Numerical Investigation of Kelvin-Helmholtz Instability in a Centrifugal Compressor Operating Near Stall

The present works details the occurrence of the Kelvin–Helmholtz instability in a centrifugal compressor operating near stall. The analysis are based on unsteady three dimensional simulations performed on a calculation domain covering the full annulus for the impeller and the vaned diffuser. A detailed investigation of the flow structure is presented, together with its evolution consequent to the mass flow reduction. It is demonstrated that this reduction leads to an enlargement of the low momentum flow region initially induced by the combination of the secondary and leakage flows. When the compressor operates near stall, the shear layer at the interface between the main flow and this low momentum flow becomes unstable and induces a periodic vortex shedding. The frequency of such an unsteady phenomenon is not correlated with the blade passing frequency. Its signature is thus easily isolated from the deterministic rotor/stator interaction. Its detection requires full-annulus simulations with an accurate resolution in time and space, which explains why it has never been previously observed in centrifugal compressors.

Uncertainty Quantification of Hot Gas Ingestion for a Gas Turbine Nozzle Using Polynomial Chaos

One of the most critical parameters in the design process of cooled hot gas components, is the Back Flow Margin (BFM). This dimensionless parameter quantifies the margin to hot gas ingestion through a cooled component wall. A correct evaluation of this parameter is crucial in order to avoid component failure. In presence of combustion chambers that exhibit low pressure losses, BFM becomes one of the most restrictive requirements in the thermal design of cooled components. In this work, a conceptual BFM assessment of the first nozzle of an HP gas turbine is described. The component is subject to the highest thermal load; complex cooling systems are required to ensure an acceptable metal temperature and to match life time requirement. Due to manufacturing tolerances and fluid dynamic uncertainties, hot gas ingestion events are possible also for a nozzle that exhibits BFM higher than zero in nominal conditions, even if with a low probability. Here, the cooling scheme of the nozzle is modeled using an in-house fluid network tool that allows a quick and accurate computation of the equivalent cooling scheme and thus the occurrence of hot gas ingestion, corresponding to a negative flow rate in one of the cooling sub-models. However, as the probability of hot gas ingestion is rather small, an accurate estimation of this event based on the standard Monte Carlo method requires a huge number of runs. A more efficient estimation of this probability can be obtained using stochastic expansion methods, such as the Polynomial Chaos Expansion. Pseudospectral approximations based on either a tensor-product expansion or the Sparse Pseudospectral Approximation Method (SPAM) are used, in order to estimate the probability of hot gas ingestion and the sensitivity to random parameters. The results are compared with those coming from Monte Carlo method, showing the superior accuracy of the stochastic expansion methods.

Development and Application of a Multi-Disciplinary Multi-Regime Design Methodology of a Low-Noise Contra-Rotating Open-Rotor

This paper focuses on the development and exploitation of a multi-disciplinary, optimization-assisted, design methodology for contra-rotating open-rotors. The design procedure relies on a two-step approach. An aero-mechanical optimization is first performed to generate a geometry with good performances over several high-speed points representative of a mission. This geometry is subsequently used as the baseline of an aero-mechanical-acoustic optimization focusing on interaction noise reduction at Cutback and Sideline low-speed points. In terms of design parameters, both rotors are modified for the first phase but only the upper part of the front rotor is altered for the noise minimization. A fully-automatic high-fidelity aero-mechanical-acoustic computational chain with fluid-structure coupling is exploited in combination with evolutionary algorithms assisted by surrogate models for the constrained-optimization process. The acoustic footprint is estimated by a simplified but fast and relevant formulation combining an unsteady lifting-line and an acoustic propagation method. The best geometry of the first design gains 1.2pt in weighted efficiency while respecting all the aero-mechanical constraints. The acoustic optimization shows that noise reduction at Sideline and Cutback points is strongly antagonistic. However, significant Sideline noise reduction from 3.5 to 5.5dB depending on the harmonics is achieved while maintaining Cutback noise and without major degradation of high-speed efficiency.

Investigation of an Asymmetric Double Entry Centrifugal Compressor With Different Radial Impellers Matching for a Wide Operating Range

The work presented here investigates the impeller matching characteristics and widens the stable operating range of front and rear impellers for an asymmetric entry double sided centrifugal compressor. A numerical approach is employed to analyze the operating characteristics of front and rear impellers, and a strategy to widen the stable operating range of double sided compressor is presented. Firstly, the performance curves of a double sided centrifugal compressor are obtained by simulating the operation of the whole-stage compressor. The result shows that the compressor operating mode switches from parallel mode to single impeller mode automatically with the decrease of the mass flow. Thus, the stable operating range of the compressor is limited. Second, the simulation of a simplified double sided compressor is conducted to reveal the mechanism of the compressor operating mode conversion. It is found that the essential reason for the conversion of the compressor operating mode is the total pressure difference between the front and rear impeller inlets. A proper increase of the rear impeller radii is helpful for improving the impeller power capability, which enables the front and rear impeller to obtain a superior matching relationship in a wider operating range and widens the stable operating range of the compressor. Furthermore, by analyzing the respective performance characteristic curves in various calculation cases, there is a critical mass flow value between the front and rear impellers for compressors with the same flow capability. When one side impeller mass flow is below the critical value, with further decrease of the flow, the pressure ratio characteristic curve of this side rises and enters the stall zone gradually. Thus, the operating mode is converted from parallel mode to single mode. This result further explains the mechanism for extending the stable operating range of a double sided compressor in a wider scope.

