Gear fault detection, determination method, device, controller and working machine

Abstract

The application provides a gear fault detection and determination method, device, controller and working machine, which is applied to the technical field of working machines, and after vibration signals and rotation speed signals of a target gear are acquired, detection samples are constructed based on the vibration signals and the rotation speed signals, and fault detection characteristic values of the target gear are extracted based on the detection samples, finally, the fault detection characteristic values of the target gear are input into a pre-trained fault detection model to obtain a fault detection result of the target gear. Through the method provided by the application, the fault detection of the gear can be completed, which provides a reference basis for timely maintenance of the gear, thereby effectively reducing the harm of the fault gear and improving the operation safety of the working machine.

Description

Technical Field

[0001] This application relates to the field of operating machinery technology, specifically to a gear fault detection and judgment method, device, controller, and operating machinery. Background Technology

[0002] Gears are among the most common mechanical structures in the engineering field. They can achieve various functions such as power transmission, changing the direction of motion, and changing the rotation speed to adjust torque. Therefore, gears are an indispensable component in most working machines.

[0003] During gear operation, various typical failure modes can easily occur, such as pitting, wear, cracks, and tooth breakage. Among these, cracks and tooth breakage pose the greatest threat to the machinery. On the one hand, cracks or tooth breakage disrupt the normal meshing process of the gears, easily leading to the failure of adjacent gears; on the other hand, tooth breakage affects the normal operation of the machinery, and the broken tooth fragments can also cause direct mechanical damage.

[0004] Therefore, how to accurately detect and diagnose gear faults, reduce the harm caused by faulty gears, and improve the operational safety of machinery has become one of the problems that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, this application aims to provide a gear fault detection and judgment method, device, controller and operating machinery, which completes the fault detection of gears based on the vibration signal and speed signal of the gears, provides a reference for timely gear repair, and thus effectively reduces the harm of faulty gears and improves the operating safety of operating machinery.

[0006] Firstly, this application provides a gear fault detection method, including: Acquire the vibration and rotational speed signals of the target gear; A detection sample is constructed based on the vibration signal and the rotation speed signal; Based on the detected samples, the fault detection feature values ​​of the target gear are extracted; The fault detection feature values ​​of the target gear are input into a pre-trained fault detection model to obtain the fault detection results of the target gear. The fault detection model is obtained by training a neural network, with the fault detection feature value of the gear as input and the fault detection result of the gear as output.

[0007] Optionally, a detection sample is constructed based on the vibration signal and the rotational speed signal, including: Extract the vibration signal corresponding to the target frequency from the vibration signal, and perform envelope processing on the vibration signal corresponding to the target frequency to obtain the processed vibration signal; Based on the rotation speed signal and the preset sampling frequency, the processed vibration signal is sampled at equal angular intervals to obtain a detection sample; The detection sample includes multiple detection sub-samples, and each detection sub-sample includes multiple vibration signal amplitudes arranged in time sequence.

[0008] Optionally, the fault detection feature values ​​include at least one of the following: rotational speed feature value, amplitude feature value, sharp impact feature value, spectral prominence feature value, and spectral multi-order feature value.

[0009] Optionally, the process of extracting the spectral salient features includes: The spectrum of the detected sample is normalized to obtain a normalized spectrum; Based on the detected samples, the fault characteristic spectrum of the target gear is determined, and the target amplitude is extracted; Wherein, the target amplitude is the amplitude corresponding to the fault feature spectral number and the spectral number that is an integer multiple of the fault feature spectral number in the normalized spectrum; The target amplitude with the largest value among all the target amplitudes is determined as the spectral salient feature value.

[0010] Optionally, the process of extracting the spectral multi-order characteristic values ​​includes: The spectrum of the detected sample is normalized to obtain a normalized spectrum, and the fault characteristic spectrum of the target gear is determined based on the detected sample. The first statistical value is obtained by counting the total number of fault feature spectral numbers and integer multiples of the fault feature spectral numbers in the normalized spectrum. The ratio of the second statistical value to the first statistical value is determined as the spectral multi-order characteristic value; The second statistical value is the total number of spectral numbers whose amplitude is greater than or equal to a preset amplitude threshold among the fault feature spectral numbers and spectral numbers that are integer multiples of the fault feature spectral numbers.

[0011] Optionally, the process of extracting the amplitude feature values ​​of the target gear includes: Construct an inverse transformation sample based on the detected sample; The inverse transform sample is divided into multiple inverse transform sub-samples; Extract the maximum value of the vibration signal amplitude from each of the inverse transform sub-samples; The average of the maximum values ​​is determined to be the amplitude characteristic value of the target gear.

[0012] Optionally, the process of extracting the sharp impact feature value of the target gear includes: Construct an inverse transformation sample based on the detected sample; The inverse transform sample is divided into multiple inverse transform sub-samples; Calculate the kurtosis index for each of the inverse transform subsamples; The number of kurtosis indices that are less than or equal to a preset kurtosis index threshold is counted to obtain a third statistical value; The ratio of the third statistical value to the total number of each inverse transform subsample is determined as the sharp impact characteristic value of the target gear.

