Intelligent rolling body distributed detection method and system based on end-side collaborative diagnosis

Abstract

The invention belongs to the technical field of mechanical fault diagnosis, and particularly discloses an intelligent rolling body distributed detection method and system based on end edge collaborative diagnosis, and the method comprises the following steps: S1, collecting the vibration and temperature data of a plurality of rolling bodies on a belt conveyor device; s2, performing unified-scale decomposition on the vibration signal of each rolling body and performing kurtosis value calculation to obtain a feature distribution vector; s3, performing mutual inspection on the feature distribution vector and the temperature feature value of each rolling body, and searching an outlier as a fault candidate point by using a K-nearest neighbor algorithm; and S4, collecting vibration data of the fault candidate points of the rolling body, performing SVMD decomposition to obtain sub-signals, performing spectral analysis and envelope spectrum analysis on the sub-signals, judging whether the fault frequency of the rolling body actually exists or not, and outputting corresponding fault information and performing fault cause analysis if the fault frequency of the rolling body actually exists. According to the technical scheme, the fault rolling body is positioned through distributed outlier detection, and the fault information is obtained.

Description

Technical Field

[0001] The present invention belongs to the technical field of mechanical fault diagnosis, and relates to an intelligent rolling element distributed detection method and system based on end-edge collaborative diagnosis. Background Art

[0002] Long-distance belt conveyors have the characteristics of long conveying distance, large transport capacity, and evenly distributed multiple rolling elements. They are widely used in coal mining, metallurgy, and construction industries. However, since long-distance belt conveyors often work in harsh environments with high speed, heavy load, vibration and impact, key components such as belts, rollers, and motors are prone to failure during operation. Therefore, in order to ensure the working stability of long-distance belt conveyors, it is of great significance to monitor their faults and discover their health status problems in a timely manner, so as to achieve safe and reliable operation of belt conveyors and gradually realize unmanned operation.

[0003] At present, the fault diagnosis methods for long-distance belt conveyors are mainly aimed at the fault detection of single components in belt conveyors, such as collecting vibration signals of components such as rollers and motors for analysis, so as to diagnose faults of these components. However, these methods have the disadvantages of not being able to fully utilize the fault information of belt conveyors and having a single detection object. In addition, the multiple rolling elements of belt conveyors have spatial differences in operating characteristics and service conditions, but similar characteristics in signals. Therefore, how to use this characteristic to perform large-span and wide-coverage fault detection on multiple rolling elements of long-distance belt conveyors is an urgent problem to be solved. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent rolling body distributed detection method and system based on end-edge collaborative diagnosis, which uses distributed outlier detection to locate faulty rolling bodies and obtain fault information.

[0005] In order to achieve the above object, the basic scheme of the present invention is: an intelligent rolling element distributed detection method based on end-edge collaborative diagnosis, comprising the following steps:

[0006] S1, collecting vibration and temperature data of multiple rolling elements on a belt conveyor device;

[0007] S2, based on wavelet packet transform, the vibration signal of each rolling element is decomposed into a unified scale, and the kurtosis value of each scale signal after decomposition of each rolling element is calculated to obtain a characteristic distribution vector;

[0008] S3, perform mutual inspection on the characteristic distribution vector and temperature data of each rolling element, and use the K nearest neighbor algorithm to find outliers as candidate fault points;

[0009] S4, collecting vibration data of the candidate fault points of the rolling body, and decomposing them to obtain sub-signals, performing spectrum analysis and envelope spectrum analysis on the sub-signals to determine whether the fault frequency of the rolling body actually exists. If so, outputting corresponding fault information and performing fault cause analysis.

[0010] The working principle and beneficial effects of this basic scheme are as follows: This technical scheme is based on the characteristics of rolling body signals. According to the differences between the signals of normal rolling bodies and faulty rolling bodies, the vibration and temperature data of multiple rolling bodies on the belt conveyor device are collected, and the distributed outlier detection method is used to locate the faulty rolling body to obtain accurate fault information.

