Variable pitch fans are of interest for future low-pressure ratio fan systems since they provide improved operability relative to fixed pitch fans. If they can also be re-pitched such that they generate sufficient reverse thrust they could eliminate the engine drag and weight penalty associated with bypass duct thrust reversers. This paper sets out to understand the details of the 3D fan stage flow field in reverse thrust operation. This study uses the Advanced Ducted Propulsor variable pitch fan test case, which has a design fan pressure ratio of 1.29. Comparison with spanwise probe measurements show that the computational approach is valid for examining the variation of loss and work in the rotor in forward thrust. The method is then extended to a reverse thrust configuration using an extended domain and appropriate boundary conditions. Computations, run at two rotor stagger settings, show that the spanwise variation in relative flow angle onto the rotor aligns poorly to the rotor inlet metal angle. This leads to two dominant rotor loss sources: one at the tip associated with positive incidence and the second caused by negative incidence at lower span fractions. The second loss is reduced by opening the rotor stagger setting, and the first increases with rotor suction surface Mach number. The higher mass flow at more open rotor settings provide higher gross thrust, up to 49% of the forward take-off value, but is limited by the increased loss at high speed.

A combined experimental and computational test program, with two low-pressure ratio aero-engine fans, has been used to identify the flow mechanisms at stall inception and the subsequent stall cell growth. The two fans have the same rotor tip clearance, annulus design, and downstream stators, but different levels of tip loading. The measurement data show that both the fans stall via spike-type inception, but that the growth of the stall cell and the final cell size is different in each fan. The computations, reproducing both the qualitative and quantitative behavior of the steady-state and transient measurements, are used to identify the flow mechanisms at the origin of stall inception. In one fan, spillage of tip leakage flow upstream of the leading edge plane is responsible. In the other, sudden growth of casing corner separation blockage leads to stall. These two mechanisms are in accord with the findings from core compressors. However, the transonic aerodynamics and the low hub-to-tip radius ratio of the fans lead to the following two findings: first, the casing corner separation is driven by shock-boundary layer interaction and second, the spanwise loading distribution of the fan determines whether the spike develops into full-span or part-span stall and both types of behavior are represented in the present work. Finally, the axial momentum flux of the tip clearance flow is shown to be a useful indicator of the leakage jet spillage mechanism. A simple model is provided that links the tip loading, stagger, and solidity with the tip clearance axial momentum flux, thereby allowing the aerodynamicist to connect, qualitatively, design parameters with the stall behavior of the fan.

This paper aims at underlining the existence of some generic properties of windmilling flows, partially spread in the literature but never clearly stated. Two kinds of axial machines are investigated from compressor mode to highly loaded windmill: a conventional fan with poor turbine performance and an optimized fan able to reach high efficiencies in both compressor and windmilling operations. Both simulations and experiments are used to perform the analysis. Three particular behaviors were identified as typical of fans operating at windmill: the inverse stacking of the speed lines visible in (Π, Q m ) turbine maps, the appearance of a slope change on the loading-to-flow coefficient diagram at windmill, and a threshold effect occurring at highly loaded windmill.

Uncertainty quantification (UQ) is an increasingly important area of research. As components and systems become more efficient and optimized, the impact of uncertain parameters is likely to become critical. It is fundamental to consider the impact of these uncertainties as early as possible during the design process, with the aim of producing more robust designs (less sensitive to the presence of uncertainties). The cost of UQ with high-fidelity simulations becomes therefore of fundamental importance. This work makes use of least-squares approximations in the context of appropriately selected polynomial chaos (PC) bases. An efficient technique based on QR column pivoting has been employed to reduce the number of evaluations required to construct the approximation, demonstrating the superiority of the method with respect to full-tensor quadrature (FTQ) and sparse-grid quadrature (SGQ). Orthonormal polynomials used for the PC expansion are calculated numerically based on the given uncertainty distribution, making the approach optimal for any type of input uncertainty. The approach is used to quantify the variability in the performance of two large bypass-ratio jet engine fans in the presence of shape uncertainty due to possible manufacturing processes. The impacts of shape uncertainty on the two geometries are compared, and sensitivities to the location of the blade shape variability are extracted. The mechanisms at the origin of the change in performance are analyzed in detail, as well as the differences between the two configurations.