[0013] Optionally, constructing an inverse transform sample based on the detected sample includes: The fault characteristic spectrum of the target gear is determined based on the detected samples; Extract the fault feature spectrum and the vibration signal amplitude corresponding to the spectrum of integer multiples of the fault feature spectrum from the detection sample to obtain the constructed sample; Perform an inverse Fourier transform on the constructed sample to obtain the inverse transform sample.

[0014] Optionally, the gear fault detection method provided in the first aspect of the present invention further includes: updating the fault detection model based on the detection sample and the actual state of the target gear.

[0015] Secondly, the present invention provides a method for determining gear faults, comprising: According to any one of the first aspects of the present invention, the gear fault detection method obtains multiple fault detection results of the target gear; Based on the multiple fault detection results, a fourth statistical value and a fifth statistical value are calculated; The fourth statistical value is the number of fault detection results that indicate a fault among the multiple fault detection results, and the fifth statistical value is the number of consecutive fault detection results that indicate a fault among the multiple fault detection results. The fourth and fifth statistical values ​​are used to determine whether the target gear is faulty.

[0016] Optionally, determining whether the target gear is faulty based on the fourth statistical value and the fifth statistical value includes: If the fourth statistical value is greater than a preset cumulative threshold, or if the fifth statistical value is greater than a preset continuous threshold, the target gear is determined to be faulty. If the fourth statistical value is less than or equal to the preset cumulative threshold, and the fifth statistical value is less than or equal to the preset continuous threshold, the target gear is determined to be normal.

[0017] Optionally, the multiple fault detection results are stored in a preset stack, and the method further includes: The fault detection results in the preset stack are updated according to the first-in, first-out (FIFO) principle.

[0018] Thirdly, the present invention provides a controller, comprising: a memory and a processor; the memory storing a program suitable for execution by the processor to implement the gear fault detection method according to any one of the first aspects of the present invention, or the gear fault determination method according to any one of the second aspects of the present invention.

[0019] Fourthly, the present invention provides a working machine, comprising: a working component and a controller as described in the third aspect of the present invention, wherein, The controller is connected to the working component, and the working component is equipped with gears.

[0020] Based on the above, the gear fault detection method provided in this application, after acquiring the vibration and rotational speed signals of the target gear, constructs a detection sample based on the vibration and rotational speed signals, extracts fault detection feature values ​​of the target gear based on the detection sample, and finally inputs the fault detection feature values ​​of the target gear into a pre-trained fault detection model to obtain the fault detection result of the target gear. The method provided by this invention can complete gear fault detection, providing a reference for timely gear repair, thereby effectively reducing the hazards of faulty gears and improving the operational safety of machinery. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart of a gear fault detection method provided in an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram illustrating the sampling effect of sampling vibration signals at equal time intervals.

[0024] Figure 3 This is a schematic diagram illustrating the sampling effect of sampling vibration signals at equal angular intervals.

[0025] Figure 4 This is a waveform diagram of a vibration signal triggered by a gear malfunction.

[0026] Figure 5 This is a schematic diagram of the spectrum of a vibration signal triggered by a gear failure.

[0027] Figure 6This is a flowchart of a gear fault determination method provided in an embodiment of the present invention.

[0028] Figure 7 This is a structural block diagram of a gear fault detection device provided in an embodiment of the present invention.

[0029] Figure 8 This is a structural block diagram of a gear fault determination device provided in an embodiment of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0031] As mentioned earlier, gears are an indispensable component of most construction machinery. Whether gears are faulty has a direct impact on ensuring the normal operation of construction machinery. In practical applications, it is necessary to detect gear faults. To this end, this invention provides a gear fault detection method that detects gear faults based on the vibration and rotation speed signals of the gears, providing a reference for timely gear repair, thereby effectively reducing the harm of faulty gears and improving the operational safety of construction machinery.

[0032] The gear fault detection method provided by this invention can be applied to a controller, which can be any controller in the working machine to which the gear to be fault detected belongs. Of course, it can also be applied to other controllers independent of the working machine, such as a server on the network side.

[0033] See Figure 1 The process of the gear fault detection method provided in this embodiment of the invention may include: S100: Acquire the vibration signal and rotational speed signal of the target gear.

[0034] In this embodiment and in subsequent embodiments, the target gear refers to the gear that needs to be fault detected. In practical applications, the target gear can be any gear in the working machinery.

[0035] During the operation of the target gear, it rotates at a certain speed under the action of the corresponding drive mechanism or power mechanism. Correspondingly, if the gear malfunctions (in this embodiment, mainly referring to cracks and broken teeth), it will repeatedly pass through the fault point during rotation, generating vibration signals. Therefore, the vibration signal generated by the gear malfunction is directly related to the gear's rotation process. Thus, this step first collects the vibration and speed signals of the target gear. It is understood that, to effectively preserve the temporal correlation between the vibration and speed signals, the vibration and speed signals should be collected synchronously, so that they correspond to the same timestamp.

[0036] In practical applications, vibration signals can be acquired by vibration sensors, and rotational speed signals can be acquired by rotational speed sensors. Of course, the above information can also be acquired by other means, which will not be listed here. This invention does not make any specific limitations on these methods.

