0

ASME Conference Presenter Attendance Policy and Archival Proceedings

2014;():V02BT00A001. doi:10.1115/GT2014-NS2B.
FREE TO VIEW

This online compilation of papers from the ASME Turbo Expo 2014: Turbine Technical Conference and Exposition (GT2014) represents the archival version of the Conference Proceedings. According to ASME’s conference presenter attendance policy, if a paper is not presented at the Conference, the paper will not be published in the official archival Proceedings, which are registered with the Library of Congress and are submitted for abstracting and indexing. The paper also will not be published in The ASME Digital Collection and may not be cited as a published paper.

Commentary by Dr. Valentin Fuster

Design Methods and CFD Modeling for Turbomachinery

2014;():V02BT39A001. doi:10.1115/GT2014-25101.

This paper deals with the simulation of steady flows in turbomachinery. Two approaches are proposed, the first one is the classical multiple-rotating frame method (MRF) by multi-zone approach where the different zones are separated by non-overlapping interfaces and solved independently. Since each zone is loaded separately, a transferring system should be properly implemented at the interface boundaries. Two techniques are considered, in the first one the conservative variables are interpolated between zones while in the second one the fluxes are transferred through the interfaces.

The other proposed approach is a new version of the MRF using a virtual interface (VMRF). This is a simplified of the previous one where the interfaces are created virtually at the solver level, rendering the method easy to implement especially for edge-based numerical schemes, and avoiding any re-meshing in case one needs to change interface position, shape or simply remove or add new one. Finally, numerical tests are performed to demonstrate the efficiency of the proposed methods by comparison with commercial codes (ANSYS FLUENT).

Commentary by Dr. Valentin Fuster
2014;():V02BT39A002. doi:10.1115/GT2014-25103.

In order to achieve greater pressure ratios, compressor designers have the opportunity to use transonic configurations. In the supersonic part of the incoming flow, shock waves appear in the front part of the blades and propagate in the upstream direction.

In case of multiple blade rows, steady simulations have to impose an azimutal averaging (mixing plane) which prevents these shock waves to extend upstream. In the present paper, several mixing plane locations are numerically tested and compared in a supersonic configuration.

An analytical method is used to describe the shock pattern. It enables to take a critical look at the CFD steady results. Based on this method, the shock losses are also evaluated. The good agreement between analytical and numerical values shows that this method can be useful to wisely forecast the mixing plane location and to evaluate the shift in performances due to the presence of the mixing plane.

Topics: Compressors
Commentary by Dr. Valentin Fuster
2014;():V02BT39A003. doi:10.1115/GT2014-25150.

In the present study, several control devices have been investigated in the framework of a high pressure compressor rotor using RANS simulations. The analysis of performance maps and of flow predictions leads to select the injection device located up-stream of the rotor tip, at the shroud in order to control the tip leakage flow. The high loss levels are reduced and the radial and azimuthal extension of the high loss area is smaller. Then the chosen flow control technique has been simulated using URANS and ZDES with the IGV passing wakes. The injection technique reduces the loss area and level, energizing and stabilizing the tip leakage vortex thanks to high momentum. So the vortex disruption is removed or at least delayed. Moreover, the influence of IGV passing wakes is reduced.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A004. doi:10.1115/GT2014-25209.

The adjoint method eliminates the dependence of the gradient of the objective function with respect to design variables on the flow field making the obtainment of the gradient both accurate and fast. For this reason, the adjoint method has become the focus of attention in recent years. This paper develops a continuous adjoint formulation for through-flow aerodynamic shape design in a multi-stage gas turbine environment based on a S2 surface quasi-3D problem governed by the Euler equations with source terms. Given the general expression of the objective function calculated via a boundary integral, the adjoint equations and their boundary conditions are derived in detail by introducing adjoint variable vectors. As a result, the final expression of the objective function gradient only includes the terms pertinent to those physical shape variations that are calculated by metric variations. The adjoint system is solved numerically by a finite-difference method with explicit Euler time-marching scheme and a Jameson spatial scheme which employs first and third order dissipative flux. Integrating the blade stagger angles and passage perturbation parameterization with the simple steepest decent method, a gradient-based aerodynamic shape design system is constructed. Finally, the application of the adjoint method is validated through a 5-stage turbine blade and passage optimization with an objective function of entropy generation. The result demonstrates that the gradient-based system can be used for turbine aerodynamic design.

Topics: Design , Turbines
Commentary by Dr. Valentin Fuster
2014;():V02BT39A005. doi:10.1115/GT2014-25230.

This article describes a nonlinear frequency domain method for the simulation of unsteady blade row interaction problems across several blade rows in turbomachinery. The capability to efficiently simulate such interactions is crucial for the improvement of the prediction of blade vibrations, tonal noise, and the impact of unsteadiness on aerodynamic performance.

The simulation technique presented here is based on the harmonic balance approach and has been integrated into an existing flow solver. A nontrivial issue in the application of harmonic balance methods to turbomachinery flows is the fact that various fundamental frequencies may occur simultaneously in one relative system, each one being due to the interaction of two blade rows. It is shown that, considering the disturbances corresponding to different fundamental frequencies as mutually uncoupled, one can develop an unsteady simulation method which from a practial view point turns out to be highly attractive. On the one hand, it is possible to take into account arbitrarily many nonlinear interaction terms. On the other, the computational efficiency can be increased considerably once it is known that the nonlinear coupling between certain subsets of the harmonics plays only a minor role.

To validate the method and demonstrate its accuracy and efficiency a multistage compressor configuration is simulated using both the method described in this article and a conventional time-domain solver.

Topics: Turbomachinery
Commentary by Dr. Valentin Fuster
2014;():V02BT39A006. doi:10.1115/GT2014-25362.

The non-axisymmetric endwall profiling has been proven to be an effective tool to reduce the secondary flow loss in turbomachinery. In this work, the aerodynamic optimization for the non-axisymmetric endwall profile of the turbine cascade and stage was presented and the design results were validated by annular cascade experimental measurements and numerical simulations. The parametric method of the non-axisymmetric endwall profile was proposed based on the relation between the pressure field variation and the secondary flow intensity. The optimization system combines with the non-axisymmetric endwall parameterization method, global optimization method of the adaptive range differential evolution algorithm and the aerodynamic performance evaluation method using three-dimensional Reynolds-Averaged Navier-Stokes (RANS) and kω SST turbulent with transition model solutions. In the part I, the optimization method is used to design the optimum non-axisymmetric endwall profile of the typical high loaded turbine stator. The design objective was selected for the maximum total pressure coefficient with constrains on the mass flow rate and outlet flow angle. Only five design variables are needed for one endwall to search the optimum non-axisymmetric endwall profile. The optimized non-axisymmetric endwall profile of turbine cascade demonstrated an improvement of total pressure coefficient of 0.21% absolutely, comparing with the referenced axisymmetric endwall design case. The reliability of the numerical calculation used in the aerodynamic performance evaluation method and the optimization result were validated by the annular vane experimental measurements. The static pressure distribution at midspan was measured while the cascade flow field was measured with the five-hole probe for both the referenced axisymmetric and optimized non-axisymmetric endwall profile cascades. Both the experimental measurements and numerical simulations demonstrated that both the secondary flow losses and the profile loss of the optimized non-axisymmetric endwall profile cascade were significantly reduced by comparison of the referenced axisymmetric case. The weakening of the secondary flow of the optimized non-axisymmetric endwall profile design was also proven by the secondary flow vector results in the experiment. The detailed flow mechanism of the secondary flow losses reduction in the non-axisymmetric endwall profile cascade was analyzed by investigating the relation between the change of the pressure gradient and the variation of the secondary flow intensity.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A007. doi:10.1115/GT2014-25364.

Aerodynamic optimization design and experimental validation for the non-axisymmetric endwall profiles of the turbine cascade have been completed in the part I of this research work. Non-axisymmetric endwall profile optimization design of the turbine stage and corresponding steady and unsteady flow characteristics were presented in the part II. Aerodynamic optimization design for the non-axisymmetric endwall profile of the turbine stage was conducted when the maximization of the total-total isentropic efficiency was set as the design objective with constraint on the mass flow rate. The aerodynamic performance of the designed turbine stage was evaluated using three-dimensional Reynolds-Averaged Navier-Stokes (RANS) solutions. The non-axisymmetric endwall profiles of the stator hub and shroud as well as the rotor hub in the turbine stage were optimized using developed endwall profile method in the part I. A total of 15 design variables were employed in the optimization for the stator and rotor endwalls. The global optimization method of the adaptive rang differential evolution algorithm was used to search the optimal non-axisymmetric endwall profile. The total-total isentropic efficiency of the turbine stage with the optimized non-axisymmetric endwall profile increases 0.26% by comparison of the referenced axisymmetric endwall design when the effects of the rotor tip clearance were also considered. The secondary flow losses of the stator and rotor were significantly reduced in the optimized non-axisymmetric endwall stage, as well as the tip leakage flow losses. In addition, the unsteady aerodynamic performance of the turbine stage with the optimized non-axisymmetric endwall profile and referenced axisymmetric endwall were numerically investigated and compared. The numerical results indicate that the fluctuating velocity in the rotor blade passage of the optimized non-axisymmetric endwall stage significantly decreases since the stator wake and secondary flow losses are reduced. Thus, the intensity of the unsteady interaction between the stator upstream flow and the flow in the rotor passage decreases. The time-averaged results indicated that the aerodynamic efficiency and output power of the turbine stage with the optimized non-axisymmetric endwall profile are higher than that of the referenced axisymmetric endwall stage. Meanwhile, the transient results at different time steps show that the periodic fluctuating amplitude of the efficiency and power of the optimized non-axisymmetric endwall stage were smaller than that of the referenced axisymmetric endwall stage due to the weaker stator/rotor unsteady interaction effects.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A008. doi:10.1115/GT2014-25379.