Three-Stage Low Pressure Compressor Modernization by Means of Optimization Methods

Low pressure compressor operation has some features. Firstly, the LPC stages work with cold air. For this reason there is transonic or subsonic flow in LPC. Secondly, the flow in LPC has complex spatial structure. Blade geometry of LPC is described by a large number of parameters. For this reason, it is difficult to pick up optimal combination of parameters manually. The solution of this problem is the usage of optimization methods to find the optimal combination of parameters. This approach was tested in this work. The main goal of this work was the LPC modernization for new parameters of gas turbine engine. Set of unimprovable solutions (Pareto set) was obtained as a result of solving optimization task. Pareto set was a compromise between the efficiency increase and the mass flow decrease. Each point from Pareto set had a correspondence with LPC unique geometry represented as an array of optimization parameters. One point of the Pareto set met all the required parameters of modernized LPC. The LPC geometry that guaranteed the efficiency increase by 1,3%, the total pressure ratio increase by 4% and mass flow rate decrease by 11% in comparison with the original LPC was obtained as a result of the investigation.

Unsteady Work and Wake Recovery due to Pressure Wave Interaction in a LP Turbine

Recent publications have demonstrated the influence of unsteady work terms on the inviscid recovery of wake momentum. So far, this so-called wake differential work effect was only validated based on selected locations and time steps in turbine rotors. The magnitude of this effect over a whole blade passing cycle and the local unsteady work mechanisms causing it are still not fully understood. Using a numerical simulation, the unsteady static pressure field of a turbine rotor is assessed. Three regions are identified in which work is transfered unsteadily to the fluid, caused by the fluid interaction with the unsteady rotor pressure field. A Lagrangian analysis is performed to validate and quantify the wake differential work concept. To be representative, a large number of wake and free stream fluid particle paths are evaluated. Overall, a 7 per cent lower wake work in the rotor is identified, averaged over a whole blade passing cycle. From a particle point of view, the rotor pressure field acts as a pressure wave propagating in circumferential direction. Due to inviscid unsteady work, this pressure wave influences the stagnation enthalpy of the fluid particles. It is shown that this effect is more dominant for wake fluid, as the wake velocity is closer to the propagation velocity of the pressure wave. A mathematical model of this so-called “wake surfing effect” and the two other unteady work mechanisms reveals how the wake momentum is recovered depending on the initial wake velocity vector. If exploited well, this unsteady work mechanism could cause a reduction of wake mixing loss, leading to an increased turbine efficiency.

Increasing Pulse Energy Recovery of Radial Turbocharger Turbines by 3D Inverse Design Method

Most radial turbines have a peak efficiency at around U/Cis (velocity ratio or jet speed ratio) of 0.7. It is a well-known fact that it is beneficial for radial turbocharger turbines to have a higher efficiency at low U/Cis region, since the pulsating engine exhaust gas at low U/Cis region (with high pressure and temperature) carries more energy compared to that at high U/Cis region (with low pressure and temperature). The improvement of Total to Static efficiency at low U/Cis region will help the turbine extract more energy from the exhaust gas and thereby increase the turbine cycle-averaged Total to Static efficiency. In the past there has been some attempts to move the peak U/Cis to lower values by using conventional or direct design approach on mixed flow impellers. But this approach usually results in reduction of stage performance. In this paper, a methodology is presented to control and move the radial turbine peak efficiency U/Cis to lower values by using a 3D inverse design method. The stage performance is measured by using steady CFD analysis. Furthermore detailed stress and vibration analysis are presented on the mechanical performance of the new design.

Unsteady Pressure Investigations of Corner Separated Flow in a Linear Compressor Cascade

The aim of this work is to present detailed unsteady pressure measurements of three-dimensional flow field in a NACA 65 linear compressor cascade. Chord-based Reynolds number of 382000 and incidence angle of 4 degrees were chosen as target configuration of the rig, which clearly presents the corner separation phenomenon at the juncture of the blade suction side and the end-wall. Concerning the experiments, a characterization of the mean and fluctuating component of wall static pressure on the surface of a specially developed blade is achieved at first. This fluctuating component is investigated utilizing nineteen high sensitivity condenser microphones plugged into blade cavities which have been carefully calibrated. Transfer functions obtained by calibration are exploited to reconstruct the time-dependent pressure signal and finally statistics, conditional ensemble averages, coherence and spectra analyses of fluctuations are presented in order to investigate the unsteady characteristics of the corner separation. High values of root mean square are individuated near the leading edge and in the separation region on the suction surface of the blade. Skewness and kurtosis show an intermittent behavior of the separation onset, which moves upstream and downstream on the suction surface. This intermittency of the separation line is probably linked with the existence of a bimodal behavior of the size of the corner separation. The analyses of coherence and conditional ensemble average between the signals at the leading edge and at the onset of the separation suggest a critical influence of angle and velocity of the incoming end-wall boundary layer on the positive pressure signatures of the shear layer, which characterize the inception of the separation.

CAD-Based Adjoint Shape Optimisation of a One-Stage Turbine With Geometric Constraints

An extension of the CAD-based parametrisation termed ‘NURBS-based parametrisation with complex constraints’, or NsPCC, is developed and applied to the aerodynamic shape optimisition of a one-stage high pressure turbine. NsPCC uses a test-point approach to impose various geometric constraints such as continuity, thickness and trailing edge radius constraint. To perform the shape optimisation using NsPCC, The surface sensitivity is first computed efficiently using a discrete adjoint solver. The displacements of the control points of the NURBS patches are used as the design variables and linked to the surface sensitivity through consistent application of Automatic Differentiation. A robust mesh deformation based on linear elasticity and further enhanced with sliding mesh capability is used to deform the mesh at each design step. Finally, the optimised rotor shape is exported as a STEP file. The method is demonstrated on a turbine stage where isentropic efficiency is improved by over 0.4% with both the inlet capacity and rotor reaction ratio deviation below the prescribed thresholds. Satisfaction of the G1 continuity, thickness and trailing radius constraints is verified.