Slewing bearing fault detection method and system and application thereof

Abstract

The invention discloses a slewing bearing fault detection method and system and application thereof.According to the scheme, working parameters in the working process of a slewing bearing are ingeniously monitored, and whether abnormity exists or not is judged by combining corresponding parameter thresholds (for example, whether the volume of audio noise exceeds a normal value or not, whether the amplitude of vibration is abnormal or not and the like); a first judgment result recording a plurality of sub-results is generated, and then a time sequence and a sliding window are combined to count the number of times that the slewing bearing is abnormal in a preset time interval in the working process, so that whether the slewing bearing has fault suspicion or not is preliminarily judged; according to the scheme, implementation is flexible and simple, and convenient and efficient abnormity feedback can be provided for field control personnel and background management personnel; according to the scheme, after the slewing bearing is adjusted to the set state, self-inspection data are collected for secondary detection and judgment, and more reliable fault judgment feedback is provided for operators and background management personnel.

Description

Technical Field

[0001] The present invention relates to the application fields of crane detection technology and deep learning network technology, and in particular to a slewing bearing fault detection method, system and application thereof. Background Art

[0002] As a popular heavy-duty transport equipment, cranes have the advantages of strong transport capacity and stable equipment use, so they are widely used in production workshops, storage workshops, ports, construction sites, etc. As an important component of cranes, the work safety and reliability research of slewing bearings has always been a hot topic in this field.

[0003] Among them, the slewing bearing is also called the turntable bearing. It has a wide range of applications in real industry, so it is often called "the joint of the machine". The slewing bearing is an important transmission component required for machinery that needs to make relative rotational motion between two objects and simultaneously bear axial force, radial force, and tipping moment. With the rapid development of the machinery industry, slewing bearings are widely used in cranes, ship equipment, engineering machinery, light industrial machinery, metallurgical machinery, medical machinery, industrial machinery and other industries.

[0004] Due to the complex working environment of the crane, the slewing bearing may be disturbed by a variety of external and internal factors during the process of being driven. Due to the complex and changeable working environment of the slewing bearing during service, there are many factors that affect its service life. Most of the existing simulation model-based or data-driven methods predict whether the slewing bearing has a fault or service life through a single or a small number of factors. In actual applications, there are often many factors that lead to slewing bearing failures and the end of service life, which leads to estimation errors. At the same time, the current full life cycle monitoring data of the slewing bearing is difficult to reuse and learn from due to statistical factors and small data volume. Therefore, how to carry out working status evaluation and abnormality detection of large slewing bearings under dynamic service, and formulate scientific and reasonable abnormality detection plans and early warning plans to ensure their safe and reliable operation are very positive and realistic topics. Summary of the invention

[0005] In view of this, the object of the present invention is to propose a slewing bearing fault detection method, system and application thereof, which are reliable in implementation, flexible in application, have good detection feedback reliability, and can reduce the interference of occasional abnormalities on fault detection results.

[0006] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0007] A slewing bearing fault detection method, comprising:

[0008] S01, responding to the signal that the slewing bearing is driven and moves, collecting the working parameters of the slewing bearing in real time when it moves, and obtaining working monitoring data;

[0009] S02, extracting target feature data from the working monitoring data, and then judging the target feature data according to preset conditions to generate a first judgment result to preliminarily determine whether it points to abnormal working of the slewing bearing, and at the same time associating the collection time of the corresponding target feature data with the first judgment result;

[0010] S03, collecting a plurality of first judgment results in time series, and when the first judgment result indicates that the slewing bearing is working abnormally, generating an abnormal information mark in the time series according to the collection time of the corresponding target feature data;

[0011] S04. Slide a sliding window with a preset time span on the time series, count the number of abnormal information marks in the time span where the sliding window is located, and when the number is greater than a preset value N, generate fault prompt information.

[0012] As a possible implementation mode, further, the working parameters of the slewing bearing during the operation of the scheme include: ring rotation speed, ring rotation amplitude, ring vibration data, audio data, and ring temperature data; the working monitoring data includes the ring rotation speed, ring rotation amplitude, ring vibration data, audio data, ring temperature data and collection time corresponding to the slewing bearing;

[0013] The slewing bearing comprises a pair of rings and rollers arranged between the pair of rings, one of the rings is provided with a gear ring and rotates under the driving force, which is set as a moving ring, and the other ring of the slewing bearing is set as a fixed ring; the moving ring is provided with an acceleration sensor, a displacement sensor and a plurality of vibration sensors; the fixed ring is provided with a plurality of audio sensors and a plurality of temperature sensors in a ring array;

[0014] The acceleration sensor is used to collect the rotation speed of the moving ring when the moving ring rotates;

[0015] The displacement sensor is used to collect the rotation amplitude of the moving ring when the moving ring rotates;

[0016] The vibration sensor is used to collect the vibration data of the moving ring when the moving ring rotates, and each vibration sensor collects the vibration data of the ring respectively, and each vibration sensor also has a unique code and position information;

[0017] The audio sensor is used to collect audio data within a preset range of the slewing bearing when the movable ring rotates, and each audio sensor collects audio data respectively, and each audio sensor also has a unique code and position information;

[0018] The temperature sensor is used to collect the ring temperature data on the fixed ring when the movable ring rotates, and each temperature sensor collects the ring temperature data respectively, and each temperature sensor also has a unique code and position information.

