• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    Method and system for the detection of defects of railway axes in fatigue tests. The present invention relates to a method and system for detecting a defect in a railway axis during a fatigue test. The method comprises the steps of: acquiring, by means of at least one sensor, a signal of vibratory type of the railway axis; calculating, in a processor, the energy of the vibratory signal acquired in a given frequency band; compare, in the processor, the energy calculated with a previously established energy threshold; determine that a failure in the axis has been detected in case the calculated energy exceeds the pre-established energy threshold; and stop the fatigue test based on the detection of a defect. (Machine-translation by Google Translate, not legally binding)
    公开号:ES2575745A1
    申请号:ES201431966
    申请日:2014-12-30
    公开日:2016-06-30
    发明作者:Juan Carlos García Prada；Cristina CASTEJÓN SISAMÓN；María Jesús GÓMEZ GARCÍA
    申请人:Universidad Carlos III de Madrid；
    IPC主号:
    专利说明:

    5
    10
    fifteen
    twenty
    25
    30
    DESCRIPTION
    Method and system for detecting railway axle defects in fatigue tests
    Object of the invention
    The present invention has application in the field of defectology and the prevention of damages. More specifically, it has application in the premature detection of fissures during fatigue tests of railway axes and the possibility of establishing a stop criterion in a fatigue testing machine.
    Background of the invention
    Currently there are many standards regulating the design, manufacture and maintenance of railway axles because the axles are one of the most important elements in a railroad vehicle in terms of safety.
    Prior to the assembly of an axle on a railway machine, it has been subjected to strict controls that include fatigue tests to verify if the axle fatigue limits actually exceed certain values specified in the regulations. Fatigue tests can be performed on machines where the axis rotates, or more frequently, on Sincotec type resonance test machines where the surrounding conditions are completely controlled. For both types of machines, the tests usually seek to verify whether a defect appears on the shaft before 107 cycles, so it is necessary to establish a safety stop criterion in case a crack appears during the test. In the case of resonance test machines, it is usually determined that the axis has cracked when the test frequency drops more than 0.5 Hz.
    Currently, during a fatigue test of a railway axis, the detection of defects, or the stopping criteria before reaching 107 cycles, is based on establishing alarm levels for the displacement of the axis center, temperature and variation of the applied load However, the stop criteria show a low sensitivity to cracks and when the machine stops, the defects are already very large.
    5
    10
    fifteen
    twenty
    25
    30
    35
    On the other hand, in the study of vibratory signals that can be acquired from an axis under analysis, it is important to highlight the modern signal processing techniques available today such as the wavelet transform (WT) and especially a of its evolutions, the so-called wavelet packet transform or transformed into wavelet packets (WPT), which allows the discrete decomposition of all frequency bands of the signal. This tool allows to decompose the signal obtaining information both in the domain of time and frequency. Because of this, it is especially useful for detecting transient phenomena and discontinuities in the signals that other tools, such as the fast Fourier transform (FFT) do not detect. In short, this tool is suitable for working with non-stationary signals such as those normally obtained from rotating machinery.
    The state of the art offers some solutions for the detection of defects in axes based on classic tools that only work in the frequency domain, such as FFT. The analysis of these signals is reduced to the first and second harmonic of the speed of rotation, which are indicated as good indicators of the presence of a defect. However, these methods have shown little effectiveness in practice in experimental work under certain load and speed conditions. Perhaps also by the presence of other defects such as misalignment or imbalance, which also affect these frequencies.
    On the other hand, in recent years, the interest in the diagnosis of defects in axes has gained a lot of interest, however, most of the techniques proposed in recent years use signals obtained from models, and the reverse identification process is rarely included of defects, so its practical application in the experimental field is more than doubtful.
    As stated above, the inventions hitherto known by the state of the
    technique for detecting fissures in shafts have not shown effectiveness in practice.
