paper, the performance of three Regenerative Pump-as-Turbine (RPaT) are analyzed using Computational Fluid Dynamics (CFD). The analysis is supported by an already validated in-house 1D code developed in cooperation with Pierburg Pump Technology Italy SPA. The obtained results are also discussed considering the theoretical behavior of the circulatory velocity in a regenerative machine as described by a widely used 1D model, which is extended in the present paper to the turbine working region. The numerical approach is validated using experimental data for both an RP (in the pump working region) and a regenerative turbine (RT) (in the turbine working region). Finally, the numerical simulation of a small-scale RP allows for the detailed characterization of both the pump and the turbine regions. The numerical analysis shows that for a RPaT it is possible to find a “switch region” where the machine turns from behaving as a pump to behaving as a turbine, the losses not being overcome by the turbine power output. The analysis of the RPaT also shows the inversion of the flow pattern and the positioning of the pivot around which the flow creates the typical helical structure that characterizes RPs.

Regenerative pumps, also referred to as “peripheral” or “side channel” pumps, are characterized by a specific speed that contextualize them between rotary positive displacement and purely radial centrifugal pumps. Although regenerative pumps are not widely distributed, they are interesting for many industrial applications. In fact, for a given flow rate they operate at lower rotational speed with respect to purely radial pumps. Furthermore, they are less affected by mechanical problems with respect to positive displacement pumps. The energy transfer mechanism is the same of centrifugal pumps, but the presence of the side channel imposes to the fluid to pass several times through the impeller, thus obtaining higher pressure rise (as for multi-stage machines) with respect to classical purely radial pumps. Unfortunately, the complexity of the flow field, the large amount of wetted surface and a disadvantageous inflow/outflow configuration contribute to limit the maximum value of hydraulic efficiency, which is also very sensitive to the design choices. Moreover, the intrinsic complexity of the helical flow path makes the theoretical performance estimation a challenging task. It is worth underlining that an accurate performance prediction using one-dimensional models would allow to accelerate greatly the design process, with a non-negligible reduction of demanding three-dimensional Computational Fluid Dynamics (CFD) campaigns. The aim of the present work is to deeply investigate the fluid dynamics of regenerative pumps and to understand how accurately the fundamental physical phenomena can be reproduced by one-dimensional theory. To comply with these aims, a systematic post-processing of the results of several steady and unsteady three-dimensional CFD simulations is exploited for the validation of the in-house one-dimensional tool DART (Design and Analysis tool for Regenerative Turbomachinery), developed at the University of Florence. The theory underlying DART is detailed, and the assumptions of the model are verified by means of comparison with the numerical results underlining the key aspects to be considered for a reliable prediction of the pump performance.

There is a need to automate stochastic uncertainty quantification codes in the digital age. Problems in turbomachinery Computational Fluid Dynamics (CFD) are characterised by non-linear and discontinuous responses and long run times. Ensuring the reliability of Uncertainty Quantification (UQ) codes in such conditions, in an autonomous way, is a challenging problem. Human involvement has always been required. In this work, we therefore suggest a new approach that combines three state-of-the-art methods: multivariate Padé approximations, Optimal Quadrature Subsampling and Statistical Learning. Its main component is the generalised least squares multivariate Padé-Legendre (PL) approximation. PL approximations are globally fitted rational functions that can accurately describe discontinuous non-linear behaviour. They need fewer model evaluations than local or adaptive methods and do not cause the Gibbs phenomenon, like continuous Polynomial Chaos methods. We describe a series of modifications of the Padé algorithm that allow us to apply it to arbitrary input points instead of optimal quadrature locations. This property is particularly useful for industrial applications, where a database of CFD runs is already available, but not in optimal parameter locations. One drawback of the PL approximation is that it is non-trivial to ensure reliability for multiple input parameters. We therefore suggest a new method to improve stability in this work: Optimal Quadrature Subsampling. Our argument is that least squares errors, caused by an ill-conditioned design matrix, are the main source of error. Finally, we use statistical learning techniques to automatically guarantee smoothness and convergence. The resulting method is shown to efficiently and correctly fit thousands of partly discontinuous response surfaces for an industrial film cooling and shock interaction problem automatically and using only 9 CFD simulations. It can be applied to any other UQ problem that is characterised by a limited amount of data and the presence of discontinuities.

