Research Papers

J. Turbomach. 2017;139(11):111001-111001-10. doi:10.1115/1.4036764.

This article discusses the development, application, and validation of an optimization method for the impellers of axial fans. The method is supposed to be quick, accurate, and applicable to optimization at an extensive range of design points (DPs). Optimality here means highest possible total-to-static efficiency for a given design point and is obtained by an evolutionary algorithm in which the target function is evaluated by computational fluid dynamics (CFD)-trained artificial neural networks (ANN) of the multilayer perceptron (MLP) type. The MLPs were trained with steady-state CFD (i.e., Reynolds-averaged Navier–Stokes (RANS)) results of approximately 14,000 distinct impellers. After this considerable one-time effort to generate the CFD dataset, each new fan optimization can be performed within a few minutes. It is shown in this article that the MLPs are reliably applicable to all typical design points of axial fans according to Cordier's diagram. Moreover, an extension of the design space toward the classic realm of mixed-flow or even centrifugal fans is observed. It is also shown that the optimization method successfully handles geometrical and operational constraints proving the high degree of universality of the method. Another focus of this article is on the application of the newly developed optimization method to numerous design points. This yields two major findings: the estimation of maximum achievable total-to-static efficiency as a function of the targeted design point (with and without geometrical constraints) as well as a quantification of the improvement over fans designed with classic methods. Both investigations are supported by flow field analyses to aerodynamically explain the findings. Experimental validation of the method was performed with a total of nine prototypes. The positive correlation between MLP, CFD, and experiment successfully validates the methodology.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111002-111002-11. doi:10.1115/1.4037056.

Winglet tips are promising candidates for future high-pressure turbine rotors. Many studies found that the design of the suction-side winglet is the key to the aerodynamic performance of a winglet tip, but there is no general agreement on the exact design philosophy. In this paper, a novel suction-side winglet design philosophy in a turbine cascade is introduced. The winglets are obtained based on the near-tip flow field of the datum tip geometry. The suction-side winglet aims to reduce the tip leakage flow particularly in the front part of the blade passage. It is found that on the casing endwall, the pressure increases in the area where the winglet is used. This reduces the tip leakage flow in the front part of the blade passage and the pitchwise pressure gradient on the endwall. As a result, the size of the tip leakage vortex reduces. A surprising observation is that the novel optimized winglet tip design eliminates the passage vortex and results in a further increasing of the efficiency. The tip leakage loss of the novel winglet tip is 18.1% lower than the datum cavity tip, with an increase of tip surface area by only 19.3%. The spanwise deflection of the winglet due to the centrifugal force is small. The tip heat load of the winglet tip is 17.5% higher than that of the cavity tip. Numerical simulation shows that in a turbine stage, this winglet tip increases the turbine stage efficiency by 0.9% mainly by eliminating the loss caused by the passage vortex at a tip gap size of 1.4% chord compared with a cavity tip.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111003-111003-11. doi:10.1115/1.4037126.

This paper presents an experimental investigation of the concept of using the combustor transition duct wall to shield the nozzle guide vane leading edge. The new vane is tested in a high-speed experimental facility, demonstrating the improved aerodynamic and thermal performance of the shielded vane. The new design is shown to have a lower average total pressure loss than the original vane, and the heat transfer on the vane surface is overall reduced. The peak heat transfer on the vane leading edge–endwall junction is moved further upstream, to a region that can be effectively cooled as shown in previously published numerical studies. Experimental results under engine-representative inlet conditions showed that the better performance of the shielded vane is maintained under a variety of inlet conditions.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111004-111004-17. doi:10.1115/1.4037028.