    Specifically, within those oriented to fatigue tests of railway axes, there are
    techniques that perform the function of the object of the invention, but in an unreliable and
    inefficient because they fail to detect the cracks that appear in a time
    railway axis, running the risk of a catastrophic failure of the axis and the machine. The state of
    art therefore demands an evolution in these techniques that allows this detection
    of defects in railway axes that, subjected to a fatigue analysis as commented
    previously, it can detect them incipiently and thus reduce the expenses of
    3
    5
    10
    fifteen
    twenty
    25
    30
    time and jan ^ a, such as the risk of catastrophic failure of the axis with the possible damage that it may imply for the machine and its surroundings.
    Description of the invention
    The present invention solves the problems discussed above by proposing a method that allows the detection of rail faults with a considerable advance compared to traditional methods. Specifically, the present invention relates to a method for detecting a defect in a railway axis during a fatigue test, comprising the steps of:
    - acquire, by at least one sensor, a vibratory type signal of the railway axis;
    - measure, in a processor, the energy of the vibratory signal acquired in a certain frequency band;
    - compare, in the processor, the measured energy with a previously established energy threshold;
    - determine that a defect in the axis has been detected in case the measured energy exceeds the pre-established energy threshold;
    - stop the fatigue test based on the detection of a defect.
    The railway axis, according to one of the embodiments of the invention, can rotate in a test bench, where the test conditions can be controlled comfortably. Alternatively, the shaft may be subjected to a resonance machine, where the shaft does not need to rotate.
    The detection of a fault implies that the current fatigue test is stopped, since a stop criterion is established for the fatigue test based on the detection of a crack. In this way, the test can be stopped with considerable anticipation compared to traditional methods.
    The energy threshold can be set according to relative energy changes, with respect to the sound axis value, detected for the certain frequency band in defective axes. In this way, the analysis of a defective axis reveals changes in the energy measured in a certain frequency band, which are recorded to establish
    5
    10
    fifteen
    twenty
    25
    30
    35
    as ^ a threshold Kmite for the following railway axes to study. In particular, a defective axis implies an increase in detectable ene in the certain frequency band.
    According to one of the embodiments of the invention, it is contemplated to determine a first critical frequency of the railway axis. Additionally, the frequency band under analysis focuses on the sub-harmonics of the first critical frequency of the axis.
    In an embodiment of the invention, the subharmonic in which it focuses on the analysis, consists of the third subharmonic of the first critical frequency of the axis. Experimental results have revealed that it has the best results in terms of premature defect detection.
    This frequency in particular is what has revealed in experimental tests to be surprising evidence for the prediction of defects. It is possible to anticipate the detection of a defect up to 150,000 cycles compared to conventional methods, which implies considerable benefits in terms of time, energy and reduction of the risk of catastrophic axis failure.
    According to one of the embodiments of the invention, it is also contemplated to analyze the energies in the harmonics of the rotation speed of the railway axis.
    The present invention contemplates applying a wavelet transform in packets to the acquired serial and, subsequently, calculating the energy of the serial from the applied transformation. The energy is distributed in packets that correspond to different frequency bands of the acquired serial.
    A second aspect of the invention relates to a system for automatically detecting faults in a railway axis, the system is characterized in that it comprises:
    - a sensor configured to acquire a vibratory type serial of the rotary axis;
    - a processor, in communication with the sensor, configured to obtain energy measurements from said series, compare the measured energies with a previously established energy threshold, determine that a defect has been detected in the axis in case the measured energy exceeds the preset energy threshold; and stop the fatigue test based on the detection of a defect in the shaft.
    5
    10
    fifteen
    twenty
    25
    30
    35
    The system of the invention contemplates, according to one of the embodiments, an axis that is being subjected to a fatigue test where the railway axis is rotated in a test bench.
    Alternatively, the system of the invention contemplates an axis that is being subjected to a fatigue test where the axis is being subjected to a resonance test machine.
    The processor may also be configured to detect increases in energy in a frequency band corresponding to the sub-harmonics of the critical frequency of the axis and the harmonics of the rotation speed.
    In one of the embodiments of the invention, it is contemplated to prioritize certain sub-harmonics where it has been detected that the effects are more advantageous, specifically the third sub-harmonic of the critical frequency of the axis.