The original Spalart–Allmaras (SA) model is known to predict premature stall when applied to fan or compressor, which is in line with the observation of other researchers who use the SA model. Therefore, to improve the prediction of the stall boundary, the original SA model was modified by scaling the source term based on the local pressure gradient and the velocity helicity of the flow. Furthermore, a generalized wall function valid for nonzero wall pressure gradient was implemented to improve the accuracy of boundary conditions at the solid wall. This work aims to produce a turbulence model which can be used to model flows near the stall boundary for the transonic fan rotors on relatively coarse grids of around 600k points per passage. Initially, two fan rotors with different design and operating speeds were used to optimize the new parameters in the modified turbulence model. The optimization was based on improving the correlation between measured and numerical radial profiles of the pressure ratio. Thereafter, steady computations were performed for two other fans (by using the same parameters), and the predictions were compared with the experimental data for all the four fan rotors. Numerical results showed a significant improvement over those obtained with the original SA model, when compared against the measured data. Finally, for completeness, it was decided to test the performance of the modified model by comparing the result with measured data for a simple canonical case.

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.

This paper aims for the analysis of experimental and numerical results of windmilling flow topologies far from freewheeling condition. Two major cooling fans were investigated: a baseline design and an innovative one meant to reach good performance in both compressor and turbine modes. Experiments are conducted with global and local characterizations to determine energy recovery potential and local loss mechanisms. Also, tests were performed on a turbofan engine to confirm some trends observed on the cooling fans. The numerical study is carried out with mixing plane steady simulations, the results of which are in fair agreement with experimental data. The difference of local topology between freewheeling and highly loaded windmill demonstrates that classical deviation rules such as Carter's are not well-suited to highly loaded windmilling flows. Finally, under certain conditions, the minor influence of the stator on the rotor topology indicates that nonrotating elements can be considered as loss generators.

For an unpowered turbofan in flight, the airflow through the engine causes the fan to freewheel. This paper considers the flow field through a fan operating in this mode, with emphasis on the effects of blade row losses and deviation. A control volume analysis is used to show that windmilling fans operate at a fixed flow coefficient which depends on the blade metal and deviation angles, while the blade row losses are shown to determine the fan mass flow rate. Experimental and numerical results are used to understand how the loss and deviation differ from the design condition due to the flow physics encountered at windmill. Results are presented from an experimental study of a windmilling low-speed rig fan, including detailed area traverses downstream of the rotor and stator. Three-dimensional computational fluid dynamics (CFD) calculations of the fan rig and a representative transonic fan windmilling at a cruise flight condition have also been completed. The rig test results confirm that in the windmilling condition, the flow through the fan stator separates from the pressure surface over most of the span. This generates high loss, and the resulting blockage changes the rotor work profile leading to modified rotational speed. In the engine fan rotor, a vortex forms at the pressure surface near the tip and further loss results from a hub separation caused by blockage from the downstream core and splitter.