An example of optimisation of high-turning outlet guide vanes for subsonic compressor stage is implemented in a new approach for blade profile topology. This topology includes types of curves on PS, SS, the number of control points and the distribution law, where several coefficients are implemented as function of the upstream flow. This functional law is defined during optimisation procedures for profiles with different parameters (Mach, inlet angle, turning angle, solidity) and further correlation analysis. As a result an 8 degree gain in turning angle for last stage guide vane is achieved. Also possible design is provided for a 12 degree gain in turning angle with higher pressure losses on 20% (relatively).

Commentary by Dr. Valentin Fuster
2014;():V02BT39A009. doi:10.1115/GT2014-25434.

Computational fluid dynamics (CFD) has become a critical tool in the design of aeroengines. Increasing demand for higher efficiency, performance and reduced emissions of noise and pollutants has focused attention on secondary flows, small scale internal flows and flow interactions. In conjunction with low order correlations and experimental data, RANS (Reynolds-Averaged Navier-Stokes) modelling has been used effectively for some time, particularly at high Reynolds numbers and at design conditions. However, the range of flows throughout an engine is vast, with most, in reality being inherently unsteady. There are many cases where RANS can perform poorly, particularly in zones characterised by strong streamline curvature, separation, transition, relaminarisation and heat transfer. The reliable use of RANS has also been limited by its strong dependence on turbulence model choice and related ad-hoc corrections. For complex flows, Large-Eddy Simulation (LES) methods provide reliable solutions, largely independent of turbulence model choice and at a relatively low cost for particular flows. LES can now be used to provide in depth knowledge of flow physics, for example in areas such as transition and real wall roughness effects. This can be used to inform RANS and lower order modelling. For some flows, LES can now even be used for design. Existing literature is used to show the potential of LES for a range of flows in different zones of the engine. Based on flow taxonomy, best practices including meshing requirements and turbulent inflow conditions are introduced, leading to the proposal of a tentative expert system for industrial use. In this way, LES becomes a well controlled tool, suitable for design use and reduces the burden on the end user. Further attention is also given to how LES can be used currently and in the future.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A010. doi:10.1115/GT2014-25499.

This paper describes the implementation of a set of nonreflecting boundary conditions of increasing approximation quality for time-accurate and time-linearized 3D RANS solvers in the time and frequency domain. The implementations are based on the computation of eigenfunctions, either analytically or numerically, of the linearized Euler or Navier-Stokes equations for increasingly complex background flows. This results in a hierarchy of nonreflecting boundary conditions based on 1D characteristics, 2D circumferential mode decomposition, and 3D circumferential and radial mode decomposition, including viscous effects in the latter, for the frequency domain solver. By applying a Fourier transform in time at the boundaries the frequency domain implementations can be employed in the time domain solver as well. The limitations of each approximation are discussed and it is shown that increasing the precision of the boundary treatment the nonreflecting property of the boundary conditions is preserved for more complex flows without incurring an excessive increase in computing time.

Results of a flutter analysis of a low pressure turbine blade obtained by time and frequency domain simulations are validated against each other and against reference results obtained with a 3D Euler frequency domain solver. The comparison of the results for different boundary conditions reveals the importance of using high quality boundary conditions.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A011. doi:10.1115/GT2014-25530.

Numerical methods have become the basis for the aerodynamic design of turbomachinery in order to reduce the time for development cycles and associated cost. Designing modern axial compressors requires high confidence in the quality of numerical predictions. In terms of the aerodynamics, the loading of the blades as well as the efficiency targets constantly increase. Losses have to be predicted precisely and the impact of three-dimensional secondary flows, separation, and laminar-turbulent transition must be taken into account. In the present paper, the aerodynamic prediction quality of the state-of-the-art turbomachinery design code TRACE is validated against experimental data from a 2.5-stage axial compressor.

The aerodynamic prediction quality is systematically investigated to determine errors and uncertainties regarding the discretization, turbulence and transition models, and importance of considering unsteady effects. Computations are performed for several operating points and the results are validated by means of the compressors integral pressure ratio as well as by means of local pneumatic probe measurements. It is shown that using the empirical γ–ReΘ model improves the prediction quality of the boundary layers and wake flows. Time-resolved computations at the design point of the compressor show that the strength and the losses of a corner separation in both vane rows are reduced to realistic levels when the periodic-unsteady interaction with the upstream wakes is considered. The generally good aerodynamic predictions for both local and integral experimental quantities qualify TRACE for aeroelastic predictions which are planned for the future.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A012. doi:10.1115/GT2014-25531.

In the use of RANS models, it is well known that the selection of the turbulence model and the numerical scheme may have a critical impact not only in terms of convergence, but also on the reliability to simulate separated or secondary flows in general. The aim of the investigation, performed using the commercial software FINE/Turbo, is the understanding and the quantification of the effects of these two numerical parameters on the performance and the stability of a state-of-the-art controlled diffusion airfoil compressor cascade. A mesh sensitivity analysis has been carried out at both design and off-design conditions. The behaviour of the main flow parameters have been investigated over the whole incidence working range, considering a variation of the inlet Mach number between 0.35 and 0.65. Five different turbulence models have been tested: Baldwin-Lomax, Spalart-Allmaras, k–ε Yang-Shih, k–ε Launder-Sharma and k–ω SST. In a specific combination of incidences and Mach numbers, the impact of turbulence model settings has been assessed imposing boundary conditions according to different criteria. Two different numerical schemes have been tested: a Jameson central scheme and a second order upwind scheme. The results between the different simulations are discussed in terms of loss coefficient distribution and incidence range; considering the turbulence model comparison, the differences are significant in the whole incidence range, specially approaching the stall limit. Baldwin-Lomax and Spalart-Allmaras simulations present the same value of last stable incidence, while Yang-Shih and SST are characterized by a reduced stall margin. In many operating conditions, simulations computed with centered scheme present negative losses in a wide area of the outlet sections. This problem is reduced if an upwind scheme is used, but causes a substantial reduction of the incidence range.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A013. doi:10.1115/GT2014-25618.

State-of-the-art aerodynamic blade design processes mainly consist of two phases: optimal design of 2D blade sections and then stacking them optimally along a three-dimensional stacking line. Such a quasi-3D approach, however, misses the potential of finding optimal blade designs especially in the presence of strong 3D flow effects. Therefore, in this paper a blade optimization process is demonstrated which uses an integral 3D blade model and 3D CFD analysis to account for three-dimensional flow features. Special emphasis is put on shortening design iterations and reducing design costs in order to obtain a rapid automatic optimization process for fully 3D aerodynamic turbine blade design which can be applied in an early design phase already.

The three-dimensional parametric blade model is determined by up to 80 design variables. At first, the most important design parameters are chosen based on a non-linear sensitivity analysis. The objective of the subsequent optimization process is to maximize isentropic efficiency while fulfilling a minimal set of constraints. The CFD model contains both important geometric features like tip gaps and fillets, and cooling and leakage flows to sufficiently represent real flow conditions.

Two acceleration strategies are used to cut down the turn-around time from weeks to days. Firstly, the aerodynamic multi-stage design evaluation is significantly accelerated with a GPU-based RANS solver running on a multi-GPU workstation. Secondly, a response surface method is used to reduce the number of expensive function evaluations during the optimization process. The feasibility is demonstrated by an application to a blade which is a part of a research rig similar to the high pressure turbine of a small civil jet engine. The proposed approach enables an automatic aerodynamic design of this 3D blade on a single workstation within few days.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A014. doi:10.1115/GT2014-25709.

This paper presents the relation between circumferential fluctuation and the geometric and flow parameters. The governing equations are derived by circumferentially averaging the three-dimensional (3D) Navier-Stokes equations. Different types of compressor cascades are simulated and the circumferential fluctuation terms are extracted according to the definition of circumferential average. Three different blade profiles are chosen, including CDA, C4 and NACA65 profile, respectively. The peak value of circumferential fluctuation terms often occurs at the leading or the trailing edge and increases as the radius grows. Meanwhile, the circumferential fluctuation terms exist at the inlet of the blade which can be accurately calculated. 0°, 15° and 30° camber angles are chosen to study the influence of camber angle. When the camber angle is smaller, the flow is more uniform and therefore, the value of circumferential fluctuation is lower. Different incidence angles are compared to discuss the relationship between circumferential fluctuation and incidence angle. For specific term of circumferential fluctuations, the distribution curves are different.

Topics: Compressors
Commentary by Dr. Valentin Fuster
2014;():V02BT39A015. doi:10.1115/GT2014-25799.

Curved blade has been widely used to reduce the endwall loss, but there is no criterion for curved blade design. Relationship of the optimum curved blade generate line (stack line) and the inlet Mach number, solidity, aspect ratio and camber angle in a linear compressor cascade were researched by optimization method in present paper. The stack line is vertically symmetrical, composed of two third-order Bezier curves and a straight line. The results show that total pressure loss coefficient decreases with the curved height increasing in the present calculate conditions at the same curved angle, and the optimum curved height is 0.5. The total pressure loss coefficient variation with curved angle presents a approximate parabola line type at the same curved height, there is an optimum curved angle, at which the total pressure loss coefficient is minimal. The optimum curved angle variation with the cascade parameters. Optimum curved angle increases with the inlet Ma and camber angle increasing, optimum curved angle variation with inlet Ma shows a polynomial curve type, optimum curved angle varied linearly with camber angle increasing. Optimum curved angle has little changes with solidity and aspect ratio increasing, optimum curved angle is about 6.5° in present conditions, but optimum curved angle will change with the blade loading. The benefit of the optimum curved blade increases with the inlet Mach number and camber angle increasing, and it has little change with solidity and aspect ratio increasing.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A016. doi:10.1115/GT2014-25816.

Transonic axial flow fan has relatively high blade tip speed and produces higher pressure ratio than the subsonic. However, considerable losses are brought about by the shock waves close to blade tip and over part of span, leading to deteriorated overall efficiency and operating flow range. It is generally acknowledged that modifications of blade stacking line (axially sweep and tangentially lean) and sectional profiles can help to control spanwise distribution of blade loading, reduce shock loss and secondary flow, and extend the operating flow range.

