This application claims priority to German Patent Application DE 102007016369.1 filed Apr. 3, 2007, the entirety of which is incorporated by reference herein.
This invention relates to a method for determining blade mistuning on integrally manufactured rotor wheels for compressors and turbines, in particular for gas-turbine engines, in which the individual blades are pulse-excited one after the other in a condition largely decoupled from the other blades, with the vibratory response of the respective blade being measured.
In compressor and turbine engineering, in particular in the aircraft engine sector, integrally formed rotor wheels, termed blisks, are increasingly used. Compared with the conventional design, integrally formed rotor wheels are capable of higher rotational speeds, thus providing for improved pressure ratio and more powerful engines. Integrally manufactured rotor wheels are increasingly employed also on compressors and turbines outside the aircraft sector. However, the one-piece design gives rise to production-inherent disturbances of the rotational symmetry, i.e. unbalance of the entire structure. This peculiarity of blisk-type rotor wheels, which is also referred to as mistuning, is reflected in specific blade natural frequencies and entails vibration amplitude and strain excesses in the blades due to aerodynamic excitation, resulting in fatigue affecting the life of the rotor wheel. Therefore, a determination of the blade natural frequencies—as closely as possible to reality—enables production-inherent disturbances of the rotational symmetry of the rotor wheel, i.e. mistuning of the entire structure, to be identified and corresponding statements on the quality of the respective blisk to be made regarding its vibratory behavior under operating conditions as well as its operational safety and service-life.
U.S. Pat. No. 7,082,371 B2 describes a method for determining mistuning of rotating blade structures, more particularly for predicting the vibratory behavior of integrally manufactured rotor wheels (blisks) of gas turbines. A software is capable of describing normal modes and natural frequencies of the blisk, actually by using tuned system frequencies and the frequency mistuning of each blade/blade sector. In this method, which is known as FFM (Fundamental Mistuning Model), normal, i.e. untuned, modes and natural frequencies of the respective blisk are used to determine sector frequencies as well as tuned system frequencies. FMM is capable of predicting how much the blisk will vibrate during rotation under operating conditions. The method includes obtaining nominal frequencies of a tuned blisk system, measuring of at least one untuned mode and calculating mistuning of at least one blade of the blisk on the basis of the nominal frequencies and of said, at least one, untuned mode and of the natural frequency. The quantities required for calculation can advantageously be obtained on a system in free, unchanged condition. However, FFM is only suitable for strongly coupled systems with a low number of blades. If the number of blades is high, natural modes corresponding to the number of blades will occur within a frequency band of a few Hz which conflict with a sufficiently precise separation of modes or require separation by extremely costly methods, respectively.
In another known method for determining mistuning of blisks, or predicting airfoil vibrations of a blisk during operation (Sever, I.A.: Experimental Validation of Turbomachinery Blade Vibration Predictions, Ph.D.-thesis, University of London (Imperial College), 2004) the disk is clamped between solid steel blocks. Additionally the blades which are currently not measured, are also restrained with considerable investment to reduce the coupling effects between the blades during excitation and vibration measurement and enable the natural frequency of the respective blade to be actually measured. According to yet another investigation into the strength design of bladed disks (Beirow, B. et. al. Experimental and numerical investigations regarding strength design optimization of bladed high-pressure compressor disks considering mistuning effects (Experimentelle und numerische Untersuchungen hinsichtlich einer Festigkeitsauslegungsoptimierung von Hochdruckverdichterschaufelscheiben unter besonderer Berücksichtigung von Mistuningeffekten), DGLR-Kongress 2003, München 2003, ISSN 0700-4083), only the disk is restrained, and the individual blades of the blisk are excited one after the other by pulse excitation, with the vibratory response being simultaneously measured by a contactless method. Thus, a relatively good identification of relative (related) mistuning distributions on blisks is principally possible on fully or partly decoupled rotor wheels in blisk design. However, with decoupling being incomplete, definite vibratory responses which correspond to the actual conditions in operation are not obtainable. Apart from the fact that such restraint incurs considerable effort and investment, the above methods can only be used on rotor wheels to which such restraint can actually be applied, i.e. where it does not conflict with blisk geometry, for example in the case of a high number of blades or with the installed state of the rotor wheels.
In a broad aspect, the present invention provides a method of the type specified at the beginning for determining blade mistuning of blisk-type rotor wheels which, irrespective of rotor wheel design and also in the installed state of the rotor wheel, ensures adequate decoupling between the respective blade under investigation and the disk as well as the other blades, with low effort, thereby providing for a definite, realistic vibratory response of the respective blade excited and, finally, permitting a definite statement to be made on the blade natural frequency distribution and, thus, rotor wheel mistuning.