[0037] S110. Construct detection samples based on vibration signals and rotational speed signals.

[0038] Understandably, the operating environment of machinery is usually harsh, with very high ambient noise levels. The vibration signals collected by vibration sensors often contain a lot of noise, which is useless for fault detection of the target gear and needs to be eliminated. Furthermore, for a specific target gear, the frequency of the vibration signal it generates under fault conditions can often be predetermined. For example, the frequency of the vibration signal generated by most target gears under fault conditions is usually between 12K±2KHz.

[0039] Based on the above, after acquiring the vibration signal, the vibration signal corresponding to the target frequency can be extracted first, and then envelope processing can be performed on the vibration signal corresponding to the target frequency to obtain the processed vibration signal. As mentioned earlier, the target frequency corresponds to the frequency of the vibration signal generated under the fault state of the target gear. In practical applications, this frequency can be determined based on the design parameters and structural characteristics of the target gear. This invention does not limit the specific selection of the target frequency. Furthermore, the process of extracting the vibration signal at the target frequency can be implemented using bandpass filtering technology, and the process of enveloping the vibration signal corresponding to the target frequency can be implemented using Hilbert or low-pass filtering technology. This invention also does not limit this process.

[0040] In practical applications, if the target gear experiences a tooth breakage, i.e., a tooth breaks and is lost, it will disrupt the normal meshing of other normal toothed gears, resulting in vibration signals generated during the rotation of the target gear. Furthermore, the target gear triggers an impact vibration once per revolution, making the characteristic frequency of the vibration signal consistent with the rotational frequency of the target gear. However, the rotational frequency of the target gear varies. If equal-interval sampling is used, multiple frequency components will appear, which is not conducive to vibration analysis. Therefore, this embodiment uses equal-angle interval sampling for the vibration signal.

[0041] After obtaining the processed vibration signal, the processed vibration signal is sampled at equal angular intervals based on the rotational speed signal and a preset sampling frequency to obtain detection samples. It can be understood that the goal of equal angular interval sampling is to ensure that the number of sampling points is consistent throughout one revolution of the target gear. Specifically, in the process of using the rotational speed signal to complete equal angular interval sampling, analog pulses are first generated according to a preset sampling frequency, such as 200 times the rotational speed, to complete the equal angular interval sampling of the vibration signal.

[0042] It should be noted that the above content is only a general introduction to the process of sampling the processed vibration signal at equal angular intervals based on the rotation speed signal. The specific implementation process can be implemented with reference to relevant technologies. This invention does not limit the specific implementation process of sampling the vibration signal at equal angular intervals.

[0043] contrast Figure 2 and Figure 3 As shown, where, Figure 2 The image shows a schematic diagram illustrating the sampling effect after sampling the vibration signal at equal time intervals. Figure 3 The diagram shows the sampling effect after sampling the vibration signal at equal angular intervals. By comparison, it can be seen that the spectral components of the vibration signal obtained after sampling at equal time intervals are more complex, which is not conducive to subsequent data analysis. Conversely, the spectral components of the vibration signal obtained after sampling at equal angular intervals are more uniform, which can effectively reduce the amount of data for subsequent fault detection.

[0044] After sampling the processed vibration signal at equal angular intervals as described above, a detection sample is obtained. This detection sample includes multiple detection sub-samples. It can be understood that one detection sub-sample corresponds to the cycle of one rotation of the target gear, and any detection sub-sample includes multiple vibration signal amplitudes arranged in time sequence.

[0045] S120. Extract fault detection feature values ​​of the target gear based on the detection samples.

[0046] Considering the gear's structure and operating characteristics, the vibration signals generated during operation when the gear has cracks or broken teeth typically exhibit the following characteristics: First, under normal circumstances, the speed of the gear is related to the magnitude of the impact vibration caused by gear cracks or broken teeth. Because broken teeth disrupt the normal meshing of the gears, the driving gear hits the driven gear teeth, causing impact vibration. When the gear speed is high, although the broken teeth cannot mesh normally, the large inertia can alleviate the impact intensity to a certain extent.

[0047] Secondly, the vibration signal generated by each vibration of the faulty gear has a relatively sharp waveform, which can be seen in the following figure. Figure 4 As shown, if the gear has multiple cracks or broken teeth, there will be multiple sharp impacts within one rotation cycle.

[0048] Furthermore, after performing a Fourier transform on the vibration signal, the resulting spectrum can be as follows: Figure 5 As shown, the gear fault spectrum in the spectrum diagram shows multiple prominent spectral lines, meaning that multiple spectral lines have high amplitude values ​​and are equally spaced, i.e., multiple prominent spectral lines.

[0049] Based on the above, the gear fault detection method provided in this embodiment of the invention extracts fault detection feature values ​​of the target gear based on the detection sample, including at least one of the following: rotational speed feature value, amplitude feature value, sharp impact feature value, spectral prominence feature value, and spectral multi-order feature value. The extraction process of each fault detection feature value will be elaborated in detail in the following content, and will not be described in detail here.