Radial inflow expander units are increasingly employed in both oil and gas upstream and downstream markets for energy saving purposes. While prominent attention is given to the performance of the expansion wheel, the downstream diffuser is scarcely covered in literature. In modern expanders, the gas entering the diffuser is typically high-speed, non-uniformly distributed from hub to shroud. Recovery factor and pressure loss coefficient are then affected by both the inlet gas conditions and the geometry of the system, including shaft end and divergence angle of the diffuser. In the present work, the application of computational methodologies to expander diffusers is initially assessed. A sensitivity analysis is then performed with respect to the inlet flow conditions and the diffuser shape. Trends of variation of recovery factor and loss coefficient are provided as a function of selected geometrical parameters and boundary conditions. It is shown that, starting from a non-optimized diffuser, an overall machine efficiency gain can be achieved. That performance improvement can be appreciated by the current market of expanders as a non-negligible competitive advantage.

A high-pressure vane equipped with a realistic film-cooling configuration has been studied. The vane is characterized by the presence of multiple rows of fan-shaped holes along pressure and suction side while the leading edge is protected by a showerhead system of cylindrical holes. Steady three-dimensional Reynolds-Averaged Navier-Stokes (RANS) simulations have been performed. A preliminary grid sensitivity analysis with uniform inlet flow has been used to quantify the effect of spatial discretization. Turbulence model has been assessed in comparison with available experimental data. The effects of the relative alignment between combustion chamber and high-pressure vanes are then investigated considering realistic inflow conditions in terms of hot spot and swirl. The inlet profiles used are derived from the EU-funded project TATEF2. Two different clocking positions are considered: the first one where hot spot and swirl core are aligned with passage and the second one where they are aligned with the leading edge. Comparisons between metal temperature distributions obtained from conjugate heat transfer simulations are performed evidencing the role of swirl in determining both the hot streak trajectory within the passage and the coolant redistribution. The leading edge aligned configuration is resulted to be the most problematic in terms of thermal load, leading to increased average and local vane temperature peaks on both suction side and pressure side with respect to the passage aligned case. A strong sensitivity of both injected coolant mass flow and heat removed by heat sink effect has also been highlighted for the showerhead cooling system.

This work shows an Uncertainty Quantification (UQ) study of film cooling with shock impingement. A numerical method is proposed to use high order polynomials for the reconstruction of the stochastic output, without the instabilities characteristic of UQ with shock dominated flows. At the same time it is shown that the region with highest uncertainty is driven by a complex flow physics involving shock–boundary layer interaction and the generation of tornado vortices that merge with kidney ones. High-pressure turbine stages are characterized by transonic conditions with the suction side of the nozzle affected by the shock shed by the trailing edge of the adjacent aerofoil. Due to manufacturing deviations and in service degradation the geometrical parameters, such as trailing edge thickness and hole diameter, are subjected to random variations, changing the shock location and the heat transfer loading on the stator nozzle. For these reasons an UQ methodology has been used in this study to model the interaction between the impinging shock and film cooling. The variability of the geometrical parameters has been represented with uniform probability distributions and the stochastic output is obtained using Probabilistic Collocation Method with Padè’s polynomials. Transonic flows are challenging in Uncertainty Quantification because a better reconstruction of the stochastic output can be achieved increasing the order of polynomials but higher order polynomials become unstable for the Runge’s phenomenon. This work proposes a method that allows the application of high order Padè’s polynomial without having instabilities in the stochastic output. The proposed methodology can be applied to other transonic configurations in gas turbines, requiring only a limited number of simulations to reconstruct the stochastic output. The results show that the maximum level of uncertainty is located downstream the region of interaction between shock and boundary layer. In particular the shock generates complex flow structures that develop into tornado vortices, highly dependent on the uncertainty input.

In this paper conjugate heat transfer analysis of the cooled vane of the MT1 research high-pressure stage is presented. Inlet boundary conditions (including non-uniform total temperature, non-uniform total pressure, swirl, turbulence intensity and turbulence length scale) are obtained considering the exit flow field of a reactive annular combustor simulator. The combustor model has been designed in order to reproduce data available in literature about exit profiles of real combustion chambers and other combustor simulators. Steady simulations are performed on a hybrid unstructured grid obtained from a grid dependence study. The transitional k T -k L -ω model by Walters and Cokljat is used as turbulent closure. Thermal fields obtained from CHT analysis of the vane considering two different clocking positions with respect to the combustor are compared. Results, including film cooling parameters and High-Pressure Vane aerodynamics, are also compared with a uniform inlet case showing the crucial importance of considering realistic boundary conditions for thermal analysis of turbine components.

