Mechanical Engineering - Research Publications

Permanent URI for this collection

Search Results

Now showing 1 - 10 of 22
  • Item
    Thumbnail Image
    Direct Numerical Simulation of Riblets Applied to Gas Turbine Compressor Blades at On- and Off- Design Incidences
    Kozul, M ; Nardini, M ; Przytarski, P ; Solomon, W ; Shabbir, A ; Sandberg, R (ASME, 2023-06-26)
    Any realizable increase in gas turbine efficiency has significant potential to reduce fuel burn and environmental impact. Streamwise micro-groove surfaces (‘riblets’) are well-known as a passive surface treatment to reduce drag, which may be useful in the context of increasing overall gas turbine efficiency. This paper presents the first direct numerical simulation of potentially performance-enhancing riblets on an axial flow high pressure compressor blade, where the micro-geometry of the riblets is fully resolved. The midspan section of a NACA6510 profile is considered at an engine-relevant true chord Reynolds number of 700,000 and Mach number 0.5 based on inlet conditions. Fixed triangular (or sawtooth) riblets are considered in the present numerical campaign. The current high-fidelity computational method permits the extraction of data such as the wall shear stress directly from the riblet surface. At the design incidence, the riblets tend to promote earlier transition to a turbulent flow over the suction side, yet significantly reduce the skin friction over the entire downstream chord to the trailing edge. The riblets reduce the viscous force over the blade by up to 18% at this nominal inflow incidence. Thus the current dataset permits new insight into the action of the riblets, since most studies of riblets on turbomachinery blades have been conducted experimentally where direct measurements of skin friction are not possible. The riblets are also able to reduce the skin friction over the high pressure compressor blade at off-design incidences, a promising result given axial flow compressors must cope with variable operating conditions.
  • Item
    No Preview Available
    Resolvent analysis-based pressure modeling for trailing edge noise prediction
    Wagner, GA ; Deuse, M ; Illingworth, SJ ; Sandberg, RD (American Institute of Aeronautics and Astronautics, 2019-05-20)
    This paper presents the first development steps of a hybrid computational aeroacoustics (CAA) method for the prediction of trailing edge (TE) noise, based on a physics-driven prediction of the surface pressure on an airfoil. Starting frommean flowdata for a given configuration, the dominant pressure modes over a foil are modeled with an incompressible formulation of the resolvent method. As a conceptual test of its suitability to predict surface pressures, the framework is used to model the unsteady surface pressure fluctuations generated by instability waves on an infinite flat plate. While a canonical test case, the flat plate is a good starting point for the investigation of airfoil TE noise. Subsequently, the framework is applied to a NACA0012 airfoil at 0° angle of attack. In the flat plate case, hydrodynamic instabilities are excited by a single frequency volume forcing and result in streamwise propagating Tollmien-Schlichting waves. The resolvent captures these instabilities and the resulting surface pressure field with good accuracy. A Mach number dependence is observed for the agreement between resolvent and DNS pressure modes, which may explain the difference in wavelengths between DNS and resolvent results in the NACA0012 airfoil case. Bearing in mind this dependence of the pressure prediction accuracy on the Mach number, the results show promise for resolvent-based TE noise predictions.
  • Item
    Thumbnail Image
    Heat Transfer Coefficient Estimation for Turbulent Boundary Layers
    Wang, S ; Xia, Y ; Abu Rowin, W ; Marusic, I ; Sandberg, R ; Chung, D ; Hutchins, N ; Tanimoto, K ; Oda, T (The University of Queensland, 2020-12-11)
    Convective heat transfer in rough wall-bounded turbulent flows is prevalent in many engineering applications, such as in gas turbines and heat exchangers. At present, engineers lack the design tools to accurately predict the convective heat transfer in the presence of non-smooth boundaries. Accordingly, a new turbulent boundary layer facility has been commissioned, where the temperature of an interchangeable test surface can be precisely controlled, and conductive heat losses are minimized. Using this facility, we can estimate the heat transfer coefficient (Stanton number, St), through measurement of the power supplied to the electrical heaters and also from measurements of the thermal and momentum boundary layers evolving over this surface. These methods have been initially investigated over a shorter smooth prototype heated surface and compared with existing St prediction models. Preliminary results suggest that we can accurately estimate St in this facility.