[0050] S130. Input the fault detection feature value of the target gear into the pre-trained fault detection model to obtain the fault detection result of the target gear.

[0051] As a preferred embodiment, this invention provides a fault detection model that takes the fault detection feature values ​​of a gear as input and the fault detection result of the gear as output, and trains a neural network to obtain the model. Of course, the fault detection model can also be trained based on other machine learning models such as random forests, which will not be discussed further here.

[0052] Based on this, after obtaining the fault detection feature values ​​of the target gear through the aforementioned steps, inputting these feature values ​​into the fault detection model directly yields the fault detection result of the target gear. As an optional implementation, the fault detection result can include both normal and faulty target gear conditions. Of course, more specific fault detection results can also be output, such as whether the target gear has a crack fault or a broken tooth fault, or the fault level of the target gear. These are also optional and, without exceeding the core concept of this invention, fall within the scope of protection of this invention.

[0053] It should be noted that before training the fault detection model, training samples need to be prepared. These samples can be obtained through two methods: historically accumulated samples and samples collected through destructive experiments. Both methods must include both faulty and non-faulty samples. Furthermore, the samples need to be manually calibrated before training, that is, their classification as either faulty or non-faulty samples must be determined. After the training samples are calibrated, the fault detection model can be trained until a fault detection model that meets the usage requirements is obtained. The specific training process of the fault detection model can be implemented based on relevant technologies, and this invention does not limit this process.

[0054] Furthermore, in the actual use of the fault detection model, it can be updated based on the detection samples obtained during use and the actual state of the target gear. This combines machine learning with periodic incremental learning, thereby continuously improving the fault detection model and increasing the accuracy of the detection results. It is understandable that the actual state of the target gear can be determined by disassembling the machinery to which it belongs and conducting an actual inspection of the target gear.

[0055] In summary, the gear fault detection method provided in this application, after acquiring the vibration and rotational speed signals of the target gear, constructs a detection sample based on the vibration and rotational speed signals, extracts fault detection feature values ​​of the target gear based on the detection sample, and finally inputs the fault detection feature values ​​of the target gear into a pre-trained fault detection model to obtain the fault detection result of the target gear. The method provided by this invention can complete gear fault detection, providing a reference for timely gear maintenance, thereby effectively reducing the hazards of faulty gears and improving the operational safety of machinery.

[0056] Furthermore, since the vibration signal contains multiple sharp impacts when there are multiple cracks or broken teeth in the gear, the aforementioned fault detection feature value can well reflect this feature. Therefore, the method provided by this invention can also be used to identify the condition of multiple faults in the gear.

[0057] The following describes the extraction process of various fault detection feature values ​​involved in S120. Assuming the detection sample is Xsam and the sample length is M (i.e., including M vibration signal amplitudes), the spectrum obtained after performing a Fourier transform on the detection sample is Fsam.

[0058] Rotational speed characteristic value: The aforementioned steps acquire the rotational speed signal of the target gear. Typically, the target gear operates at a stable speed. Therefore, the average rotational speed value carried in the signal can be used as the current rotational speed of the target gear, and the ratio of the current rotational speed to the maximum allowable rotational speed of the target gear can be used as the rotational speed characteristic value. Alternatively, this process can be viewed as a normalization of the target gear's current rotational speed, resulting in a rotational speed characteristic value that is any value within the range [0, 1].

[0059] Amplitude characteristic value: First, the fault characteristic spectrum of the target gear is determined based on the test samples. Specifically, the fault characteristic spectrum of the target gear can be calculated according to the following formula.

[0060] Np=(M / Nc)×(f1 / f2) (1) Wherein, Np represents the fault characteristic spectrum of the target gear; Nc represents the number of sampling points in each sampling period; f1 represents the fault characteristic frequency of the target gear; f2 represents the rotational frequency of the target gear.

[0061] It should be noted that for a given gear, its fault characteristic frequency can be calculated precisely, and can be implemented with reference to relevant technologies. This invention does not limit the specific calculation process of the fault characteristic frequency of the target gear. When applying this method to a given working machine, the fault characteristic frequency of the target gear can be pre-stored in the memory, and when running to this step, the fault characteristic frequency can be directly called.

[0062] Secondly, the vibration signal amplitudes corresponding to the fault feature spectrum and integer multiples of the fault feature spectrum in the test samples are extracted, and the results are used as the constructed samples. Specifically, a Fourier transform is first performed on the test samples, and the resulting spectrum is Fsam. The vibration signal amplitudes corresponding to the fault feature spectrum and integer multiples of the fault feature spectrum in Fsam are set to zero, and all other values ​​are set to zero to obtain the constructed sample, labeled Fc. It should be noted that the integer multiples mentioned above include zero multiples, that is, the constructed sample includes the vibration signal amplitude corresponding to spectrum 0.

[0063] Next, an inverse Fourier transform is performed on the constructed sample to obtain the inverse transform sample Xc. It is understood that the goal of the aforementioned steps is to remove non-gear-type fault impact information from the sample Xsam, and the sample length of the resulting inverse transform sample Xc is also M. As an optional implementation, normalization can be achieved by dividing the amplitude of each vibration signal in Xc by the maximum range of the vibration sensor.

