METHOD FOR DETERMINING AT LEAST ONE MECHANICAL PARAMETER OF A MECHANICAL OBJECT

DE502020011185D1Active Publication Date: 2025-06-18MTU AERO ENGINES GMBH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
DE502020011185
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Priority Date
2019-12-03
Filing Date
2020-12-01
Publication Date
2025-06-18
Estimated Expiration
2040-12-01

AI Technical Summary

Technical Problem

Accurately measuring and determining mechanical parameters, such as damping and stiffness, of integrally manufactured paddle wheels, which exhibit complex vibration modes and time-variant properties due to mistuning and aerodynamic effects.

Method used

A method involving two excitation stages: first, using a multi-frequency acoustic sweep signal to measure the system response, and second, selectively exciting specific frequencies with a sine signal to determine mechanical parameters like damping and stiffness based on amplitude and phase responses.

Benefits of technology

This method allows for precise determination of mechanical parameters, effectively addressing the challenges of mistuning and time-variant properties in integrally manufactured paddle wheels, thereby improving the accuracy of vibration analysis and reducing fatigue-causing vibrations.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader
Need to check novelty before this filing date? Find Prior Art

Description

Technical field

[0001] Embodiments are disclosed below which relate to a method for determining at least one mechanical parameter of a mechanical object, for example a stiffness or a vibration behavior of an integrally manufactured impeller. background

[0002] The measurement of mechanical objects, e.g., compressor wheels and / or turbine wheels for turbomachinery, can traditionally be performed by applying an excitation to the object, which triggers a mechanical response from the object that can be measured as a time-dependent signal. To determine the mechanical properties of the object, e.g., its mass, damping, and / or its natural frequencies and vibration modes, the time signal can be converted to the frequency domain, thus generating a corresponding transfer function. This transfer function can then be used as a model to provide information about the mechanical properties of the object and can be simulated. A measuring device for such a method is disclosed in DE 102009010375 A1.

[0003] Blade wheels, e.g., compressor wheels and / or turbine wheels for turbomachinery, were previously manufactured from a wheel disk and specially manufactured blades. The blade wheels were positively connected to a wheel hub, usually by a dovetail joint. Recently, blade wheels have also been manufactured integrally. Integral blade wheels (also called "blade-integrated disks" or "blisks") have a one-piece construction. The positive connection between the wheel hub and blades is eliminated, e.g., by milling the blades and hub from a single blank. In contrast to non-integrally manufactured blade wheels, blisks generally exhibit lower and therefore more difficult to determine damping, since no energy dissipation through friction can occur at the positive contact surfaces between the blades and wheel hub.

[0004] Differences between individual blades result in resonance passages appearing differently for each blade. This property is also referred to as mistuning. Causes of mistuning in blade wheels can be inherent in the manufacturing process, for example, caused by a tool wearing during blade milling. Material inhomogeneities are also conceivable causes of mistuning. In system dynamics, mistuning can be defined by the fact that the natural frequencies of the individual blades differ from one another.

[0005] In addition to these inhomogeneities, paddle wheels exhibit a multitude of different natural vibration modes, which can be excited to vibrations during operation depending on the design and the interaction of the paddle wheel components.

[0006] In general, the natural frequencies and natural vibration shapes of a blade wheel can also change depending on the rotational speed, aerodynamic pressure, and / or temperature of a mechanical object, especially a blade wheel. For example, the natural frequencies of the blades typically increase with increasing rotational speed of the blade wheel. Such a system is therefore time-variant during operation.

[0007] Measuring paddle wheels using the purely frequency-based methods mentioned above can lead to unsatisfactory results. To achieve the required low failure rates for these objects, precise determination of the mechanical parameters is necessary, especially to minimize fatigue-causing vibrations. Description

[0008] This gives rise to the problem of accurately measuring and determining mechanical parameters of a mechanical object, in particular the determination of damping and stiffness of a paddle wheel as well as the position and amplitude of resonance frequencies that can be excited at a paddle wheel.

