Study on the Influence of Inlet Asymmetry on Aerodynamic Noise of Cooling Fan

Experimental and numerical studies on the aerodynamic noise characteristics of a variable-speed axial fan commonly used for electronic device heat dissipation were conducted. First, the far-field noise spectrum of the fan was measured using a microphone array on the contour plane of the fan axis. The spectral analysis indicated that the discrete single-tone noise energy ratio was high, which indicates that it was the dominant aerodynamic noise. Afterwards, the double-uniform sampling point mode correction technique, which is based on the circumferential acoustic mode measurement method, was used to obtain the modal distribution on the inlet and outlet sides of the cooling fan. The influence of inlet unevenness on the cooling fan was identified. The traditional Tyler-Sofrin rotor-stator interaction formula was modified to account for the non-axisymmetric shape of the fan inlet bellmouth. The validity of the modified formula was verified by measuring the circumferential acoustic modes of three cooling fans with different rotor and strut counts. Furthermore, a CFD numerical study was conducted using Fluent to understand the influence of uneven inlet flow. The results showed that uneven inlet flow significantly affects the size and distribution of unsteady pulses on the rotor blades, which cause regular, periodic changes as the rotor blades rotate. Interactions between rotor blades and inlet unevenness were observed via the POD method as well. The discussion of the circumferential modes from aerodynamic noise of an axial flow cooling fan can act as a reference for further cooling fan noise reduction measures.

Towards Primary Breakup Simulation of a Complete Aircraft Nozzle at Realistic Aircraft Conditions

Predicting details of aircraft engine combustion by means of numerical simulations requires reliable information about spray characteristics from liquid fuel injection. However, details of liquid fuel injection are not well documented. Indeed, standard droplet distributions are usually utilized in Euler-Lagrange simulations of combustion. Typically, airblast injectors are employed to atomize the liquid fuel by feeding a thin liquid film in the shear zone between two swirled air flows. Unfortunately, droplet data for the wide range of operating conditions during a flight is not available. Focusing on numerical simulations, Direct Numerical simulations (DNS) of full nozzle designs are nowadays out of scope. Reducing numerical costs, but still considering the full nozzle flow, the embedded DNS methodology (eDNS) has been introduced within a Volume of Fluid framework (Sauer et al., Atomization and Sprays, vol. 26, pp. 187–215, 2016). Thereby, DNS domain is kept as small as possible by reducing it to the primary breakup zone. It is then embedded in a Large Eddy Simulation (LES) of the turbulent nozzle flow. This way, realistic turbulent scales of the nozzle flow are included, when simulating primary breakup. Previous studies of a generic atomizer configuration proved that turbulence in the gaseous flow has significant impact on liquid disintegration and should be included in primary breakup simulations (Warncke et al., ILASS Europe, Paris, 2019). In this contribution, an industrial airblast atomizer is numerically investigated for the first time using the eDNS approach. The complete nozzle geometry is simulated, considering all relevant features of the flow. Three steps are necessary: 1. LES of the gaseous nozzle flow until a statistically stationary flow is reached. 2. Position and refinement of the DNS domain. Due to the annular nozzle design the DNS domain is chosen as a ring. It comprises the atomizing edge, where the liquid is brought between inner and outer air flow, and the downstream primary breakup zone. 3. Start of liquid fuel injection and primary breakup simulation. Since the simulation of the two-phase DNS and the LES of the surrounding nozzle flow are conducted at the same time, turbulent scales of the gas flow are directly transferred to the DNS domain. The applicability of eDNS to full nozzle designs is demonstrated and details of primary breakup at the nozzle outlet are presented. In particular a discussion of the phenomenological breakup process and spray characteristics is provided.

Development, Analysis, and Validation of a Simultaneous Inlet Total Pressure and Swirl Distortion Generator

Over the last decade, the Turbomachinery and Propulsion Research Laboratory at Virginia Tech has researched, invented, developed, computationally analyzed, experimentally tested, and improved turbofan engine inlet distortion generators. This effort began with modernizing and improving inlet total pressure distortion screens originally conceived over half a century ago; continued with the invention of inlet swirl distortion generators (StreamVanes™) made possible only through advances in modern additive manufacturing technology; and has, thus far, culminated in a novel combined device (ScreenVanes™) capable of simulating realistic flight conditions of coupled inlet total pressure and swirl distortion in a ground-test turbofan engine research platform. The present research focuses on the methodology development, computational analysis, and experimental validation of a novel simultaneous inlet total pressure and swirl distortion generator. A case study involving a single bend S-duct inlet distortion profile demonstrates the ability to generate a high-fidelity profile simulation, yet outlines a design process sufficiently generic for application to any arbitrary inlet geometry or distortion profile. A computational fluid dynamics simulation of the S-duct inlet provided the target profile extracted at the aerodynamic interface plane. Next, utilizing a method of inverse propagation, the planar distortion profile was propagated upstream to yield a flow field that could be manufactured by a distortion generator adequately isolated from turbomachinery effects. The total pressure distortion screen and swirl distortion StreamVane components were then designed and computationally analyzed. Upon successful computational reproduction of the S-duct inlet distortion profile, experimental hardware was fabricated and tested to validate the ScreenVane methodology and distortion generating device. Comparison of the S-duct manufactured distortion and the ScreenVane manufactured distortion was used as the primary criterion for profile replication success. Results from a computational analysis of both the S-duct and ScreenVane indicated excellent agreement in distortion pattern shape, extent, and intensity with full-field total pressure recovery and swirl angle profiles matching within approximately 0.80% and 2.6°, respectively. Furthermore, experimental validation of the ScreenVane indicated nearly identical full-field total pressure recovery and swirl angle profile replication of approximately 1.10% and 2.6°, respectively, when compared to the computational results. The investigation concluded that not only was the ScreenVane device capable of accurately simulating a complex inlet distortion profile, but also produced a viable device for full-scale turbofan engine ground test.