[0019] As a preferred implementation option, preferably, in the present scheme S02, the target feature data extracted from the working monitoring data is one or more of the ring vibration data, audio data, and ring temperature data; they are respectively compared and judged with the corresponding preset thresholds to generate judgment sub-results, and then aggregated to form the first judgment result.

[0020] As a preferred implementation option, preferably, in the present solution S02, when one of the target feature data exceeds a preset threshold, its judgment sub-result is defined as indicating that the slewing bearing is operating abnormally.

[0021] As a preferred implementation option, preferably, in the present solution S02, the preset thresholds of the ring vibration data and the audio data are determined by the mapping relationship model between the operating state of the slewing bearing and the audio and vibration, respectively, and the mapping relationship model is established based on a query table, which includes:

[0022] (1) After lubricating the roller of the slewing support, construct the operation of the slewing support under different ring rotation speeds and ring rotation amplitudes, and then collect sound wave data S and vibration data Z of preset time lengths under different conditions; wherein the sound wave data includes volume decibel data, and the vibration data includes amplitude data;

[0023] (2) Based on the acoustic wave data S and vibration data Z, calculate the mean of the collected time to obtain the acoustic wave data mean S 均值 , vibration data mean Z 均值 ;

[0024] (3) According to the mean value S of the acoustic wave data 均值 , vibration data mean Z 均值 Set the thresholds respectively, the sound wave data threshold is N×S 均值 , the vibration data threshold is M×Z 均值 , where N and M are both greater than 1;

[0025] (4) The ring rotation speed and ring rotation amplitude of the slewing bearing under different conditions are taken as motion parameters, and the corresponding acoustic wave data mean S 均值 , vibration data mean Z 均值 The sound wave data threshold N×S formed based on 均值 , vibration data threshold M×Z 均值As limiting parameters, a lookup table of threshold values ​​of the slewing bearing operation status and audio and vibration is established, i.e., a mapping relationship model;

[0026] In S02, when the ring vibration data and audio data are compared with the corresponding preset thresholds respectively, the sound wave data threshold and vibration data threshold are extracted from the threshold query table through the ring rotation speed and ring rotation amplitude in the working monitoring data. When there are no corresponding ring rotation speed and ring rotation amplitude items in the threshold query table, the sound wave data threshold and vibration data threshold corresponding to the item closest to the ring rotation speed are extracted as the preset thresholds for comparison and judgment.

[0027] As a preferred implementation option, preferably, in the present scheme S03, when one of the judgment sub-results in the first judgment result points to abnormal operation of the slewing bearing, an abnormal information mark is generated in the time series according to the collection time of its corresponding target feature data.

[0028] As a preferred implementation option, preferably, in solution S04, the time span of the sliding window is 10 to 120 seconds.

[0029] Based on the above, the present invention further provides a slewing bearing fault self-detection method, which includes the slewing bearing fault detection method described above, and further includes:

[0030] A01. Respond to the fault self-check instruction made by the operator according to the fault prompt information, make the slewing bearing work with preset motion parameters to the preset state, collect the working parameters of the slewing bearing in real time, and obtain self-check monitoring data;

[0031] A02. Extract target data from the self-test monitoring data and pre-process it according to preset conditions;

[0032] A03. Input the preprocessed target data and other data in the self-test monitoring data as input items into the trained detection neural network for detection to generate a detection result.

[0033] As a preferred implementation option, preferably, the preset motion parameters of this solution include the ring rotation speed and the ring rotation amplitude, and the preset state is that the slewing support performs reciprocating motion or rotational motion within the preset ring rotation amplitude at the preset ring rotation speed;

[0034] The real-time data collection of the working parameters of the slewing bearing includes: ring rotation speed, ring rotation amplitude, ring vibration data, audio data and ring temperature data.

[0035] As a preferred implementation option, preferably, in the present solution A01, before the slewing bearing is operated with preset motion parameters, grease is added to the rollers of the slewing bearing through an external lubrication filling device.

[0036] As a preferred implementation option, preferably, in this scheme A02, the target data includes ring vibration data and audio data, and the corresponding vibration sensors and audio sensors are preprocessed according to preset conditions after collecting signals to obtain ring vibration data and audio data, and the preprocessing includes: signal denoising and time-frequency analysis. The data after time-frequency analysis are sequentially input into the convolutional neural network CNN and the long short-term memory network LSTM to obtain processed signal data.