[0009] This problem is solved by the invention defined by the subject matter of independent claim 1. The dependent claims relate to corresponding developments. Various aspects and embodiments of these aspects are disclosed below, providing additional features and advantages.

[0010] Some embodiments solve the problem that the blades of a blisk were manufactured with an unknown detuning, making it unclear which resonances the blisk has that must be avoided during operation. For this purpose, the blades are mounted in a stationary position of the blisk and excited by an acoustic sweep signal in a first measuring stage. The system response is measured at selected measuring points on the blades of the blisk and by laser vibrometry. From the information obtained, specific frequencies are selected, which are then acoustically excited with a sine signal in a second measuring stage. From the amplitude and phase responses recorded in this way, the mechanical parameters of damping and / or stiffness of the various blades and / or the forces on the blades generated by the acoustic signal are identified on a model-based basis.

[0011] A first aspect concerns a procedure comprising the steps: First, exciting the object with a multi-frequency signal; detecting a first response signal of the object at one or more measurement points on the object; transforming the first response signal from a time domain to a frequency-dependent domain; selecting one or more frequencies based on the frequency-dependent domain; secondly exciting the object based on the selected frequencies; detecting a second response signal of the object at one or more measurement points on the object; determining a mechanical parameter based on the second response signal.

[0012] A mechanical parameter can, in particular, be damping. Additionally or alternatively, a mechanical parameter can be stiffness. Additionally or alternatively, a mechanical parameter can be a natural frequency, a vibration mode, in particular a parameter of a vibration mode, for example a node circumference or a node diameter. Additionally or alternatively, a mechanical parameter can also represent a resonance frequency at which the object oscillates or vibrates more intensively when excited thereby. A mechanical parameter can comprise one or more of the individual mechanical parameters described above. In particular, a mechanical parameter can comprise several mechanical parameters of the same type that relate to different locations on the mechanical object. For example, a mechanical parameter can comprise stiffnesses of different blades of a blade wheel and / or stiffnesses under different operating parameters.A mechanical parameter can be a scalar, a vector or a tensor.

[0013] A mechanical object can in particular be a disk or a disk-shaped object, which can in particular be out of tune. In particular, a mechanical object can consist of a metal or of a plastic. In particular, a mechanical object can comprise a metal or a plastic. A mechanical object can in particular be a part of a turbomachine or a turbocharger. In particular, a mechanical object can be an impeller or comprise an impeller and in particular can be an impeller of a compressor stage or turbo stage of a turbomachine. In particular, a mechanical object can be a propeller or an impeller. However, a mechanical object can also be a musical instrument, in particular a percussion instrument, e.g. a tom-tom or a hi-hat.

[0014] A multifrequency excitation signal can, in particular, be a Dirac pulse or a step pulse, although it should be noted that a mechanical implementation of such a multifrequency signal results in a low-pass filtered Dirac pulse or a low-pass filtered step pulse. Alternatively, a multifrequency excitation signal can also be implemented using a sweep signal or a chirp signal. In another alternative, a multifrequency signal can be implemented using white noise or pink noise. In another alternative, a specially configured excitation signal can be used, which is defined, for example, in a frequency range. With the exception of a step signal, these input signals have the disadvantage that the excited mechanical object may have difficulty oscillating to these input signals.Thus, the measurement results for such input signals will not reflect a static state of the mechanical system. In particular, a multifrequency excitation signal can be configured to the highest node diameter of the mechanical object, so that it specifically excites this natural vibration.

[0015] The acquisition of a first response signal from the object at one or more measuring points on the object can, in particular, be preceded by a determination of the corresponding measuring points. In particular, this can be based on a finite element model of the mechanical object. For example, a finite element model of a paddle wheel or a blade can be created. By simulating the finite element model or by a corresponding simulation of the mechanical object, mechanical parameters of the mechanical object can be predetermined to a certain extent. For example, it is possible to identify the ranges in which damping or stiffness of the mechanical object can vary. Additionally or alternatively, it is possible to determine the range in which the mechanical object exhibits natural vibrations and / or which and / or how many natural vibration modes the mechanical object exhibits.In particular, the proper motion modes of a blade of a paddle wheel can be determined by simulation.