A laboratory experimental method and an analysis technique are presented for evaluation of individual film-cooling row flow capacity characteristics. The method is particularly suited to complex systems such as hot section nozzle guide vanes (NGV) with lossy feed system characteristics. The method is believed to be both more accurate and more experimentally efficient than previous techniques. The new analysis technique uses an experimentally calibrated network model to represent the complex feed system and replaces the need for internal loss measurements, which are both demanding and inaccurate. Experiments are performed in the purpose-built University of Oxford Coolant Capacity Rig (CCR), a bench-top, blow-down type facility with atmospheric back-pressure. The design of the CCR is informed by the requirements to assess engine-scale film-cooled components rapidly, accurately, and precisely. Improvements in the experimental method include a differential mass flow rate measurement method (which eliminates the effect of leaks and minimizes the number of rows that must be blanked, ensuring that the internal coupling is as close as possible to the engine condition) and a variable bypass flow which ensures the mass flow measurement nozzle always operates within its calibrated range. We demonstrate the method using two high-pressure (HP) NGV designs: an engine part with relatively uncoupled (in terms of internal loss) cooling rows; and a laser-sintered part with highly coupled cooling rows. We show that the individual-row flow capacity of a high-pressure nozzle guide vane (HPNGV) can be evaluated in the CCR in a single day to a 2σ precision of approximately 0.5% and a 2σ accuracy (bias) of 0.6%. The importance of performing individual-row capacity measurements is demonstrated: failure to scale flow capacity on a row-by-row basis introduces an error of 30% in the engine situation.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111005-111005-11. doi:10.1115/1.4037029.

This paper describes the effects of coolant-to-mainstream density ratio and specific heat capacity flux ratio (the product of blowing ratio and specific heat capacity ratio) on the overall cooling effectiveness of high pressure (HP) turbine vanes. Experimental measurements have been conducted at correct engine-matched conditions of Mach number, Reynolds number, turbulence intensity, and coolant-to-mainstream momentum flux ratio. Vanes tested were fully cooled production parts from an engine currently in service. A foreign gas mixture of SF6 and Ar was selected for injection as coolant in the facility so that density and blowing ratios were also matched to the engine situation. The isentropic exponent of the foreign gas mixture coincides with that of air. Full-coverage surface maps of overall cooling effectiveness were acquired by an infrared (IR) thermography technique at a range of mainstream-to-coolant temperature ratios. Measurements were subsequently scaled to engine conditions by employing a new theory based on the principle of superposition and a recovery and redistribution temperature demonstrated in previous papers. It is shown that the two aerodynamically matched situations of air- and foreign-gas-cooled experiments give virtually the same effectiveness trends and patterns. Actual levels differ, however, on account of specific heat capacity flux ratio differences. The effect is described and quantified by a one-dimensional analytical model of the vane wall. Differences in Biot number with respect to engine conditions are discussed as they also influence the scaling of turbine metal temperatures.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111006-111006-8. doi:10.1115/1.4037127.

This paper will compare two approaches of sensitivity analysis, namely (i) the adjoint method which is used to obtain an initial estimate of the geometric sensitivity of the gas-washed surfaces to aerodynamic quantities of interest and (ii) a Monte Carlo type simulation with an efficient sampling strategy. For both approaches, the geometry is parameterized using a modified NACA parameterization. First, the sensitivity of those parameters is calculated using the linear (first-order) adjoint model. Since the effort of the adjoint computational fluid dynamics (CFD) solution is comparable to that of the initial flow CFD solution and the sensitivity calculation is simply a postprocessing step, this approach yields fast results. However, it relies on a linear model which may not be adequate to describe the relationship between relevant aerodynamic quantities and actual geometric shape variations for the derived amplitudes of shape variations. Second, in order to better capture nonlinear and interaction effects, a Monte Carlo type simulation with an efficient sampling strategy is used to carry out the sensitivity analysis. The sensitivities are expressed by means of the coefficient of importance (CoI), which is calculated based on modified polynomial regression and therefore able to describe relationships of higher order. The methods are applied to a typical high-pressure compressor (HPC) stage. The impact of a variable rotor geometry is calculated by three-dimensional (3D) CFD simulations using a steady Reynolds-averaged Navier–Stokes model. The geometric variability of the rotor is based on the analysis of a set of 400 blades which have been measured using high-precision 3D optical measurement techniques.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111007-111007-10. doi:10.1115/1.4037362.