[0064] Furthermore, the amplitude feature value of the target gear is extracted based on the obtained inverse transform samples. Specifically, the inverse transform samples are first divided into multiple inverse transform sub-samples. As an optional implementation, the inverse transform samples are divided into Q = M / N inverse transform sub-samples according to the number of sampling points Nc in each rotation cycle of the target gear. For example, if the number of sampling points in each rotation cycle of the target gear is 200 and the sample length is 2000, then Q = 2000 / 200 = 10. Next, the maximum value of the vibration signal amplitude in each inverse transform sub-sample is extracted, and the average of the maximum values ​​is determined as the amplitude feature value of the target gear. For example, if the maximum values ​​of the vibration signal amplitude in Q = 5 inverse transform sub-samples are [0.15, 0.2, 0.17, 0.16, 0.18], then the amplitude feature value A = (0.15 + 0.2 + 0.17 + 0.16 + 0.18) / 5 = 0.172.

[0065] Sharp impact characteristic value: Based on dividing the inverse transform sample into multiple inverse transform sub-samples, the kurtosis index of each inverse transform sub-sample is calculated. The calculation of the kurtosis index for any inverse transform sub-sample can be implemented with reference to relevant technologies, and this invention does not impose specific limitations on it. Furthermore, the number of kurtosis indices less than or equal to a preset kurtosis index threshold is counted to obtain a third statistical value. Finally, the ratio of the third statistical value to the total number of all inverse transform sub-samples is determined as the sharp impact characteristic value of the target gear.

[0066] For example, if the total number of inverse transform subsamples is 5, and the kurtosis indices of each inverse transform subsample are [0.2, 0.25, 0.3, 0.26, 0.21], and the preset kurtosis threshold is 0.26, then the corresponding third statistic is 4, and the sharp impact feature value I = 4 / 5 = 0.8.

[0067] It should be noted that in practical applications, the preset kurtosis index threshold can be flexibly selected according to the actual detection requirements and gear characteristics, and this invention does not impose specific limitations on it.

[0068] Key spectral features: First, the vibration signal amplitude of each spectral line in the spectrum Fsam obtained after Fourier transforming the test sample is divided by the largest vibration signal amplitude in Fsam to complete the normalization process of the spectrum Fsam. The result is denoted as Fo. Further, the amplitude of the fault characteristic spectral number of the target gear and its multiples are taken from Fo to obtain the target amplitude. The maximum value among the obtained target amplitudes is used as the prominent feature value of the spectrum.

[0069] For example: Fsam = [0, 0.01, 0.03, 0.06, 1.5, 0.07, 0.04, 0.02, 1.1, 0.03, 0.06, 0.01, 0.07], normalized to obtain Fo = [0, 0.0067, 0.02, 0.04, 1, 0.0467, 0.0267, 0.0133, 0.7333, 0.02, 0.04, 0.0067, 0.0467], fault characteristic spectral index Pgear = 4, then the target amplitudes corresponding to Pgear and its multiples are 1, 0.7333 and 0.0467, and the spectral prominence characteristic value Peig = 1.

[0070] It should be noted that the main function of the spectral highlight feature value is to ensure that a fault in a regular sample or other component is not identified as a gear fault.

[0071] Spectral multi-order eigenvalues: After obtaining the statistically normalized spectrum Fo through the aforementioned steps, the total number K of the spectra corresponding to the fault feature spectral number Pgear and its multiples in Fo is counted. Then, the number S of spectral lines with amplitudes not lower than a preset amplitude threshold in the spectrum corresponding to the fault feature spectral number Pgear and its multiples in the normalized spectrum Fo is counted. The ratio of S to K is used as the spectral multi-order characteristic value of the target gear.

[0072] Following the previous example, the spectrum corresponding to the fault characteristic spectrum number Pgear and its multiples are numbers 4, 8 and 12, a total of 3, so K = 3. The amplitudes corresponding to these spectra are 1, 0.7333 and 0.0467, respectively. If the threshold is set to 0.155 and S = 2, then the multi-order characteristic value of the spectrum of the target gear = 2 / 3 = 0.667.

[0073] It should be noted that the main function of the spectral multi-order eigenvalue is to ensure that conventional gear faults are not identified as cracks or broken teeth.

[0074] Furthermore, based on the foregoing embodiments, the present invention also provides a gear fault determination method. This method is mainly used to determine whether a target gear is faulty based on the detection results of the fault detection method provided in the foregoing embodiments. Similar to the gear fault detection method provided in the foregoing embodiments, the gear fault determination method provided in this embodiment is applied to a controller. This controller can be any controller in the working machinery to which the gear to be fault detected belongs. Of course, it can also be applied to other controllers independent of the working machinery, such as a server on the network side.

[0075] See Figure 6 The process of the gear fault determination method provided in this embodiment of the invention may include: S200. According to the gear fault detection method, obtain multiple fault detection results of the target gear.