    A final aspect of the invention relates to an informatic program comprising program code means adapted to perform the steps of the method of the invention, when said program is executed in a general purpose processor, a digital serial processor, an FPGA , an ASIC, a microprocessor, a microcontroller, or any other form of programmable hardware.
    Thus, the present invention therefore proposes, based on experimental results, the premature detection of defects in axes subjected to fatigue testing by monitoring energy in certain frequency bands related to harmonics of the rotation speed and critical frequencies of the axis, which are especially revealing for the third subharmonic of the first critical frequency of the axis. The energy monitoring in this frequency band allows the detection of a defect approximately 150,000 cycles before conventional methods, saving time, energy and reducing risks for the machine and its surroundings, which represents a considerable advance for its technical field.
    Description of the drawings
    To complement the description that is being made and in order to help a
    better understanding of the features of the invention, according to a preferred example
    6
    5
    10
    fifteen
    twenty
    25
    30
    35
    of practical realization of the same, it is accompanied as an integral part of said description, a set of drawings where, with an illustrative and non-limiting nature, the following has been represented:
    Figure 1.- Shows a diagram of the decomposition procedure for the WPT up to decomposition level 3.
    Figure 2.- Represents the evolution of the Jan ^ a, as a function of the number of cycles during a fatigue test, of the first harmonic of the rotation speed.
    Figure 3.- Represents the evolution of energy, as a function of the number of cycles during a fatigue test, of the third subharmonic of the first resonance frequency.
    Figure 4.- Shows an enlarged view of Figure 3, where the area of interest can be seen in more detail.
    Detailed description of the invention
    For the development of the present invention, it is important to start by mentioning the experimental tests that motivate it. It can be observed that according to one of the possible embodiments, a conventional fatigue test is first proposed to which the railway axes are subjected to study which, in this particular case, are three hollow railway axes made of steel. Then, the critical frequencies of the axis are determined by performing an experimental modal analysis, for example by accelerometry and a dynamical hammer, and for this case the critical frequencies are obtained according to the following table:
    Resonance frequency (N) Frequency (Hz)
     one  194.2
     2  485.2
     3  886.0
     4  1335.5
     5  1857.2
    Once a conventional test is launched, the main stop criteria for the
    tests are, on the one hand, if the number of cycles reaches a value of 107 and on the other hand, if
    detects a defect before the final number of cycles is reached. Setting the
    Detection of a defect depending on whether certain parameters exceed a threshold value
    7
    5
    10
    fifteen
    twenty
    25
    30
    previously established for the displacement of the center of the axis, temperature and variation in the applied load.
    For this specific case, in which the three axes discussed above have been submitted to the test, it is obtained as a result that two of them have been broken, because they have not exceeded the applied load, specifically test 1 and 3. The following table reflects the test conditions:
     Test  Speed (RPM) Load (kN) Stop criteria Number of cycles
     one  509 240 threshold value reached 1688594
     2  509 216 maximum number of cycles 10000000
     3  509 240 threshold value reached 1467172
    The results are fully confirmed by ultrasound and magnetic particles, to corroborate that they are indeed defective axes that have to be removed, since they do not meet the requirements of fatigue resistance necessary to be mounted on a railway machine. At the time of the test stop, the appearance of cracks in the axis is checked with the naked eye, which demonstrates the low sensitivity of these traditional methods.
    However, during these conventional tests, according to an embodiment of the invention, a triaxial accelerometer is installed in the central shaft bearing housing, a signal conditioner and a data acquisition card to send the data to a computer by USB connection In this way, you can acquire the vibration signals produced during the tests, whose parameters are represented below for the specific case we are describing:
    Sampling frequency (Hz) Signal length Data storage
    6000 16384 1 acquisition every 3 minutes
    Next, the vibration signals are processed to identify significant changes that may be associated with the presence of faults in the railway axis. Particularly and according to the preferred embodiment of the present invention, the signals are processed using the WPT, a variation of the DWT, which is especially efficient for carrying
    5
    10
    fifteen
    twenty
    25
    30
    It carries out local analysis of non-stationary signals and allows obtaining correlation coefficients between the signal and a mother wavelet.