[0037] As a preferred implementation option, preferably, the method for preprocessing the signals collected by the vibration sensor and the audio sensor in this solution includes:

[0038] (1) Signal denoising

[0039] Use a bandpass filter to denoise the signals collected by the vibration sensor and audio sensor to remove low-frequency and high-frequency interference signals so that the signal is concentrated in the frequency band containing fault information; assuming that the original signal is x(t) and the filtered signal is y(t), it can be expressed as:

[0040] y(t)=x(t)*h(t)

[0041] Where h(t) is the impulse response function of the filter, and * represents the convolution function;

[0042] (2) Time-frequency analysis

[0043] The short-time Fourier transform (STFT) is used to transform the signal from the time domain to the time-frequency domain to capture the transient characteristics of the signal. The formula is defined as follows:

[0044]

[0045] Among them, X(t,ω) is the complex representation in the time-frequency domain, which is the STFT result, that is, the complex value representation of the signal at time t and frequency ω, which represents the amplitude and phase information of the signal at ω within a small time window at time t; x(τ) is the time domain representation of the original signal x(t), τ represents the time variable, ω(τ-t) is the time window function, which weights the signal within a period of time near time t, τ in the time window function represents the time variable, ω is the frequency variable, e -jωτis a complex exponential function, which is the core part of Fourier transform and is used to convert signals from time domain to frequency domain. j in the complex exponential function is an imaginary unit, which represents the imaginary part of a complex number. dτ is an integral variable, which represents the integration of time τ. The integration is to accumulate the frequency components of the signal weighted by the window function on the entire time axis.

[0046] (3) Convolutional Neural Network (CNN) processing

[0047] The spatial features of time-frequency data are extracted through the convolutional neural network CNN. The time-frequency feature matrix X(t,ω) obtained by time-frequency analysis is used as input. The output of the lth convolutional layer of the convolutional neural network CNN can be expressed as:

[0048]

[0049] in, is the output feature at position (i, j); is the weight matrix of the convolution kernel, the size of the convolution kernel is k×k; b l is the bias term, σ(·) is the activation function;

[0050] (4) Long Short-Term Memory Network (LSTM) Processing

[0051] The convolution kernel of the convolutional neural network CNN is used to extract features layer by layer, and the extracted features are used as input items of the long short-term memory network LSTM; the long short-term memory network LSTM is used to capture the time dependency of the data, which includes the following calculation steps:

[0052] Forget gate calculation:

[0053] f t =σ(W f ·[h t-1 , x t ]+b f )

[0054] Input gate calculation:

[0055] i t =σ(W i ·[h t-1 , x t ]+b i )

[0056] Candidate memory cell state calculation:

[0057]

[0058] Memory unit status update:

[0059]

[0060] Output gate calculation and final output:

[0061] o t =σ(W o ·[h t-1 , x t ]+b o )

[0062] h t =o t ⊙tanh(C t )

[0063] Among them, x t is the input feature sequence of the long short-term memory network LSTM; h t is the output state of the long short-term memory network LSTM; W f , W i , W C , W o is the weight matrix; b f 、b i 、b c 、b o is the bias term; σ(·) is the Sigmoid activation function, tanh(·) is the hyperbolic tangent activation function; ⊙ represents element-wise multiplication.

[0064] Based on the above, the present invention also provides a slewing bearing fault detection system, which includes:

[0065] A working monitoring module is used to respond to the signal that the slewing bearing is driven and moves, collect the working parameters of the slewing bearing in real time when it moves, and obtain working monitoring data;

[0066] A data extraction module, used to extract target feature data from work monitoring data;

[0067] A data judgment module is used to judge the target characteristic data according to preset conditions, generate a first judgment result, and preliminarily determine whether it points to abnormal operation of the slewing bearing, and at the same time associate the collection time of the corresponding target characteristic data with the first judgment result;

[0068] An information marking module is used to collect multiple first judgment results in time series, and when the first judgment result points to abnormal operation of the slewing bearing, an abnormal information mark is generated in the time series according to the collection time of the corresponding target characteristic data;

[0069] The information feedback module is used to slide a sliding window with a preset time span on the time series, count the number of abnormal information tags in the time span of the sliding window, and generate fault prompt information when the number is greater than a preset value N.

[0070] By adopting the above technical scheme, the present invention has the following beneficial effects compared with the prior art: the slewing bearing fault detection method of the present scheme cleverly monitors the working parameters of the slewing bearing during operation, and then combines the corresponding parameter thresholds to determine whether there is an abnormality (for example, whether the volume of the audio noise exceeds the normal value, whether the amplitude of the vibration is abnormal, whether the local temperature of the slewing bearing ring has abnormalities, etc.); by generating a first judgment result recording multiple sub-results, and then combining the time series and the sliding window to count the number of abnormalities of the slewing bearing within a preset time interval during operation, it is preliminarily determined whether there is a suspected fault in the slewing bearing; the scheme is not only flexible and simple to implement, but also can be used to detect whether there is a suspected fault in the slewing bearing. It is simple and can provide convenient and efficient abnormal feedback for on-site control personnel and background management personnel; on this basis, this solution further responds to the fault self-check instruction made by the operator according to the fault prompt information, so that the slewing bearing works to the preset state with the preset motion parameters, and then collects the working parameters of the slewing bearing in real time to obtain self-check monitoring data; in this way, the slewing bearing can be subjected to secondary fault detection. After adjusting its working condition to the set state, it is re-detected by using the detection neural network, so that the fault judgment of the slewing bearing can better combine the previous historical data to make a secondary judgment on whether the slewing bearing is faulty, thereby improving the reliability of abnormal monitoring of the slewing bearing. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0072] Figure 1 It is a schematic diagram of a brief implementation process of the slewing bearing fault detection method of this scheme;

