Computational investigation of aeromechanical HCF effects in a compressor rotor

Paul King

Computational investigation of aeromechanical HCF effects in a compressor rotor

2001, 37th Joint Propulsion Conference and Exhibit

Abstract

High-Cycle Fatigue is a major problem facing the gas turbine industry today. It has been investigated by many researchers, using many different methods. Due to its highly complex nature, designers still do not have adequate tools to accurately predict the onset of high-cycle fatigue. A three-dimensional Navier-Stokes program was used to perform a study of the unsteady aerodynamics on a compressor rotor. The effect of aerodynamic detuning on the forced response of a rotor blade was compared to a baseline tuned rotor case. Detuning consisted of a ten percent decrease in circumferential spacing between alternate pairs of blades. The high-cycle fatigue effects of this detuning were investigated by examining the unsteady forces and moments on the rotor blades and inlet guide vanes. Computations were performed using a three-dimensional NASA research code (ADPAC) on a cluster of five desktop PCs. Computational times were on the order of several days for a grid of approximately 500,000 cells. These computations showed that detuning of the rotor blade could result in a reduction in the forced response of the IGV and rotor blades. This reduction came without much loss in overall performance (less than ten percent) and therefore may be a viable option to reduce high-cycle fatigue.

This research helps address one of the grand challenges of turbomachinery i.e., the accurate prediction of the forced response in multi-row compressors subjected to various crossings and operating points. Specifically, this focuses on understanding the impact of multi-row interaction on the unsteady aerodynamics and mistuned forced response behavior of a subsonic axial compressor. The phenomena of forced response remain one of the most challenging areas of turbomachinery aeromechanics. This thesis helps address some of the shortcomings in current literature related to unsteady aerodynamics and mistuned forced response predictions. The flow is inherently unsteady due to the complex flow field, blade row interactions, and secondary flows. Predicting the forced response behavior is a challenging task. Blade failures due to aeromechanical problems have resulted in fatalities and severe engine/aircraft damage, with some of the recent incidents being on Air France Flight 66 and Southwest Airlines Flight 1380. The experimental compressor studied herein is the Purdue 3.5 stage compressor, representing the rear stages of a modern high-pressure compressor (HPC). The focus of this research is on the vibratory response of rotor 2 (R2). One interesting feature of this configuration is that three rows have the same vane count i.e., the inlet guide vanes (IGV), stator 1, and stator 2. All contribute to the forcing function simultaneously. Also, the difference in blade count between the embedded R2 and the other rotors is the same. Computational data obtained using a commercial computational fluid dynamics (CFD) code, CFX, and an in-house mistuning response code MISER are compared against experimental data to understand the physical phenomena, determine the predictions' accuracy, and develop methods to improve predictions further. The first part of this research presents results from the torsional mode (1T) and a higher-order mode (1CWB) for the case where the stator count (44) of both neighboring stators is the same. Since both contribute to the forcing simultaneously, wake and potential field effects cannot be easily distinguished. The impact of physical wave reflection from downstream (Rotor 3) and the upstream influence from the IGV is also determined. The influence of operating conditions on the forcing function is also investigated. This is further fed into an in-house mistuning code, which predicts the response of all blades. The computational results are compared with experimental data. Finally, the effect of sideband traveling wave excitations (both amplitude and phase) on the blade response prediction was determined. The second part of the thesis deals with the study extended to a more realistic case in which the stator count of the embedded stators is different. Since the upstream and downstream influences are at different frequencies, we can separate the effects. This creates two torsional mode crossings (1T/44 and 1T/38) at different rotational speeds. Once again, the impact of operating conditions on the forced response behavior and the individual blade responses are determined. Further, this research contributes to the future development of model reduction methods and quantifies the error induced by utilizing model reduction techniques under different circumstances. The third section of the thesis deals with a configuration in which the stator is asymmetric i.e., has a different stator count on either side of the “split line.” The idea of having an asymmetric configuration originated in a NASA report [45] but has received little attention in the literature. Although current literature provides an insight into the steady aerodynamic performance of such configurations, no work to date explains the complex unsteady blade row interactions occurring in such configurations. This research describes the forcing function reduction phenomena due to asymmetry, provides general guidance on modeling techniques for such cases, and investigates possible scenarios and outcomes. The thesis then dives into determining the impact of stator hub cavities on the forced response prediction. Currently, the research on stator hub cavities only involves determining their influence on steady aerodynamics. The current work helps fill up the gap in the literature by determining its influence on unsteady aerodynamics and mistuned blade predictions. The fourth section discusses the impact of hub cavities on the steady flow in multiple locations around the blade passage and the impact of hub cavity flow on the unsteady aerodynamics, which determines the magnitude of the forcing function. The last chapter of the thesis quantifies the individual blade responses for all multi-row cases described in the previous sections. This section also discusses the impact of veering region modes and mode localization on the mistuned response prediction. The idea of perturbing the system mode frequency in a probabilistic manner was introduced in this thesis for the first time. Physical responses and dependencies have never been seen in the literature. The concept of strain energy-based mistuning models was expanded. For the first time in two decades, two new mistuning models were introduced, which were developed under the framework of the FMM. Also, the idea of perturbing structural damping in a probabilistic manner was introduced for the first time in this thesis. This thesis contributes extensively to understanding the various steady and unsteady aerodynamic interactions of multi-row configurations and some of the key findings are: 1. The impact of a downstream rotor (R3) cannot be neglected in forced response computations. The modal force prediction was within 10% accuracy, which was achieved by adding the downstream row. 2. The work also highlights the significance of having a downstream row that does not contribute to the forcing function at the same frequency but acts as a wall to reflect waves, contributing to the forcing. 3. The impact of spurious wave reflections on the forcing function was also quantified. In the absence of non-reflecting boundary conditions, these spurious waves can have a tremendous influence on the forcing function 4. The fidelity of model reductions techniques, particularly the time transformation method, is highlighted. This can not only serve as a guiding tool for the development of methods in the future but also reduce computational time significantly (3-4X reduction) 5. The impact of having an asymmetric stator exciting an embedded rotor was determined at multiple operating conditions. The benefit of asymmetry was limited to how the asymmetric stator excited the embedded rotor and not when any other rows excited the rotor. The asymmetry also results in the creation of sideband excitation responses, the magnitude of which is comparable to the dominant response. Also, the influence of stator hub cavities on the unsteady aerodynamic flow field was quantified, and the modal force prediction was found to improve by 10% for a 3-row case 6. Finally, the mistuned blade response was predicted using the modal forces obtained earlier, system modes obtained computationally, and blade frequencies obtained experimentally. This work contains several new insights into mistuned predictions. The mistuning work described here provides guidance, including sideband traveling wave excitations in the mistuning model. The thesis also introduced the concept of system mode and structural damping perturbations in a probabilistic manner, and the result was found to be deterministic. Several new plotting methods were introduced to represent data in a novel manner. Two new high fidelity strain energy-based mistuning models helped improve the blade response prediction and provided the most accurate date under the FMM framework. This work guides mistuning computations, including the effect of sideband excitations on mistuning parameters.