This work newly proposes an uncertainty quantification (UQ) method named sparse approximation of moment-based arbitrary polynomial chaos (SAMBA PC) that offers a single solution to many current problems in turbomachinery applications. At the moment, every specific case is characterized by a variety of different input types such as histograms (from experimental data), normal probability density functions (PDFs) (design rules) or fat tailed PDFs (for rare events). Thus, the application of UQ requires the adaptation of ad hoc methods for each individual case. A second problem is that parametric PDFs have to be determined for all inputs. This is difficult if only few samples are available. In gas turbines, however, the collection of statistical information is difficult, expensive, and having scarce information is the norm. A third critical limitation is that if using nonintrusive polynomial chaos (NIPC) methods, the number of required simulations grows exponentially with increasing numbers of input uncertainties: the so-called “curse of dimensionality.” It is shown that the fitting of parametric PDFs to small data sets can lead to large bias and the direct use of the available data is more accurate. This is done by propagating uncertainty through several test functions and the computational fluid dynamics (CFD) simulation of a diffuser, highlighting the impact of different PDF fittings on the output. From the results, it is concluded that the direct propagation of the experimental data set is preferable to the fit of parametric distributions if data is scarce. Thus, the suggested method offers an alternative to the maximum entropy theorem to handle scarce data. SAMBA simplifies the mathematical procedure for many different input types by basing the polynomial expansion on moments. Its moment-based expansion automatically takes care of arbitrary combinations of different input data. It is also numerically efficient compared to other UQ implementations. The relationship between the number of random variables and number of simulation is linear (only 21 simulations for ten input random variables are required). It is shown in this paper that SAMBA's algorithm can propagate a high number of input distributions through a set of nonlinear analytic test functions. Doing this, the code needs a very small number of simulations and preserve a 5% error margin. SAMBA's flexibility to handle different forms of input distributions and a high number of input variables is shown on a low-pressure turbine (LPT) blade-based on H2 profile. The relative importance of manufacturing errors in different location of the blade is analyzed.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111008-111008-11. doi:10.1115/1.4037575.

Temporal variation of components' performance is becoming a crucial parameter in turbomachinery design process. The main physical mechanism driving the time-dependent behavior is the unsteady bladerow interaction as stator–rotor relative motion due to rotating frame of reference. However, so far unsteady effects have been ignored in design processes in common engineering practice. In fact, steady approach has been generally employed for computational fluid dynamics (CFD)-based turbomachinery design. Moreover, conventional blade design has been based on single operating point considerations. Taking into account multiple time-dependent phenomena, as the unsteady performance parameters variation, might be beneficial in making a further improvement on component performance. In quantitative terms, first of all it is important to investigate the relative effect of unsteady variation, compared to the standard steady approach, and to create a capability for calculating temporal sensitivity variation, while keeping a reasonable computing cost. This work investigates the unsteady variation of turbomachinery performance on quasi-three-dimensional (3D) geometries: single-stage turbine and single-stage compressor. Steady flow solutions using mixing plane method are compared to the unsteady flow solutions using a direct unsteady calculation, while assessing the introduction of the space–time gradient (STG) method. The results clearly show how the unsteady variation is a non-negligible effect in performance prediction and blade design. Then, a new computational technique to quantify temporal sensitivity variation is introduced, based on the STG method, with an extension to adjoint-based sensitivity analysis. The relation between time and space in multipassage-multirow domain, the fundamental assumption of the STG method, is applied within the adjoint operator formulation, which gives unsteady sensitivity information on a broad range of design parameters, at the cost of a single computation. Finally, the unsteady sensitivities are compared to the ones resulting from steady solution in the two quasi-3D cases. This work introduces a coherent and effective mathematical formulation for accounting deterministic unsteadiness on component design, while reducing computational cost compared to standard unsteady optimization techniques.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111009-111009-8. doi:10.1115/1.4037670.