    DWT can be implemented using filters, thus decomposing a temporary signal x (t) by two filters, a low pass filter and another high pass filter. The low pass filter provides information on the low frequencies, called approximation information (A), while the high frequency information is obtained using the high pass filter and is called detail information (D).
    The wavelet packet transform (WPT) is a variation of the DWT decomposition, which is obtained by applying the DWT in a recursive manner, as shown in Figure 1, where W (k; j) represents the signal coefficients in each package, "k" being the decomposition level and "j" the position of the package within the decomposition level. Thus, each correlation vector W (k; j) has the following structure:
    W (k, j) = {wi (k; j); ..., WN (k; j)} = {wi (k; j)}
    The concept of energy that is used in a WPT analysis is similar to the Fourier theona. The energy of the packages is obtained from the sum of the squares of the coefficients of each package, that is:
    Ekj = Ei (wi (k; j)} 2
    The level of decomposition used determines the frequency resolution of each packet. Taking into account that the signal can only be reconstructed up to the pi frequency, limited by the sampling frequency, the frequency resolution of each fr packet can be calculated using the following equation:
    Applying these transformations on the vibratory signals acquired previously, the energy packages obtained according to a mother wavelet "Daubechies 6" are analyzed, which is reasonable to choose since their effectiveness has been proven in this field by previous studies of the state of the art . According to an embodiment of the invention,
    5
    10
    fifteen
    twenty
    25
    30
    35
    They analyze different levels of decomposition, specifically from 5 to 9. For decomposition level 5, a total of 32 (25) packages, or energy levels, are obtained for each signal. Although this number is easily manageable to detect significant changes as the number of fatigue test cycles progresses, each of the packages refers to a very wide frequency band (93.8 Hz). Therefore, once increments are detected in one of the packages with the number of cycles, this package is broken down to higher levels to fine tune the frequency value affected by the defect. The maximum decomposition level value chosen in this case is 9 to ensure a balance between the precision of the frequency and a low computational cost, but according to other embodiments, this analysis can be considered with other parameters.
    Experimental tests reflect as the number of cycles progresses, the evolution of energy in the different frequency bands. The analysis of the vibration signals of the three tests reveals a great increase in the energy in several frequency bands precisely for the tests that resulted in the breakage of the axis, while this phenomenon does not occur when the axis endures the complete test without present any defect, that is, it reaches 107 cycles.
    Significant changes in the evolution of the energy related to harmonics of the rotation speed can be seen, as for example in Figure 2, where the evolution of the energy is shown specifically in the frequency band 5.9-11.7 Hz, which contains the first harmonic of the speed of rotation (8.5 Hz) and in the last 20,000 cycles of the test, before the test is stopped by traditional methods, a change in the tendency of the energy motivated by the axis defects under study, which in this case is test 3.
    Instead, the most surprising thing happens in the observation of the subharmonics of
    critical axis frequencies, where significant changes are also detected,
    specifically, in the frequency band 64.5-70.3 Hz, which corresponds to the third
    sub-harmonic of the first critical frequency of the axis (which corresponds according to the test
    experimental of the particular realization under study with 194.19 Hz, and its third
    subharmonic with 64.73 Hz). Figure 3 shows the evolution of energy in this band of
    frequencies in the transverse direction and it can be seen that changes in the trend of
    energy of this frequency band are observed before those of the harmonics of the
    turning speed. This surprising effect on the energy trend is used by the
    10
    5
    10
    fifteen
    twenty
    25
    30
    35
    present invention to advantageously detect the defects in an axis under study during a fatigue test, since, in view of the figures, it can be affirmed that the defects can be detected much earlier and establishing a threshold value for the jan ^ a, the The test machine would stop approximately 150,000 cycles before following the conventional stop criteria, as shown in the detection area (40) of Figure 4, which corresponds to an enlarged view of Figure 3 for the area of interest.
    The present invention therefore monitors the vibrations emitted by the axis continuously and for the detection of defects the phenomena observed during the dynamic analysis of the axis in subharmonics of the critical speed of the axis are sought, as well as harmonics of the rotation speed. .