[0076] The gear fault detection method mentioned in this step can be any of the gear fault detection methods provided in the foregoing embodiments. The target gear is tested multiple times by the gear fault detection method to obtain multiple fault detection results.

[0077] As an optional implementation, a preset stack can be provided to store the fault detection results. Based on the data storage principle of a stack, which stores data according to the first-in, first-out (FIFO) principle, in practical applications, the fault detection results in the preset stack need to be updated according to the FIFO principle.

[0078] S210. Calculate the fourth and fifth statistical values ​​based on multiple fault detection results.

[0079] In this embodiment, the fourth statistic is the number of fault detection results that indicate a fault among multiple fault detection results, and the fifth statistic is the number of consecutive fault detection results that indicate a fault among multiple fault detection results.

[0080] The fourth statistical value reflects the cumulative nature of the target gear's fault points, while the fifth statistical value reflects the continuous nature of the target gear's fault points.

[0081] S220. Determine whether the target gear is faulty based on the fourth and fifth statistical values.

[0082] Specifically, if the fourth statistical value is greater than the preset cumulative threshold, or the fifth statistical value is greater than the preset continuous threshold, the target gear is determined to be faulty; conversely, if the fourth statistical value is less than or equal to the preset cumulative threshold and the fifth statistical value is less than or equal to the preset continuous threshold, the target gear is determined to be normal.

[0083] It should be noted that in practical applications, the setting of the preset cumulative threshold and the preset continuous threshold needs to be comprehensively considered in conjunction with factors such as the detection accuracy requirements, the computing power of the controller hardware, and the structural characteristics of the target gear. This invention does not limit the specific values ​​of the two.

[0084] Understandably, in order to provide accuracy in the judgment results, especially when using a preset stack to store the detection results, the target gear can be finally determined to be faulty after the preset stack is full of a preset number of fault detection results, according to steps S210 and S220. Correspondingly, when the preset stack is not full, the fault detection results can continue to be accumulated according to the first-in-first-out principle.

[0085] In summary, the gear fault determination method provided by the embodiments of the present invention, based on the fault detection method provided in the aforementioned embodiments, obtains multiple fault detection results of the target gear and ultimately determines whether the target gear is faulty. This can provide maintenance personnel with a basis for inspection and repair, which helps them to repair and replace the gears, effectively reducing the probability of accidents and improving the operational safety of the machinery.

[0086] The gear fault detection device provided by this invention is described below. This device belongs to the same concept as the gear fault detection method provided in the embodiments of this application, and can execute the gear fault detection method provided in any embodiment of this application. It possesses the corresponding functional modules and beneficial effects for executing the gear fault detection method. Technical details not described in detail in this embodiment can be found in the gear fault detection method provided in the embodiments of this application, and will not be repeated here.

[0087] See Figure 7 , Figure 7 This is a structural block diagram of a gear fault detection device provided in an embodiment of the present invention. The detection device provided in this embodiment of the present invention includes: The first acquisition unit 10 is used to acquire the vibration signal and rotation speed signal of the target gear; Construction unit 20 is used to construct detection samples based on vibration signals and rotational speed signals; Extraction unit 30 is used to extract fault detection feature values ​​of the target gear based on the detection sample; The detection unit 40 is used to input the fault detection feature values ​​of the target gear into the pre-trained fault detection model to obtain the fault detection result of the target gear. The fault detection model takes the fault detection feature values ​​of the gear as input and the fault detection results of the gear as output, and trains a neural network to obtain the model.

[0088] Optionally, the building unit 20, used to construct a detection sample based on vibration and rotational speed signals, includes: Extract the vibration signal corresponding to the target frequency from the vibration signal, and perform envelope processing on the vibration signal corresponding to the target frequency to obtain the processed vibration signal; Based on the rotation speed signal and the preset sampling frequency, the processed vibration signal is sampled at equal angular intervals to obtain the detection sample; The detection sample includes multiple detection sub-samples, and each detection sub-sample includes multiple vibration signal amplitudes arranged in time sequence.

[0089] Optionally, the extraction unit 30 is used for the process of extracting the spectral salient features, including: The spectrum of the detected sample is normalized to obtain a normalized spectrum; Based on the detected samples, the fault characteristic spectrum of the target gear is determined, and the target amplitude is extracted; Wherein, the target amplitude is the amplitude corresponding to the fault feature spectral number and the spectral number that is an integer multiple of the fault feature spectral number in the normalized spectrum; The target amplitude with the largest value among all the target amplitudes is determined as the spectral salient feature value.

[0090] Optionally, the extraction unit 30 is used to extract the spectral multi-order characteristic values, including: The spectrum of the detected sample is normalized to obtain a normalized spectrum, and the fault characteristic spectrum of the target gear is determined based on the detected sample. The first statistical value is obtained by counting the total number of fault feature spectral numbers and integer multiples of the fault feature spectral numbers in the normalized spectrum. The ratio of the second statistical value to the first statistical value is determined as the spectral multi-order characteristic value; The second statistical value is the total number of spectral numbers whose amplitude is greater than or equal to a preset amplitude threshold among the fault feature spectral numbers and spectral numbers that are integer multiples of the fault feature spectral numbers.

