Luca di Mare

Forced Response of Mistuned Bladed Disks in Gas Flow: A Comparative Study of Predictions and Full-Scale Experimental Results

An integrated experimental-numerical study of forced response for a mistuned bladed disk has been performed. A full chain for the predictive forced response analysis has been developed including data exchange between the computational fluid dynamics code and a code for the prediction of the nonlinear forced response for a bladed disk. The experimental measurements are performed at a full-scale single stage test rig with excitation by aerodynamic forces from gas flow. The numerical modeling approaches and the test rig setup are discussed. Comparison of experimentally measured and predicted values of blade resonance frequencies and response levels for a mistuned bladed disk without dampers is performed. A good correspondence between frequencies at which individual blades have maximum response levels is achieved. The effects of structural damping and underplatform damper parameters on amplitudes and resonance frequencies of the bladed disk are explored. It is shown that the underplatform damper significantly reduces scatters in values of the individual blade frequencies and maximum forced response levels.

Virtual Gas Turbines: Geometry and Conceptual Description

This paper documents the construction of a Virtual Engine, with particular reference to its geometry and conceptual description.. The phrase Virtual Engine denotes a system which allows simulations of whole gas-turbine engines to be undertaken at any desired level of fidelity or physical modeling. In order to be of any practical use, the system must allow the computations to be setup as automatically as possible and needs to contain provisions for the exchange of boundary data between adjacent computational domains — e.g. solid-gas interfaces. The paper illustrates the application of the system to the representation and analysis of a modern commercial turbofan engine.Copyright

Modern multicore and manycore architectures: Modelling, optimisation and benchmarking a multiblock CFD code

Abstract Modern multicore and manycore processors exhibit multiple levels of parallelism through a wide range of architectural features such as SIMD for data parallel execution or threads for core parallelism. The exploitation of multi-level parallelism is therefore crucial for achieving superior performance on current and future processors. This paper presents the performance tuning of a multiblock CFD solver on Intel SandyBridge and Haswell multicore CPUs and the Intel Xeon Phi Knights Corner coprocessor. Code optimisations have been applied on two computational kernels exhibiting different computational patterns: the update of flow variables and the evaluation of the Roe numerical fluxes. We discuss at great length the code transformations required for achieving efficient SIMD computations for both kernels across the selected devices including SIMD shuffles and transpositions for flux stencil computations and global memory transformations. Core parallelism is expressed through threading based on a number of domain decomposition techniques together with optimisations pertaining to alleviating NUMA effects found in multi-socket compute nodes. Results are correlated with the Roofline performance model in order to assert their efficiency for each distinct architecture. We report significant speedups for single thread execution across both kernels: 2-5X on the multicore CPUs and 14-23X on the Xeon Phi coprocessor. Computations at full node and chip concurrency deliver a factor of three speedup on the multicore processors and up to 24X on the Xeon Phi manycore coprocessor.

NUMERICAL STUDIES INTO INTAKE FLOW FOR FAN FORCING ASSESSMENT

Recent trends in design for civil intakes lead towards shorter diffuser sections, unorthodox installations and more loaded lips. All these features increase the risk of lip stall in flight at incidence or in cross wind and increase the level of forcing seen by the fan blades because of the interaction with non-uniform flow from the intake.In this study we analyze the behavior of prediction tools for intake distortion. In particular we compare the performance of popular turbulence models for standard intake flows and we discuss their behavior on the grounds of their behavior for elementary flows.We conclude our study by comparing forcing and distortion figures of merit from different models.Copyright

Fan Forced Response Due to Ground Vortex Ingestion

This paper presents a computational study of the formation and ingestion of ground vortices and resulting fan forced response levels in a large turbofan operating near the ground. The model is based on an integrated aeroelasticity numerical method; the aerodynamic part is based on a 3D unstructured Reynolds-Averaged Naveir-Stokes solver. The mechanical model uses linear modal model for the structure, allowing for the direct computation of the structural response during the unsteady simulations. The analysis shows that under certain fan speed and mass flow rate conditions, for a given fan and intake combination, situated at a fixed distance from the ground, an inlet vortex can form near the ground. This inlet vortex is drawn into the intake causing inlet distortions that could excite several low engine order harmonics of the fan. Predictions are compared with measured data showing good agreement in general. Ability to predict the level of response at the design stage allows for implementing design solutions preventing possible failure due to high cycle fatigue.Copyright