The present study is to maximize the comprehensive benefits of simultaneously optimizing the sectional profiles and stack line by means of a global optimization method with reduced cost. In contrast with previous studies, it is of two distinguished features. First, in blade geometry parameterization, both sectional profiles and stacking line are varied to provide more flexible blade shape variation and subsequently permit more optimization performance gains. Secondly, with simultaneous variation of sectional profiles and stacking line, number of optimization variables and nonlinearity of optimization problem will increase largely. How to obtain a global optimal solution and also reduce the computation become the major concerns. For this purpose, a global optimization method proposed by us is used. It includes an improved CCEA (Cooperative Co-Evolution Algorithm) optimizer, adaptively updated kriging surrogate model, and one-stage Expected Improvement (EI) approach that permits adaptive sampling. At initial stage, a coarse surrogate model is constructed with small number of samples. During the optimization process, some new samples are identified, evaluated, and then used to refine the model and conduct further optimal searching. In the optimization process, the accuracy of the surrogate model is improved based on its own characteristics of optimization problem and this permits the optimizer to conduct the aim-oriented optimal searches. In such a manner, the surrogate model sustains high-level of accuracy while uses fewer samples, thus the blade optimization and computations are significantly reduced.

The optimization is conducted for NASA Rotor67 at design flow rate with a single workstation of DELL 7500. It is demonstrated that the optimized blade design produces significant performance gains at design condition (where the overall efficiency and pressure ratio are increased respectively by 1.27 and 6.53 points) and also at off-design conditions.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A017. doi:10.1115/GT2014-25869.

The CFD assisted design of modern single- and multi-stage turbomachines is usually performed with the mixing plane approach in order to assess the components matching. While the mixing-plane state-of-the-art is based on a boundary-condition based approach, hereafter called explicit, the authors presented last year a novel, fully implicit method, which shows considerable advantages compared to the explicit one. In the present paper the quality and advantages of the novel approach compared to the state-of-the-art will be shown through a variety of detailed examples. The main issues discussed are the built-in ability to reduce incoming disturbances and to manage backflow at the interface due to the implicit formulation. With selected cases it is shown that no special care has to be taken to avoid reflections at the interface also for inviscid transonic or fully-supersonic cases. Moreover, detailed results for a high-pressure centrifugal compressor are presented, showing that the proposed approach is able to capture both, the global behavior as well as local flow-features over the complete speed-line, while the explicit approach partially fails on the same test case.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A018. doi:10.1115/GT2014-25876.

The use of Computational Fluid Dynamics (CFD) tools for integrated simulations of gas turbine components has emerged as a promising way to predict undesired component interactions thereby giving access to potentially better engine designs and higher efficiency. In this context, the ever-increasing computational power available worldwide makes it possible to envision integrated massively parallel combustion chamber-turbomachinery simulations based on Large-Eddy Simulations (LES). While LES have proven their superiority for combustor simulations, few studies have employed this approach in complete turbomachinery stages. The main reason for this is the known weaknesses of near wall flow modeling in CFD. Two approaches exist: the wall-modeled LES, where wall flow physics is modeled by a law-of-the-wall, and the wall-resolved LES where all the relevant near wall physics is to be captured by the grid leading to massive computational cost increases. This work investigates the sensitivity of wall-modeled LES of a high-pressure turbine stage. The code employed, called TurboAVBP, is an in-house LES code capable of handling turbomachinery configurations. This is possible through an LES-compatible approach with the rotor/stator interface treated based on an overset moving grids method. It is designed to avoid any interference with the numerical scheme, allow the proper representation of turbulent structures crossing it and run on massively parallel platforms. The simulations focus on the engine-representative MT1 transonic high-pressure turbine, tested by QinetiQ. To control the computational cost, the configuration employed is composed of 1 scaled stator section and 2 rotors. The main issues investigated are the effect of mesh resolution and the effect of sub-grid scale models in conjunction with wall modeling. The pressure profiles across the stator and rotor blades are in good agreement with the experimental data for all cases. Radial profiles at the rotor exit (in the near and far field) show improvement over RANS predictions. Unsteady flow features, inherently present in LES, are, however, found to be affected by the modeling parameters as evidenced by the obtained shock strengths and structures or turbulence content of the different simulations.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A019. doi:10.1115/GT2014-25967.

In this paper we present a fully coupled algorithm for the resolution of compressible flows at all speed. The pressure-velocity coupling at the heart of the Navier Stokes equations is accomplished by deriving a pressure equation in similar fashion to what is done in the segregated SIMPLE algorithm except that the influence of the velocity fields is treated implicitly. In a similar way, the assembly of the momentum equations is modified to treat the pressure gradient implicitly. The resulting extended system of equations, now formed of matrix coefficients that couples the momentum and pressure equations, is solved using an algebraic multigrid solver.

The performance of the coupled approach and the improved efficiency of the novel developed code was validated comparing results with experimental and numerical data available from reference literature test cases as well as with segregated solver as exemplified by the SIMPLE algorithm. Moreover the reference geometries considered in the validation process cover the typical aerodynamics applications in gas turbine analysis and design, considering Euler to turbulent flow problems and clearly indicating the substantial improvements in terms of computational cost and robustness.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A020. doi:10.1115/GT2014-25977.

A validation study of a variety of compressible flow turbomachinery cases is presented with comparisons to test data using OpenFOAM. OpenFOAM is open-source code consisting of various solvers and computational libraries focused on CFD. The study used a particular solver version with a density based approach that was derived from the “extended” branch of OpenFOAM. The example cases all consisted of single blade row designs at steady state and were run fully viscous (unless noted otherwise) with various turbulence models.

The results showed a definite superiority of the density based solver over other OpenFOAM solvers in a test suite of simplified cases as well as in more complex examples in actual turbomachinery designs. A typical Laval nozzle case and transonic bump case are presented demonstrating the basic ability of the solver to capture shocks and to handle transonic flow in general. Actual turbomachinery applications consisted of a two-dimensional transonic compressor cascade, a moderately supersonic two-dimensional turbine cascade, two radial compressor cases, and a radial inflow turbine.

The results showed the solver to be very capable of capturing pressure distributions and, most importantly, aerodynamic loss through the machines. The ability of the solver to accurately model performance in a wide range of different designs and across the entire performance map was demonstrated. Detailed comparisons to highly regarded test data are shown.

Special examination was made of the computational costs of the solver which were quite high with run times coming in at about 10 times longer than other commercial compressible flow solvers. Several acceleration methods are discussed which significantly improved run time performance.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A021. doi:10.1115/GT2014-26044.

Accurate numerical simulations depend on the correct prediction of all relevant flow phenomena. For many aeronautical devices such as turbomachinery the behaviour of boundary layers, wall shear stress and wall heat transfer are significant for the performance. Turbulence and transition may influence such flow characteristics. Onset and the extent of transition can therefore be of high importance for the design process.

In this paper three-dimensional steady-state simulations of a two-stage turbine with and without modeling of transition are performed. The configuration consists of a transonic high pressure turbine stage followed by an S-shaped turning mid turbine frame and a counter-rotating low pressure turbine rotor. The in-house Reynolds-averaged Navier-Stokes solver has been applied to this configuration with the correlation based γ–Reθ transition model developed by Menter et al. which has been added to the SST turbulence model. Also, calculations without transition with the SST and the Spalart-Allmaras turbulence model have been performed. This configuration is also the subject of experimental investigations at the Institute for Thermal Turbomachinery and Machine Dynamics at Graz University of Technology, so that measurement data at two planes are used for the verification of the simulations. After verification and flow analysis, the results are also discussed to evaluate the effect of modeling transition phenomena. An analysis of the overall efficiency of the turbine and the efficiency losses in the blade and vane rows are finally presented.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A022. doi:10.1115/GT2014-26119.

Large-Eddy Simulations of wall bounded, low Mach number turbulent flows are conducted using an unstructured finite-volume solver of the compressible flow equations. The numerical method employs linear reconstructions of the primitive variables based on the least-squares approach of Barth. The standard Smagorinsky model is adopted as the subgrid term. The artificial viscosity inherent to the spatial discretization is maintained as low as possible reducing the dissipative contribution embedded in the approximate Riemann solver to the minimum necessary. Comparisons are also discussed with the results obtained using the implicit LES procedure.

Two canonical test-cases are described: a fully developed pipe flow at a bulk Reynolds number Reb = 44 × 103 based on the pipe diameter, and a confined rotor-stator flow at the rotational Reynolds number ReΩ = 4 × 105 based on the outer radius. In both cases the mean flow and the turbulent statistics agree well with existing DNS or experimental data.

Topics: Turbulence
Commentary by Dr. Valentin Fuster
2014;():V02BT39A023. doi:10.1115/GT2014-26131.

Compressor maps of aero engines show the relation between corrected inlet mass flow and total pressure ratio for various engine speeds. Different speed lines represent different operating conditions of the compressor, where especially operating bounds like surge and choke are important for the design process. Typically, 3D CFD compressor maps are computed with the so called hot geometry given for the aerodynamic design point. However, in reality airfoil shapes will change for different engine speeds and gas loads resulting in twisted airfoils and changed tip clearances. Thus, using the nominal hot geometry for the whole compressor map is not fully correct. In order to obtain higher quality performance maps these effects need to be considered. The paper shows a process for computing compressor maps with 3D CFD, where strucural deformations of the blade due to varying speeds and gas loads are taken into account by blade morphing. This process is applied to a 1.5-stage compressor showcase.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A024. doi:10.1115/GT2014-26220.

A density-based solver for turbomachinery application is developed based on the central-upwind schemes of Kurganov and Tadmor using the open source CFD-library OpenFOAM. Preconditioning of Weiss and Smith is utilized to extend the applicability down to the incompressibility limit. Implicit residual averaging, bulk viscosity damping and local time stepping are employed to speed up the simulations. A low-storage 4-stage Runge-Kutta scheme and dual time-stepping are used for time integration.