A local, intermittency-function-based transition model was developed for the prediction of laminar-turbulent transitional flows with freestream turbulence intensity Tu at low (Tu < 1%), moderate (1% < Tu < 3%), and high Tu > 3% levels, and roughness effects in a broad range of industrial applications such as turbine and helicopter rotor blades, and in nature. There are many mechanisms (natural or bypass) that lead to transition. Surface roughness due to harsh working conditions could have great influence on transition. Accurately predicting both the onset location and length of transition has been persistently difficult. The current model is coupled with the k–ω Reynolds-averaged Navier–Stokes (RANS) model, that can be used for general computational fluid dynamics (CFD) purpose. It was validated on the ERCOFTAC experimental zero-pressure-gradient smooth flat plate boundary layer with both low and high leading-edge freestream turbulence intensities. Skin friction profiles agree well with the experimental data. The model was then tested on ERCOFTAC experimental flat plate boundary layer with favorable/adverse pressure gradients cases, periodic wakes, and flows over Stripf's turbine blades with roughness from hydraulically smooth to fully rough. The predicted skin friction and heat transfer properties by the current model agree well with the published experimental and numerical data.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111010-111010-9. doi:10.1115/1.4037492.

Natural convection is an important heat transfer mode for flexible operations of gas turbines and steam turbines. Its prediction presents considerable challenges. The strong interdependence between fluid and solid parts points to the need for coupled fluid–solid conjugate heat transfer (CHT) methods. The fundamental fluid–solid time scale disparity is further compounded by the long-time scales of practical turbine flexible operations. In addition, if a high-fidelity flow model (e.g., large eddy simulation (LES)) is adopted to resolve turbulence fluctuations, extra mesh dependency on solid domain mesh may arise. In this work, understanding of the extra solid mesh dependency in a directly coupled LES based CHT procedure is gained by an interface response analysis, leading to a simple and clear characterization of erroneously predicted unsteady interface temperatures and heat fluxes. A loosely coupled unsteady CHT procedure based on a multiscale methodology for solving problems with large time scale disparity is subsequently developed. A particular emphasis of this work is on efficient and accurate transient CHT solutions in conjunction with the turbulence eddy resolved modeling (LES) for natural convection. A two-scale flow decomposition associated with a corresponding time-step split is adopted. The resultant triple-timing formation of the flow equations can be solved efficiently for the fluid–solid coupled system with disparate time scales. The problem statement, analysis, and the solution methods will be presented with case studies to underline the issues of interest and to demonstrate the validity and effectiveness of the proposed methodology and implemented procedure.

Commentary by Dr. Valentin Fuster
J. Turbomach. 2017;139(11):111011-111011-10. doi:10.1115/1.4037758.

This paper investigates the impacts of uneven wall heating conditions under different buoyancy numbers on flow field and heat transfer performance of a rotating channel with one side smooth and one side roughened by 45 deg inclined ribs. Parametric Reynolds-averaged Navier–Stokes (RANS) simulations were conducted under two different wall heating conditions: only ribbed wall heated, as in experiment setup, and all walls heated, under three different buoyancy numbers. Results are compared, when available, with experimental results. Numerical results show that uneven wall heating has only a minor impact on nonrotating cases and very low buoyancy rotating cases. However, it has a significant influence, on both, the heat transfer behavior and the flow field, when the buoyancy number is large. In the ribbed trailing rotating tests, the all walls heated cases show significantly higher heat transfer rate than only the ribbed wall heated cases. The discrepancy is enlarged as buoyancy number increases. The heat transfer in the all walls heated case increases monotonically with the buoyancy number, whereas in the ribbed wall, heated case is slight reduced. In the ribbed leading rotating tests, the heat transfer sensitivity to the heating conditions is not conspicuous, and for both cases, the heat transfer level slightly reduced as the buoyancy number increased. The flow field investigation shows that there is a significant displacement of main flow in the all walls heated cases than only the ribbed wall heated cases under high buoyancy numbers. This displacement is due to the buoyancy effect and responsible for the heat transfer differences in uneven heating problems. According to the results obtained in the paper, we conclude that when buoyancy effects are relevant, the heating settings can play a significant role in the heat transfer mechanisms and therefore in the experimental and numerical results.

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