Band 30, Heft 1

The Neural Networks is mostly used to engine fault diagnostic system due to its good learning performance, but it has a drawback due to low accuracy and long learning time to build learning data base. This work builds inversely a base performance model of a turboprop engine to be used for a high altitude operation UAV using measuring performance data, and proposes a fault diagnostic system using the base performance model and artificial intelligent methods such as Fuzzy and Neural Networks. Each real engine performance model, which is named as the base performance model that can simulate a new engine performance, is inversely made using its performance test data. Therefore the condition monitoring of each engine can be more precisely carried out through comparison with measuring performance data. The proposed diagnostic system identifies firstly the faulted components using Fuzzy Logic, and then quantifies faults of the identified components using Neural Networks leaned by fault learning data base obtained from the developed base performance model. In leaning the measuring performance data of the faulted components, the FFBP (Feed Forward Back Propagation) is used. In order to user's friendly purpose, the proposed diagnostic program is coded by the GUI type using MATLAB.

Unsteady numerical simulations have been conducted to investigate the effect of axial spacing between the stator vanes and the rotor blades on the performance of a transonic, single-stage, high-pressure, axial turbine. Three cases were considered, the normal case, which is based on the geometry of a commercial jet engine and has an axial spacing at 50% blade span equal to 42% of the vane axial chord, as well as two other cases with axial spacings equal to 31 and 52% vane axial chords, respectively. Present interest has focused on the effect of axial gap size on the instantaneous and time-averaged flows as well as on the blade loading and the turbine performance. Decreasing the gap size reduced the pressure and increased the Mach number in the core flows in the gap region. However, the flows near the two endwalls did not follow monotonic trends with the gap size change; instead, the Mach numbers for both the small gap and the large gap cases were lower than that for the normal case. This Mach number decrease was attributed to increased turbulence due to the increased wake strength for the small gap case and an increased wake width for the large gap case. In all considered cases, large pressure fluctuations were observed in the front region of the blade suction side. These pressure fluctuations were strongest for the smaller spacing. The turbine efficiencies of the cases with the larger and smaller spacings were essentially the same, but both were lower than that of the normal case. The stator loss for the smaller spacing case was lower than the one for the larger spacing case, whereas the opposite was true for the rotor loss.

In this study, aerodynamic and structural design of the composite propeller blade for a regional turboprop aircraft is performed. The thin and wide chord propeller blade of high speed turboprop aircraft should have proper strength and stiffness to carry various kinds of loads such as high aerodynamic bending and twisting moments and centrifugal forces. Therefore the skin-spar-foam sandwich structure using high strength and stiffness carbon/epoxy composite materials is used to improve the lightness. A specific design procedure is proposed in this work as follows; firstly the aerodynamic configuration design, which is acceptable for the design requirements, is carried out using the in-house code developed by authors, secondly the structure design loads are determined through the aerodynamic load case analysis, thirdly the spar flange and the skin are preliminarily sized by consideration of major bending moments and shear forces using both the netting rule and the rule of mixture, and finally, the stress analysis is performed to confirm the structural safety and stability using finite element analysis commercial code, MSC. NASTRAN/PATRAN. Furthermore the additional analysis is performed to confirm the structural safety due to bird strike impact on the blade during flight operation using a commercial code, ANSYS. To realize the proposed propeller design, the prototype blades are manufactured by the following procedure; the carbon/epoxy composite fabric prepregs are laid up for skin and spar on a mold using the hand lay-up method and consolidated with a proper temperature and vacuum in the oven. To finalize the structural design, the full-scale static structural test is performed under the simulated aerodynamic loads using 3 point loading method. From the experimental results, it is found that the designed blade has a good structural integrity, and the measured results agree well with the analytical results as well.

The present work is a study of the effects of rotor blade scaling of a single-stage high pressure turbine on the time-averaged turbine performance and on parameters that influence vibratory stresses on the rotor blades and stator vanes. Three configurations have been considered: a reference case with 36 rotor blades and 24 stator vanes, a case with blades upscaled by 12.5%, and a case with blades downscaled by 10%. The present results demonstrate that blade scaling effects were essentially negligible on the time-averaged turbine performance, but measurable on the unsteady surface pressure fluctuations, which were intensified as blade size was increased. In contrast, blade torque fluctuations increased significantly as blade size decreased. Blade scaling effects were also measurable on the vanes.

Turbomachines are exposed to many factors which may cause failure of its components. One of these, high cycle fatigue, can be caused by blade flutter. This paper evaluates the use of an air injection nozzle as a means of exciting vibrations on the first stage rotor blades of a rotating axial compressor. Unsteady simulations of the excitation velocity perturbations were performed on the Computational Fluid Dynamics (CFD) software, Numeca FINE™/Turbo. Experimental testing on a three-stage, low Mach number axial flow compressor provided data that was used to implement boundary conditions and to verify certain aspects of the unsteady simulation results.