The presented solver is compared with results from ANSYS Fluent 13.0 and measurement data. Three different test cases are conducted to analyze different flow conditions: The circular bump for low and high speed inviscid flows and computational performance assessment, the two-dimensional VKI turbine guide vane for viscous flows and the the three-dimensional DLR high speed centrifual compressor validating the performance for rotating turbo-machinery. All three test cases show a very good agreement between OpenFOAM and ANSYS Fluent.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A025. doi:10.1115/GT2014-26338.

A numerical investigation of incompressible turbomachinery and the comparison of two CFD packages are presented within this paper. A ducted single rotor fan is simulated with OpenFOAM and ANSYS FLUENT by applying methods as comparable as possible. The characteristic maps and flow fields are analyzed and the results from the CFD codes are compared to examine differences regarding accuracy and efficiency. Additionally the influence of the turbulence model is determined. It is found that the CFD programs show a good agreement especially at the machines design point.

The information about the flow field of this fan is used for the modelling of a high-performance and energy-efficient ducted contra-rotating fan (CRF). Comparing the CRF simulation results to those of the single rotor fan, a doubling of the total pressure rise and a significant reduction of the swirl in the wake flow can be noticed.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A026. doi:10.1115/GT2014-26363.

The blade geometry design process is integral to the development and advancement of compressors and turbines in gas turbines or aeroengines. An airfoil section design feature has been added to a previously developed open source parametric 3D blade design tool. The second derivative of the mean-line (related to the curvature) is controlled using B-splines to create the airfoils. This is analytically integrated twice to obtain the mean-line. A smooth thickness distribution is then added to the airfoil with two options either the Wennerstrom distribution or a quartic B-spline thickness distribution. B-splines have also been implemented to achieve customized airfoil leading and trailing edges. Geometry for a turbine, compressor, and transonic fan are presented along with a demonstration of the importance of airfoil smoothness.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A027. doi:10.1115/GT2014-26365.

Particles contained in air can deposit on the blade surface to cause fouling when lubricating oil and water steam are existed. Fouling changes blade geometry and blade surface roughness is increased, thus aerodynamic performance is affected. Many researchers simulated axial flow compressor fouling by adding constant surface roughness and modifying blade thickness which can’t reflect the real status of fouled compressor. In this paper, reverse technology is introduced to reconstruct the solid model of fouled compressor which is imported into fluid flow simulation software. The flow of gas phase and gas-solid coupling phase are implemented to reveal the nature of flow in fouled axial flow compressor. Based on Euler-Lagrange model, this paper made numerical simulation of gas-solid two phase flow in the axial flow compressor rotor cascade. Simulation result shows that fouling causes the decrease of effective flow area, thus thermodynamic performance is degraded. Gas-solid phase flow shows that particles are not uniformly deposited on the blade surface. When particle is smaller and rotor blade is rough, it is more easily deposited on the surface. And particle mass concentration is affected by ambient conditions such as inlet temperature, rotational speed, particle diameter, particle mass flow rate.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A028. doi:10.1115/GT2014-26371.

Despite many advances in both optimization methods and computational fluid dynamics, the timely automatic selection and refinement, via physics-based and empirical methods, of “optimal” configurations of compression systems remains challenging. This is due, in part, to the large number of design parameters (with associated high computational cost) operating over wide ranges that can be non-smooth, if not discontinuous (to which many optimization algorithms, developed for smooth problems, are ill-suited). It is further complicated by the phasic nature of turbomachinery design and the associated need to balance the amount of time and computational resource devoted to selecting the most promising configurations with that expended in their refinement. This paper compares a number of combinations of a multi-fidelity approach for configuration selection with a high-fidelity method for design refinement. The system is tested on the aerodynamic design of a complete two-spool core compression system for a generic high bypass ratio turbofan. The resulting designs are obliged to meet familiar constraints for overall design point pressure rise and surge margin together with a number of mechanical constraints including maximum shaft speeds. Through the configuration phase, the number of stages and the duty split between the spools are permitted to change. It is shown that the performance of the design refinement phase is only a weak function of the preceding configuration phase provided that the latter is well into diminishing returns with respect to approaching a converged solution. It is hence shown possible to obtain equally good designs in around half the computational run-time by exploiting this weak dependence by effectively decoupling the configuration and refinement phases and starting the latter before the former has apparently finished. It is also shown that if either configuration or refinement is allowed to dominate the design process, inferior designs result. The best designs are associated with between half and three-quarters of the design effort being devoted to configuration selection.

Topics: Design , Compression
Commentary by Dr. Valentin Fuster
2014;():V02BT39A029. doi:10.1115/GT2014-26504.

Conjugate heat transfer is a key feature of modern gas turbine, as cooling technology is widely applied to improve the turbine inlet temperature for high efficiency. Impact of conjugate heat transfer on heat loads and thermodynamic efficiency is a key issue in gas turbine design. This paper presented a through flow calculation method to predict the impact of heat transfer on the design process of a convective cooled turbine. A cooling model was applied in the through flow calculations to predict the coolant requirements, as well as a one-dimensional mixing model to evaluate some key parameters such as pressure losses, deviation angles and velocity triangles because of the injection cooling air. Numerical simulations were performed for verification of the method and investigation on conjugate heat transfer within the blades. By comparing these two calculations, it is shown that the through flow calculation method is a useful tool for the blade design of convective cooled turbines because of its simplicity and flexibility.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A030. doi:10.1115/GT2014-26515.

This paper presents an in-house CFD package and its derivations for turbomachinery flow simulation as well as aerodynamic design optimization which have been employed in several primary Chinese aero-engine institutions. The package contains TurboMesh, a highly automated mesh generation code for turbomachinery, and MAP, a CFD solver for general purpose. Besides the programming strategies, the numerical schemes, and the parallelization methods adopted in the codes are outlined. Emphasis is placed on a novel mixing plane model used in MAP. The proposed model satisfies flux conservation property and very robust in actual usage. Additionally, further improved performance of the model can be achieved by applying a technique similar to the perfect matching layer for non-reflecting boundary conditions. On the basis of MAP, a set of derivations have also been developed. They include several versions of MAP which are based on specific flow models, respectively, an inverse code for the design optimization of 3D blade shape, an improved-delayed-detached-eddy-simulation based code, and a scale-adaptive-simulation based code. Some of these versions of MAP are briefly introduced and demonstrated through a few examples except for the inverse code in which a direct method proposed by the author is explained with a few more words. Illustrations show the applicability of the inverse code for the design of compressor blades in practical multi-blade row environment. By embedding MAP with an in-house numerical optimization package, the numerical optimization of the 2D/3D blade shape can be realized. Some examples for 3D aerodynamic optimizations of compressors are presented.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A031. doi:10.1115/GT2014-26604.

This paper develops the discrete adjoint equations for a turbomachinery RANS solver and proposes a framework for fully-automatic gradient-based constrained aerodynamic shape optimization in a multistage turbomachinery environment. The systematic approach for the development of the discrete adjoint solver is discussed. Special emphasis is put on the development of the turbomachinery specific features of the adjoint solver, i.e. on the derivation of flow-consistent adjoint inlet/outlet boundary conditions and, to allow for a concurrent rotor/stator optimization and stage coupling, on the development of an exact adjoint counterpart to the non-reflective, conservative mixing-plane formulation used in the flow solver. The adjoint solver is validated by comparing its sensitivities with finite difference gradients obtained from the flow solver. A sequential quadratic programming algorithm is utilized to determine an improved blade shape based on the objective function gradient provided by the adjoint solution. The functionality of the proposed optimization method is demonstrated by the redesign of a single-stage transonic compressor. The objective is to maximize the isentropic efficiency while constraining the mass flow rate and the total pressure ratio.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A032. doi:10.1115/GT2014-26707.

Two applications of the non-linear eddy-viscosity model EARSM are presented in the simulations of transonic turbulent flow involving shock waves and other related complex features. The simulations are implemented applying an in-house CFD program based on the unstructured discontinuous Galerkin method, an alternative discretization method of the classical finite volume one to precisely capture the flow features. A series of turbulence feature variables in boundary layers are comparatively observed and analyzed. For the first case of transonic flow over a bump, the redistribution effect of Reynolds stress components rooted in the non-linear constitution relation promotes streamwise turbulence fluctuation and suppresses the normal one in boundary layer, comparing with the traditional linear constitution relation, especially when passing the shocks. The production magnitudes of the turbulence shear stress and kinetic energy for the non-linear model show slightly more sensitive to perturbations, such as the occurrence of shock front or compression corner, than the linear one. For the second case of a transonic turbine vane, similar redistribution effect of the non-linear model is also verified on suction surface around the strong shock. The straightforward redistribution effect is absent on pressure surface around middle part of the vane with favorable pressure gradients. There the non-linear model evaluates higher magnitudes of streamwise, normal and shear Reynolds stress components than the linear one, thus resulting locally stronger heat convection and higher surface temperature.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A033. doi:10.1115/GT2014-26735.

In an effort to provide accurate simulations of fluid-structure interactions in turbomachinery, this paper describes a powerful method to deform mesh, using interpolation based on radial basis functions (RBF). It has been assessed on a 3D annular turbine, including a tip gap. The main difficulty of this method is to define number and position of control points. A greedy algorithm is proposed to address this issue and is tested on the annular turbine and a deforming panel placed in a shock tube. Finally, the method is slightly adapted to take into account periodic boundary conditions, which allow mesh morphing for a unique interblade channel by preserving constant pitch on lateral boundaries.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A034. doi:10.1115/GT2014-26739.

An implicit time integration, high-order discontinuous Galerkin method is assessed on the DNS of the flow in the T106C cascade at low Reynolds number. This code, aimed at providing high orders of accuracy on unstructured meshes for DNS and LES simulations on industrial geometries, was previously successfully assessed on fundamental, academic test cases. The computational results are compared to the experimental values and literature, and the obtained flow field characteristics are discussed. Although adequate resolution is supposed to be attained, discrepancies with respect to the experiment are found. These differences were furthermore consistently found by all authors in the workshop on high-order methods for CFD. The origins are therefore conjectured to result from insufficient adequation between computational setup and experiments, as no modeling is assumed. A plan for further investigation is proposed.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A035. doi:10.1115/GT2014-26762.