Simulation of Multi-Stage Compressor at Off-Design Conditions

Computational Fluid Dynamics (CFD) has been widely used for compressor design, yet the prediction of performance and stage matching for multi-stage, high-speed machines remain challenging. This paper presents the authors’ effort to improve the reliability of CFD in multistage compressor simulations. The endwall features (e.g. blade fillet and shape of the platform edge) are meshed with minimal approximations. Turbulence models with linear and non-linear eddy viscosity models are assessed. The non-linear eddy viscosity model predicts a higher production of turbulent kinetic energy in the passages, especially close to the endwall region. This results in a more accurate prediction of the choked mass flow and the shape of total pressure profiles close to the hub. The non-linear viscosity model generally shows an improvement on its linear counterparts based on the comparisons with the rig ∗Corresponding author (feng.wang@eng.ox.ac.uk) TURBO-17-1165 Wang 1 data. For geometrical details, truncated fillet leads to thicker boundary layer on the fillet and reduced mass flow and efficiency. Shroud cavities are found to be essential to predict the right blockage and the flow details close to the hub. At the part speed the computations without the shroud cavities fail to predict the major flow features in the passage and this leads to inaccurate predictions of massflow and shapes of the compressor characteristic. The paper demonstrates that an accurate representation of the endwall geometry and an effective turbulence model, together with a good quality and sufficiently refined grid result in a credible prediction of compressor matching and performance with steady state mixing planes. Introduction Three-dimensional CFD simulations of multiple blade rows were developed by Adamczyk [1] using the average passage approach. Each blade row was modelled seperately and only one passage of each blade row was required in the simulation. Blade row interactions were evaluated by including additional force terms in the equations. This approach was further simplified by Denton and the resulting approach was termed as the mixing plane approach [2]. The mixing plane assumes a mixed-out state at the bladerow interface by conserving mass, momentum and energy fluxes. The drawback of this approach is that the unsteady blade interactions are removed. Despite its simplicity, the performance of turbomachines is predicted reasonably well using mixing planes. Consequently, steady state RANS simulations have been the work-horse of turbomachinery design. The capability of RANS simulations in assisting turbomachinery design was already demonstrated by early researchers [3, 4], even if limitations in computer power, meshing techniques and physical modelling restricted the spatial resolution as well as the amount of geometric detail present in the simulations. Typical examples of features omitted from the computational models in the early days were blade fillets, stator shrouds, Variable Stator Vane (VSV) penny gaps etc. Typical grids used were H-grids with pinched tip gaps. Fully meshed tip gaps were introduced and comparisons with pinched tip gaps were summarized by Denton [5]. As the meshing techniques evolved, multiblock structured meshes were introduced into the turbomachinery design and allowed the generation of optimal grids on the blade-to-blade section [6,7]. However, geometrical approximations (e.g. truncated fillets) are still commonplace because of the limitations in the way the geometries of the endwalls and of the blades are represented in the mesh technique. The effect of endwall geometric features is important in predicting the endwall flows and contributes to setting the flow capacity. One of the most important endwall features for modern machines is the shroud cavities. A typical configuration is shown in Fig. 1. Shabbir et al [8] studied the effect of the hub leakage flow on high speed compressor rotors and the pressure deficit close to the hub was well captured by including the hub leakage flow. Wellborn and Okashi [9] reported the effect of shroud cavities on the performance of multi-stage compressor experimentally. Their data showed that the shroud cavities have considerable negative impact on the compressor performance. In their numerical study, the shroud cavities were not meshed as CFD domains, instead they were introduced through a Knife-to-Knife (K2K) model [10]. Another important element in the construction of steady models for turbomachinery is the turbulence model. The flows in multistage compressor passages are highly viscous and three-dimensional around endwall regions. Furthermore the flow is prone to separation due to strong adverse pressure gradient even at design conditions. Predicting the complex flow features TURBO-17-1165 Wang 2