    The proposed invention, according to a particular embodiment, offers a criterion of early stopping of the fatigue test from the study of the signals emitted by the railway axis, since during said test the vibratory type signals are acquired by means of a sensor, for example an accelerometer, which are collected by an acquisition system and sent to a signal processor. After processing the signals and applying wavelet transform in packages to obtain a measure of the energy (7) of the signals, an exhaustive analysis by frequency bands is performed and the energy differences in certain bands are detected that indicate the presence of a defect . The main disadvantage is that the changes in the natural frequencies are very small, and can be covered due to experimental errors, but the present invention proposes the analysis of the sub-harmonics of the first critical frequency of the axis and the harmonics of the speed of rotation , the third subharmonic of the critical frequency being especially relevant, where the energy trend is more evident and allows the detection of a fault to be anticipated considerably. Therefore, energy standards are established for which it is determined that the axis is defective and, to said energy threshold, the test stop is linked, which guarantees the early stop of said test in the event that it occurs the detection of a fault in the axis.
    The energy patterns acquired during the previous analysis are stored in a memory
    of the processor and, according to one of the embodiments of the invention, when submitted
    a new railway axis to a fatigue test, the pattern is recovered from said memory
    corresponding to said type of railway axis. It is determined that there is a defect
    establishing an energy threshold and monitoring the frequency bands of interest already
    eleven
    5
    10
    fifteen
    twenty
    25
    30
    35
    mentioned, which allows establishing a stop criterion for the fatigue test based on the detection of said defects.
    权利要求:
    Claims (13)
    [1]
    5
    10
    fifteen
    twenty
    25
    30
    35
    1. - Method for detecting defects in a railway axis during a fatigue test, which includes the following steps:
    a) acquire, by at least one sensor, a vibratory type signal of the railway axis;
    b) calculate, in a processor, the energy of the vibratory signal acquired in a given frequency band;
    c) compare, in the processor, the calculated energy with a previously established threshold of jan ^ a;
    d) determine that a defect in the axis has been detected in case the calculated energy exceeds the preset energy threshold;
    e) stop the fatigue test based on the detection of a defect in step d).
    [2]
    2. - Method according to the previous claim where the axis rotates in a test bench.
    [3]
    3. - Method according to claim 1 wherein the axis is being subjected to a resonance test machine.
    [4]
    4. - Method according to any of the preceding claims wherein the energy threshold is set according to some energy changes detected for the frequency band determined on defective axes.
    [5]
    5. - Method according to claim 4 wherein a defective axis implies an increase in detectable energy in certain frequency bands.
    [6]
    6. - Method according to any of the preceding claims which further comprises determining a first critical frequency of the railway axis and where the frequency band is a subharmonic of the first critical frequency of the axis.
    [7]
    7. - Method according to claim 6 where the subharmonic corresponds to the third subharmonic of the first critical frequency of the axis.
    [8]
    8.- Method according to claim 2 which also includes calculating a speed
    of rotation of the railway axis and where the certain frequency band is a harmonic of
    13
    5
    10
    fifteen
    twenty
    25
    said rotation speed.
    [9]
    9. - Method according to any of the preceding claims which further comprises applying a wavelet transform in packets to the acquired signal, to calculate its ene ^ a in certain frequency bands.
    [10]
    10. - Method according to claim 9 where the energy is distributed in packages corresponding to the determined frequency bands of the acquired signal.
    [11]
    11. - System to automatically detect defects on a railway axis during a fatigue test, the system is characterized by:
    - a sensor configured to acquire a vibratory signal of the railway axis;
    - a processor, in communication with the sensor, configured to calculate the energy of said signals, compare the calculated energies with a previously established energy threshold, determine that a defect has been detected in the axis in case the calculated energy exceeds the threshold of preset energy and stop the fatigue test based on the detection of a fault.
    [12]
    12. - System according to claim 11 wherein the sensor is an accelerometer.