In order to increase the performance of a modern gas turbine, compressors are required to provide higher pressure ratio and avoid incurring higher losses. The tandem aerofoil has the potential to achieve a higher blade loading in combination with lower losses compared to single vanes. The main reason for this is due to the fact that a new boundary layer is generated on the second blade surface and the turning can be achieved with smaller separation occurring. The lift split between the two vanes with respect to the overall turning is an important design choice.

In this paper an automated three-dimensional optimisation of a highly loaded compressor stator is presented. For optimisation a novel methodology based on the Multipoint Approximation Method (MAM) is used. MAM makes use of an automatic design of experiments, response surface modelling and a trust region to represent the design space. The CFD solutions are obtained with the high-fidelity 3D Navier-Stokes solver HYDRA. In order to increase the stage performance the 3D shape of the tandem vane is modified changing both the front and rear aerofoils. Moreover the relative location of the two aerofoils is controlled modifying the axial and tangential relative positions. It is shown that the novel optimisation methodology is able to cope with a large number of design parameters and produce designs which performs better than its single vane counterpart in terms of efficiency and numerical stall margin.

One of the key challenges in producing an automatic optimisation process has been the automatic generation of high-fidelity computational meshes. The multi block-structured, high-fidelity meshing tool PADRAM is enhanced to cope with the tandem blade topologies. The wakes of each aerofoil is properly resolved and the interaction and the mixing of the front aerofoil wake and the second tandem vane are adequately resolved.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A036. doi:10.1115/GT2014-26776.

Based on flat plate results, mean velocity and friction coefficient estimation methods are proposed for rough surface turbulent boundary layers on axial compressor and turbine blades. The ratio of the displacement thickness to boundary layer thickness (δ*/δ) was first suggested by Zagarola and Smits (1998) for smooth pipe flows. The same parameter is proposed in this paper to scale the normalized mean velocity defect of smooth and rough surface flat plate turbulent boundary layers with zero, favorable, and adverse pressure gradients. The available mean velocity defect profiles of smooth and rough surface boundary layers from axial compressor and turbine blades are also scaled and compared to the flat plate results. Irrespective of the Reynolds number (Reθ), pressure gradient (K), and roughness (k), δ*/δ provides appropriate scaling for collapsing the flat plate and turbomachinery data. From the results, a new one-variable power law based on δ*/δ is proposed to estimate the mean velocity profile. The proposed power law can accurately estimate boundary layers on flat plates, compressor blades, and turbine blades. Finally, a new empirical Cf correlation is proposed for rough surface turbulent boundary layers under pressure gradients. The proposed Cf correlation is based on that of Bergstrom et al. (2005) and newly incorporates the acceleration parameter K. It can accurately estimate Cf in turbulent boundary layers of rough surface flat plates as well as those of smooth turbine blades.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A037. doi:10.1115/GT2014-26868.

The leading technology of the Siemens SGT5-4000F heavy duty gas turbine is demonstrated through recent engine measurements at customer sites. The 4000F fleet comprises more than 240 engines that exceed performance targets because of well established, non-heuristic design processes that are continuously enhanced.

On-site measurements of pressure, temperature and flow were performed at various locations in the flow path of the 4-stage turbine to support validation of Siemens’s proprietary 3D Computational Fluid Dynamics (CFD) analysis suite. Despite inherent but well understood modeling deficiencies, these advanced CFD prediction capabilities surpassed other tools and increased the accuracy of the overall turbine design process.

The performance of each turbine stage and flow features related to the exhaust diffuser were captured by the calculations. Overall performance characteristic shapes coincided with heat balances based on measurements. Radial traverses at the turbine exit, static pressure along the engine axis, and temperature sensors were matched well. The level of accuracy in delta predictions exceeded industry standards.

The design suite was able to predict performance parameters prior to measurement within respective confidence levels. Therefore, this advanced 3D CFD design suite, validated with the test data, will form the basis for future turbine development programs.

Topics: Design
Commentary by Dr. Valentin Fuster
2014;():V02BT39A038. doi:10.1115/GT2014-26892.

Numerical simulations were performed of experiments from a cascade of stator blades at three low Reynolds numbers representative of flight conditions. Solutions were assessed by comparing blade surface pressures, velocity and turbulence intensity along blade normals at several stations along the suction surface and in the wake. At Re = 210,000 and 380,000 the laminar boundary layer over the suction surface separates and reattaches with significant turbulence fluctuations. A new 3-equation transition model, the k-kL-ω model, was used to simulate this flow. Predicted locations of the separation bubble, and profiles of velocity and turbulence fluctuations on blade-normal lines at various stations along the blade were found to be quite close to measurements. Suction surface pressure distributions were not as close at the lower Re. The solution with the standard k-ω SST model showed significant differences in all quantities. At Re = 640,000 transition occurs earlier and it is a turbulent boundary layer that separates near the trailing edge. The solution with the Reynolds stress model was found to be quite close to the experiment in the separated region also, unlike the k-ω SST solution. Three-dimensional computations were performed at Re = 380,000 and 640,000. In both cases there were no significant differences between the midspan solution from 3D computations and the 2D solutions. However, the 3D solutions exhibited flow features observed in the experiments — the nearly 2D structure of the flow over most of the span at 380,000 and the spanwise growth of corner vortices from the endwall at 640,000.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A039. doi:10.1115/GT2014-26958.

The rotation-curvature correction proposed by Spalart and Shur is implemented in the Display FormulaγRe~θt transition model of Langtry and Menter. The correction term modifies the turbulence production such that it is damped in convex curvatures and enhanced otherwise. The curvature corrected transition model is first validated on a U-shaped channel flow for which experimental data and reference numerical results are available. The improved prediction capability is then assessed on a series of well documented two-dimensional turbomachinery problems. A better agreement with experimental data is observed in the simulation of the transition onset, which leads to an improved estimation of parameters of practical interest such as the heat transfer and pressure coefficient.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A040. doi:10.1115/GT2014-26976.

A robust mixing plane method satisfying interface flux conservation, non-reflectivity and retaining interface flow variation; valid at all Mach numbers and applicable for any machine configuration is formulated and implemented in a vertex based finite volume solver for flow analysis and inverse design of turbomachinery stage configurations. The formulation is based on superposing perturbed flow variables in the form of 3D characteristics obtained along the flow direction on the exchanged mixed out average quantities at the stage interface. A condition is derived in the mixed-out averaging procedure to distinguish between the subsonic and supersonic flow conditions at the interface. Using preconditioning technique, the new functionality is demonstrated to be applicable for a wide range of interface conditions and over different machine configurations with small spatial gap across the blade rows. The method is shown to satisfy flux conservation across the interface without generating spurious oscillations in the flow field at the domain boundaries and validated against available commercial solvers.

Subsequently, a blade re-design approach in a multi-row configuration is conceptualised and demonstrated by the application of the 3D inverse design method on a single stage Low Pressure Turbine. Meridional load variation, stage reaction and blade stacking angle are considered as the design variables to explore the design space. Conducting design runs at a fixed mass flow boundary condition and similar overall loading condition; the optimised configuration is shown to satisfy redistributed meridional load, providing performance improvement while maintaining a similar level of flow rate and work extraction as the baseline configuration.

Topics: Design , Turbines
Commentary by Dr. Valentin Fuster
2014;():V02BT39A041. doi:10.1115/GT2014-27019.

The primary focus of this work is to validate a CFD model intended to be used for transonic compressor design purposes. This design model includes a coarse grid using wall functions and mixing planes at interfaces connecting the compressor components. The computations are compared with experimental data from the transonic highly loaded 1.5 stage compressor test case Hulda. Additional comparisons are done with higher complexity CFD models accounting for the rotor-stator interaction. The performance of Hulda has been measured with both a small and a large tip clearance. These two configurations are used to investigate the necessity of resolving the tip clearance gap in the design model. The comparison is presented in terms of the overall performance at two rotational speeds as well as radial distribution of total pressure and total temperature at stations downstream of the rotor. The predictive capability at these speeds is assessed in terms of mass flow, pressure ratio and efficiency. Furthermore, the response of the predicted radial flow distributions with respect to the throttle setting along the two rotational speeds is qualitatively compared with the measurements.

The validation of the small tip clearance test shows that the design model, with or without tip gap modeling, is in good agreement with the measurements at both speeds. As for the large tip clearance test a design model resolving the tip clearance was able to predict trends but the penalty related to the increased tip gap was overestimated compared to the measured.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A042. doi:10.1115/GT2014-27026.

An approach for estimation of the turbulence length scale at the inflow boundary is proposed and presented. This estimation yields reasonable turbulence decay, supporting the transition model in accurately predicting the laminar-turbulent transition location and development. As an additional element of the approach, the sensitivity of the turbulence model to free-stream values is suppressed by limiting the eddy viscosity in non-viscous regions. Therefore the well known realizability constraint after Durbin [1] is modified. The method is implemented in DLR’s turbomachinery flow solver TRACE in the framework of the k–ω turbulence model by Wilcox [2] and the γ–Reθ transition model by Langtry and Menter [3]. The improved model is tested to the T106A turbine testcase and validated at the T161 turbine cascade under low speed conditions and T170 turbine cascade at high speed conditions.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A043. doi:10.1115/GT2014-27064.

An adjoint-based shape optimization approach for supersonic turbine cascade is proposed. The algorithm is based on a discrete adjoint method, state-of-the-art parametrization techniques (NURBS) and a preconditioned steepest descent optimizer to search the optimal point. The potential of the optimization approach is verified on two different design problems. Initially the design methodology is applied to the re-design of an existing supersonic turbine cascade operating at nominal conditions, with the aim of obtaining a more uniform flow at blade outlet section. Then, an original extension of the algorithm for treating off-design conditions is envisaged. The method combines a standard multi-point optimization technique with an uncertainty quantification algorithm to assess the design points and the weights of the multi-point problem. The capability of the novel approach in providing robust designs is finally investigated by maximizing the performances of the same baseline configuration working under a relatively wide range of operating conditions. In both tests remarkable outcomes are achieved in terms of improvement of blade performances and computational efficiency.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A044. doi:10.1115/GT2014-27127.

