Rolls-Royce (Germany)

Volutes downstream of radial compressor impellers and their respective diffusors have proven to be useful flow geometries, as they collect and deflect the swirling mass flow efficiently and feature a compact shape. Whilst a good preliminary design of a volute can be found by means of the conservation of mass and angular momentum at the diffusor outlet, the pursuit of ever-increasing efficiencies raises the question of which design methods are best suited to find these designs. The combination of automatized optimization and reliable computational fluid dynamics appears promising. Hence, three different optimization strategies are tested, their pros and cons discussed and their results compared. Two of the methods exploit the adjoint method to determine gradients. They both prove to be superior in terms of computational effort and design improvement. Both algorithms suggest a prominent design change that concerns the volute tongue. It is moved away from the impeller in a way that it homogenizes the static pressure at the diffusor vane inlets and leads to an overall reduction in pressure drop.

The design of fan-blisks is a multi-criterion optimisation challenge primarily involving the aerodynamic shape optimisation of blade profiles and subsequent blade balancing, i.e., shifting profiles in axial and circumferential directions to avoid stress hotspots. Although it is already known that blade balancing affects aerodynamic properties, this correlation is often not taken into account, which is why it is usually performed as an independent step after solving the aerodynamic design problem. However, this assumption is questionable because such shifts should be used to control both secondary flow effects and resulting stress. Therefore, the paper proposes problem formulations combining both aspects. Since optimisation requires costly numerical evaluations, machine learning methods are investigated to predict aerodynamic performance and stress constraints more efficiently. This enables a global optimisation process by reducing computational costs. Various different surrogate types are investigated, where stress constraints are formulated either as regression task predicting stress maxima, or as classification problem directly assessing design feasibility.

Future aero engine designs must address environmental challenges and meet noise and emissions regulations. To increase efficiency and reduce size, axial compressors require higher pressure ratios and a more compact design, leading to necessary modifications in the variable stator vanes, especially in the stator hub region. This study examines the impact of a variable cantilevered stator on the performance and aerodynamics of a 1.5-stage transonic compressor, representative of a high-pressure compressor front stage. Experimental tests at the transonic compressor test rig at Technical University of Darmstadt involved two variable stators with identical airfoil designs but different hub configurations, using the same inlet guide vane and rotor. Detailed aerodynamic analysis was conducted using steady and unsteady instrumentation. The cantilevered stator achieved a 2% increase in efficiency and a 1% increase in total pressure ratio, due to higher aerodynamic loading and reduced pressure losses. The primary performance gain comes from the reduction of the hub blockage area. The cantilevered stator also performed well at near stall conditions, unlike the shrouded stator. Time-resolved measurements indicated that loss mechanisms are closely linked to the rotor wake phase. Overall, variable cantilevered stators outperformed shrouded stators in this compressor stage.

There exists numerous techniques for connecting an aircraft’s power plant to its supporting structure, the pylon, and for connecting the pylon to the wing. This paper investigates how two connection strategies employed in wide body commercial aircraft with fan mounted and core mounted power plants affect the sizing of the wing and pylon for the same loading. Sizing of the pylon in isolation is investigated and compared to results for when sized with the wing. Eighty-eight load cases, including flight, landing and those generated by the power plant are selected for the sizing studies. Static aeroelastic analysis of a reduced order aircraft model provides the flight loads. Free sizing the thickness of components is then performed. Results show that the sized pylon and wing with connections for a fan mounted power plant system are 1.7% lighter than that for a core mounted system. Pylons sized with the wing for fan and core mounted power plants are found to be 10.3% and 12.7% heavier than those sized in isolation. These investigations contribute towards wider work to develop whole aircraft models for static and dynamic aeroelastic analysis, to support combined aircraft and engine design.

Impedance spectroscopy is one of the continuously developing tools deployed for the inspection and evaluation of the health condition of electrical machines and machine components. Being a practical implementation and expansion of the Broadband Dielectric Spectroscopy (BDS) and the Electrochemical Impedance Spectroscopy (EIS), it allows the characterization of components by their amplitude and phase responses over frequency and the system's resonances. To this direction, this paper presents a study on the effect of thermal degradation on the insulation system of concentrated stator coils from the stator winding of a 2 MW permanent magnet synchronous generator purposed for aerospace applications. The thermal aging pattern and degradation mechanisms of the coil insulation are discussed through experimental results by ex-situ impedance spectroscopy.