[0073] Figure 2 It is a schematic diagram of the principle of realizing statistics of abnormal information mark A within a preset time range by sliding a sliding window H on a time series T in the slewing bearing fault detection method of the present scheme;

[0074] Figure 3 It is one of the simplified implementation structure diagrams of the slewing bearing mentioned in this scheme;

[0075] Figure 4 This is a brief implementation flow diagram of the slewing bearing fault self-checking of this scheme;

[0076] Figure 5 This is a schematic diagram of the unit module connection of the slewing bearing fault detection system of this scheme. DETAILED DESCRIPTION

[0077] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples. It is particularly noted that the following examples are only used to illustrate the present invention, but are not intended to limit the scope of the present invention. Similarly, the following examples are only partial embodiments of the present invention rather than all embodiments, and all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0078] like Figure 1 As shown, this embodiment provides a slewing bearing fault detection method, which includes:

[0079] S01, responding to the signal that the slewing bearing is driven and moves, collecting the working parameters of the slewing bearing in real time when it moves, and obtaining working monitoring data;

[0080] S02, extracting target feature data from the working monitoring data, and then judging the target feature data according to preset conditions to generate a first judgment result to preliminarily determine whether it points to abnormal working of the slewing bearing, and at the same time associating the collection time of the corresponding target feature data with the first judgment result;

[0081] S03, collecting a plurality of first judgment results in time series, and when the first judgment result indicates that the slewing bearing is working abnormally, generating an abnormal information mark in the time series according to the collection time of the corresponding target feature data;

[0082] S04. Slide a sliding window with a preset time span on the time series, count the number of abnormal information marks in the time span where the sliding window is located, and when the number is greater than a preset value N, generate fault prompt information.

[0083] Combination Figure 2 As shown, in this scheme, T is taken as the timeline. When the first judgment result points to the abnormal operation of the slewing bearing, A is marked on the timeline T as the abnormal information mark according to the collection time of its corresponding data, and then the sliding window H is slid on the timeline T. At the same time, the number of marks A in the time window H is counted, and the generation trigger of the fault prompt information is realized in combination with the preset mark number threshold.

[0084] In terms of working parameter monitoring, the working parameters of the slewing bearing during operation described in this scheme include: ring rotation speed, ring rotation amplitude, ring vibration data, audio data, and ring temperature data; the working monitoring data includes the ring rotation speed, ring rotation amplitude, ring vibration data, audio data, ring temperature data and collection time corresponding to the slewing bearing.

[0085] Combination Figure 3As shown, in this scheme, the slewing bearing 100 includes a pair of rings 101, 102 and a roller arranged between the pair of rings, one of the rings 102 is provided with a gear ring 1021 and rotates under the driving force, which is set as the moving ring 102, and the other ring 101 of the slewing bearing 100 is set as the fixed ring 101; the moving ring 102 is provided with an acceleration sensor 3, a displacement sensor 4 and a plurality of vibration sensors 5; the fixed ring 101 is provided with a plurality of audio sensors 6 and a plurality of temperature sensors 7 in a ring array.

[0086] The acceleration sensor 3 is used to collect the rotation speed of the moving ring 102 when the moving ring 102 rotates;

[0087] The displacement sensor 4 is used to collect the rotation amplitude of the moving ring 102 when the moving ring 102 rotates;

[0088] The vibration sensor 5 is used to collect the ring vibration data when the moving ring 102 rotates, and each vibration sensor collects the ring vibration data respectively, and each vibration sensor also has a unique code and position information;

[0089] The audio sensor 6 is used to collect audio data within the preset range of the slewing bearing when the moving ring 102 rotates, and each audio sensor collects audio data respectively, and each audio sensor also has a unique code and position information;

[0090] The temperature sensor 7 is used to collect the ring temperature data on the fixed ring 101 when the movable ring 102 rotates, and each temperature sensor collects the ring temperature data respectively, and each temperature sensor also has a unique code and position information.

[0091] As a preferred implementation option, preferably, in this solution S02, the target feature data extracted from the working monitoring data is one or more of the ring vibration data, audio data, and ring temperature data; they are respectively compared and judged with the corresponding preset thresholds to generate judgment sub-results, and then aggregated to form a first judgment result. In this solution S02, by extracting the target feature data (ring vibration data, audio data, ring temperature data) from the working monitoring data, and then judging the target feature data (ring vibration data, audio data, ring temperature data) according to preset conditions, that is, comparing and judging the preset parameter thresholds corresponding to the ring vibration data, audio data, and ring temperature data to determine whether they exceed them, and then aggregate to generate a first judgment result to preliminarily determine whether it points to abnormal working of the slewing bearing, and at the same time, the collection time of the corresponding target feature data is associated with the first judgment result.

[0092] For the convenience of distinction, as a better implementation option, preferably, in the present scheme S02, when one of the target feature data exceeds the preset threshold, its judgment sub-result is defined as pointing to abnormal operation of the slewing bearing, that is, the first judgment result includes multiple sub-results, which respectively correspond to whether the ring vibration data, audio data, and ring temperature data are abnormal.