Generation of the grid for blade passages with packaging using universal grid generators usually takes much time. The paper is devoted to grid generation in turbo machine blade passages with packaging in automatic mode. The main requirement to the approach is to obtain the grid with minimum engineer participant. In the developed procedure engineer must specify only general input data: number of nodes, cell size near solid bodies and geometrical data.

Multiblock structured grids are considered. All grid blocks have node-to-node attachment between each other; periodicity is also specified from node to node. The grid in blade passage consists of two blocks: “O” grid around blade and “H” grid in blade passage. Additional blocks are used to describe different ZR-effects such us tip clearances, leakage seals and bleed air systems.

A variational method of constructing three dimensional grids composed of hexahedral cells is applied. The combination of the energy density functional and cell size functional is used. The first functional lets us control the shapes and the second functional lets us control the sizes of grid cells. Grid untangling procedure is also developed.

Developed approach was tested using the blades of axial and centrifugal compressors and axial turbines. Results of grid generation are presented.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A045. doi:10.1115/GT2014-27219.

Numerical prediction of the Stage 67 transonic fan stage employing wall jet tip injection flow control and study of the physical mechanisms leading to stall suppression and stability enhancement afforded by endwall recirculation/injection is the focus of this paper. Reynolds averaged Navier-Stokes computations were used to perform detailed analysis of the Stage 67 configuration experimentally tested at NASA’s Glenn Research Center in 2004. Time varying predictions of the stage plus recirculation and injection flowpath were executed utilizing the Nonlinear Harmonic approach. Significantly higher grid resolution per passage was achieved than what has been generally employed in prior reported numerical studies of spike stall phenomena in transonic compressors. This paper focuses on characterizing the physics of spike stall embryonic stage phenomena and the influence of tip injection, resulting in experimentally and numerically demonstrated stall suppression.

Commentary by Dr. Valentin Fuster
2014;():V02BT39A046. doi:10.1115/GT2014-27239.

Film cooling is a very effective cooling method for protecting the turbine blades exposed to hot gas from the heat. Since its cooling effectiveness is highly dependent on the shape of the hole, a wide variety of concepts and design parameters regarding hole shapes have been researched. However, there are no well-defined ways to determine the optimum shape of a film cooling hole.

The CFD is a powerful tool for film cooling hole optimization. But with the number of parameters that define the film cooling hole shapes being so numerous, analytical optimization with CFD often requires computational resources that are unrealistic for the average design environment. Accordingly, for CFD to be effective in the optimization process, it is necessary to reduce the number of computations or shorten the calculation time per computation.

In order to solve this problem, this paper presents a novel approach of applying 3D-POD (3D-Proper Orthogonal Decomposition) to the optimization of film cooling holes. POD is one of the most important component analysis methods and has the potential to reduce the number of parameters.

From the computation results, a solution group was made by the RSM (Response Surface Method) and assessment functions, i.e., film cooling effectiveness, heat transfer coefficient, mixing loss, concentration of stress and robustness were considered first. In the end, however, considering the sensitivity of each objective function, the optimal hole shapes were obtained with only the film effectiveness being evaluated.

In the following sections, this method and its results are described in detail.

Commentary by Dr. Valentin Fuster

Multidisciplinary Design Approaches, Optimization, and Uncertainty Quantification (With Structures and Dynamics and Heat Transfer)

2014;():V02BT45A001. doi:10.1115/GT2014-25081.

This paper presents a fully automated procedure to estimate the uncertainty of compressor stage performance, due to impeller manufacturing variability. The methodology was originally developed for 2D stages, i.e., stages for which the impeller blade angle and thickness distribution are only defined at the hub end-wall. Here, we extend the procedure to general 3D stages, for which blade angle and thickness distributions can be prescribed independently at the shroud and hub endwalls. Starting from the probability distribution of the impeller geometrical parameters, 3D sample geometries are generated and 1D/2D aerodynamic models are created, which are used to predict the performance of each sample geometry. The original procedure used the Monte Carlo method to propagate uncertainty. However, this requires a large number of samples to compute accurate performance statistics. Here we compare the results from Monte Carlo, with those obtained using Sparse Grid Polynomial Chaos Expansion (PCE) and a Multidimensional Cubature Rule for uncertainty propagation. PCE has exponential convergence in the stochastic space for smooth functions, and the use of sparse grids mitigates the increase of sample points due to the increase in the number of uncertain parameters. The cubature rule has accuracy limitations, but sample points increase only linearly with the number of parameters. For a 3D stage, the probability distributions of the performance characteristics are computed, as well as the sensitivity to the design parameters. The results show that PCE and Multidimensional Cubature give similar results to MC computations, with a much lower computational effort.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A002. doi:10.1115/GT2014-25184.

Traditional multi-fidelity surrogate models require that the output of the low fidelity model be reasonably well correlated with the high fidelity model and will only predict scalar responses. The following paper explores the potential of a novel multi-fidelity surrogate modelling scheme employing Gappy Proper Orthogonal Decomposition (G-POD) which is demonstrated to accurately predict the response of the entire computational domain thus improving optimization and uncertainty quantification performance over both traditional single and multi-fidelity surrogate modelling schemes.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A003. doi:10.1115/GT2014-25293.

Although game-theoretical models to study social and economic problems have existed for a long time, they have been sparsely used for the design of engineering systems. This is due to the significant theoretical hurdles posed by game formulations for real engineering environments /problems. In this study we show our first attempt at adapting the frame-work of game-theoretical models for engineering problems, in particular the aero-mechanical optimization of a notional turbine blade. We pose the design problem as a series of games, starting with the determination of the Pareto front, the non-cooperative (disagreement) point and the optimal solution as the tangent intersection of the Pareto front and contours of the overall system objective. We present gradient-based algorithms that determine the Pareto front, the non-cooperative solution and the tangent solution. The solution to this series of games provides the basis of a new equilibrium concept namely, System Optimal Cooperative Solution (SOCS), which is the central theme of this paper. Finally we compare the SOCS solution against other cooperative solutions like Nash-Bargaining [1]. The results of this study show that in engineering environments previously known cooperative solutions like Nash-Bargaining and Kalai-Smordinsky [2] are not that important while the notion of a System Optimal Cooperative Solution, SOCS, is the equilibrium solution of relevance. For the particular example we consider, the SOCS is shown to be more favoring the aerodynamic performance when compared against the Nash-Bargaining equilibrium solution.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A004. doi:10.1115/GT2014-25385.

The development of Supercritical CO2 (S-CO2) power cycles is currently a major focus of the engineering and scientific community. The reason for such a growing interest in this type of power can be explained by the significant benefits in size and efficiency of power cycles, which use S-CO2 as a working fluid, as compared to conventional steam power generation. Many areas of application such as nuclear, solar, waste heat, energy storage, and clean coal combustion, are being studied for S-CO2 power production. Most of the publications discussing S-CO2 are concentrated on optimization of the cycle’s thermodynamic characteristics, topping and bottoming and have been conceptualized based on the heat source. At the same time, numerous aspects of turbomachinery design are often overlooked or are not well understood. This article discusses some specific engineering aspects of the design of turbine flow path which uses S-CO2 as a working fluid. The following design options have been studied to determine the best turbine configuration: number of stages, rotational speed, impulse versus reaction, types of stages, and radial clearance influence. The effect of larger bending loads, resulting from high power density on nozzles and blade chords size and, consequently, turbine length, has also been studied. The authors hope that the results presented in the article will help the engineering community design better S-CO2 turbomachinery.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A005. doi:10.1115/GT2014-25495.

This paper presents a multi-objective and multi-disciplinary design optimization and data mining of gas turbine blade profile and cooling system by using conjugate heat transfer analysis. A 3D multi-disciplinary aerothermal optimization and data mining is proposed and developed by integrating the global optimization method of self-adaptive multi-objective differential evolution (SMODE) algorithm based on constraint-handling method, the CHT method for aerothermal performance evaluation of gas turbine blade, the 3D blade parameterization method and the self-organization map (SOM) based data mining technique. Using CHT, a numerical investigation was carried out to evaluate the aerothermal performance of C3X model, which consists of the blade passage, the blade solid domain and the internal coolant flow passages. The results calculated by the CHT method were validated by the experimental results. A new parameterization method for modeling the blade profile and cooling system has been developed. The optimization is intended to minimize the maximum blade temperature and the temperature gradient with constraints on the coolant mass flow rate, total mass flow rate and total pressure recovery coefficient of the blade. 27 Pareto solutions are obtained after the multidisciplinary design optimization for the gas turbine blade. Detailed aerothermal analysis shows that the thermal performance of the blade is significantly improved without deteriorating the related aerodynamic performance, thereby the correctness and effectiveness of our proposed optimization method are demonstrated. The SOM-based data mining on optimization design space is also applied to explore the trade-off relations between objective functions and correlations among design variables and objective function.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A006. doi:10.1115/GT2014-25787.

In order to meet the requirements of rising energy demand, one goal in the design process of modern steam turbines is to achieve high efficiencies. A major gain in efficiency is expected from the optimization of the last stage and the subsequent diffuser of a low pressure turbine (LP). The aim of such optimization is to minimize the losses due to separations or inefficient blade or diffuser design. In the usual design process, as is state of the art in the industry, the last stage of the LP and the diffuser is designed and optimized sequentially. The potential physical coupling effects are not considered. Therefore the aim of this paper is to perform both a sequential and coupled optimization of a low pressure steam turbine followed by an axial radial diffuser and subsequently to compare results. In addition to the flow simulation, mechanical and modal analysis is also carried out in order to satisfy the constraints regarding the natural frequencies and stresses. This permits the use of a meta-model, which allows very time efficient three dimensional (3D) calculations to account for all flow field effects.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A007. doi:10.1115/GT2014-25795.