[0016] Based on the simulation results, measurement points can be distributed on the mechanical object. In particular, measurement points can be arranged at an exposed tip of the mechanical object, for example, at a corner or at the end of a blade of a fan wheel. This is because vibrations of most of the object's natural vibration modes have an effect on exposed ends of a mechanical object. In particular, one or more measurement points on the mechanical object can be arranged such that they fulfill an optimization criterion, in particular so that information obtained through the measurement is optimized. For example, such an optimization can be based on maximizing the Fisher information. Additionally or alternatively, the measurement points can be distributed such that linearly independent information can be observed for different modes.Accordingly, the number of measuring points can depend on the number of natural vibration modes determined in the simulation. For example, if seven natural vibration modes are determined, seven measuring points can be arranged on the mechanical object. A measurement can be carried out using a laser, for example a laser Doppler vibrometer. Additionally or alternatively, a measurement can be carried out using strain gauges arranged at the corresponding measuring points on the mechanical object. Additionally or alternatively, a measurement can be carried out contactlessly using imaging techniques, for example using a camera that observes the measuring points. Accordingly, the measuring points must also be arranged so that they can be observed by the sensor used.

[0017] An embodiment of the first aspect relates to a method, wherein the object is a paddle wheel, in particular one manufactured in an integrated manner, or one or more blades of a paddle wheel.

[0018] Such a blade wheel can, in particular, be an integrally manufactured blade wheel. Integrally manufactured blade wheels are also called "blade-integrated disks," or "blisks" for short. Mechanical parameters of a blade wheel can, in particular, relate to disk vibrations and / or blade vibrations and / or vibrations generated by a coupling between blades and disks. In particular, a mechanical parameter can also relate to a mode configuration consisting of linearly and circularly arranged nodes of a blade wheel. Nodes are the points that remain at rest during a blade or blade wheel oscillation. Mechanical parameters can, in particular, be time-variant. This can be the case, for example, for blade wheels whose stiffness increases with increasing speed and is thus dependent on a speed and thus, accordingly, dependent on time.A mechanical parameter can also relate to the detuning of a mechanical object, in particular a paddle wheel. Detuning is defined by different natural frequencies of the paddle wheel's blades. Paddle wheels are usually detuned because only in the ideal case is a paddle wheel completely cyclic or symmetrical. Such detuning can be caused in particular by production-specific time variations, for example by tool wear during the milling of an integrated paddle wheel or by a tool change or manual reworking. These are just a few of the causes that can lead to differences in the paddle wheel's blades. Accordingly, a mechanical parameter can in particular relate to a standard deviation of the paddle's natural frequencies.Additionally or alternatively, a mechanical parameter can relate to one or more absolute differences between different blades of a blade wheel. If the blade natural frequencies differ from one another, locally limited vibration amplitudes arise, for example, over the circumference of a blade wheel. These differences lead to additional Fourier coefficients in the spectrum of the individual blades. These additional coefficients or the different spectra of different blades can also form the basis for a mechanical parameter. Accordingly, a mechanical parameter can also relate to the mode fullness of a blade wheel. Additionally or alternatively, a mechanical parameter can also relate to a degree of localization, which, for example, relates the ratio of the largest blade amplitude to the root mean square of all blade amplitudes of a blade wheel.This mechanical parameter can be used to compare different paddle wheels with regard to their detuning.

[0019] According to the invention, the first excitation is carried out by a sound signal.