Numerical techniques are commonly used during both design and analysis processes, mainly considering single components. Technological progress asks for advanced approaches that include real-machine conditions and analyze components interaction, especially considering the combustor/turbine coupling. Modern combustors operate with strong swirl motions in order to obtain an adequate flame stabilization, generating a very complex flow field characterized by high turbulence level. These aspects affect performance of downstream components which are subjected to very aggressive inlet flow conditions: non-uniform total temperature, non-uniform total pressure, swirl and high turbulence intensity. For these reasons coupled analysis of combustor and turbine is necessary to accurately predict aero-thermal aspects that influence performance and reliability of these two components. From a numerical point of view the simulation of a single domain characterized by a reactive flow with very different Mach number regimes (from low-Mach flow in combustion chamber to transonic flow in turbine) is problematic due to the different numerical requirements needed, especially concerning stability and accuracy. These problems could be overcome using coupled methods to simultaneously simulate combustor and turbine in separated domains which are managed by different solvers that communicate with each other. A coupling method for the study of combustor/turbine interaction using the RANS methodology is proposed. In the first part of the paper the method is described and validated. The second part is dedicated to the application of the proposed coupling methodology to a realistic test case consisting of a model annular combustor and the Nozzle Guide Vane (NGV) of the MT1 high-pressure turbine stage. A commercial solver and an in-house code are respectively used for the simulation of combustor and NGV. Results are presented and analyzed highlighting the importance of such type of simulations in understanding aero-thermal phenomena that characterize combustor/vane interaction.

High-pressure turbines are subjected to high expansion ratios which may result in supersonic vane or rotor outlet. In the supersonic regime, shock waves are formed at the vane trailing edges that periodically impact on the downstream rotor blades and the neighboring vane. Pulsating cooling was proposed to modulate the trailing edge shocks to diminish the detrimental efficiency abatement and structural problems. A large test section equipped with three airfoils, replicated the actual loading of a transonic vane. Tests were performed in a short duration compression tube facility at four Mach numbers (0.8, 0.95, 1.1 and 1.2) and two different Reynolds numbers (4 and 6 million). Coolant was fed to the trailing edge of the central airfoil through a siren valve, that allowed to study four different coolant blowing cases: no-blowing, continuous blowing at three pressure levels and a pulsating coolant blowing condition. Unsteady numerical simulations of the flow over the airfoil model were performed using ANSYS 14 (Fluent) flow solver. In order to understand the impact of the different steady and unsteady cooling schemes on the efficiency of high pressure turbine bladings, the individual contribution of trailing edge and profile losses were calculated. The coolant ejection generated a significant reduction of the trailing edge loss. The overall losses also diminished by the introduction of cooling as compared to no blowing case. Improvements in loss levels owing to pulsating cooling observed to be more pronounced for the engine representative cooling rates (∼3%). The trends in loss variation with respect to cooling scheme show that pulsating cooling may become superior for high blowing cases.

Numerical techniques are commonly used during both design and analysis processes, mainly considering separated components. Technological progress asks for advanced approaches that allow to analysing the interaction between the components, especially when considering combustor/turbine interaction. Hot spots and inlet swirl profiles generated by the combustor have been demonstrated to affect high-pressure turbine performances and reliability. This work deals with the investigation of the effects of realistic boundary conditions for the high-pressure turbine vane, also proposing an approach for coupled simulation of the combustor/vane interaction. The method consists in a loosely coupled approach for the data exchange on the combustor/vane interface section. Data from the combustor exit section (stagnation conditions, velocity profile and turbulent quantities) are provided to the vane inlet and vice versa (for the static pressure). The proposed method is applied to a test case consisting of a redesigned combustor and the vane of the MT1 test case from QinetiQ. A preliminary analysis was dedicated to define the combustor geometry and the operating conditions. Then, the MT1 working conditions have been rescaled and coupled with the combustor, maintaining the stage geometry and the experimental non-dimensional parameters. Second order accurate steady simulations were performed for both combustor and high-pressure turbine vane. Calculations with a uniform profile and a theoretical nonuniform inlet profile (deriving from the EU funded TATEF2 project) have been considered as representative of commonly used approaches. The results obtained for the stator in terms of isentropic Mach number and Nusselt number along blades surfaces and inner end-wall are compared with each other and with the available experimental data. Due to the large dimension of computational grids a parallel approach was applied. The activity was carried out using the IBM PLX supercomputer in the frame of the FrUIT project supported by CINECA.

A series of tests on a specific designed linear nozzle guide vane (NGV) cascade with trailing edge coolant ejection was carried out to investigate the influence of the trailing edge bleeding (TEB) on the loss behaviour of the profile. Wake traverses with a five-hole probe and measurements of the pressure distribution on the profile were taken varying the ejection rate under reference main flow conditions, namely Re 2 th = 1.056·10 6 and Ma 2 th = 0.8 (Re 2 th based on the true chord). Wake total pressure losses and isentropic Mach number distributions on the profile were compared to measurements without coolant ejection, showing a significant influence of the TEB both on the wake development and on the flow in the vane passage. Numerical simulations of the experiments showed good agreement with the measured data and provided a deeper understanding of the flow phenomena, revealing the differences in the development of the wake with and without trailing edge coolant ejection and illustrating the blockage effect of the TEB on the flow in the vane passage.