Mark G. Turner

Assessment of Computational Fluid Dynamics and Experimental Data for Shock Boundary-Layer Interactions

A workshop on the computational fluid dynamics (CFD) prediction of shock boundary-layer interactions (SBLIs) was held at the 48th AIAA Aerospace Sciences Meeting. As part of the workshop, numerous CFD analysts submitted solutions to four experimentally measured SBLIs. This paper describes the assessment of the CFD predictions. The assessment includes an uncertainty analysis of the experimental data, the definition of an error metric, and the application of that metric to the CFD solutions. The CFD solutions provided very similar levels of error and, in general, it was difficult to discern clear trends in the data. For the Reynolds-averaged Navier-Stokes (RANS) methods, the choice of turbulence model appeared to be the largest factor in solution accuracy. Scale-resolving methods, such as large-eddy simulation (LES), hybrid RANS/LES, and direct numerical simulation, produced error levels similar to RANS methods but provided superior predictions of normal stresses. Copyright

Radial Migration of Shed Vortices in a Transonic Rotor Following a Wake Generator: A Comparison Between Time Accurate and Average Passage Approaches

This paper shows a comparison of an unsteady simulation using turbo and an average passage simulation for a two blade row configuration consisting of a wake generator followed by a transonic rotor. Two spacings were simulated, both close and far. The unsteady results compare well with experiment especially for the profile of efficiency difference between close and far. An analysis of results helps to explain the unusual profile seen experimentally that is due to the radial migration of wake generator shed vortices with negative radial velocities near the tip. In addition, different components of the average passage body forces (deterministic stresses) are explored that shows the main terms are the axial momentum and the metal blockage.

Multistage Turbine Simulations With Vortex–Blade Interaction

The average passage approach of Adamczyk et al. (1990) has been used to simulate the multistage environment of the General Electric E{sup 3} low-pressure turbine. Four configurations have been analyzed and compared to test data. These include the nozzle only, the first stage, the first stage and a half, and the first two stages. A high casing slope on the first-stage nozzle causes the secondary flow vortex to separate off the casing and enter the downstream rotor. The detrimental effect on performance due to this vortex interaction has been predicted by the above approach, whereas isolated blade row calculations cannot simulate this interaction. The unsteady analysis developed by Chen et al. (1994) has also been run to understand the unsteady flow field in the first-stage rotor and compare with the average passage model and test data. Comparisons of both the steady and the unsteady analyses with data are generally good, although in the region near the casing of the shrouded rotors, the predicted loss is lower than that shown by the data.

Linear Deterministic Source Terms for Hot Streak Simulations

Steady state surface rothalpy over specific heat at constant pressure (temperature units) results obtained with a linear unsteady solution-based lumped deterministic source term are compared with results obtained from a traditional, nonlinear, inviscid unsteady solution for an aircraft engine first stage high-pressure turbine rotor configuration. The relationship between the source terms and traditional solution variables is explored to offer a unique insight into comparing the two approaches. Boundary condition/potential field effects and the order of accuracy of the available schemes are also explored and show a significant effect on surface rothalpy results. The new technique demonstrates a significant potential for approximately including unsteady hot streak effects in time average calculations with minimal computer effort.

Investigation of Loss Generation in an Embedded Transonic Fan Stage at Several Gaps Using High Fidelity, Time-Accurate CFD

The Blade-Row Interaction (BRI) rig at the Air Force Research Laboratory (AFRL), Compressor Aero Research Laboratory (CARL), has been simulated at three axial gaps between the highly loaded upstream stator row and the downstream transonic rotor using TURBO. Previous work with the Stage Matching Investigation (SMI) demonstrated a strong dependence of mass flow rate, efficiency, and pressure ratio on the axial spacing between an upstream wake generator and the downstream rotor through the variation of the axial gap. Several loss producing mechanisms were discovered and related to the spacings, referred to as close, mid, and far. In the SMI work, far spacing had the best performance. The BRI experiments were a continuation of the SMI work with the wake generator replaced by a swirler row to turn the flow and a deswirler row to create a wake by diffusion. Results of the BRI experiments showed a performance degradation between mid and far spacing which was not observed in SMI. This trend is seen in the numerical work as well, and the time-averaged data shows that the majority of this performance change occurred in the rotor. An unsteady separation bubble periodically forms and collapses as shocks reflect through the stator passage, creating additional aerodynamic blockage. The shed vortices induced by the unsteady loading and unloading of the stator trailing edge are chopped, with a frequency related to the spacing, by the rotor leading edge and ingested by the rotor. Once ingested the vortices interact in varying degrees with the rotor boundary layer. A treatment of the loss production in the BRI rig is given based on the time-accurate and time-averaged, high-fidelity TURBO results.Copyright