The paper presents the modelling of unsteady performance of a degraded 4-stage compressor and single stage gas generator turbine blades of PT6T turboshaft aero-engine of a helicopter. The two sections were set as control volumes for analytical and numerical modeling. Numerically, The blade specimens (NACA 65 series) were developed using SOLIDWORKS 20 and simulations performed with FLUENT in ANSYS 20.0. The RANS (Reynolds-averaged Navier-Stokes) equations with Shear Stress Transport model SST (k-w) were chosen for the unsteadiness of pressure and temperature distributions over different levels of reductions in surface area of the blades' pressure side. 900 x 10 3 mesh elements size were selected and the boundary conditionsinlets for the two control volumes were 295-325 K and 1083-1245 K for compressor and turbine respectively. Analytically, equations for different levels of degradations (surface area reductions) were developed to determine their flow performance at new pressure and temperature for compressor (and turbine (with change in time and the corresponding rise in centrifugal stress. Results from FLUENT predicts the performance of the sections for 10% surface area reduction with complex structure in the turbulent flow imposes high fatigue stress, hence shows the highest closeness to surge margin. For the compressor, the result emphasizes the impact of inlet conditions on degraded blades over exit conditions. Also in the turbine, velocity contour shows adverse/backward flow as a result of high turbulence formation and rising fatigue due to change in exit pressure flow from stage to stage in the compressor. This exit pressure determines the TIT in the turbine which is a function of efficiency of the single stage gas generator turbine and is crucial to the overall efficiency of the engine and the safety of the engine as a whole. In conclusion, the inter-component flow behaviour between the degraded compressor and turbine as revealed in this study shows the near real-life situation of the engine performance. Summarily, the accurate engine life estimation can be deduced from TIT rising from 1100-1200 K and centrifugal stress 60MN/mm 2 .