Preliminary Research on Future Smart Engine Architecture

Aero engines that fit the future have now increasingly attracted the attention of aerospace industry and academia. With this trend, many research projects have been carried out to explore future aero engine technologies. This paper focuses on engine design field, and aims to satisfy the future flight missions that may be unpredictably varying. However, the intrinsic strong coupling of engine component matching mechanism blocks acceleration of engine design. Under this condition, this paper comes up with the concept of smart engine architecture that via a series of engine decoupling strategies, the components can be decoupled to an extent that by properly selecting and assembling them, an engine that satisfies certain flight mission can be designed, this is named mission-oriented pluggable design mode in this paper. Following this idea, a multi-purpose engine design scheme is presented to demonstrate the potential of this engine design mode, and further value of smart engine architecture is discussed.

On-Line Component Map Adaptive Procedure Based on Sensor Data

A key challenge in the gas turbine community is to adapt the engine model by matching measured data with simulation data. This study presents a procedure aiming to calibrate a certain type of gas turbine for power generation. To reproduce degradation, disturbance is injected into the healthy components maps at different time. Subsequently, six correction factors along with measured data and unmeasured parameters are coupled together using cooperative working equations and optimized based on primal-dual interior point method. When performing the adaptive procedure, Jacobian and hessian matrices are calculated using finite difference since the component maps have external, mapped, functions implemented as lookup-tables, and mode-switching statements. To improve the accuracy of first-order and second-order partial derivatives, the finite difference is enhanced by Richardson extrapolation method. The search scope of correction factors and unmeasured parameters are determined by the whole working conditions. Meanwhile, an adaptive update method of initial solution is proposed to make sure the convergence of the optimization procedure as quickly as possible. Finally, the proposed method is further applied to the on-line adaptation in case of performance degradation. The influence of measurement noise on optimization is also studied. It is demonstrated that the procedure is capable of refining the component maps progressively, which is significant for the model-based gas path diagnostics and prognostics.

Improvement of Turboshaft Restart Time Through an Experimental and Numerical Investigation

Inflight shutdown of one engine for twin-engine helicopters have proven beneficial for fuel consumption. A new flight mode is then considered, in which one engine is put into sleep mode (the gas generator is kept at a stabilized, sub-idle speed by means of an electric motor, with no combustion), while the second engine runs almost at nominal load. The ability to restart the engine in sleep mode is then critical for safety reasons. Indeed, the certification of this flight mode involves ensuring a close-to-zero failure rate for in-flight restarts as well as a fast restart capability of the shutdown engine. In this paper, the focus is made on improving the restart time of the shutdown turboshaft engine. Fast restart capability is necessary for flight management reasons. Indeed, in case of a failure of the engine operating close to nominal load while the other one is in sleep mode, there is no more power available and the helicopter can lose up to 15–20 meters per second during autorotation. The restart time becomes a critical parameter to limit the loss of altitude. In the configuration studied, the fast restart is achieved thanks to the electric motor designed to deliver a high torque to the gas generator shaft. This electric motor is powered by an additional battery, more powerful than the conventional one dedicated for standard restarts. The aim of the paper is to assess the potential restart time saving using an approach combining test rig data analysis and numerical results generated by a thermodynamic model able to simulate at very low rotational speed. A gas turbine engine starting process is composed of two main phases: the light-up phase and the acceleration phase. It is important to understand the detailed phenomenology of these two phases as well as the various sub-systems involved, first to highlight the influencing parameters of both phases and then to establish an exhaustive listing of the possible time optimizations. From the test rig campaign, conducted at Safran Helicopter Engines on a high power free turbine turboshaft engine, we are able to accurately break down the phases of the start-up sequence, which helps us to identify what steps of the sequence worth shortening. With the engine performance thermodynamic model, we can then use the information gathered from the test rig analysis to further predict how to save time and to give guidelines for developing new control strategies. The results of this study show that a fast restart going from sleep mode to max power speed can be up to 60% faster than a conventional restart going from sleep mode to idle speed. This is significantly faster, especially if one takes into account the higher final speed targeted by the fast restart.