Abstract The characterisation and quantification of runway foreign object ingestion (FOI) at the end of the landing run with a variable pitch fan (VPF) in reverse thrust mode are described in this work. This is achieved by developing a novel, fully coupled Computational Fluid Dynamics - Discrete Phase Model (CFD-DPM) methodology that allows the tracking of individual debris, which is placed on the runway. The flow field solutions are computed by the three-dimensional (3D) Reynolds-averaged Navier-Stokes (RANS) equations and the particulate phase is tracked in a Lagrangian way. A modern 300-seater aircraft is considered that features a high bypass ratio geared turbofan engine architecture on a conventional twin-engine airframe in landing configuration that rolls on the runway at low speeds and a typical high engine power mode. The aerodynamic reverse thrust flow field at such low aircraft speeds is characterised by the development of a “horseshoe” vortex formed in front of the engine that significantly affects the debris trajectories. Based on this, two FOI mechanisms are identified and discussed in detail in the paper. The effect of key debris properties on FOI percentage, such as size, material density and shape, are investigated and reported. Non-linear, non-intuitive trends are obtained due to the two-way interaction of flow features and particles. Moreover, a non-dimensional analysis that accounts for the changes in debris variables of interest is presented and provides general trends for aircraft with a VPF as the main thrust reversal unit. Generally, debris ingestion increases with lower aircraft speed and debris material density. The benefits of having a VPF are highlighted since it acts as an additional “protective layer” to the engine core and stops and redirects incoming debris from the nominal intake. The assessment of ingestion of foreign objects in realistic end-of-landing-run conditions as described in this study is important to examine the feasibility of reverse thrust-capable VPF engines for future sustainable aircraft.

Abstract This paper introduces several techniques that merge engineering simulation with machine learning for the thermomechanical design of a double-sided seal in the secondary air system of an aero engine turbine subassembly. The focus is on Variational Auto-Encoders (VAEs) integrated with Residual Networks (ResNets), exploring various applications including synthetic image generation, image simulation predictions, and multi-modal predictions. The models were trained using meticulously crafted images of thermo-mechanical simulation results for the double-sided seal assembly, designed to encapsulate essential performance inputs and outputs in a single frame. Additionally, double-encoded variables were incorporated for self-quality monitoring of generated images. The results demonstrate high accuracy across a range of design applications. The synthetic image generation model enhances the performance of other models and enables image synthesis, while the image simulation predictions models facilitate simulation predictions under fully constrained design parameters, as well as design exploration given under-constrained output targets. Furthermore, these models can be integrated with existing surrogate-based frameworks to expedite design optimisation. Notably, the final model could eliminate the need for feature embeddings. This paper shows that, with slight architectural modifications, the VAE-ResNets can be adapted for a diverse range of design applications.

Abstract The trailing edge of a high-pressure turbine blade presents a challenge to turbine cooling engineers. Its thin nature requires that film cooling holes are manufactured through electrode discharge machining (EDM), a process which requires high precision in order to prevent extreme variations in hole geometry and location. Previous studies on manufacturing tolerances have looked at fan-shaped holes and only onflat plates. This study aims to understand the extremes of possible manufacturing variations in the trailing edge cutback region of a high pressure turbine blade to help set appropriate manufacturing limits. A row of cylindrical film holes with a radial inclination of 30° was tested and used as the benchmark. Several film hole configurations were then tested over a range of blowing ratios. These included film hole shifts between cutback section and ejection plane as well as cutback step shifts, resulting in a change in the thickness of the cutback thickness. The experiments were carried out in the high velocity 2D linear cascade experimental facility using pressure sensitive paint (PSP) to measure film effectiveness. At low blowing ratios, most manufacturing variations were found to have very little effect on film cooling effectiveness. At higher blowing ratios, higher offsets between the cutback section and the film hole ejection plane were found to have greater detrimental effects on film cooling effectiveness. A numerical study was completed to complement the experimental findings and support the understanding of manufacturing limits in trailing edge film holes.

Abstract This article presents a novel micro-mix fuel injection concept designed specifically for hydrogen combustion, “direct hydrogen injection” (DHI). In this arrangement, fuel and air are injected into the combustor as individual discrete jets. By varying the angle of the fuel and air holes, the flow field may be shaped, altering mixing processes and flame stabilization. The DHI concept offers significant advantages for the combustion of hydrogen when compared to conventional fuel injection arrangements. The small size of the injection sites is advantageous as it results in smaller flame volumes, reduced combustor residence times and, as a direct consequence, reduced emissions of pollutants — principally the oxides of nitrogen (NOx). This article presents the first optical measurements of the DHI concept, with a subset of injector configurations tested at larger than engine scale. These clearly demonstrate the ability to control the structure of the flame and flow field by altering air jet angle leading to the formation of substantially changed reaction zones, both a conical flame and a V-flame have been created by changing only the air/fuel injection angle. The influence of the number of fuel injection sites on the flame structure is also investigated, with a significant reduction in flame length achieved by increasing the number of injection sites. Steady-state RANS predictions are also performed and compared to the experimental results, with the CFD found to reasonably recover the overall structure of the flame for most cases. Despite reasonable first order performance, fuel jet penetration is typically over predicted, leading to longer jet flames, whilst the flame thickness is possibly under-predicted, suggesting opportunity for improvement. The combined experimental, numerical study, however, shows significant potential for the direct hydrogen injection concept. Compact, stable flames are generated, without ignition challenges or stability issues. The flame structures are further unchanged by air velocity and/or fuel flow rate.