  • Item
    Thumbnail Image
    Data-driven combustion modeling for a turbulent flame simulated with a computationally efficient solver
    Talei, M ; Ma, D ; Sandberg, R (ASME: The American Society of Mechanical Engineers, 2020)
    The use of machine learning (ML) for modeling is on the rise. In the age of big data, this technique has shown great potential to describe complex physical phenomena in the form of models. More recently, ML has frequently been used for turbulence modeling while the use of this technique for combustion modeling is still emerging. Gene expression programming (GEP) is one class of ML that can be used as a tool for symbolic regression and thus improve existing algebraic models using high-fidelity data. Direct numerical simulation (DNS) is a powerful candidate for producing the required data for training GEP models and validation. This paper therefore presents a highly efficient DNS solver known as HiPSTAR, originally developed for simulating non-reacting flows in particular in the context of turbomachinery. This solver has been extended to simulate reacting flows. DNSs of two turbulent premixed jet flames with different Karlovitz numbers are performed to produce the required data for training. GEP is then used to develop algebraic flame surface density models in the context of large-eddy simulation (LES). The result of this work introduces new models which show excellent performance in prediction of the flame surface density for premixed flames featuring different Karlovitz numbers.
  • Item
    Thumbnail Image
    DNS of a turbulent premixed jet flame with fully developed turbulent inlet
    Ma, MC ; Talei, M ; Sandberg, RD (Australasian Fluid Mechanics Society, 2018-01-01)
    A Direct Numerical Simulation (DNS) study is conducted of a premixed round jet flame with a fully turbulent flow exiting a long pipe as inlet condition. The inclusion of the pipe in the simulation with a fully turbulent flow inside ensures that the turbulence-flame interaction, in particular near the inlet region, is well represented. The orientation of turbulence structures in different regions of the domain is identified. The interaction between the turbulence structures and the flame is examined using vector alignments. Streamwise oriented vortical structures formed in the pipe boundary layer are convected into the base of the flame. Preferential alignment of the flame normal with the compressive eigenvector of the strain rate tensor is observed close to the pipe exit suggesting that the turbulence characteristics of the jet inlet have an effect on the flame front close to the flame base.
  • Item
    Thumbnail Image
    LARGE EDDY SIMULATIONS OF HIGH ROSSBY NUMBER FLOW IN THE HIGH PRESSURE COMPRESSOR INTER-DISK CAVITY
    Saini, D ; Sandberg, R (ASME: The American Society of Mechanical Engineers, 2021)
    The focus of the present study is to understand the effect of Rayleigh number on a high Rossby number flow in a high pres- sure compressor (HPC) inter-disk cavity. These cavities form be- tween the compressor disks of a gas turbine engine, and they are an integral part of the internal air cooling system. We perform highly resolved large eddy simulations for two Rayleigh numbers of 0.76 × 10^8 and 1.54 × 10^8 at a fixed Rossby number of 4.5 by solving the compressible Navier–Stokes equations. The results show a flow structure dominated by a toroidal vortex in the inner region of the cavity. In the outer region, the flow is observed to move radially outwards by Ekman layers formed on the side disks and to move radially inwards through the central core region of the cavity. An enhancement in the in- tensity of the radial flares is observed in the outer region of the cavity for the high Rayleigh number case with no perceivable effect in the inner region. The near shroud region is mostly dom- inated by the centrifugal buoyancy-induced flow and the wall Nusselt number calculated at the shroud is in close agreement with centrifugal buoyancy-induced flow without an axial bore flow.
  • Item
    Thumbnail Image
    DEVELOPMENT AND USE OF MACHINE-LEARNT ALGEBRAIC REYNOLDS STRESS MODELS FOR ENHANCED PREDICTION OF WAKE MIXING IN LPTS
    Akolekar, HD ; Weatheritt, J ; Hutchins, N ; Sandberg, RD ; Laskowski, G ; Michelassi, V (AMER SOC MECHANICAL ENGINEERS, 2018-01-01)
    Non-linear turbulence closures were developed that improve the prediction accuracy of wake mixing in low-pressure turbine (LPT) flows. First, Reynolds-averaged Navier-Stokes (RANS) calculations using five linear turbulence closures were performed for the T106A LPT profile at exit Mach number 0.4 and isentropic exit Reynolds numbers 60,000 and 100,000. None of these RANS models were able to accurately reproduce wake loss profiles, a crucial parameter in LPT design, from direct numerical simulation (DNS) reference data. However, the recently proposed kv2w transition model was found to produce the best agreement with DNS data in terms of blade loading and boundary layer behavior and thus was selected as baseline model for turbulence closure development. Analysis of the DNS data revealed that the linear stress-strain coupling constitutes one of the main model form errors. Hence, a geneexpression programming (GEP) based machine-learning technique was applied to the high-fidelity DNS data to train non-linear explicit algebraic Reynolds stress models (EARSM). In particular, the GEP algorithm was tasked to minimize the weighted difference between the DNS and RANS anisotropy tensors, using different training regions. The trained models were first assessed in an a priori sense (without running any CFD) and showed much improved alignment of the trained models in the region of training. Additional RANS calculations were then performed using the trained models. Importantly, to assess their robustness, the trained models were tested both on the cases they were trained for and on testing, i.e. previously not seen, cases with different flow features. The developed models improved prediction of the Reynolds stress, TKE production, wake-loss profiles and wake maturity, across all cases, in particular those trained on just the wake region.
