A labyrinth seal is commonly used to decrease the flow leakage loss between rotating and static components in aero engines. It is susceptible to aeroelastic instability because of its low stiffness. The aim of this study was to establish methods to predict and suppress it effectively. To achieve this, both numerical and experimental investigations are conducted using ansyscfx and ansys mechanical. These solvers are coupled to simulate the flutter precisely. Also, to assess the accuracy of the simulation qualitatively and quantitatively, a test rig is built. In the first part of this study, the accuracy of the numerical method is confirmed for several test cases with different seal clearance variations. Flutter inception is evaluated in detail for various pressure ratios and rotation speeds. The numerical results show good agreement with the experimental results. It is also confirmed that the aeroelastic instability is very sensitive to the seal clearance variations. These results show the same tendency as those in previous works. In the second part of this study, this paper tries to develop a flutter suppression method with higher leakage performance. This is achieved by changing the seal geometry. To detect the important geometric parameters, the contribution of each geometric component to aeroelastic instability is carefully analyzed. On the basis of this, the seal geometry is modified and its performance is evaluated. The optimized labyrinth seal shows good performance in terms of flow leakage and aeroelastic stability. Through this study, a new flutter suppression method is established.

A labyrinth seal is commonly used to decrease the flow leakage loss between rotating and static components in aero engines. It is susceptible to aeroelastic instability because of its low stiffness. The aim of this study is to clarify the physical mechanism of labyrinth seal flutter and to establish a method to predict and suppress it effectively. To achieve this, both numerical and experimental investigations are conducted using ANSYS CFX and ANSYS Mechanical. These solvers are coupled to simulate the flutter precisely. Also, to assess the accuracy of the simulation qualitatively and quantitatively, a test rig is built. In the first part of this study, the accuracy of the numerical method is confirmed for several test cases with different seal clearance variations. Both one-way and two-way fluid structure interaction (FSI) analyses are performed. Flutter inception is evaluated in detail for various pressure ratios and rotation speeds. The numerical results show reasonably good agreement with the experimental results. It is also confirmed that the aeroelastic instability is very sensitive to the seal clearance variations. These results show the same tendency as those in previous works [1–5]. In the second part of this study, this paper tries to develop a flutter suppression method without deteriorating the flow leakage performance. This is achieved by changing the seal geometry without changing the seal clearance variation. To detect the important geometric parameters, the contribution of each geometric component to aeroelastic instability is carefully analyzed. On the basis of this, the seal geometry is modified and its performance is also evaluated. The optimized labyrinth seal shows good performance in terms of flow leakage and aeroelastic stability. Through this study, a new method to suppress the flutter with low flow leakage is established.

Centrifugal compressors employed in the oil and gas industry are operated at high gas pressure conditions and are used in a wide operation range. Accurate prediction of the rotating stall and the destabilizing aerodynamic force is one of the key technologies for these compressors, because rotating stall can sometimes cause severe problems with subsynchronous shaft vibration and limit its operation range. Thus, the aim of this study is to establish a method of accurately predicting the inception of rotating stall and its effect on shaft vibration. To achieve this, numerical investigations are carried out by unsteady Reynolds-averaged Navier-Stokes (RANS) simulation with a full annulus model of the compressor stage. Also, to assess the accuracy of the simulation qualitatively and quantitatively, a high-pressure compressor test rig that contains a shrouded impeller and a vaneless diffuser is built. To investigate the effect of the rotating stall on the shaft vibration, an experiment is carried out at relatively high gas pressure with the inlet pressure level exceeding 30 barA. In the first part of the study, the accuracy of compressor performance prediction is studied by steady computational fluid dynamics (CFD) simulation. It is found that by taking the wall roughness effect into account, the predicted performance shows good agreement with the experimental result. Thus, a subsequent study of the rotating stall is also carried out by considering its effect. In the second part of the study, the accuracy of predicting the rotating stall is studied. In the experiment, two types of rotating stall are measured. One is a multiple-cell stall induced in the vaneless diffuser, for which the speed of rotation is relatively low and the other is a one-cell stall induced in the impeller region. The properties of the multiple-cell stall agree with the previous experimental and numerical studies, and the rotating stall has the limited effect on shaft vibration. Conversely, the one-cell stall shows severe subsynchronous vibration. In this study, both types of stall prediction are examined by CFD simulation. It is found that the simulation can predict the inception of the rotating stall with relatively high accuracy as the predicted results show good agreement with the experimental results in terms of cell count, rotation speed, pressure fluctuation level and the effect on shaft vibration. Through this study, the effectiveness of unsteady CFD simulation is confirmed for these types of stall and vibration prediction.