Integration of Computational Fluid-Particle Dynamics Techniques for the Instantaneous Estimation of Particle Erosion Damage on Axial Fan Blade Sections

In the last decade, the authors focused their research in the development and implementation of accurate numerical tools and methods able to predict the erosion and deposit on turbomachinery blades operating with particle-laden flows. These models and methods give complete three-dimensional description of the phenomenon, but their application is limited to a single working condition of the blade. The present paper covers the first step in the definition of a general methodology to extend the applicability of these tools to a full range of the machines operating conditions. The method aims to obtain an instantaneous prediction of the expected damage pattern for a blade section, given its local working condition in terms of relative fluid-particle flow. The final result is based on a precomputed database associated to the blade section, where the single element is obtained by computing the erosion damage using the aforementioned numerical tools. This paper will show the methodology to obtain the database associated to the midspan section of an induced draft fan subjected to erosion due to coal ash particle. The final database is then used to predict the damage state of the section associated to a given point in the characteristic curve of the fan.

Acceleration Method for Evolutionary Optimization of Variable Cycle Engine

Variable cycle engine (VCE) is considered as one of the best options for advanced military or commercial supersonic propulsion system. Variable geometries enable the engine to adjust performance over the entire the flight envelope but add complexity to the engine. Evolutionary algorithms (EAs) have been widely used in the design of VCE. The initial guesses of the engine model are generally set using design point information during evolutionary optimization. However, the design point information is not suitable for all situations. Without suitable initial guesses, the Newton-Raphson solver will not be able to reach the solution quickly, or even get a convergent solution. In this paper, a new method is proposed to obtain suitable initial guesses of VCE model during evolutionary optimization. Differential evolution (DE) algorithm is used to verify our method through a series of optimization cases of a double bypass VCE. The result indicates that the method can significantly reduce the VCE model call number during evolutionary optimization, which means a dramatic reduction in terms of evolution time. And the robustness of the optimization is not affected by the method. The method can also be used in the evolutionary optimization of other engines.

Flow Distortion Into the Core Engine for an Installed Variable Pitch Fan in Reverse Thrust Mode

The flow distortion at core engine entry for a Variable Pitch Fan (VPF) in reverse thrust mode is described from a realistic flow field obtained using an integrated airframe-engine model. The model includes the VPF, core entry splitter, complete bypass nozzle flow path wrapped in a nacelle and installed to an airframe in landing configuration through a pylon. A moving ground plane to mimic the rolling runway is included. 3D RANS solutions are generated at two combinations of VPF stagger angle and rotational speed settings for the entire aircraft landing run from 140 to 20 knots. The internal reverse thrust flow field is characterized by bypass nozzle lip separation, pylon wake and recirculation of flow turned back from the VPF. A portion of the reverse stream flow turns 180° with separation at the splitter leading edge to feed the core engine. The core engine feed flow exhibits circumferential and radial non-uniformities that depend on the reverse flow development at different landing speeds. The temporal dependence of the distorted flow features is also explored by an URANS analysis. Total pressure and swirl angle distortion descriptors, as defined by the Society of Automotive Engineers (SAE) S-16 committee, and, total pressure loss into the core engine are described for the core feed flow at different operating conditions and landing speeds. It is observed that the radial intensity of total pressure distortion is critical to core engine operation, while the circumferential intensity is within acceptable limits. Therefore, the baseline sharp splitter edge is replaced by two larger rounded splitter edges of radii, ∼0.1x and ∼0.2x times the core duct height. This was found to reduce the radial intensity of total pressure distortion to acceptable levels. The description of the installed core feed flow distortion, as described in this study, is necessary to ascertain stable core engine operation, which powers the VPF in reverse thrust mode.

Methodology for Simulation of Soak-Back in a Helicopter Engine Bay Using Lattice Boltzmann Method

Soak-back conditions are a crucial challenge for engine manufacturers to reduce the turnaround time between flights and improve engine durability and security. The natural convection conditions at stake make the numerical predictions of the flow and thermal behaviors rather difficult, when not prohibitively expensive. Consequently, the topic is still assessed with tests which can only be performed at a very late stage of the engine development and do not provide a good overview of the physics behind. Failures discovered so late are extremely expensive and complex to solve. This paper presents the first phase of a methodology development to tackle the soak-back of an engine with the SIMULIA PowerFLOW Suite Computational Fluid Dynamics code. Comparisons for validation are made with the tests on one hand, and with ANSYS Fluent RANS (Reynolds Averaged Navier-Stokes) simulations whenever possible. The work conducted so far focuses on the engine bay only with coupled fluid-thermal simulations while the core flow is simplified into a 1D fluid nodes network. A first simplified approach has proved to recover some of the phenomena observed in both the tests and RANS simulations. It failed however to match the initial temperatures of the soak phase, which is consistent with the choice of modelization made. Improvements to the model were therefore brought by adding more complexity and fidelity to the geometry, environment and tests scenario. The new results significantly improve the comparisons with the tests and RANS simulations. Some differences on absolute temperature levels and evolution rates remain here and there and highlight the necessity to improve the core flow modelization, which is what the next phase of the methodology development will focus on.