[0020] A first excitation with a multi-frequency signal can occur in particular through acoustic excitation. Additionally or alternatively, a first excitation can also occur through a mechanical unit, for example a modal hammer, with which, for example, a blade of a stationary airfoil is excited. Additionally or alternatively, a mechanical excitation can also occur through a rotation of the mechanical object, in particular a rotation of a blade wheel. In this case, excitation can occur through aerodynamic effects, in particular through turbulence caused by the blade wheels when they move in a fluid, in particular in a gas, for example air.

[0021] According to the invention, the multifrequency signal is a sweep signal or a chirp signal.

[0022] In the case of a sweep signal or chirp signal, a signal length can be ten seconds and a change in the corresponding pulse can be 1 kHz / s, so that a corresponding bandwidth of frequencies can be mapped per excitation signal.

[0023] A sweep pulse and / or chirp pulse can also be selected to target and excite a large node diameter, particularly the largest node diameter. This can be achieved, in particular, by appropriately parameterizing the pulse's phase position. Advantageously, the large node diameters of a paddle wheel often have a low coupling between the disk and the paddle wheel blades, thus improving the detection of a mechanical parameter of a blade and, in particular, allowing better detection of detuning.

[0024] According to the invention, the length of the sweep signal or the chirp signal is shorter than a mechanical settling time of the object.

[0025] In particular, the length of the sweep signal or chirp signal can be 10 ms and / or comprise a frequency interval of up to 10 kHz. Additionally or alternatively, the duration of a sweep pulse can depend on a node diameter family to be detected or on the number of node diameter families to be detected. In particular, the duration can be between 0.5 s and 3 s per node diameter family.

[0026] An embodiment of the first aspect relates to a method, wherein the frequency-dependent range into which the first response signal is transformed is time-dependent, speed-dependent and / or position-dependent.

[0027] Transforming the first output signal from a time domain to a frequency-dependent domain can, in particular, involve local or temporal resolution. For example, the first output signal can be transformed into a spectrogram using a short Fourier transformation. Additionally or alternatively, transformation can be performed using a wavelet transformation or a chirplet transformation. This advantageously allows, in particular, time-variant or position-variant aspects of the mechanical parameter of the mechanical object to be captured.

[0028] An embodiment of the first aspect relates to a method wherein the transformation of the first response signal is performed by a wavelet transformation or a chirplet transformation. Advantageously, these transformations are evaluated along a relationship between time and excitation frequency.

[0029] A wavelet transformation can be performed, in particular, based on a continuous wavelet transformation. Alternatively, a wavelet transformation can be performed based on a discrete wavelet transformation. Wavelets that can be used include, in particular, a Morlet wavelet, a Haar wavelet, a Meyer wavelet, or a Daubechies wavelet. For a paddle wheel, in particular, a wavelet transformation can be performed for different blades.

[0030] An embodiment of the first aspect relates to a method, wherein selecting one or more frequencies based on the frequency-dependent range maximizes information about an amplitude or phase response over the frequency-dependent range.

[0031] The selection of one or more frequencies based on the frequency-dependent range can, in particular, comprise a plurality of sampling points in a frequency-dependent range of the mechanical object or of a part of the mechanical object, in particular a blade of a paddle wheel. Additionally or alternatively, the sampling points can be set for different times, in which the time-dependent range is resolved, for one or more parts of the mechanical object, in particular for one or more blades of a paddle wheel. In particular, a number of sampling points can be selected such that a system of equations to be used later to determine the mechanical parameter is determined or overdetermined. In particular, the sampling points can be arranged in a frequency-dependent range such that they capture as much information as possible about this frequency response.In particular, sampling points can be selected where a first derivative or a second derivative of the frequency-dependent curve is high or maximum, respectively. Additionally or alternatively, sampling points can be selected where the frequency-dependent curve exhibits maximum values. This makes it possible, in particular, to capture a resonance characteristic of the mechanical object, in particular a resonance characteristic of a blade of a paddle wheel.

[0032] An embodiment of the first aspect relates to a method wherein a plurality of frequencies are selected such that a system of equations with respect to the mechanical parameter is determined or overdetermined.

