John Dodds

In this paper an efficient numerical strategy for computational fluid dynamics (CFD) simulation of surge events in a three-shaft engine compression system is presented. Numerical results are compared against measured data and the sources of discrepancies and uncertainties are addressed. It is discovered that the engine bleed system has a major impact in reducing the aerodynamic loading during surge in the core system. To the best knowledge of the authors, this is the first time that such a complex CFD computation is attempted and compared against real engine measured data and will provide valuable information to other CFD users as well as gas turbine manufacturers.

Compressor surge imposes a limit on aero-engine operability and can compromise integrity due to significant aerodynamic loads imparted on the engine components. The aim of this paper is to use 3D unsteady CFD to predict the surge loadings on a 3 shaft engine. The computations are performed using a whole assembly approach. In this work, the effect of two types of surge trigger on

Surge in a high-speed compressor can lead to violent disruption of flow, damages to the blade structures and eventually engine shutdown. The flow disruption caused by rotating stall and surge could potentially couple with blade vibration modes which can often not be identified during the design phase. As turbomachinery blade aerofoils are thinned to improve performance and reduce

Compressor surge imposes a limit on aero-engine operability and can compromise integrity because of significant aerodynamic loads imparted on the engine components. The aim of this article is to use 3D unsteady computational fluid dynamics (CFD) to predict the surge loadings on a compression system of a modern three-spool engine. The compression system is matched at a high power condition and computations are performed using a whole-assembly approach. In this study, the effect of two types of surge initiation on the maximum loading recorded during surge are studied: throttling of the high pressure compressor (HPC) and turning of the intermediate pressure compressor (IPC) variable stator vanes (VSV). An explanation of the main physical phenomena that contribute to those loadings is offered. It was found that in an aero-engine surge event, the maximum overpressure is driven by a combined effect of the surge shock wave passing and high pressure gas blown toward the front of the engine during depressurization. The amplitude of maximum overpressure is dictated by the compression system exit pressure at the moment of surge inception. The surge initiation via HPC throttling produces larger overpressure and therefore should be considered for design intent.

Stall followed by surge in a high-speed compressor can lead to violent disruption of the flow, damage to the blade structures and eventually engine shutdown. Knowledge of unsteady blade loading during surge is crucial for compressor design such as axial gap optimization. The aim of this paper is to demonstrate the feasibility of using three-dimensional full assembly unsteady Reynolds-averaged Navier–Stokes (URANS) CFD for modeling surge cycles of an eight-stage high-speed compressor rig. Results from this work show stalling of the mid-stages is the surge trigger. During the flow reversal, a strong acoustic reflection occurs when the convected entropy perturbations reach the intake opening, which increase the blade loading significantly. During recovery, a hysteresis loop was recorded due to hot air re-ingestion, which led to a strong shear at mid-span of the inlet guide vane (IGV)/R1 domain and the formation of rotating helical flow structures. The final phase of recovery was accompanied by a four-cell multi-row tip rotating stall, which was cleared as the compressor recovered to the forward flow characteristic. It was also shown that the single passage model, despite its limitations and shortcomings in modeling recovery, can predict reasonably accurate transient flow features during surge and thus provide considerable insight to the flow behavior, which can be used to obtain a first approximation of casing and blade loading.

This paper presents the interaction between blade vibration and part-span rotating stall in a multi-stage high speed compressor. Unsteady aerodynamic and aeroelastic simulations were conducted using URANS CFD. Steady state computations showed short length scale disturbances formed local to the...

Reynolds-averaged Navier–Stokes (RANS) equations are employed for aerodynamic and aeroelastic modeling in axial compressors. Their solutions are highly dependent on the turbulence models for closure. The main objective of this work is to assess the widely used Spalart–Allmaras model suitability for high-speed compressor flows. For this purpose, an extensive investigation of the sources of uncertainties in a high-speed multi-stage compressor rig was carried out. The grid resolution near the casing end wall, which affects the tip leakage flow and casing boundary layer, was found to have a major effect on the stability limit prediction. Refinements in this region led to a stall margin loss prediction. It was found that this loss was exclusively due to the destruction term in the SA model.

This paper investigates the flow near the stall boundary for a low-speed/low-pressure ratio fan. Three-dimensional, Reynolds-averaged Navier–Stokes computations are performed for a modern low speed fan rig for which extensive measured data are available. Simulations are conducted at 80% corrected speed, for which the measured constant speed characteristic contains a part with positive slope. It is shown in this paper that by using an unsteady whole assembly approach, it is possible to predict the flow for all the points on the measured constant speed characteristic (including those on the positive slope part), which is not achievable by using a single passage strategy as it would result in premature “numerical stall.” The results of the computations reveal that for the operating points on the positive slope part of the characteristic, the flow structure becomes asymmetric and hence requires a whole assembly numerical model. The type of asymmetry which appears at lower flow coefficients is similar to the multicell, part span rotating stall, which can occur on the front stages of core compressors at stable operating conditions. The numerical results showed a good correlation with the measured data in terms of stall characteristics.

In this two-part paper the phenomenon of part span rotating stall is studied. The objective is to improve understanding of the physics by which stable and persistent rotating stall occurs within high speed axial flow compressors. This phenomenon is studied both experimentally (Part I) and numerically (Part II). The experimental observations reported in Part I are now explored through the use of 3D unsteady Reynolds-averaged Navier–Stokes (RANS) simulation. The objective is to both validate the computational model and, where possible, explore some physical aspects of the phenomena. Unsteady simulations are presented, performed at a fixed speed with the three rows of variable stator vanes adjusted to deliberately mismatch the front stages and provoke stall. Two families of rotating stall are identified by the model, consistent with experimental observations from Part I. The first family of rotating stall originates from hub corner separations developing on the stage 1 stator vanes. These gradually coalesce into a multicell rotating stall pattern confined to the hub region of the stator and its downstream rotor. The second family originates from regions of blockage associated with tip clearance flow over the stage 1 rotor blade. These also coalesce into a multicell rotating stall pattern of shorter length scale confined to the leading edge tip region. Some features of each of these two patterns are then explored as the variable stator vanes (VSVs) are mismatched further, pushing each region deeper into stall. The numerical predictions show a credible match with the experimental findings of Part I. This suggests that a RANS modeling approach is sufficient to capture some important aspects of part span rotating stall behavior.

In this two-part paper, the phenomenon of part span rotating stall is studied. The objective is to improve understanding of the physics by which stable and persistent rotating stall occurs within high speed axial flow compressors. This phenomenon is studied both experimentally (Part I) and through the use of unsteady RANS simulations (Part II). In this paper, the behavior of an eight stage high speed compressor is studied during slow acceleration maneuvres along a fixed working line. Casing mounted pressure transducers and rotor mounted strain gages are used to examine the spectral content of any unsteadiness in the flow and its behavior across the operating range. By deliberate aerodynamic mismatching of the front stages through adjustment of three rows of variable stator vanes (VSVs), stable rotating stall is initiated. The observed behavior falls into two “families” of high and low frequency when tracked on the instrumentation. Further analysis based on the Doppler shift between the static and rotating measurements confirms that these respective phenomena are due to rotating stall of high and low cell count. Acoustic modes resulting from stall/rotor interaction are also identified. Strong correlation of the stall intensity with simple 1D meanline predicted loading parameters suggests that these families of behavior are independently linked to the stalling of different regions within the compressor.