In other words, the basic idea of the present invention is that, in a method, the blades of the integrally manufactured rotor wheel—except for that blade that is excited for measurement of the vibratory response—are further mistuned by temporarily fitting each with an identical extra mass. Thus, the vibration frequency of the blades which are not under investigation is shifted from the natural frequency-near range of the blades, thereby essentially avoiding disturbing or falsifying coupling effects in the vibratory response of the blade under investigation. In this way, the natural frequency of the individual blades, and thus blade natural frequency distribution on the rotor wheel and mistuning of the entire structure, are realistically determinable. Amplitude and strain increases in operation involved with unavoidable, production-inherent mistuning are detectable beforehand, enabling reliable statements regarding the problem of fatigue as well as service-life and operational safety of the blisk to be made.
In accordance with a further important feature of the present invention, a specific extra mass is required to obtain additional blade mistuning, with the size of this mass depending on the design blade mass, the design blade frequency of the blade modes and the expected magnitude of blade mistuning.
In a further embodiment of the present invention, transfer functions are determined on the basis of the measured vibratory response signals for each blade and mode family from which the maximum frequencies pertaining to the respective amplitude maxima can be read. From these, an average value will be calculated which finally serves as reference for representing the blade natural frequency distribution or mistuning of the rotor wheel, respectively.
In an advantageous development of the present invention, the average values, which are calculated slightly too small, are correctable in dependence of blade stiffness and design blade mass using a mass-spring-damper model as an equivalent rotor wheel model.
In an embodiment of the present invention, the extra mass is provided in the form of small metallic cylinders fixed with beeswax to corresponding locations on the blades which are not under investigation. The impact pulse for blade excitation is in each case produced at the same location and under the same angle using a miniature hammer. The vibratory response of the excited blade is recorded contactless by laser vibrometry.
The method is applied to blisk-type compressor and turbine rotor wheels, in particular compressor rotor wheels of aircraft gas turbines, actually in the removed state in virtually free condition, or also in the installed state, of the rotor wheel.
The present invention is more fully described in the light of the accompanying drawings showing a preferred embodiment. In the drawings:
As shown in
The size of the extra mass Δm (g) is determined in dependence of the design blade mass mb (g), the design blade natural frequency of the ith blade mode f0.i (Hz) and the expected magnitude of blade mistuning, i.e. the proportional mistuning span of the ith blade mode Δfspan=Δfmax+Δfmin (%), with Δfmax=(fi,max−f0.i)/fi,max (%) being the positive maximum of mistuning and Δfmin=(fi,min−f0.i),/fi,min (%) being the negative minimum of mistuning of the respective ith blade mode, established as follows:
This is a conservative estimate since the effectively co-vibrating blade mass assumes a smaller value compared to the actual design blade mass mb.
A transfer function is now calculated from the force input signals and the vibratory response signals for each of the, for example, 80 blades of the blisk 11 of a high-pressure compressor and each family of modes (1st bending, 2nd bending, 1st torsion, 2nd torsion etc.). The maximum frequencies fi pertaining to the amplitude maxima (
To illustrate the disadvantages of the state of the art,
Since the average values fi,mean calculated in the above for the representation of the related blade frequency distribution are calculated slightly too small due to unavoidable coupling effects and the extra masses used, correction of the average value to even more accurately represent the blade natural frequency distribution is provided by using a mass-spring-damper model. (Beirow, B. et. al. Localisation phenomenon on high-pressure compressor rotor disks (Lokalisierungsphänomen bei Hochdruckverdichterschaufelscheiben) VDI-Reports No. 168, 2006, ISBN 3-18-091968-X). Such an equivalent blisk model (EBM) is depicted in
The distribution of the blade stiffnesses kb,i adapted in this way, together with the known design blade mass mb, defines the blade natural frequency distribution of the blisk with regard to the family of modes under examination via:
which simultaneously corresponds to the model specification of the last iteration step. With the blade natural frequency distribution so determined, blade mistuning is described with sufficient accuracy.
1 Rotor wheel
2 Foamed plastic mat
4 Blades with extra mass
4.1 Blades without extra mass
6 Blade leading edge
7 Extra mass
8 Control unit
9 Impulse hammer
10 Single-point laser vibrometer
Δm Size of extra mass
mb Design blade mass
f0.i Design blade natural frequency
Δfspan Expected blade mistuning
fi,max Positive maximum of mistuning
fi,min Negative minimum of mistuning
kb,i Blade stiffness
Number | Date | Country | Kind |
---|---|---|---|
10 2007 016 369.1 | Apr 2007 | DE | national |