The present work is a contribution to understanding the windmilling operation of low-speed fans. Such an operating situation is described in the literature, but the context (mainly windmilling of aero-engines) often involves system dependence in the analysis. Most of the time, only regimes very close to the free-windmilling are considered. A wider range is analyzed in the present study, since the context is the examination of the energy recovery potential of fans. It aims at detailing the isolated contribution of the rotor, which is the only element exchanging energy with the flow. Other elements of the system (including the stator) can be considered as loss generators and be treated as such in an integrated approach. The evolution of the flow is described by the use of theoretical and experimental data. A theoretical model is derived to predict the operating trajectories of the rotor in two characteristic diagrams. A scenario is proposed, detailing the local evolution of the flow when a gradual progression toward free and load-controlled windmilling operation is imposed. An experimental campaign exerted on two low-speed fans aims at the analysis of both the local and global aspects of the performance, for validation. From a global point of view, the continuity of the operating trajectory is predicted and observed across the boundary between the quadrants of the diagrams. The flow coefficient value for the free-windmilling operation is fairly well predicted. From a local point of view, the local co-existence of compressor and turbine operating modes along the blade span is observed as previously reported. It is further demonstrated here that this configuration is not exclusive to free-windmilling operation and occurs inside a range that can be theoretically predicted. It is shown that for a given geometry, this local topology strongly depends on the value of the flow coefficient and is very sensitive to the inlet spanwise velocity distribution.

In modern turbomachinery blade design, nonradial stacking of the profiles is often assumed to be one of the ways to improve the performance of a machine. Instead of stacking the profiles radially, the stacking line is changed by several modifications such as sweep, dihedral, lean, or a combination of these. Nonradial stacking influences secondary flows that have effects on the aerodynamic parameters such as efficiency, pressure rise, blade loading, and stall margin. However, many of the studies in literature are limited by the comparison of two or three cases. This situation leads to conflicting results because a modification may cause a positive effect in one study while in another one, the same modification may have a negative effect. In this study, a modified free vortex axial fan (named as base fan (BF) for this study) is designed first and the profiles of the blades are stacked radially by joining the centroids of the profiles. Second, 45 deg, 30 deg forward sweep (FS) and backward sweep (BS) modifications are applied. The effects of these modifications on aerodynamic performance of the fans are investigated by means of numerical calculations. The results show that FS and BS do not significantly affect the overall performance of the fan at the design flowrate in spite of the occurring modifications of the local blade pressure distribution. However, at low flowrates, FS and BS have positive and negative effects on the fan performance, respectively.

The noise emitted by axial fans plays an integral role in product design. When conventional design procedures are applied, the aeroacoustic properties are controlled via an extensive trial-and-error process. This involves building physical prototypes and performing acoustic measurements. In general, this procedure makes it difficult for a designer to gain an understanding of the functional relationship between the noise and geometrical parameters of the fan. Hence, it is difficult for a human designer to control the aeroacoustic properties of the fan. To reduce the complexity of this process, we propose an inverse design methodology driven by a genetic algorithm. It aims to find the fan geometry for a set of given objectives. These include, most notably, the sound pressure frequency spectrum, aerodynamic efficiency, and pressure head. Individual bands of the sound pressure frequency spectrum may be controlled implicitly as a function of certain geometric parameters of the fan. In keeping with inverse design theory, we represent the design of axial fans as a multi-objective multiparameter optimization problem. The individual geometric components of the fan (e.g., rotor blades, winglets, guide vanes, shroud, and diffusor) are represented by free-form surfaces. In particular, each blade of the fan is individually parameterized. Hence, the resulting fan is composed of geometrically different blades. This approach is useful when studying noise reduction. For the analysis of the flow field and associated objectives, we utilize a standard Reynolds averaged Navier–Stokes (RANS) solver. However, for the evaluation of the generated noise, a meshless lattice-Boltzmann solver is employed. The method is demonstrated for a small axial fan, for which tonal noise is reduced.

In this study, experiments and numerical simulations were carried out to evaluate the effects of radially distorted incident flow on the performance of axial-flow fans, which were equipped with three kinds of forward-skewed blades. Loss coefficient and velocity components at exit section as well as overall performance are discussed, both at the design point and the lower mass flow rate point. Furthermore, rotating stall was also observed by use of three dynamic pressure sensors. All results reveal that there are beneficial effects of forward-skewed blades on the fan performance, but the fan performance and its unsteady aerodynamic characteristics are quite affected by the radially distorted incident flow.