The work characteristics and loss-generation mechanism of a single-stage axial flow compressor in windmilling operation were investigated via experiments and computational fluid dynamics analyses. The windmilling state occurs when air flowing through an unlit engine drives the compressor rotor blades, similar to a turbine. This phenomenon applies mostly to aircraft engines, where it is caused by ram pressure. When the inlet flow coefficient is gradually increased in the design, the rotor blades gradually enters the windmilling operation from the tip toward the hub. This research has focused on two windmilling operations: free windmilling (FW) and highly loaded (HL) windmilling. In the case of FW, the net work performed by the rotor blades to the fluid is canceled out (zero), and the rotor is in an idle state. In the HL windmilling condition, the work performed to the rotor blades by the fluid increases, the compressor acts as a turbine, and power is generated. According to the detailed numerical results, the total-pressure loss under the free and HL windmilling conditions was mainly caused by three flow structures: (1) tip leakage flow from the suction surface (SS) to the pressure surface (PS) near the leading edge and that from the PS to the SS near the trailing edge; (2) the interaction of leading-edge separation vortices due to the highly negative incidence and the rotor leading-edge vortex; and (3) the boundary-layer separation near the hub wall. Surface-pressure measurement on a rotating rotor blade revealed that the distribution of the rotor operating mode existed not only in the spanwise direction but also in the chordwise direction under the windmilling operations. The turbine mode region was observed near the leading edge, while the compressor mode region was observed near the trailing edge, even in the HL windmilling condition. Therefore, the driving force of the windmilling was dominated not by the area of the turbine mode on the rotor surface but by the strength of the operating mode, i.e., the static-pressure difference between the SS and PS on the rotor. Finally, the unsteady flow field within blade-to-blades passages was investigated via an unsteady detached eddy simulation, and the differences in the loss-generation mechanism between the FW and HL windmilling conditions were examined. : Free windmilling : Highly loaded

A full two-way coupling aeroelasticity analysis of rotating machinery and their main components requires considerable computational analysis and central processing unit time. This is why the most common methodology for this type of problem is a one-way coupling between fluid and structure. In this case, forces acting on rotor blades that result from a flow are introduced into a structural model of a blade as a local spot load in the centre of gravity or a series of local spot points deployed along the blade length. However, this method does not take into account how the forces change in a chordwise direction (along chord). An alternative way is to use the method of directly transferring the pressure field as a surface load from the computational fluid dynamics analysis to a structural model of a blade. In this case, various interpolation methods are needed to transfer the results from the computational fluid dynamics mesh on to the structural mesh nodes. In this paper, the authors compare how both methods of load transfer affect rotor blade stress and displacement levels during one period of rotation. Calculations were carried out for the first-stage compressor blade of an aircraft engine. Ansys 12.1 was used to calculate the entire structure. Unsteady computational fluid dynamics calculations were carried out using Fluent for a 1.5-stage axial compressor model. For the numerical calculations a nonviscous flow was used. The unsteady forces were calculated on 10 control cross-sections of a rotor blade. The transient results obtained from the computational fluid dynamics calculations were transferred onto a structural rotor blade model using APDL -an ANSYS language script. For both methods of load transference, transient displacements and transient stresses for the rotor blade were calculated. The harmonic analysis results were compared. Mesh sensitivity analysis was also carried out for the structural model. In the last part of the paper, based on described above methodology a study involved to bird strike was shown in the way of how foreign object debris affects the dynamic stress level of a rotor blade. A comparison between an undisturbed engine inlet and one with an ingested foreign object was carried out. The analysis focused on the first-stage compressor blade of an aircraft engine with a partially blocked radial inlet and will be based on pressure method transferring described earlier.

Log In

Computational investigation of aeromechanical HCF effects in a compressor rotor

Sign up for access to the world's latest research

Abstract

Related papers

Related topics