[0033] After detecting a second response signal of the object at one or more measuring points of the object, wherein these measuring points may in particular differ from the measuring points when detecting the first response signal of the object, a selection of the information obtained can in particular be carried out.

[0034] In particular, the selection can be made with respect to specific regions of the mechanical object. If the mechanical object is a paddle wheel or a blade of a paddle wheel, a veering region or a crossing region, for example, can be selected as regions. Additionally or alternatively, the results can be selected with respect to specific frequencies, in particular because interesting vibration characteristics of the analyzed mechanical object can be observed at these specific frequencies. In particular, frequencies can be selected at which interesting node diameter families appear. For example, a node diameter family that appears particularly at a high frequency can be selected and, in addition, a node diameter family that appears particularly at a low frequency.

[0035] An embodiment of the first aspect relates to a method wherein the second excitation of the object is carried out on the basis of one or more sinusoidal excitation signals.

[0036] In particular, because the quality of the previous evaluations can still be improved, the second excitation of the object is carried out based on the formed support points. The second excitation can be carried out in a similar way to the first excitation and in particular also with a sweep signal or a chirp signal. Alternatively, the second excitation can be carried out in particular with a sinusoidal oscillation, for example with a sine oscillation or a cosine oscillation. This has the advantage that the system can stabilize in response to the excitation, so that a steady-state behavior of the mechanical object can be observed as a response of the mechanical object to the second excitation. By measuring the selected frequencies in the analysis carried out on the basis of the first excitation, the previously determined results can be analyzed better and with higher resolution.

[0037] An embodiment of the first aspect relates to a method, wherein determining the mechanical parameter comprises selecting the second response signals on the basis of which the mechanical parameter is determined.

[0038] A mechanical parameter can be determined from the second response signal, in particular, by adapting the parameters of a model, in particular a dynamic model. For example, the damping and stiffness parameters from a corresponding equation of motion for the blade of a blisk can be adapted such that the response of the equation of motion, when subjected to an excitation similar to the second excitation, produces the same system response, or at least a system response whose differences from the measured result have been minimized. Of course, the equation of motion can also be adapted depending on different rotational speeds. Additionally or alternatively, the excitation force acting on a blade can be calculated, in particular if mechanical parameters identified on one blade are transferred to another blade, and the equation of motion is solved accordingly on this basis.

[0039] The mechanical parameter can also be determined using a so-called reduced-order model. In particular, the model can incorporate additional operating conditions, such as the temperature prevailing during operation, the aerodynamic conditions prevailing during operation, and / or the forces prevailing during operation, particularly centrifugal forces. This can further increase the accuracy of the obtained mechanical parameter.

[0040] To implement the method according to the invention, a device can be configured to: to carry out a first excitation of the object with a multi-frequency signal; to record a first response signal of the object at one or more measuring points on the object; to transform the first response signal from a time domain into a frequency-dependent domain; to carry out a second excitation of the object based on selected frequencies; to record a second response signal of the object at one or more measuring points on the object; to determine the mechanical parameter based on the second response signal.

[0041] The device may further comprise further structures and functions in order to be able to implement the method steps disclosed above in the context of the first aspect.

[0042] The mechanical object can be supported by one or more springs with a given stiffness.

[0043] The suspension can be achieved, in particular, by springs from which the mechanical object is suspended. Additionally or alternatively, the suspension can be achieved such that the mechanical object is at least partially supported by the suspension. The stiffness created by the one or more springs can, in particular, be lower than the stiffness of the mechanical object. In particular, the difference between the stiffness of the suspension and the stiffness of the mechanical object can be so great that the mechanical object is practically floating, thus avoiding, in particular, coupling effects between the device and the mechanical object to be measured.

[0044] The first excitation and / or the second excitation can be carried out by a plurality of loudspeakers, each of which excites a part of the mechanical object, in particular one of several blades of a blade wheel, and wherein a calibration of a sound level and a phase for one or more loudspeakers is carried out successively.