  • Item
    Thumbnail Image
    MACHINE LEARNING FOR THE DEVELOPMENT OF DATA DRIVEN TURBULENCE CLOSURES IN COOLANT SYSTEMS
    Hammond, J ; Montomoli, F ; Pietropaoli, M ; Sandberg, R ; Michelassi, V (ASME: The American Society of Mechanical Engineers, 2020-06-22)
    This work shows the application of Gene &pression Pro­gramming to augment RANS turbulence closure modelling for flows through complex geometries, designed for additive manu­facturing. Specifically, for the design of optimised internal cool­ing channels in turbine blades. One of the challenges in internal coolant design is the heat transfer accuracy of the RANS formu­lation in comparison to higher fidelity methods, which are still not used in design on account of their computational cost. How­ever, high fidelity data can be extremely valuable for improving cu"ent lower fidelity models and this work shows the application of data driven approaches to develop turbulence closures for an internally ribbed duct. Different approaches are compared, and the results of the improved model are illustrated. The work shows the potential of using data driven models for accurate heat transfer predictions even in non-conventional configurations.
  • Item
    Thumbnail Image
    INTEGRATION OF MACHINE LEARNING AND COMPUTATIONAL FLUID DYNAMICS TO DEVELOP TURBULENCE MODELS FOR IMPROVED TURBINE WAKE MIXING PREDICTION
    Akoleka, HD ; Zhao, Y ; Sandberg, R ; Pacciani, R (Search Results Web results ASME: The American Society of Mechanical Engineers, 2021)
    This paper presents development of accurate turbulence closures for wake mixing prediction by integrating a machine-learning approach with Reynolds Averaged Navier-Stokes (RANS)-based computational fluid dynamics (CFD). The data-driven modelling framework is based on the gene expression programming (GEP) approach previously shown to generate non-linear RANS models with good accuracy. To further improve the performance and robustness of the data-driven closures, here we exploit that GEP produces tangible models to integrate RANS in the closure development process. Specifically, rather than using as cost function a comparison of the GEP-based closure terms with a frozen high fidelity dataset, each GEP model is instead automatically implemented into a RANS solver and the subsequent calculation results compared with reference data. By first using a canonical turbine wake with inlet conditions prescribed based on high-fidelity data, we demonstrate that the CFD-driven machine-learning approach produces non-linear turbulence closures that are physically correct, i.e. predict the right downstream wake development and maintain an accurate peak wake loss throughout the domain. We then extend our analysis to full turbine blade cases and show that the model development is sensitive to the training region due to the presence of deterministic unsteadiness in the near wake region. Models developed including this region have artificially large diffusion coefficients to overcompensate for the vortex shedding steady RANS cannot capture. In contrast, excluding the near wake region in the model development produces the correct physical model behavior, but predictive accuracy in the near-wake remains unsatisfactory. We show that this can be remedied by using the physically consistent models in unsteady RANS, implying that the non-linear closure producing the best predictive accuracy depends on whether it will be deployed in RANS or unsteady RANS calculations. Overall, the models developed with the CFD assisted machine learning approach were found to be robust and capture the correct physical behavior across different operating conditions.
  • Item
    Thumbnail Image
    HIGH-FIDELITY SIMULATIONS OF A HIGH-PRESSURE TURBINE VANE SUBJECT TO LARGE DISTURBANCES: EFFECT OF EXIT MACH NUMBER ON LOSSES
    Zhao, Y ; Sandberg, R (Search Results Web results ASME: The American Society of Mechanical Engineers, 2021)
    We report on a series of highly resolved large-eddy simulations of the LS89 high-pressure turbine (HPT) vane, varying the exit Mach number between Ma = 0:7 and 1:1. In order to accurately resolve the blade boundary layers and enforce pitchwise periodicity, we for the first time use an overset mesh method, which consists of an O-type grid around the blade overlapping with a background H-type grid. The simulations were conducted either with a synthetic inlet turbulence condition or including upstream bars. A quantitative comparison shows that the computationally more efficient synthetic method is able to reproduce the turbulence characterictics of the upstream bars. We further perform a detailed analysis of the flow fields, showing that the varying exit Mach number significantly changes the turbine efficiency by affecting the suction-side transition, blade boundary layer profiles, and wake mixing. In particular, the Ma = 1:1 case includes a strong shock that interacts with the trailing edge, causing an increased complexity of the flow field. We use our recently developed entropy loss analysis (Zhao and Sandberg, GT2019-90126) to decompose the overall loss into different source terms and identify the regions that dominate the loss generation. Comparing the different Ma cases, we conclude that the main mechanism for the extra loss generation in the Ma = 1:1 case is the shock-related strong pressure gradient interacting with the turbulent boundary layer and the wake, resulting in significant