HIGH FIDELITY 3D TURBOFAN ENGINE SIMULATION WITH EMPHASIS ON TURBOMACHINERY-COMBUSTOR COUPLING

Introduction The 3-dimensional flow in the primary flow path of the GE90-94B high bypass ratio turbofan engine has been achieved. The simulation of the compressor components, the cooled high pressure turbine and the low pressure turbine was performed using the APNASA turbomachinery flow code. The combustor flow and chemistry were simulated using the National Combustor Code, NCC. The engine simulation matches the engine thermodynamic cycle for a sea-level takeoff condition. The fan, booster and OGV are corrected to the cycle condition from component simulations, whereas the high pressure compressor and turbines have been simulated at the cycle condition and coupled to the NCC code by passing profiles. Details of this coupling are presented. Significant gains in parallel computing are demonstrated which allow simulations to take place that can impact design. One of the goals of the Numerical Propulsion System Simulation (NPSS) Program at NASA Glenn Research Center has been to demonstrate a high-fidelity 3D Turbofan Engine Simulation. This simulation will support the multi-dimensional, multi-fidelity, multidiscipline concept of the design and analysis of propulsion systems for the future. This paper describes the current status of one major part of that goal: the complete turbofan engine simulation using an advanced 3-D Navier-Stokes turbomachinery solver, APNASA, coupled with the National Combustion Code, NCC. A production engine has been chosen for this demonstration: the GE90 turbofan engine shown in Fig. 1. A sea level, Mach 0.25, takeoff condition has been chosen for the simulation. The main reason is that detailed cooling flows for the turbine are well known at takeoff since this represents the cooled turbine design condition. Since the cooling flow represents a significant amount of the boundary condition information required for the simulation, it was felt this was a good point for the simulation. It also represents a condition where there are the highest temperatures and most stress in the engine, and is therefore a practical point to gain further understanding. _ C A S b A The GE90 development program included component testing of all the turbomachinery as well as the combustor. The full engine simulation effort has taken advantage of this. All the turbomachinery components 38th AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit 7-10 July 2002, Indianapolis, Indiana AIAA 2002-3769 Copyright

Investigation of Loss Generation in an Embedded Transonic Fan Stage at Several Gaps Using High-Fidelity, Time-Accurate Computational Fluid Dynamics

The blade-row interaction (BRI) rig at the Air Force Research Laboratory, Compressor Aero Research Laboratory, has been simulated at three axial gaps between the highly loaded upstream stator row and the downstream transonic rotor using TURBO. Previous work with the stage matching investigation (SMI) demonstrated a strong dependence of mass flow rate, efficiency, and pressure ratio on the axial spacing between an upstream wake generator and the downstream rotor through the variation of the axial gap. Several loss producing mechanisms were discovered and related to the spacings, referred to as close, mid, and far. In the SMI work, far spacing had the best performance. The BRI experiments were a continuation of the SMI work with the wake generator replaced with a swirler row to turn the flow and a deswirler row to create a wake by diffusion. Results of the BRI experiments showed a performance degradation between mid- and far spacings, which was not observed in SMI. This trend is seen in the numerical work as well, and the time-averaged data show that the majority of this performance change occurred in the rotor. An unsteady separation bubble periodically forms and collapses as shocks reflect through the stator passage, creating additional aerodynamic blockage. The shed vortices induced by the unsteady loading and unloading of the stator trailing edge are chopped, with a frequency related to the spacing, by the rotor leading edge and ingested by the rotor. Once ingested the vortices interact in varying degrees with the rotor boundary layer. A treatment of the loss production in the BRI rig is given based on the time-accurate and time-averaged, high-fidelity TURBO results.

A Wall Function for Calculating the Skin Friction With Surface Roughness

Wall functions are used in CFD to provide skin friction values and hence the boundary conditions for flow variables along solid surfaces. In this paper the effect of surface roughness on skin friction is incorporated in a wall function approach which uses Spalding’s formula. The use of Spalding’s formula extends the method to a wider range of wall distance than the logarithmic friction law. For CFD applications the results are then re-formulated for explicit calculation throught the use of an additional variable — the ratio of the surface roughness height to the distance from the solid surface. For a given computational grid this information is readily available in a CFD calculation. This methodology is applied to a linear compressor cascade that has been experimentally measured for different roughness values. The comparison of the simulation followed the same trends as the experiment, but with less overall effect.Copyright

Educational Software for Blade and Disk Design

One problem with many introductory level turbomachinery courses is a lack of easy to use design and visualization tools. Oftentimes students will become too focused on the underlying math and never develop a good understanding of the physical systems they are working with. To combat this problem, a series of GUI based design and visualization codes have been created. The codes are intended primarily for educational purposes, but in many cases they are robust enough for actual design use. All of the new codes have been designed to complement and to a small extent connect to the existing T-Axi suite of codes. This paper will focus on two new freely available codes: T-Axi Blade and T-Axi Disk. T-Axi Blade was created to help visualize the key design features of a single rotor, with special emphasis on vector triangles, and the ability to design compressor, turbine, axial and centrifugal rotors with a universal approach. T-Axi Blade can also output files for use with T-Axi Disk, a code to aid in the design of a lightweight disk to support the blade row. This code allows the user to design a disk interactively with instantaneous feedback in the form of weights, stresses, and a series of 2D and 3D visualizations. Taken together these codes offer a simple introduction to multidisciplinary engineering. In this paper the structure of these codes and the numerical models are discussed. Ideas are presented of how these codes can be used as a classroom tool and as an actual design tool. An example analysis of the third stage GE EEE HPC axial rotor is presented to demonstrate the features of these codes.Copyright

Optimization of a 3-Stage Booster: Part 2—The Parametric 3D Blade Geometry Modeling Tool

A parametric approach for blade geometry design has been developed to obtain 3D blade models. The geometry of the blade is defined by a basic set of parameters that are first obtained from an axisymmetric solver. These parameters include the leading edge meridional coordinates, flow angles, axial chord, and the meridional coordinates of streamlines. Other parameters such as thickness to chord ratio need to be defined. Using these parameters the 2D airfoils are created and are stacked radially using one of the many multiple options that define the stacking axes from several additional parameters. The tool produces the desired number of 2D sections in a normalized coordinate system. Each blade section is then transformed to a 3D Cartesian coordinate system. Using Unigraphics-NX (CAD package), these sections are lofted and a 3D blade model is obtained. Parametric update of the spline points defining the 3D blade sections results in new blade shapes without going directly back into the CAD system. The importing of the geometry into a CFD solver, and a finite element solver to determine mode shapes and stresses is demonstrated. Full details of the blade procedure is presented for a 3-Stage Booster design. This parametric approach for defining blade geometry and how it lays a groundwork for a high-fidelity optimization procedure is described.Copyright