[0093] Since the slewing bearing may cause aggravated vibration and increased noise due to factors such as rotation speed and temperature under different working conditions, and the slewing bearing itself may not have any faults, in order to improve the reliability of the first judgment result during comparative judgment, as a preferred implementation option, preferably, in this solution S02, the preset thresholds of the ring vibration data and the audio data are determined by the mapping relationship model between the operating state of the slewing bearing and the audio and vibration, respectively, and the mapping relationship model is established based on a query table, which includes:

[0094] (1) After lubricating the roller of the slewing support, construct the operation of the slewing support under different ring rotation speeds and ring rotation amplitudes, and then collect sound wave data S and vibration data Z of preset time lengths under different conditions; wherein the sound wave data includes volume decibel data, and the vibration data includes amplitude data;

[0095] (2) Based on the acoustic wave data S and vibration data Z, calculate the mean of the collected time to obtain the acoustic wave data mean S 均值 , vibration data mean Z 均值 ;

[0096] (3) According to the mean value S of the acoustic wave data 均值 , vibration data mean Z 均值 Set the thresholds respectively, the sound wave data threshold is N×S 均值 , the vibration data threshold is M×Z 均值 , where N and M are both greater than 1;

[0097] (4) The ring rotation speed and ring rotation amplitude of the slewing bearing under different conditions are taken as motion parameters, and the corresponding acoustic wave data mean S 均值 , vibration data mean Z 均值 The sound wave data threshold N×S formed based on 均值 , vibration data threshold M×Z 均值 As the limiting parameters, a lookup table of threshold values ​​of the slewing bearing operation status and audio and vibration is established, namely, a mapping relationship model.

[0098] On this basis, in the present scheme S02, when the ring vibration data and audio data are compared and judged with the corresponding preset thresholds respectively, the sound wave data threshold and vibration data threshold are extracted from the threshold query table through the ring rotation speed and ring rotation amplitude in the working monitoring data. When there are no corresponding ring rotation speed and ring rotation amplitude items in the threshold query table, the sound wave data threshold and vibration data threshold corresponding to the item closest to the ring rotation speed are extracted as the preset thresholds for comparison and judgment.

[0099] As a preferred implementation option, preferably, in the present solution S03, when one of the judgment sub-results in the first judgment result points to abnormal operation of the slewing bearing, an abnormal information mark is generated in the time series according to the collection time of the corresponding target feature data.

[0100] Regarding the time span of the sliding window, as an example of a preferred implementation option, preferably, in the present solution S04, the time span of the sliding window is 10 to 120 seconds.

[0101] The slewing bearing fault detection method of this scheme cleverly monitors the working parameters of the slewing bearing during its operation, and then combines the corresponding parameter thresholds to determine whether there is an abnormality (for example, whether the volume of the audio noise exceeds the normal value, whether the amplitude of the vibration is abnormal, whether the local temperature abnormality of the slewing bearing ring occurs, etc.); by generating a first judgment result with multiple sub-results recorded, and then combining the time series and the sliding window to count the number of abnormalities that occur in the slewing bearing within a preset time interval during its operation, it is preliminarily determined whether the slewing bearing is suspected of failure; this scheme is not only flexible and simple to implement, but also can provide convenient and efficient abnormality feedback for on-site control personnel and background management personnel.

[0102] Combination Figure 4 As shown, based on the above, this solution also provides a slewing bearing fault self-detection method, which includes the slewing bearing fault detection method described above, and further includes:

[0103] A01. Respond to the fault self-check instruction made by the operator according to the fault prompt information, make the slewing bearing work with preset motion parameters to the preset state, collect the working parameters of the slewing bearing in real time, and obtain self-check monitoring data;

[0104] A02. Extract target data from the self-test monitoring data and pre-process it according to preset conditions;

[0105] A03. Input the preprocessed target data and other data in the self-test monitoring data as input items into the trained detection neural network for detection to generate a detection result.

[0106] In this scheme, as a preferred implementation option, preferably, the preset motion parameters in this scheme include the ring rotation speed and the ring rotation amplitude, and the preset state is that the slewing support performs reciprocating motion or rotational motion within the preset ring rotation amplitude at the preset ring rotation speed.

[0107] In addition, the real-time data collection of the working parameters of the slewing bearing includes: ring rotation speed, ring rotation amplitude, ring vibration data, audio data and ring temperature data.

[0108] The slewing bearing fault self-inspection method of this scheme serves as a secondary detection and judgment when the slewing bearing is working normally. It mainly adjusts the working state of the slewing bearing to a preset standard state or set state, collects the working parameters again, and then judges the working parameters under a specific state to determine whether the abnormality in the previous work is accidental or real.

[0109] On the basis of the slewing bearing fault detection method, this scheme further responds to the fault self-check instruction made by the operator according to the fault prompt information, makes the slewing bearing work to the preset state with the preset motion parameters, and then collects the working parameters of the slewing bearing in real time to obtain self-check monitoring data; in this way, the slewing bearing can be subjected to secondary fault detection, and after adjusting its working condition to the set state, it is re-detected by using the detection neural network, so that the fault judgment of the slewing bearing can better combine the previous historical data to make a secondary judgment on whether the slewing bearing is faulty, thereby improving the reliability of abnormal monitoring of the slewing bearing.