Part I of this paper validated the ability of the unsteady Reynolds-Averaged Navier-Stokes (RANS) solver PTURBO to accurately simulate distortion transfer and generation through selected blade rows of two multistage fans. In this part, unsteady RANS calculations were successfully applied to predict the 1/rev inlet total pressure distortion transfer in the entirety of two differently designed multistage fans. This paper demonstrates that high-fidelity computational fluid dynamics (CFD) can be used early in the design process for verification purposes before hardware is built and can be used to reduce the number of distortion tests, hence reducing engine development cost. The unsteady RANS code PTURBO demonstrated remarkable agreement with the data, accurately capturing both the magnitude and the profile of total pressure and total temperature measurements. Detailed analysis of the flow physics identified from the CFD results has led to a thorough understanding of the total temperature distortion generation and transfer mechanism, especially for the spatial phase difference of total pressure and total temperature profiles. The analysis illustrates that the static parameters are more revealing than their stagnation counterpart and that pressure and temperature rise are more revealing while the pressure and temperature ratio could be misleading. The last stage is effectively throttled by the inlet distortion even though the overall engine throttle remains unchanged. The total temperature distortion generally grows as flow passes through the fan stages.

Demands for improved performance and operability of advanced propulsion systems require an understanding of the physics of inlet flow distortion transfer and generation and the subsequent engine response. This also includes developing a high-fidelity characterization capability and suitable tools/rules for the design of distortion tolerant engines. This paper describes efforts to establish a high-fidelity prediction capability of distortion transfer and fan response via high-performance computing. The current CFD capability was evaluated with a focus of predicting the transfer of prescribed inlet flow distortions. Numerical simulations, comparison to experimental data, and analysis of two selected three-stage fans are presented. The unsteady Reynolds-Averaged Navier-Stokes (RANS) code PTURBO demonstrated remarkable agreement with data, accurately capturing both the magnitude and profile of total pressure and total temperature measurements. Part I of this paper describes the establishment of the required numerical simulation procedures. The computational domains are limited to the first three blade rows for the first multistage fan and the last three blade rows for the second fan. This paper presents initial validation and analysis of the total pressure distortion transfer and the total temperature distortion generation. Based on the established ground work of Part I, the entire two multistage fans were simulated with inlet distortion at normal operating condition and near stall condition, which is Part II of this paper. Part II presents the full range validation against engine test data and in-depth analysis of distortion transfer and generation mechanisms throughout the two fans.

This paper reports on quantitative tests of passive techniques for rotor-tip noise control in low-speed axial flow fans, based on blade-tip modifications involving the addition of antivortex appendages as end-plates. The end-plate thickness chordwise distribution is determined to control the chordwise evolution of the leakage vortex rotation number. The results confirm that the new end-plate configurations provide a mechanism by which leakage vortex bursting can be avoided. As such, the modified rotors represent an effective means of passive control of vortex breakdown.

An improved construction of air-separator device, which has radial-vanes embedded within its inlet circumferential opening with their leading-edges facing the moving tips of the fan rotor-blades so as to scoop the tip flow, was investigated with respect to the stall-prevention effect on a low-speed, single-stage, lightly loaded, axial-flow fan. Stall-prevention effects by the separator layout, relative location of the separator to the rotor-blades, and widths of the openings of the air-separator inlet and exit were parametrically surveyed. As far as the particular fan is concerned, the device together with the best relative location has proved to be able to eliminate effectively the stall zone having existed in the original solid-wall characteristics, which has confirmed the promising potential of the device. Guidelines were obtained from the data for optimizing relative locations of the device to the rotor-blades, maximizing the stall-prevention effect of the device, and minimizing the axial size of the device for a required stall-prevention effect, at least for the particular fan and possibly for fans of similar light-load fans. The data suggest the changing internal flow conditions affected by the device conditions.