[0045] In particular, the excitation can be in phase and / or with a predetermined phase shift between the individual loudspeakers. Short description of the characters

[0046] Further advantages and features will become apparent from the following embodiments, which refer to the figures. The figures do not always show the embodiments to scale. The dimensions of the various features may be enlarged or reduced accordingly, particularly for the sake of clarity of the description. The following shows, partially schematically: Fig. 1 a coupling diagram; Fig. 2a signal representation of a wavelet; Fig. 3 a frequency response of a mechanical object; Fig. 4 a coupling diagram; Fig. 5 a coupling diagram; and Fig. 6 a block diagram of a method according to an embodiment of the disclosure.

[0047] In the following descriptions, identical reference symbols refer to identical or at least functionally equivalent features.

[0048] In the following description, reference is made to the accompanying drawings, which form a part of this disclosure, and in which is shown by way of illustration specific aspects in which the present disclosure may be understood. It should be understood that other aspects and / or features may be utilized and functional, structural, or logical changes may be made without departing from the scope of the present disclosure. The following detailed description, therefore, is not to be taken in a limiting sense, since the scope of the present invention is defined by the appended claims.

[0049] In general, a disclosure about a described method also applies to a corresponding device for carrying out the method or a corresponding system comprising one or more devices, and vice versa. If, for example, a specific method step is described, a corresponding device may comprise a feature for carrying out the described method step, even if this feature is not explicitly described or illustrated in the figure. On the other hand, if, for example, a specific device is described on the basis of functional units, a corresponding method may comprise a step that carries out the described functionality, even if such steps are not explicitly described or illustrated in the figures. Likewise, a system may be provided with corresponding device features or with features for carrying out a specific method step.It is understood that features of the various exemplary aspects and embodiments described above or below may be combined with one another unless expressly stated otherwise. Description of the characters

[0050] The Fig. 1 refers to the initial excitation of a mechanical object. In this case, the mechanical object is an integrally manufactured paddle wheel. To better illustrate the complex vibration behavior of a paddle wheel, the coupling diagram can be used. Here, the natural frequencies of the paddle wheel are plotted against the corresponding number of node diameter lines. Fig. 1 shows such a coupling diagram 100 and, in addition, various node diameter families 102, each characterized by similar natural frequencies. Fig. 1The non-integer node diameters shown are interpolations that serve solely to clarify the position of a node diameter family. The natural frequencies considered range from 0-10 kHz. A total of seven node diameter families are shown. A first node diameter family oscillates at less than 1 kHz. A second node diameter family oscillates at approximately 2 kHz. Other node diameter families oscillate with natural frequencies between 2 kHz and 8 kHz. A top node diameter family oscillates at approximately 9 kHz. The integrally manufactured paddle wheel disc under consideration is excited by the method for recording at least one mechanical parameter with a first multi-frequency signal. This excitation occurs because the paddle wheel disc is stably mounted at rest.A loudspeaker is arranged beneath each blade of the paddle wheel disc as an acoustic actuator, which can emit the multi-frequency signal in such a way that it impinges on each blade of the paddle wheel disc in the same way and causes an initial excitation. To emulate a rotation of the paddle wheel disc, the loudspeakers are controlled with a corresponding phase offset, so that each loudspeaker emits a corresponding excitation for its blade with a phase shift, and thus the corresponding blades are also excited with a phase shift. The phase-shifted control is produced by a corresponding control unit, which also generates the multi-frequency signal. To avoid undesirable coupling effects between the blades or the paddle wheel disc and the device for recording the at least one mechanical parameter of the paddle wheel disc, the paddle wheel disc is mounted on three springs by the device.The springs are designed as spiral springs and are located below the paddle wheel disc, so that the paddle wheel disc rests its weight on the three springs. The springs have a combined stiffness that is significantly lower than the stiffness of the paddle wheel disc or that is expected of the paddle wheel disc and its blades. The first excitation with a multi-frequency signal 101 causes the paddle wheel disc to vibrate via its blades. In this case, the multi-frequency signal 101 represents a sweep pulse. The sweep pulse is generated from 3.84*10^6 sampling points in the time domain. The sweep signal sweeps through frequencies that span a bandwidth of 10 kHz from the lowest frequency to the highest frequency. Alternatively, the method was also implemented with an excitation signal of 22 s and a clock rate of 375 kHz with 80 loudspeakers for all blades of a paddle wheel.In contrast to the method disclosed in DE 102009010375 A1, the excitation is implemented not only with individual frequencies, but also with a continuous sequence of multiple frequencies as a continuous frequency sweep. Furthermore, the phase control is no longer realized with a delay device, but with a signal output device on which the phase-shifted and calibrated signals are statically stored. Calibration is performed sequentially for sound level and phase. Furthermore, instead of a single measurement step, the disclosed method comprises two steps with a first excitation and a second excitation to increase the accuracy of the recorded mechanical parameter.