[0091] Optionally, the extraction unit 30, for the process of extracting the amplitude feature values ​​of the target gear, includes: Construct an inverse transformation sample based on the detected sample; Divide the inverse transform samples into multiple inverse transform sub-samples; Extract the maximum value of the vibration signal amplitude from each inverse transform sub-sample; The average of the maximum values ​​is determined as the amplitude characteristic value of the target gear.

[0092] Optionally, the extraction unit 30, for extracting the sharp impact feature values ​​of the target gear, includes the following process: Construct an inverse transformation sample based on the detected sample; The inverse transform sample is divided into multiple inverse transform sub-samples; Calculate the kurtosis index for each inverse transform subsample; The number of kurtosis indices that are less than or equal to the preset kurtosis index threshold is counted to obtain the third statistical value; The ratio of the third statistical value to the total number of each inverse transform subsample is determined as the sharp impact characteristic value of the target gear.

[0093] Optionally, the extraction unit 30 is used to construct an inverse transform sample based on the detected sample, including: The fault characteristic spectrum of the target gear is determined based on the detected samples; Extract the fault feature spectrum and the vibration signal amplitude corresponding to the spectrum of integer multiples of the fault feature spectrum from the detection sample to obtain the constructed sample; Perform an inverse Fourier transform on the constructed sample to obtain the inverse transform sample.

[0094] The gear fault determination device provided by this invention is described below. This device belongs to the same conceptual framework as the gear fault determination method provided in the embodiments of this application, and can execute the gear fault determination method provided in any embodiment of this application. It possesses the corresponding functional modules and beneficial effects for executing the gear fault determination method. Technical details not described in detail in this embodiment can be found in the gear fault determination method provided in the embodiments of this application, and will not be repeated here.

[0095] See Figure 8 , Figure 8 The diagram shown is a structural block diagram of a gear fault determination device provided in an embodiment of the present invention. The determination device provided in this embodiment may include: The second acquisition unit 50 is used to acquire multiple fault detection results of the target gear according to the gear fault detection method described above. Statistical unit 60 is used to calculate a fourth statistical value and a fifth statistical value based on multiple fault detection results; Among them, the fourth statistic is the number of fault detection results that indicate a fault among multiple fault detection results, and the fifth statistic is the number of consecutive fault detection results that indicate a fault among multiple fault detection results. The determination unit 70 is used to determine whether the target gear is faulty based on the fourth statistical value and the fifth statistical value.

[0096] Optionally, the determining unit 70 is used to determine whether the target gear is faulty based on the fourth statistical value and the fifth statistical value, including: If the fourth statistical value is greater than the preset cumulative threshold, or if the fifth statistical value is greater than the preset continuous threshold, the target gear is determined to be faulty. If the fourth statistical value is less than or equal to the preset cumulative threshold and the fifth statistical value is less than or equal to the preset continuous threshold, the target gear is determined to be normal.

[0097] Optionally, embodiments of the present invention also provide a controller, including: a memory and a processor; the memory stores a program suitable for execution by the processor to implement the gear fault detection method provided in the foregoing embodiments, or the gear fault determination method provided in any of the foregoing embodiments.

[0098] Optionally, embodiments of the present invention also provide a working machine, including: a working component and a controller provided in the foregoing embodiments, wherein, The controller is connected to the working component, and the working component is equipped with gears.

[0099] Those skilled in the art will understand that the contents disclosed herein can be varied and modified in many ways. For example, the various devices or components described above can be implemented in hardware, or in software, firmware, or a combination of some or all of the three.

[0100] Furthermore, while this disclosure makes various references to certain units in systems according to embodiments of this disclosure, any number of different units may be used and operated on clients and / or servers. The units are merely illustrative, and different aspects of the system and method may use different units.

[0101] This disclosure uses flowcharts to illustrate the steps of a method according to embodiments of this disclosure. It should be understood that the preceding or following steps are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes.

[0102] Those skilled in the art will understand that all or part of the steps in the above methods can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium, such as a read-only memory. Optionally, all or part of the steps in the above embodiments can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiments can be implemented in hardware or as a software functional module. This disclosure is not limited to any particular combination of hardware and software.

[0103] Unless otherwise defined, all terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. It should also be understood that terms such as those defined in a common dictionary should be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and not as having an idealized or highly formalized meaning, unless expressly defined herein.

[0104] The foregoing description is intended to illustrate the present disclosure and should not be construed as limiting it. While several exemplary embodiments of the present disclosure have been described, those skilled in the art will readily understand that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present disclosure. Therefore, all such modifications are intended to be included within the scope of the present disclosure as defined by the claims. It should be understood that the foregoing description is intended to illustrate the present disclosure and should not be construed as limiting it to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present disclosure is defined by the claims and their equivalents.