The stall-prevention effect of air separators incorporating radial vanes in place of the existing axial vanes was investigated on a low-speed, single-stage, lightly loaded axial-flow fan for effective and compact air separators of a simplified structure. From the survey, paying attention to several geometrical dimensions of the device, the following conclusions are obtained: (1) Simplified radial vanes made of flat plates could show strong stall-prevention effect comparable to those of the curved-vane type one. The most favorable ones showed no stall up to the fan shut-off conditions. (2) Radial heights of the recirculation passage within the air separator showed significant influences on the stall improvement. It should be larger than some critical size experimentally given in the study. (3) The axial length of the device should be larger than some critical size given experimentally in the study. Too much reduced axial length could give rise to an abrupt loss in the effect. (4) The optimum axial locations of the rotor-tip blade leading edge within the device inlet opening were found to lie near the center of the width of the inlet opening from both aspects of stall improvement and fan efficiency.

This paper describes the observation, measurement, and computation of vortex shedding behind a cascade of turbine nozzle guide vanes that have a blunt trailing edge. At subsonic discharge speeds, periodic wake vortex shedding was observed at all times at a shedding frequency in the range 7–11 kHz. At high subsonic speeds the wake was susceptible to strong energy redistribution. The effect was greatest around an exit Mach number of 0.95 and results are presented for that condition. An unusually cold flow on the wake centerline and hot spots at the edges of the wake were measured. These were found to be a manifestation of Eckert–Weise effect energy separation in the shed vortex street. Experimental identification of these phenomena was achieved using a new stagnation temperature probe of bandwidth approaching 100 kHz. Using phase-averaging techniques, it was possible to plot contours of time-resolved entropy increase at the downstream traverse plane. Computational work has been undertaken that gives qualitative confirmation of the experimental results and provides a more detailed explanation of the fine scale structure of the vortex wake. The topology of the wake vortical structures behind blunt trailing-edged turbine blades is becoming clearer. These measurements are the first instantaneous observations of the energy separation process occurring in turbine blade wake flows. This was also the first demonstration of the use of the probe in the frequency, Mach number, and temperature ranges typical of operation behind the rotors of high-performance turbomachines such as transonic fans.

A numerical analysis of the flow in axial flow fans with skewed blades has been conducted to study the three-dimensional flow phenomena pertaining to this type of blade shape. The particular fans have a low pressure rise and are designed without stator. Initial studies focused on blades skewed in the circumferential direction, followed by investigations of blades swept in the direction of the blade chord. A Navier–Stokes code was used to investigate the flow. The simulation results of several fans were validated experimentally. The three-dimensional velocity field was measured in the fixed frame of reference with a triple sensor hot-film probe. Total pressure distribution measurements were performed with a fast response total pressure probe. The results were analyzed, leading to a design method for fans with swept blades. Forward swept fans designed accordingly exhibited good aerodynamic performance. The sound power level, measured on an acoustic fan test facility, improved.

The bending mode Flutter of a modern transonic fan has been studied using a quasi-three-dimensional viscous unsteady CFD code. The type of flutter in this research is that of a highly loaded blade with a tip relative Mach number just above unity, commonly referred to as transonic stall flutter. This type of Flutter is often encountered in modern wide chord fans without a part span shroud. The CFD simulation uses an upwinding scheme with Roe’s third-order flux differencing, and Johnson and King’s turbulence model with the later modification due to Johnson and Coakley. A dynamic transition point model is developed using the e n method and Schubauer and Klebanoff’s experimental data. The calculations of the flow in this fan reveal that the source of the flutter of IHI transonic fan is an oscillation of the passage shock, rather than a stall. As the blade loading increases, the passage shock moves forward. Just before the passage shock unstarts, the stability of the passage shock decreases, and a small blade vibration causes the shock to oscillate with a large amplitude between unstarted and started positions. The dominant component of the blade excitation force is due to the foot of the oscillating passage shock on the blade pressure surface.