State-of-the-art axial compressors of gas turbines employed in power generation plants and aero engines should have both high efficiency and small footprint. Thus, in many cases, axial compressors are designed to have thin rotor blades and stator vanes with short axial distances, and are driven at high rotational speed. Recently, problems of high cycle fatigue (HCF) associated with forced response excitation have gradually increased as a result of these trends. Rotor blade fatigue can be caused not only by the wake and potential effect of the adjacent stator vane, but also by the stator vanes of two, three or four compressor stages away. Thus, accurate prediction and suppression method of them under the resonance condition are necessary in the design process. Concerning the forced response excitation associated with the adjacent stator vanes, there are many previous studies on simulating the vibration by fluid structure interaction (FSI) simulation. In these studies, the aerodynamic force acting on the blade is simulated by an efficient unsteady computational fluid dynamics (CFD) method such as the nonlinear harmonic (NLH) method. These methods can be available in commercial CFD solvers and can significantly reduce computational cost. However, there are few examples of the problems associated with the stator vanes from two and three compressor stages away and no efficient simulation method is available. In this study, the problem of rotor blade vibration caused by the stator vanes of two and three compressor stages away is studied. Ways to accurately predict and effectively control the vibration are also investigated. In the first part of the study, one-way FSI simulation is carried out using a full annulus CFD model. To validate the accuracy of the simulation, experiments are also conducted using a gas turbine test facility. The vibration level of the blade is measured using a blade tip timing (BTT) measurement system and the obtained results are compared with the simulated data. It is found that one-way FSI simulation can accurately predict the order of the vibration level. In the second part of the study, a method of controlling the forced response excitation is investigated by optimizing the clocking of the stator vanes. It is confirmed that by controlling the clocking of the stator vanes, the vibration amplitude can be effectively suppressed without reducing the compressor performance. Through this study, ways to evaluate and control the unsteady pressure force and vibration response of the rotor blade are validated. By optimizing the clocking of stator vanes, the blade vibration level can be effectively reduced.

The accurate prediction of high cycle fatigue (HCF) is becoming one of the key technologies in the design process of state-of-the-art axial compressors. If they are not properly designed, both rotor blades and stator vanes can be damaged. There are two main factors to cause HCF. One is low engine order (LEO) and the other is high engine order (HEO) excitation by fluid force associated with adjacent rotor-stator interaction. For the front stages of axial compressors for power generations and aero engines, the inlet Mach number of a rotor tip typically exceeds the speed of sound and strong shock waves tend to be induced. This can be the source of HEO excitation fluid force, and adjacent stator vanes are sometimes severely damaged. Thus, the aim of this study is to establish an efficient method for predicting the vibration response in this type of problem with high accuracy. To achieve this, numerical investigations are carried out by one-way fluid structure interaction (FSI) simulation. To validate the accuracy of FSI simulation, experiments are also conducted using a gas turbine engine for power generation. In the experiment, the vibration level is measured with strain gauges mounted on the surface of stator vanes and the data are compared with the predicted results. In the first part of the study, efficient prediction methods of excitation fluid force on the stator vane are investigated by time transformation (TT) and harmonic balance (HB) methods. Their accuracies are evaluated by comparing the results with those calculated by transient rotor stator (TRS) simulation whose pitch ratio is one between rotor and stator computational domains. It is found that the TT method can accurately predict the excitation fluid force with lower computation load even when there are pitch differences between rotor and stator regions. In the second part of the study, forced response analyses are carried out using the excitation fluid force obtained in the unsteady flow simulation. To obtain the total damping of the system, both hammering test and flutter simulations are carried out. Computed results are validated with experimental data and it is found that the predicted vibration level is in good agreement with experimental results. Through this study, the effectiveness of one-way FSI simulation is confirmed for this type of forced response prediction. By utilizing the combination of efficient unsteady computational fluid dynamics (CFD) methods and harmonic response analysis, vibration amplitude can be predicted accurately and efficiently.