[0051] The Fig. 2shows a signal curve 200 of a mother wavelet, with the aid of which one or more response signals resulting from the first excitation with a sweep signal are transformed into a frequency- and time-dependent domain. In this case, the mother wavelet represents a Morlet wavelet. With the aid of this, a wavelet transformation is carried out. The Morlet mother wavelet has a real part 201 and an imaginary part 202. An amplitude 203 of the wavelet is marked as an envelope. The signals processed by the wavelet transformation, which are recorded as response signals to the first excitation with the sweep pulse 101, are recorded laser vibrometrically at the individual blades and are determined at a plurality of measuring points. The measuring points are arranged at an upper tip of each blade. The measuring points are distributed such that previously simulated node diameter families can be recorded.

[0052] The Fig. 3reveals an amplitude response 300, i.e., an amplitude plotted against a frequency. The amplitude response 301 represents the vibration behavior of the third blade of the impeller at a rotational speed of 5000 revolutions / min and exhibits resonance. This rotational speed is reached when the impeller is started up after a defined operating time and represents a typical operating state in which the impeller operates. In addition to the amplitude response of the third blade of the impeller, an ideal amplitude response 302 is shown at the corresponding rotational speed. The detuning results, for example, from the difference between the ideal amplitude response and the amplitude response of the corresponding blade.

[0053] The Fig. 3further shows the selection of multiple frequencies 303 as sampling points from the measured frequency-dependent range. The selected frequencies are represented by vertical lines, each ending with a round point on the measured curve. Accordingly, the sampling points are selected for the third blade and at a speed of 5000 rpm. Additionally or alternatively, one or more sampling points can also be selected for other speeds. Fig. 3only shows the selection of the support points for the third blade. In this example, this step is carried out for all blades of the impeller in order to determine the differences in the mechanical parameters between the individual blades and thus the detuning of the impeller with respect to a node family. In particular, enough support points are required for a blade that the system of equations used later, from which the mechanical parameters are determined, is at least determined, i.e. enables a unique solution. In this case, these are two support points per blade and per node family. In particular, the same support points can be used for all blades. Alternatively, more support points can be selected, in particular three, four, five or ten, so that the system of equations is overdetermined.An overdetermined system of equations can be solved, for example, by the method of least squares, in particular by using the Moore-Penrose pseudoinverse, with respect to the mechanical parameter(s).

[0054] The Fig. 4 shows the coupling diagram 100 from Fig. 1 . In addition, Fig. 4 The selected frequencies 303 are now plotted along the Y-axis. They are located in the lower square on the right edge of the diagram. Furthermore, further mode families are shown in Fig. 4, which result from additional resonances not shown, which can also be represented with at least five support points each. A second excitation of the integrated paddle wheel is carried out with these frequencies. Unlike the first excitation, a single multifrequency signal is not used; instead, the paddle wheel or the blades are excited with individual monofrequency sinusoidal oscillations that only include the selected frequencies. The type of excitation and the experimental implementation remain the same as for the first excitation of the paddle wheel.