The effect of the small scale turbulence on the development of the round jet is experimentally studied using hot wire anemometry. The small scale turbulence is generated by placing a grid having grid spacing near the nozzle exit. Experiments are carried out for the range of Reynolds numbers (4000, 6000 and 8000). It is observed that the small scale structures present in the jet slows down the jet mixing characteristics and there by retards the decay rate in the near-field region. It is observed that in the grid disturbed jets, the ratio of the turbulent kinetic energy to mean-flow kinetic energy is relatively lower than that of the undisturbed jet without any grid. From the spectral analysis, it is observed that the grid disturbed jets has higher energy content than that of the undisturbed jets.

A numerical investigation of shock wave control over a highly-loaded transonic compressor blade is presented in the current work. Two types of active flow control methods, namely blowing and suction, are employed. Their effectiveness in controlling the shock wave boundary layer interaction in the intended cascade passage has been investigated numerically. The three-dimensional RANS equations in Cartesian coordinate system are solved using the cell-centred control volume approach. Blowing and suction at four locations along the blade are simulated. Aerodynamic performance is evaluated for each case in terms of total pressure recovery coefficient, deviation angle, etc. Results show that blowing is not capable of providing desired control outcome, as the effects can only be seen in the neighbourhood of the blowing slot. The reconfiguration of shock waves in cascade is realized by suction through the slots located downstream after shock waves. Both the right branch of leading edge shock and the passage shock wave move downstream toward the suction slot. Compared with the baseline case, the suction with a suction ratio of 1.2% from slot D (at the 51% Chord) increased the total pressure recovery coefficient by 10%, and decreased the deviation angle by 5°.

In order to study the flow field in the impeller in the accelerating start-up process of centrifugal compressor, the 3-D and 1-D transient accelerated flow governing equations along streamline in the impeller of the centrifugal compressor are derived in detail, the assumption of pressure gradient distribution is presented, and the solving method for 1-D transient accelerating flow field is given based on the assumption. The solving method is achieved by programming and the computing result is obtained. It is obtained by comparison that the computing method is met with the test result. So the feasibility and effectiveness for solving accelerating start-up problem of centrifugal compressor by the solving method in this paper is proven.

This paper deals with the numerical simulation of sonic underexpanded axi-symmetric jets, issuing from a convergent nozzle, at different nozzle pressure ratios. The simulations have been carried out by solving the unsteady Reynolds Average Navier-Stokes (RANS) equations utilizing the realizable k - ε (epsilon) two equations turbulence model in the commercial software FLUENT (version 6.3.26). The flow is assumed to be axi-symmetric and turbulent. The computational domain included a convergent nozzle of length 50 mm and a rectangular domain of size 25 times the nozzle exit diameter (D) verses 4D, which have been modeled and meshed in GAMBIT (version 2.46). The grid consisted of 81020 cells, while gradient grid adaptions were performed later to pronounce the accuracy of the results and visualize the shock-cell structures. From the simulations, it is inferred that, for correctly expanded jets, the potential core extends upto 6D, where D is the exit diameter of the nozzle. For underexpanded jets, the shock-cell structures in the jet flow field were captured properly. A pattern of ‘shock-diamonds’ or shock-cells were observed in the flow at low NPR. The formation of shock-cell is found to change considerably and also its strength increases with increase in nozzle pressure ratio (NPR). From this study, it is clear that the effect of favorable pressure gradient in the underexpanded jet influences the jet characteristics significantly. At high NPRs, ‘Mach disc’ formation is observed in the flow field. The flow after the Mach disc becomes subsonic and regains energy from the surrounding fluid.

In this paper the flow near blade tip of gas turbines was investigated for blowing ratios M = 0, 0.9, 1.7 and 3, and gap 0.3, 1 and 1.5 percent of blade span. The structure of the flow near blade tip was identified using PIV. Velocity profiles from vertical PIV planes around the blade show the effects of blowing ratio on flow field distribution. Vortex flows were observed in the pressure side of leading edge for M = 0 or larger. Instead, the trailing edge region has high velocity magnitude but attached to blade wall. Numerical simulations were conducted in order to investigate the effects of varying the tip gap. Results show that leakage flow in suction side of trailing edge does not perturb too much the flow structure as in leading edge. It is observed that part of cooling flow redistributes as leakage from the tip cavity towards the suction side as expected. However, leakage on pressure side with high momentum rates was found as well, mainly on leading edge. It is confirmed that leakage mass flow distributes unevenly through blade chord as a function of the tip-gap size, cooling flow and passage conditions of the mainstream with maximum value on trailing edge. The paper shows that increments of blowing ratio reduce the contribution of coolant to leakage flow.

This study systematically investigates the slip phenomena in the centrifugal air compressor impellers by CFD. Eight impeller blades for different specific speeds, wrap angles and exit blade angles are designed by compressor design software to analyze their flow fields. Except for the above three variables, flow rate and number of blades are the other two. Results show that the deviation angle decreases as the flow rate increases. The specific speed is not an important parameter regarding deviation angle or slip factor for general centrifugal compressor impellers. The slip onset position is closely related to the position of the peak value in the blade loading factor distribution. When no recirculation flow is present at the shroud, the variations of slip factor under various flow rates are mainly determined by difference between maximum blade angle and exit blade angle, Δβ max-2 . The solidity should be of little importance to slip factor correlations in centrifugal compressor impellers.