[0110] Among them, the trained detection neural network can use the existing slewing bearing working data set or the data set of different operating conditions constructed in the laboratory for labeling, and then input into the neural network for training until the model converges. The training principle will not be elaborated here.

[0111] In order to reduce the noise during signal collection, as a better implementation option, preferably, in the present scheme A01, before the slewing bearing is operated with preset motion parameters, grease is added to the rollers of the slewing bearing through an external lubrication filling device.

[0112] In the optimization of signal acquisition of vibration sensors and audio sensors, as a better implementation option, preferably, in this scheme A02, the target data includes ring vibration data and audio data, and the corresponding vibration sensors and audio sensors are preprocessed according to preset conditions after collecting signals to obtain ring vibration data and audio data, and the preprocessing includes: signal denoising and time-frequency analysis. The data after time-frequency analysis are sequentially input into the convolutional neural network CNN and the long short-term memory network LSTM to obtain processed signal data.

[0113] The specific method of preprocessing the signals collected by the vibration sensor and the audio sensor in this solution includes:

[0114] (1) Signal denoising

[0115] Use a bandpass filter to denoise the signals collected by the vibration sensor and audio sensor to remove low-frequency and high-frequency interference signals so that the signal is concentrated in the frequency band containing fault information; assuming that the original signal is x(t) and the filtered signal is y(t), it can be expressed as:

[0116] y(t)=x(t)*h(t)

[0117] Where h(t) is the impulse response function of the filter, and * represents the convolution function;

[0118] (2) Time-frequency analysis

[0119] The short-time Fourier transform (STFT) is used to transform the signal from the time domain to the time-frequency domain to capture the transient characteristics of the signal. The formula is defined as follows:

[0120]

[0121] Among them, X(t,ω) is the complex representation in the time-frequency domain, which is the STFT result, that is, the complex value representation of the signal at time t and frequency ω, which represents the amplitude and phase information of the signal at ω within a small time window at time t; x(τ) is the time domain representation of the original signal x(t), τ represents the time variable, ω(τ-t) is the time window function, which weights the signal within a period of time near time t, τ in the time window function represents the time variable, ω is the frequency variable, e -jωτ is a complex exponential function, which is the core part of Fourier transform and is used to convert signals from time domain to frequency domain. j in the complex exponential function is an imaginary unit, which represents the imaginary part of a complex number. dτ is an integral variable, which represents the integration of time τ. The integration is to accumulate the frequency components of the signal weighted by the window function on the entire time axis.

[0122] (3) Convolutional Neural Network (CNN) processing

[0123] The spatial features of time-frequency data are extracted through the convolutional neural network CNN. The time-frequency feature matrix X(t,ω) obtained by time-frequency analysis is used as input. The output of the lth convolutional layer of the convolutional neural network CNN can be expressed as:

[0124]

[0125] in, is the output feature at position (i, j); is the weight matrix of the convolution kernel, the size of the convolution kernel is k×k; b l is the bias term, σ(·) is the activation function;

[0126] (4) Long Short-Term Memory Network (LSTM) Processing

[0127] The convolution kernel of the convolutional neural network CNN is used to extract features layer by layer, and the extracted features are used as input items of the long short-term memory network LSTM; the long short-term memory network LSTM is used to capture the time dependency of the data, which includes the following calculation steps:

[0128] Forget gate calculation:

[0129] f t =σ(W f ·[h t-1 , x t ]+b f )

[0130] Input gate calculation:

[0131] i t =σ(W i ·[h t-1 , x t ]+b i )

[0132] Candidate memory cell state calculation:

[0133]

[0134] Memory unit status update:

[0135]

[0136] Output gate calculation and final output:

[0137] o t =σ(W o ·[h t-1 , x t ]+b o )

[0138] h t =o t ⊙tanh(C t )

[0139] Among them, x t is the input feature sequence of the long short-term memory network LSTM; h t is the output state of the long short-term memory network LSTM; W f , W i , W C , W o is the weight matrix; bf 、b i 、b c 、b o is the bias term; σ(·) is the Sigmoid activation function, tanh(·) is the hyperbolic tangent activation function; ⊙ represents element-wise multiplication.

[0140] Combination Figure 5 As shown, based on the above, this solution also provides a slewing bearing fault detection system, which includes:

[0141] A working monitoring module is used to respond to the signal that the slewing bearing is driven and moves, collect the working parameters of the slewing bearing in real time when it moves, and obtain working monitoring data;

[0142] A data extraction module, used to extract target feature data from work monitoring data;

[0143] A data judgment module is used to judge the target characteristic data according to preset conditions, generate a first judgment result, and preliminarily determine whether it points to abnormal operation of the slewing bearing, and at the same time associate the collection time of the corresponding target characteristic data with the first judgment result;

[0144] An information marking module is used to collect multiple first judgment results in time series, and when the first judgment result points to abnormal operation of the slewing bearing, an abnormal information mark is generated in the time series according to the collection time of the corresponding target characteristic data;

[0145] The information feedback module is used to slide a sliding window with a preset time span on the time series, count the number of abnormal information tags in the time span of the sliding window, and generate fault prompt information when the number is greater than a preset value N.