The manufacturing processes used to create compressor blades inevitably introduce geometric variability to the blade surface. In addition to increasing the performance variability, it has been observed that introducing geometric variability tends to reduce the mean performance of compressor blades. For example, the mean adiabatic efficiency observed in compressor blades with geometric variability is typically lower than the efficiency in the absence of variability. This “mean-shift” in performance leads to increased operating costs over the life of the compressor blade. These detrimental effects can be reduced by using robust optimization techniques to optimize the blade geometry. The impact of geometric variability can also be reduced by imposing stricter tolerances, thereby directly reducing the allowable level of variability. However, imposing stricter manufacturing tolerances increases the cost of manufacturing. Thus, the blade design and tolerances must be chosen with both performance and manufacturing cost in mind.

This paper presents a computational framework for performing simultaneous robust design and tolerancing of compressor blades subject to manufacturing variability. The manufacturing variability is modelled as a Gaussian random field with non-stationary variance to simulate the effects of spatially varying manufacturing tolerances. The statistical performance of the compressor blade system is evaluated using the Monte Carlo method. A gradient based optimization scheme is used to determine the optimal blade geometry and distribution of manufacturing tolerances.

Topics: Compressors , Design , Blades
Commentary by Dr. Valentin Fuster
2014;():V02BT45A008. doi:10.1115/GT2014-25854.

Given the ever increasing demands on turbomachinery performance, various advanced blade shape optimizations have been actively developed and applied in modern blading designs. Multidisciplinary and concurrent optimizations have attracted considerable attention, offering the advantage of disciplinary interactions being included more simultaneously in a design process.

This paper presents the development of a multidisciplinary optimization algorithm for the concurrent blade aerodynamic and aeromechanic shape optimization of realistic 3D turbine stages. A non-gradient algorithm is enhanced by a new re-scaled response surface (RSM) model. This meta-model is able to rescale the design space and redefine the response surface during a blade shape optimization process, leading to a much enhanced convergence compared to a standard RSM approach. The optimization algorithm is developed in conjunction with an efficient nonlinear harmonic phase solution method solving the unsteady flow equations in the frequency domain, combined with a finite element analysis (FEA) to extract the structural dynamic characteristics of the blades.

The effectiveness of the concurrent method is examined for an optimized design of a realistic LP turbine stage. The optimization goals are the maximization of the isentropic stage efficiency and aeroelastic flutter stability (aero-damping). Two sets of cases are considered. In the first set, the shaping is applied only to stator blades, while for the second set, both stator and rotor blades are shaped. The concurrent cases are compared with their single-disciplinary counterparts. For both sets of the cases, the advantages of the concurrent treatment are clearly demonstrated.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A009. doi:10.1115/GT2014-25858.

The design of an aero-engine is traditionally divided into three levels: conceptual design, preliminary design and detailed design. This three-step design process is inherently iterative, which can slow the design process and overall productivity. Additionally, as an integrated systems engineering analysis, aero-engine design involves multiple-disciplines. The complex coupled-relationship among multiple-disciplines and multiple-components gives rise to severe conflict with performance requirements when designing, especially when it comes to high-performance aero-engine. Traditionally, designers need to empirically balance all kinds of requirements, which lead to a longer design cycle. So it is necessary to apply Multidisciplinary Design Optimization (MDO) to organize and manage the process of design system which sufficiently utilizes the effect of interaction of multidisciplines for the optimal solution. The MDO of a turbine flow path is one of the key multidisciplinary optimization technologies in aeroengine overall design. The problem studied and presented in this paper consists in optimizing a turbine modeled by a multidisciplinary system of two coupled disciplines: turbine aerodynamics and structural strength, with temperature limited by the materials. In the present work, three modules are established to conduct the MDO research of turbine flow path: flow path design, turbine strength calculation and MDO. The aeroengine turbine flow path, including high and low pressure turbine flow path, is designed in the first module, with its efficiency estimated. In the second module, turbine rotors consisting of blades, discs and the low spool shaft are parametric modeled so as to analyze the structural aspects of turbine rotors, such as weight and stresses. MDO is conducted using multi-island genetic algorithm optimization (MIGA) optimization algorithm provided in iSIGHT software. Fully Integrated Optimization (FIO) strategy is studied to deal with the multidisciplinary analysis. The complex coupling relations between aerodynamic performance and turbine strength are analyzed to establish turbine multidisciplinary optimization system. The optimal values of loading coefficient, rotational speed, bore diameter of rotor discs defined by the shaft size, and other independent design variables are obtained in order to achieve minimum weight of turbine rotors while simultaneously meeting the strength and aerodynamics efficiency requirements. This method presented in this paper can greatly shorten turbine design cycle, improve aeroengine design ability, and is prospective to be widely applied to engineering field.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A010. doi:10.1115/GT2014-25872.

The understanding of aeroelastic phenomena is fundamental for the structural integrity of many applications in aerospace and mechanical engineering and even in some other disciplines (e.g. civil engineering) where flexible structures possibly undergo unsteady fluid-dynamic loads. Therefore the availability of accurate analysis tools for the study of the aeroelastic interaction between aerodynamic and elastic forces is an important asset for the design of modern, high performance turbomachinery.

Together with the more and more powerful computing resources, current trends pursue the adoption of high-fidelity tools and state-of-the-art technology within the research fields of Computational Structural Dynamics (CSD) and Computational Fluid Dynamics (CFD). This choice is somehow obliged when dealing with highly non-linear aeroservoelastic phenomena.

The approach typically used for turbomachinery aeroelastic analysis features the so-called “one-way coupling”, i.e. the loads predicted by the aerodynamic model are transferred to the structural model to evaluate relevant stresses and displacements.

The objective of the present work is to illustrate the design and implementation of a platform for solving multidisciplinary non-linear Fluid-Structure Interaction (FSI) problems with a “two-way coupling” or fully coupled approach, that is linking together high-fidelity state-of-the-art CSD and CFD tools by means of a robust, flexible aeroelastic interface scheme.

The credibility of the proposed aeroservoelastic analysis toolbox is assessed by tackling a set of aeronautical and turbomachinery-oriented benchmark test problems such as: the evaluation of the fully coupled non-linear aeroelastic trim of HIRENASD (HIgh REynolds Number AeroStructural Dynamics) wing and the identification of the aerodynamic damping coefficient of Standard Configuration 10, high subsonic/transonic, 2D/3D compressor cascade. The results are compared with reference experimental and numerical data available in literature.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A011. doi:10.1115/GT2014-26014.

The following paper proposes an accelerated medial object transformation for the tip clearance optimisation of whole engine assemblies. A considerable reduction in medial object generation time has been achieved through two different mechanisms. Faces leading to unnecessary branches in the medial mesh are removed from the model and parallelisation of the medial object generation is improved through the subdivision of the original 3D CAD model. The time savings offered by these schemes are presented with respect to the generation of the medial objects of two complex gas turbine engine components. It is also demonstrated that the utilization of these techniques within a design optimisation may result in a considerable reduction in wall time.

Topics: Engines , Optimization
Commentary by Dr. Valentin Fuster
2014;():V02BT45A012. doi:10.1115/GT2014-26038.

This paper presents the application of a viscous adjoint method in the optimization of the endwall contour of a turning mid turbine frame (TMTF). The adjoint method is a gradient based optimization method that allows for the computation of the complete gradient information by solving the governing flow equations and their corresponding adjoint equations only once per function of interest (objective and constraints), so that the computation time of the optimization is nearly independent of the number of parameters. With the use of a greater number of parameters a more detailed definition of endwall contours is possible, so that an optimum can be approached more precisely. A Navier-Stokes flow solver coupled with Menter’s SST k–ω turbulence model is utilized for the CFD simulations, whereas the adjoint formulation is based on the constant eddy viscosity approximation for turbulence. The total pressure ratio is used as the objective function in the optimization. The effect of contouring on the secondary flows is evaluated and the performance of the axisymmetric TMTF is calculated and compared with the optimized design.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A013. doi:10.1115/GT2014-26167.

In order to improve product development cycle, design engineers use multi-disciplinary analysis tools which allow better productivity. This paper covers the development of new tools to improve the preliminary design phase of turbine disc, being a critical part of aircraft engines. First, a new single platform D&A (Design & Analysis) tool integrating commercial CAD (Computer Aided Design) and FEA (Finite Element Analysis) software processing in batch mode is presented. This integrated architecture leads to a real improvement enabling a cohesive single integrated simulation environment that offers significant time reduction on user manipulation and execution. An optimization of disc geometry is then performed by using different optimization algorithms and configurations for a given disc parameterized model. The results show potential improvement over the current preliminary rotor discs for life and burst limited design. Finally, optimal curves obtained by developing HPT (High Pressure Turbine) disc reference charts, indicate how to get the minimum weight for given mechanical performance without running any structural analysis. These new tools supporting disc design have allowed improvement of disc life and durability leading to reduction of preliminary design phase duration.

Topics: Design , Turbines , Disks
Commentary by Dr. Valentin Fuster
2014;():V02BT45A014. doi:10.1115/GT2014-26320.

Aero-mechanical design of the turbine section of a small scale turboshaft engine is presented in this paper. A single stage high-pressure turbine (HPT) and the power turbine stage (PT), have been designed by means of automated optimization. This study demonstrates how multi-disciplinary optimization can be used effectively in today’s industrial development cycles with respect to timeframe and computational resources. Both, the aerodynamic performance and the mechanical blade behavior were subject to the optimization in a very high dimensional design space expressed by well above 100 free design parameters for the annular duct and the bladings of two axial stages. In the first part, this paper describes the design task and constraints in order to meet the requested thermodynamic cycle performance and fabricational requirements. In the second part, the optimization strategy is explained with focus on geometry parameterization, simulation setups for flow and structural analysis and acceleration techniques for the optimization itself.

Finally, a very promising resulting design is reviewed in terms of a detailed aerodynamic and mechanical assessment and regarding to the overall engine concept. This work contributes to the development of a highly efficient, light-weight propulsion system powering, beside a wide range of other possible applications, for example small aerial systems such as helicopter drones.