Claims

1. A method for detecting gear faults, characterized in that, include: Acquire the vibration and rotational speed signals of the target gear; A detection sample is constructed based on the vibration signal and the rotation speed signal; Based on the detected samples, the fault detection feature values ​​of the target gear are extracted; The fault detection feature values ​​of the target gear are input into a pre-trained fault detection model to obtain the fault detection results of the target gear. The fault detection model is obtained by training a neural network, with the fault detection feature value of the gear as input and the fault detection result of the gear as output. The fault detection feature values ​​include spectral multi-order feature values. The process of extracting the spectral multi-order feature values ​​includes: The spectrum of the detected sample is normalized to obtain a normalized spectrum, and the fault characteristic spectrum of the target gear is determined based on the detected sample. The first statistical value is obtained by counting the total number of fault feature spectral numbers and integer multiples of the fault feature spectral numbers in the normalized spectrum. The ratio of the second statistical value to the first statistical value is determined as the spectral multi-order characteristic value; The second statistical value is the total number of spectral numbers whose amplitude is greater than or equal to a preset amplitude threshold among the fault feature spectral numbers and spectral numbers that are integer multiples of the fault feature spectral numbers.

2. The method according to claim 1, characterized in that, A detection sample is constructed based on the vibration signal and the rotational speed signal, including: Extract the vibration signal corresponding to the target frequency from the vibration signal, and perform envelope processing on the vibration signal corresponding to the target frequency to obtain the processed vibration signal; Based on the rotation speed signal and the preset sampling frequency, the processed vibration signal is sampled at equal angular intervals to obtain a detection sample; The detection sample includes multiple detection sub-samples, and each detection sub-sample includes multiple vibration signal amplitudes arranged in time sequence.

3. The method according to claim 1, characterized in that, The fault detection feature values ​​also include at least one of the following: rotational speed feature value, amplitude feature value, sharp impact feature value, and spectral prominence feature value.

4. The method according to claim 3, characterized in that, The process of extracting the spectral salient features includes: The spectrum of the detected sample is normalized to obtain a normalized spectrum; Based on the detected samples, the fault characteristic spectrum of the target gear is determined, and the target amplitude is extracted; Wherein, the target amplitude is the amplitude corresponding to the fault feature spectral number and the spectral number that is an integer multiple of the fault feature spectral number in the normalized spectrum; The target amplitude with the largest value among all the target amplitudes is determined as the spectral salient feature value.

5. The method according to claim 3, characterized in that, The process of extracting the amplitude feature values ​​of the target gear includes: Construct an inverse transformation sample based on the detected sample; The inverse transform sample is divided into multiple inverse transform sub-samples; Extract the maximum value of the vibration signal amplitude from each of the inverse transform sub-samples; The average of the maximum values ​​is determined to be the amplitude characteristic value of the target gear.

6. The method according to claim 3, characterized in that, The process of extracting the sharp impact feature value of the target gear includes: Construct an inverse transformation sample based on the detected sample; The inverse transform sample is divided into multiple inverse transform sub-samples; Calculate the kurtosis index for each of the inverse transform subsamples; The number of kurtosis indices that are less than or equal to a preset kurtosis index threshold is counted to obtain a third statistical value; The ratio of the third statistical value to the total number of each inverse transform subsample is determined as the sharp impact characteristic value of the target gear.

7. The method according to claim 5 or 6, characterized in that, Constructing inverse transform samples based on the detected samples includes: The fault characteristic spectrum of the target gear is determined based on the detected samples; Extract the fault feature spectrum and the vibration signal amplitude corresponding to the spectrum of integer multiples of the fault feature spectrum from the detection sample to obtain the constructed sample; Perform an inverse Fourier transform on the constructed sample to obtain the inverse transform sample.

8. The method according to any one of claims 1-6, characterized in that, Also includes: The fault detection model is updated based on the test samples and the actual state of the target gear.

9. A method for determining gear faults, characterized in that, include: According to any one of claims 1-8, the gear fault detection method obtains multiple fault detection results of the target gear; Based on the multiple fault detection results, a fourth statistical value and a fifth statistical value are calculated; The fourth statistical value is the number of fault detection results that indicate a fault among the multiple fault detection results, and the fifth statistical value is the number of consecutive fault detection results that indicate a fault among the multiple fault detection results. The fourth and fifth statistical values ​​are used to determine whether the target gear is faulty.

10. The method according to claim 9, characterized in that, Determining whether the target gear is faulty based on the fourth and fifth statistical values ​​includes: If the fourth statistical value is greater than a preset cumulative threshold, or if the fifth statistical value is greater than a preset continuous threshold, the target gear is determined to be faulty. If the fourth statistical value is less than or equal to the preset cumulative threshold, and the fifth statistical value is less than or equal to the preset continuous threshold, the target gear is determined to be normal.

11. The method according to claim 9, characterized in that, The multiple fault detection results are stored in a preset stack, and the method further includes: The fault detection results in the preset stack are updated according to the first-in, first-out (FIFO) principle.

12. A controller, characterized in that, include: Memory and processor; The memory stores a program suitable for execution by the processor to implement the gear fault detection method according to any one of claims 1 to 8, or the gear fault determination method according to any one of claims 9 to 11.

13. A type of operating machinery, characterized in that, include: The operating component and the controller according to claim 12, wherein, The controller is connected to the working component, and the working component is equipped with gears.

Images