[0146] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0147] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0148] The above descriptions are only some embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A slewing bearing fault detection method, characterized in that: It includes: S01, responding to the signal that the slewing bearing is driven and moves, collecting the working parameters of the slewing bearing in real time when it moves, and obtaining working monitoring data; S02, extracting target feature data from the working monitoring data, and then judging the target feature data according to preset conditions to generate a first judgment result to preliminarily determine whether it points to abnormal working of the slewing bearing, and at the same time associating the collection time of the corresponding target feature data with the first judgment result; S03, collecting a plurality of first judgment results in time series, and when the first judgment result indicates that the slewing bearing is working abnormally, generating an abnormal information mark in the time series according to the collection time of the corresponding target feature data; S04. Slide a sliding window with a preset time span on the time series, count the number of abnormal information marks in the time span where the sliding window is located, and when the number is greater than a preset value N, generate fault prompt information.

2. The slewing bearing fault detection method according to claim 1, characterized in that: The working parameters of the slewing bearing during operation include: ring rotation speed, ring rotation amplitude, ring vibration data, audio data, and ring temperature data; the working monitoring data includes the ring rotation speed, ring rotation amplitude, ring vibration data, audio data, ring temperature data, and acquisition time corresponding to the slewing bearing; The slewing bearing comprises a pair of rings and rollers arranged between the pair of rings, one of the rings is provided with a gear ring and rotates under the driving force, which is set as a moving ring, and the other ring of the slewing bearing is set as a fixed ring; the moving ring is provided with an acceleration sensor, a displacement sensor and a plurality of vibration sensors; the fixed ring is provided with a plurality of audio sensors and a plurality of temperature sensors in a ring array; The acceleration sensor is used to collect the rotation speed of the moving ring when the moving ring rotates; The displacement sensor is used to collect the rotation amplitude of the moving ring when the moving ring rotates; The vibration sensor is used to collect the vibration data of the moving ring when the moving ring rotates, and each vibration sensor collects the vibration data of the ring respectively, and each vibration sensor also has a unique code and position information; The audio sensor is used to collect audio data within a preset range of the slewing bearing when the movable ring rotates, and each audio sensor collects audio data respectively, and each audio sensor also has a unique code and position information; The temperature sensor is used to collect the ring temperature data on the fixed ring when the movable ring rotates, and each temperature sensor collects the ring temperature data respectively, and each temperature sensor also has a unique code and position information.

3. The slewing bearing fault detection method according to claim 2, characterized in that: In S02, the target feature data extracted from the working monitoring data is one or more of the ring vibration data, the audio data, and the ring temperature data; They are compared and judged with corresponding preset thresholds respectively to generate judgment sub-results, which are then aggregated to form a first judgment result; In S02, when one of the target characteristic data exceeds a preset threshold, its judgment sub-result is defined as indicating that the slewing bearing is operating abnormally.

4. The slewing bearing fault detection method according to claim 3, characterized in that: In S02, the preset thresholds of the ring vibration data and the audio data are determined by a mapping relationship model between the operating state of the slewing bearing and the audio and vibration, respectively. The mapping relationship model is established based on a query table, which includes: (1) After lubricating the roller of the slewing support, construct the operation of the slewing support under different ring rotation speeds and ring rotation amplitudes, and then collect sound wave data S and vibration data Z of preset time lengths under different conditions; wherein the sound wave data includes volume decibel data, and the vibration data includes amplitude data; (2) Based on the acoustic wave data S and vibration data Z, calculate the mean of the collected time to obtain the acoustic wave data mean S 均值 , vibration data mean Z 均值 ; (3) According to the mean value S of the acoustic wave data 均值 , vibration data mean Z 均值 Set the thresholds respectively, the sound wave data threshold is N×S 均值 , the vibration data threshold is M×Z 均值 , where N and M are both greater than 1; (4) The ring rotation speed and ring rotation amplitude of the slewing bearing under different conditions are taken as motion parameters, and the corresponding acoustic wave data mean S 均值 , vibration data mean Z 均值 The sound wave data threshold N×S formed based on 均值 , vibration data threshold M×Z 均值 As limiting parameters, a lookup table of threshold values ​​of the slewing bearing operation status and audio and vibration is established, i.e., a mapping relationship model; In S02, when the ring vibration data and audio data are compared with the corresponding preset thresholds respectively, the sound wave data threshold and vibration data threshold are extracted from the threshold query table through the ring rotation speed and ring rotation amplitude in the working monitoring data. When there are no corresponding ring rotation speed and ring rotation amplitude items in the threshold query table, the sound wave data threshold and vibration data threshold corresponding to the item closest to the ring rotation speed are extracted as the preset thresholds for comparison and judgment.

5. The slewing bearing fault detection method according to claim 3 or 4, characterized in that: In S03, when one of the judgment sub-results in the first judgment result indicates that the slewing bearing is working abnormally, an abnormal information mark is generated in the time series according to the collection time of the corresponding target feature data; In S04, the time span of the sliding window is 10 to 120 seconds.