[0055] The Fig. 5 also shows the coupling diagram 100 from Fig. 1 or Fig. 4 After the paddle wheel has been excited a second time with the selected frequencies via sinusoidal pulses, frequencies or frequency ranges are selected a second time. These frequency ranges include, as shown in Fig. 5Shown are individual node diameter families 502. Using the corresponding amplitudes or phases at the selected frequency ranges, a model can now be calculated from which the desired parameters—in this case, the Lehr damping ratio, stiffness changes, damping properties, and the excitation—are determined. A so-called reduced-order code (ROC) can be used as the basis for the system of equations.

[0056] The Fig. 6shows an overview of the disclosed method according to one embodiment. The identification 601 of various mechanical parameters 602 of a mechanical object is carried out by means of a model-based adaptation 605 of corresponding parameters with regard to experimentally determined information 603 and calculated using an ROC 604. The parameter δλ represents a change in the stiffness of a blade, e.g., relative to an average value or to the stiffness of the previous blade. The parameter ξ represents a damping of a blade, e.g., according to Lehr's damping measure. The parameter ξ_m_B represents the average damping of the blades. This parameter relates to the average damping of the disk ξ_m_S. The parameter δξ_B indicates a deviation of the damping of a blade from ξ_m_B. The parameter f_B indicates the excitation for a blade or for each blade.A set of parameters for each blade of the impeller allows a comprehensive description of the impeller's detuning.

[0057] The system of equations thus includes the change in the stiffness of each blade. Furthermore, the system of equations includes the damping properties of the disk and / or the blades. The damping of each blade can be determined separately and / or as an average value. The system of equations also includes the excitation of the blades and / or the disk according to the excitation during the measurement. The excitation can be applied to one or more blades, in particular to each blade, and / or to the disk.

Claims

1. Method (600) for determining at least one mechanical parameter (δλ, ξ_m_B, ξ_m_S, δξ_B, f_B) of a mechanical object, having the steps of: - first excitation of the object with a multifrequency signal (101); - detecting a first response signal of the object at one or more measuring points on the object; - transforming the first response signal from a time range into a frequency-dependent range (300); - selecting one or more frequencies (303) on the basis of the frequency-dependent range; - second excitation of the object on the basis of the selected frequencies (303); - detecting a second response signal of the object at one or more measuring points of the object; - determining a mechanical parameter (δλ, ξ_m_B, ξ_m_S, δξ_B, f_B) on the basis of the second response signal, characterized in that the first excitation is carried out by a sound signal and the multifrequency signal is a sweep signal or a chirp signal, the length of the sweep signal or the chirp signal being shorter than a mechanical settling time of the object.

2. Method according to the preceding claim, wherein the object is a blade wheel, in particular one manufactured in an integrated manner, or one or more blades of a blade wheel.

3. Method according to either of the preceding claims, wherein the frequency-dependent range (300) into which the first response signal is transformed is time dependent, speed dependent and / or position dependent.

4. Method according to the preceding claim, wherein the transformation of the first response signal is carried out by a wavelet transformation (200) or by a chirplet transformation.

5. Method according to any of the preceding claims, wherein selecting one or more frequencies (301) on the basis of the frequency-dependent range maximizes information about an amplitude- or phase response over the frequency-dependent range.

6. Method according to any of the preceding claims, wherein a plurality of frequencies (301, 303) are selected such that a system of equations is determined or overdetermined with respect to the mechanical parameter (δλ, ξ_m_B, ξ_m_S, δξ_B, f_B).

7. Method according to any of the preceding claims, wherein the second excitation of the object is carried out on the basis of one or more sinusoidal excitation signals (301).

8. Method according to any of the preceding claims, wherein determining the mechanical parameter comprises selecting the second response signals (502) on the basis of which the mechanical parameter (δλ, ξ_m_B, ξ_m_S, δξ_B, f_B) is determined.