    [13]
    13. - Computer program comprising program code means adapted to perform the steps of the method according to any of claims 1-10, when said program is executed in a general purpose processor, a digital signal processor, an FPGA, a ASIC, a microprocessor, a microcontroller, or any other form of programmable hardware.
    类似技术:
    公开号 | 公开日 | 专利标题
    ES2800319T3|2020-12-29|Fault detection system and associated method
    ES2397084T3|2013-03-04|Method and apparatus for monitoring and analysis of vibrations in rotary machines
    BR112012010702B1|2019-11-26|PROCESS OF DETECTION OF DAMAGE AT LEAST ONE BEARING BEARING
    Saucedo-Dorantes et al.2016|Multiple-fault detection methodology based on vibration and current analysis applied to bearings in induction motors and gearboxes on the kinematic chain
    ES2575745B1|2017-04-12|Method and system for detecting railway axle defects in fatigue tests
    JP2008292288A|2008-12-04|Bearing diagnostic device for reduction gear
    ES2549652B1|2016-10-07|Method and system to automatically detect faults in a rotary axis
    US9267925B2|2016-02-23|Pole integrity meter and method of determining pole integrity
    JP2010256352A|2010-11-11|Structural integrity monitoring system
    SE0701428L|2008-12-13|measuring device
    Liu et al.2018|Wheel tread defect detection for high-speed trains using FBG-based online monitoring techniques
    WO2017103294A1|2017-06-22|Method for detecting defects in railway axles during fatigue tests
    Jüngert et al.2014|Local acoustic resonance spectroscopy | for glass fiber-reinforced polymer applications
    Li et al.2005|Crack damage detection in beam-like structures using RBF neural networks with experimental validation
    BR112019014810A2|2020-02-27|METHOD FOR DETECTION, USE OF METHOD FOR DETECTION, DETECTION INSTALLATION AND DETECTION TROLLEY
    Gonçalves et al.2015|Diagnosis of bearing faults in induction motors by vibration signals-Comparison of multiple signal processing approaches
    Jiang et al.2015|Rolling bearing quality evaluation based on a morphological filter and a Kolmogorov complexity measure
    Zhang et al.2017|Experimental study of normal contact force between a rolling pneumatic tyre and a single asperity
    Babu et al.2014|High frequency acceleration envelope power spectrum for fault diagnosis on journal bearing using DEWESOFT
    ES2290107T3|2008-02-16|PROCEDURE AND DEVICE FOR DETECTING A DEFECTIVE BEARING OF A RAILWAY VEHICLE.
    Zamorano et al.2019|Analysis in the time-frequency domain of different depths of a crack located in a change of section of a shaft
    CN101526417B|2011-09-14|Method for diagnosing structural damage by musical tone rule
    JP2014152624A5|2015-10-01|
    Siew et al.2015|Fault severity trending in rolling element bearings
    ES2716324A1|2019-06-11|DEVICE AND PROCEDURE FOR MONITORING THE STATE OF CONCRETE TUNNEL COATING |
    同族专利:
    公开号 | 公开日
    ES2575745B1|2017-04-12|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US6192758B1|1998-12-14|2001-02-27|Kang Huang|Structure safety inspection|
    US20040078160A1|2002-10-11|2004-04-22|Frei Mark G.|Method, computer program, and system for intrinsic timescale decomposition, filtering, and automated analysis of signals of arbitrary origin or timescale|WO2017103294A1|2015-12-16|2017-06-22|Universidad Carlos Iii De Madrid|Method for detecting defects in railway axles during fatigue tests|
    CN110849753A|2019-12-09|2020-02-28|大连理工大学|Metal material fatigue strength prediction method based on micro scratches|
    法律状态:
    2017-04-12| FG2A| Definitive protection|Ref document number: 2575745 Country of ref document: ES Kind code of ref document: B1 Effective date: 20170412 |
    优先权:
    申请号 | 申请日 | 专利标题
    ES201431966A|ES2575745B1|2014-12-30|2014-12-30|Method and system for detecting railway axle defects in fatigue tests|ES201431966A| ES2575745B1|2014-12-30|2014-12-30|Method and system for detecting railway axle defects in fatigue tests|
    [返回顶部]