6. A slewing bearing fault self-detection method, comprising the slewing bearing fault detection method according to any one of claims 3 to 5, characterized in that: It also includes: A01. Respond to the fault self-check instruction made by the operator according to the fault prompt information, make the slewing bearing work with preset motion parameters to the preset state, collect the working parameters of the slewing bearing in real time, and obtain self-check monitoring data; A02. Extract target data from the self-test monitoring data and pre-process it according to preset conditions; A03. Input the preprocessed target data and other data in the self-test monitoring data as input items into the trained detection neural network for detection to generate a detection result.

7. The slewing bearing fault self-checking method according to claim 6, characterized in that: The preset motion parameters include the ring rotation speed and the ring rotation amplitude, and the preset state is that the slewing support performs reciprocating motion or rotational motion within the preset ring rotation amplitude at the preset ring rotation speed; The real-time data collection of the working parameters of the slewing bearing includes: ring rotation speed, ring rotation amplitude, ring vibration data, audio data and ring temperature data.

8. The slewing bearing fault self-checking method according to claim 6 or 7, characterized in that: In A01, before the slewing bearing is operated with preset motion parameters, grease is added to the rollers of the slewing bearing through an external lubrication filling device; In A02, the target data includes ring vibration data and audio data. After collecting signals, the corresponding vibration sensors and audio sensors are preprocessed according to preset conditions to obtain ring vibration data and audio data. The preprocessing includes: signal denoising and time-frequency analysis. The data after time-frequency analysis are sequentially input into the convolutional neural network CNN and the long short-term memory network LSTM to obtain processed signal data.

9. The slewing bearing fault self-checking method according to claim 8, characterized in that: The method for preprocessing the signals collected by the vibration sensor and the audio sensor includes: (1) Signal denoising Use a bandpass filter to denoise the signals collected by the vibration sensor and audio sensor to remove low-frequency and high-frequency interference signals so that the signal is concentrated in the frequency band containing fault information; assuming that the original signal is x(t) and the filtered signal is y(t), it can be expressed as: y(t)=x(t)*h(t) Where h(h) is the impulse response function of the filter, and * represents the convolution function; (2) Time-frequency analysis The short-time Fourier transform (STFT) is used to transform the signal from the time domain to the time-frequency domain to capture the transient characteristics of the signal. The formula is defined as follows: Among them, X(t,ω) is the complex representation in the time-frequency domain, which is the STFT result, that is, the complex value representation of the signal at time t and frequency ω, which represents the amplitude and phase information of the signal at ω within a small time window at time t; x(τ) is the time domain representation of the original signal x(t), τ represents the time variable, ω(τ-t) is the time window function, which weights the signal within a period of time near time t, τ in the time window function represents the time variable, ω is the frequency variable, e -jωτ is a complex exponential function, which is the core part of Fourier transform and is used to convert signals from time domain to frequency domain. j in the complex exponential function is an imaginary unit, which represents the imaginary part of a complex number. dτ is an integral variable, which represents the integration of time τ. The integration is to accumulate the frequency components of the signal weighted by the window function on the entire time axis. (3) Convolutional Neural Network (CNN) processing The spatial features of time-frequency data are extracted through the convolutional neural network CNN. The time-frequency feature matrix X(t,ω) obtained by time-frequency analysis is used as input. The output of the lth convolutional layer of the convolutional neural network CNN can be expressed as: in, is the output feature at position (i, j); is the weight matrix of the convolution kernel, the size of the convolution kernel is k×k; b l is the bias term, σ(·) is the activation function; (4) Long Short-Term Memory Network (LSTM) Processing The convolution kernel of the convolutional neural network CNN is used to extract features layer by layer, and the extracted features are used as input items of the long short-term memory network LSTM; the long short-term memory network LSTM is used to capture the time dependency of the data, which includes the following calculation steps: Forget gate calculation: f t =σ(W f ·[h t-1 ,x t ]+b f ) Input gate calculation: i t =σ(W i ·[h t-1 ,x t ]+b i ) Candidate memory cell state calculation: Memory unit status update: Output gate calculation and final output: the t =σ(W o ·[h t-1 ,x t ]+b o ) h t =o t ⊙tanh(C t ) Among them, x t is the input feature sequence of the long short-term memory network LSTM; h t is the output state of the long short-term memory network LSTM; W f , W i , W C , W o is the weight matrix; b f , b i , b c , b o is the bias term; σ(·) is the Sigmoid activation function, tanh(·) is the hyperbolic tangent activation function; ⊙ represents element-wise multiplication.

10. A slewing bearing fault detection system, characterized in that: It includes: A working monitoring module is used to respond to the signal that the slewing bearing is driven and moves, collect the working parameters of the slewing bearing in real time when it moves, and obtain working monitoring data; A data extraction module, used to extract target feature data from work monitoring data; A data judgment module is used to judge the target characteristic data according to preset conditions, generate a first judgment result, and preliminarily determine whether it points to abnormal operation of the slewing bearing, and at the same time associate the collection time of the corresponding target characteristic data with the first judgment result; An information marking module is used to collect multiple first judgment results in time series, and when the first judgment result points to abnormal operation of the slewing bearing, an abnormal information mark is generated in the time series according to the collection time of the corresponding target characteristic data; The information feedback module is used to slide a sliding window with a preset time span on the time series, count the number of abnormal information tags in the time span of the sliding window, and generate fault prompt information when the number is greater than a preset value N.

Images