AN ENERGY BASED FOULING MODEL FOR GAS TURBINES: EBFOG

Fouling is a major problem in gas turbines for aeropropulsion because the formation of aggregates on the wet surfaces of the machine affects aerodynamic and heat loads. The representation of fouling in CFD is based on the evaluation of the sticking probability, i.e. the probability a particle touching a solid surface has to stick to that surface. Two main models are currently available in literature for the evaluation of the sticking coefficient: one is based on a critical threshold for the viscosity, the other is based on the normal velocity to the surface. However, both models are application specific and lack generality. This work presents an innovative model for the estimation of the sticking probability. This quantitiy is evaluated by comparing the kinetic energy of the particle with an activation energy which describes the state of the particle. The sticking criterion takes the form of an Arrhenius-type equation. A general formulation for the sticking coefficient is obtained. The method, named EBFOG (Energy Based FOulinG), is the first ”energy” based model presented in the open literature able to account any common deposition effect in gas turbines. The EBFOG model is implemented into a Lagrangian tracking procedure, coupled to a fully three-dimensional CFD solver. Particles are tracked inside the domain and equations for the momentum and temperature of each particle are solved. The local geometry of the blade is modified accordingly to the deposition ∗Address all correspondence to this author, email: nicola.casari@unife.it, nicola.casari15@imperial.ac.uk rate. The mesh is modified and the CFD solver updates the flow field. The application of this model to particle deposition in high pressure turbine vanes is investigated, showing the flexibility of the proposed methodology. The model is particularly important in aircraft engines where the effect of fouling for the turbine, in particular the reduction of the HP nozzle throat area, influences heavily the performance by reducing the core capacity. The energy based approach is used to quantify the throat area reduction rate and estimate the variation in the compressor operating condition. The compressor operating point as a function of the time spent operating in a harsh environment can be in this way predicted to estimate, for example, the time that an engine can fly in a cloud of volcanic ashes. The impact of fouling on the throat area of the nozzle is quantified for different conditions.

Virtual Gas Turbine: Pre-Processing and Numerical Simulations

The design process of a gas turbine engine involves interrelated multi-disciplinary and multi-fidelity designs of engine components. Traditional component-based design process is not always able to capture the complicated physical phenomenon caused by component interactions. It is likely that such interactions are not resolved until hardware is built and tests are conducted. Component interactions can be captured by assembling all these components into one computational model. Nowadays, numerical solvers are fairly easy to use and the most time-consuming (in terms of man-hours) step for large scale gas turbine simulations is the preprocessing process. In this paper, a method is proposed to reduce its time-cost and make large scale gas turbine numerical simulations affordable in the design process. The method is based on a novel featured-based in-house geometry database. It allows the meshing modules to not only extract geometrical shapes of a computational model and additional attributes attached to the geometrical shapes as well, such as rotational frames, boundary types, materials, etc. This will considerably reduce the time-cost in setting up the boundary conditions for the models in a correct and consistent manner. Furthermore, since all the geometrical modules access to the same geometrical database, geometrical consistency is satisfied implicitly. This will remove the time-consuming process of checking possible mismatching in geometrical models when many components are present. The capability of the proposed method is demonstrated by meshing the whole gas path of a modern three-shaft engine and the Reynold’s Averaged Navier-Stokes (RANS) simulation of the whole gas path.© 2016 ASME

Reduced Passage Method for Multirow Forced Response Computations

This paper presents a time-domain Fourier method for modeling steady and unsteady nonaxisymmetric flows in turbomachinery on a reduced computational domain. The method extends well-established single-passage multirow methods, which efficiently model periodic unsteadiness in single stages, to assemblies with stationary circumferential perturbations with periodicity different from the blade count. Such perturbations are caused, for example, by rotor–rotor/stator–stator interaction or geometric circumferential variations. The method is therefore suitable to study low-engine-order forcing problems, flow past nonuniform assemblies, and clocking problems. The proposed method solves the flow inside several discrete passages, located at different circumferential positions, using a time-accurate scheme. Boundary conditions at the azimuthal and interrow surfaces are approximated via time–space Fourier series and couple the individual passages. The reduced passage model is validated against the whole annulus solutio...

Automated Hex Meshing for Turbomachinery Secondary Air System

In this paper, we present a novel process of creating hexahedral meshes for the turbomachinery secondary air system. The meshing process is automated from the geometry import to the mesh setup and requires minimum human interventions. The core of the process is a hexahedral meshing algorithm with a boundary layer mesh automatically created. The hex meshing algorithm combines the pave-sweep, general sheet-insertion and a novel technique which creates the boundary layer mesh by carefully placing, maintaining and dicing a buffer layer around a geometry. After the mesh is created, relevant boundary conditions for the mesh are also assigned automatically. The whole meshing process is systematically automated and has the potential to considerably reduce the time cost in meshing the turbomachinery secondary air system.