The engine prototype is expected to be tested the first time in 2014.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A015. doi:10.1115/GT2014-26368.

Today’s UAV helicopter industry faces a lack of highly reliable, SFC optimized turboshaft engines in the 40kW to 100kW class, resulting in a significant drawback for the overall flight envelope and the system availability of these aircraft. This paper describes the design process for a turboshaft engine with a shaft power output of about 80kW. A thermodynamic cycle model is derived from the flight envelope of the Swiss UAV NEO S-350 helicopter drone. Different compressor configurations are analysed and discussed with regard to the power output and SFC of the engine as well as to manufactural constraints. Combining a high flow density with a high isentropic efficiency and pressure ratio, a three stage compressor configuration was selected. The design is based on two axial front stages with a total pressure ratio of 1.55 and 1.45, respectively, and a diagonal last stage with a total pressure ratio of 2.8. Finally, the aero-mechanical design and optimization process of the compressor is depicted and the manufactural process is described. The engine prototype is expected to be tested the first time in 2014.

Topics: Engines , Compressors , Design
Commentary by Dr. Valentin Fuster
2014;():V02BT45A016. doi:10.1115/GT2014-26431.

This paper presents numerical optimization of a compressor rotor, to deepen the knowledge of endwall flow in the large-scale axial subsonic compressor, accordingly reduce its endwall loss and improve its aerodynamic performance. With numerical simulation and numerical optimization tools, three-dimensional stacking principle is optimized to improve the design operation point performance for the rotor. Results show that, hub region of the rotor cannot undertake large blade loading; compared to the prototype rotor, obvious aerodynamic performance improvements locate near the hub area, and a certain degree of positive dihedral in this region effectively helps to reduce its flow loss. The effect of “loaded leading edge and unloaded trailing edge” due to positive dihedral was shown, which suppresses flow separation near the trailing edge, consequently obviously reduces the flow loss and largely improves the rotor aerodynamic performance.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A017. doi:10.1115/GT2014-26465.

In this paper a 3D optimization process for the aerodynamic design of a centrifugal multistage compressor stage is presented using the commercial software NUMECA/Design3D. The exercise, starting from a given single stage configuration with vaned diffuser, consists in the automatic design of the vaned diffuser with a return channel in order to obtain a repetitive compressor stage for multistage compressor architecture. The design process uses the meta-model approach coupled to 3D Navier-Stokes simulations; the optimization algorithm drives the automatic design process to minimize a prescribed goal function. Only the design condition has been considered and the effect of the starting DoE population on the optimization process has been investigated. In the paper a critical analysis on the design process using the meta-model approach is performed and the most crucial issues to set up a best practice for the effective use of such technology are discussed in some detail.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A018. doi:10.1115/GT2014-26490.

Gas-liquid two-phase flows are encountered in a variety of applications such as turbo-machinery flows, gas-turbines, ram-jet and scram-jets, automotive engines and aircraft engines. Designing systems to control such flows is enormously challenging owing to the addition of new non-dimensional groups that characterize the two-phase flow system compared to a single-phase flow. Additionally, two-phase flows can exhibit non-linear hydrodynamic instabilities that determine the overall behavior of the system.

In this study, we choose a generic two-phase flow configuration that exhibits known complexities in realistic two-phase flow systems. The goal of the study is to optimize the geometry of the two-phase flow configuration with minimal computational cost. We propose a probabilistic approach to model the stochastic system and optimize the two-phase flow system under uncertain inputs. The potential benefits of the approach are highlighted along with future directions for using probabilistic design techniques to optimize two-phase flow systems.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A019. doi:10.1115/GT2014-26571.

In simple optimization problem, direct searching methods are most accurate and practical enough. However, for more complicated problem which contains many design variables and demands high computational costs, surrogate model methods are recommendable instead of direct searching methods. In this case, surrogate models should have reliability for not only accuracy of the optimum value but also globalness of the solution. In this paper, the Kriging method was used to construct surrogate model for finding aerodynamically improved three dimensional single stage turbine. At first, nozzle was optimized coupled with base rotor blade. And then rotor was optimized with the optimized nozzle vane in order. Kriging method is well known for its good describability of nonlinear design space. For this reason, Kriging method is appropriate for describing the turbine design space, which has complicated physical phenomena and demands many design variables for finding optimum three dimensional blade shapes. To construct airfoil shape, Prichard topology was used. The blade was divided into 3 sections and each section has 9 design variables. Considering computational cost, some design variables were picked up by using sensitivity analysis. For selecting experimental point, D-optimal method, which scatters each experimental points to have maximum dispersion, was used. Model validation was done by comparing estimated values of random points by Kriging model with evaluated values by computation. The constructed surrogate model was refined repeatedly until it reaches convergence criteria, by supplying additional experimental points. When the surrogate model satisfies the reliability condition and developed enough, finding optimum point and its validation was followed by. If any variable was located on the boundary of design space, the design space was shifted in order to avoid the boundary of the design space. This process was also repeated until finding appropriate design space. As a result, the optimized design has more complicated blade shapes than that of the baseline design but has higher aerodynamic efficiency than the baseline turbine stage.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A020. doi:10.1115/GT2014-26683.

Typically, complex components are designed with a CAD system, where free-form surfaces are usually described by B-Spline or NURBS surfaces. Especially in turbo machinery, design is performed for hot working conditions. However, for manufacturing a cold, unloaded CAD model is required. A deformed component with smooth surfaces may also be needed e.g. for flow calculations at different operating points. Engine-part deformations caused by loads and temperature changes are calculated by a finite element program which produces a displacement field as output. In such cases the original B-Spline description may be used and the shape of the relevant surfaces may be adjusted via displacement of the B-Spline control-points. The procedure is to decompose the solid into part-surfaces, identify corresponding FE nodes, apply the FE displacement field to the projected points of the FE nodes, and determine displacements of the control points by solving linear fitting problems. The required position and tangential continuity can be achieved by special treatment of the edge curves and their neighboring control points. Treatment of singularities of the linear fitting problem at trimmed surfaces and inadequate FE meshes requires special focus. The presented generic morphing method was implemented as an extension to the CAD system Unigraphics NX 7.5 and can now be used as the basis for automated hot-to-cold geometry transformations. Applications to a compressor aerofoil and an exhaust mixer segment demonstrate the robustness and accuracy of the method.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A021. doi:10.1115/GT2014-26937.

Small scale wind turbines can meet a substantial part of the electricity demand of residential buildings and facilities in isolated areas. It is a curious fact, however, that for many of these systems the actual power output has been dramatically overestimated. This can be partially explained by the very high rated wind speeds at which the design power output applies. The current work depicts the pathway to an aerodynamically optimized design of a small scale horizontal axis wind turbine in the 1kW class, optimized for wind speeds between 3.5 m/s and 5.5 m/s, a typical range of the energetic average of urban wind speeds. The aerodynamic stability of the blade has been a particular focus leading to a nearly constant efficiency over a range of wind speeds. The rotating speed of the system is adjusted to the optimal tip speed ratio at wind speeds up to maximum power via active control of the aerodynamic torque of the rotor blades. This is realized by adapting the generator torque to the current wind speed guaranteeing optimal efficiency and power output. The rotor blade optimization has been conducted unconventionally, in a turbomachinery-inspired 3D-blade design optimization campaign, using high-fidelity compressible CFD. This approach is described in detail, focussing on geometry parametrization and the numerical model with reasonable boundary conditions. Finally, the aerodynamic performance of the rotor blade is assessed at different wind speeds and pitching angles.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A022. doi:10.1115/GT2014-26961.

Design of centrifugal compressors in different applications from industrial to turbochargers to aeroengine is subject to difficult multi-disciplinary ( aerodynamics and mechanical) and multipoint/multi-objective requirements. These multi-disciplinary and multi-point requirements have to be met by iterations between aerodynamics and mechanical design, leading to long development times and bottlenecks in the design process. In this paper, for the first time, a commercially available solution, compatible with industrial development times, is presented for 3D multi-disciplinary and multi-point design optimisation of turbomachinery blades. The methodology combines 3D inverse design method, automatic optimizers, 3D CFD and 3D FEA codes. The key aspect of the approach is to parameterise the 3D geometry through the blade loading distribution used in 3D inverse design code TURBOdesign1, which results in ability to access large part of design space with very few design parameters. The Design of Experiments method is used to generate a number of geometries which are then analysed by 3D CFD code STAR-CCM+ and 3D FEA code Abaqus. Different performance parameters related to aerodynamics (efficiency, stable operating range etc) and structural integrity (maximum principal stress, etc) are then evaluated. The data is then used to create a response surface. The validity and accuracy of the response surface is evaluated by CFD and FEA and then once confirmed a Multi-objective Genetic Algorithm is run on the response surface to explore the trade-offs between different design parameters, such as peak efficiency, stable operating range and mechanical stress. In this paper the methodology is applied to the redesign of the well-known Eckardt centrifugal compressor impeller.

Commentary by Dr. Valentin Fuster
2014;():V02BT45A023. doi:10.1115/GT2014-27229.

A new approach for adaptively sampling a design parameter space using an error estimate through the reconstruction of flow field by a combination of proper orthogonal decomposition (POD) and radial basis function network (RBFN) is presented. It differs from other similar approaches in that it does not use the reconstructed flow field by POD for the evaluation of objective functions, and thus it can be a subset of the flow field. Advantages of this approach include the ease of constructing a chain of simulation codes as well as the flexibility of choosing where and what to reconstruct within the solution domain. An improvement in achieving a good prediction quality, with respect to other adaptive sampling methods, has been demonstrated using supersonic impulse turbine optimization as the test case. A posteriori validation of the surrogate models were also carried out using a set of separately-evaluated samples, which showed a similar trend as the Leave-One-Out (LOO) cross-validation. The progressively enriched surrogate model was then used to achieve the more uniformly populated Pareto front with fewer number of function evaluations.

Commentary by Dr. Valentin Fuster

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In