[0011] Furthermore, based on wavelet packet transform, the vibration signals of each rolling element are decomposed into a unified scale, and the kurtosis value of each scale signal after decomposition of each rolling element is calculated to obtain the characteristic distribution vector. The specific steps are as follows:

[0012] At a certain moment i , assuming that the vibration signal collected by the wireless vibration and temperature sensor nodes of N rolling elements is f k (t), k = 1, 2, ..., N; decompose each vibration signal using WPT to obtain m sub-signals of each vibration signal;

[0013] Based on the kurtosis value, the characteristic distribution vector M of the sub-signal after each vibration signal is decomposed is calculated. k =(ξ1,ξ2,…,ξ m ),ξ m is the kurtosis value of sub-signal m, and the kurtosis value is:

[0014]

[0015] Among them, kurt is the kurtosis value, E means expectation, μ means the sequence mean, σ means the sequence standard deviation, and x means the signal sequence.

[0016] The characteristic distribution vector of the sub-signal after each vibration signal is decomposed is calculated based on the kurtosis value, and the operation is simple.

[0017] Furthermore, the characteristic distribution vectors and temperature characteristic values ​​of each rolling element are mutually checked, and the method of finding outliers using the K nearest neighbor algorithm is as follows:

[0018] For a sample set X = [x1, x2, ..., x m ];

[0019] In each sample set x i With x j The Euclidean distance d between i,j for:

[0020] d i,j =||xi -x j || 2

[0021] Where, i = 1, 2, ... m, j = 1, 2, ... m;

[0022] According to the Euclidean distance difference, find each sample point x i The first K nearest neighbors of x oj , j = 1, 2, ... K, and then calculate the average distance D between each sample point and its K nearest neighbor samples i :

[0023]

[0024] D i It is x i The K nearest neighbor distance of the feature distribution vector and the temperature feature value is set, and the sample points exceeding the K nearest neighbor distance threshold are judged as outlier sample points;

[0025] The outlier sample points of the feature distribution vector sample set and the temperature feature sample set are calculated respectively by the K-nearest neighbor algorithm, and these outlier sample points are the fault candidate points.

[0026] The characteristic distribution vectors and temperature characteristic values ​​of each intelligent rolling element wireless data transmission node are mutually checked to obtain candidate fault points, which is conducive to fault screening.

[0027] Further, in step S4, vibration and temperature data of the rolling body at L consecutive moments are collected, and steps S2 and S3 are executed to find outliers of the wireless vibration and temperature sensor transmission nodes at each moment. If a wireless vibration and temperature sensor transmission node is continuously an abnormal outlier, the vibration data signal of the wireless vibration and temperature sensor transmission node is obtained, the optimal penalty coefficient α of SVMD is set, and SVMD is used to perform adaptive decomposition diagnosis on the vibration data signal;

[0028] The sub-signals after SVMD decomposition are subjected to spectrum analysis and envelope spectrum analysis to determine whether a fault actually exists and to perform fault diagnosis analysis.

[0029] If the characteristic frequency of the rolling element exists in the spectrum analysis and envelope spectrum of the sub-signal, it means that a real fault exists, and the cause of the fault can be analyzed.

[0030] Furthermore, the method for setting the optimal penalty coefficient α of SVMD is:

[0031] When performing SVMD decomposition on the vibration signal f(t), the initial parameters α and ε are input, and each mode and modal center frequency are obtained through SVMD decomposition and iteratively updated until the convergence condition is met:

[0032]

[0033] Among them, α is the penalty coefficient, ε is the convergence threshold, is the nth mode, is the n+1th mode;

[0034] After the iteration stops, the kurt value of the signal reconstructed by SVMD decomposition under the α parameter is calculated, the α value is updated, and the next SVMD decomposition iteration is performed until the α group is traversed;

[0035] Compare the kurtosis values ​​of the decomposed and reconstructed signals under different α values, select the decomposition parameter α with the largest kurtosis value, and output the corresponding filter parameter group;

[0036] For the optimal decomposition parameters, a set of modal components IMF and corresponding center frequencies ω are obtained through SVMD decomposition. k , calculate the upper and lower frequency cutoff points corresponding to 0.1 times the center frequency amplitude of each modal component spectrum [f 下 ,f 上 ], the required bandwidth is B = f 上 -f 下 .

[0037] The parameters of SVMD are only the penalty coefficient α and the convergence coefficient ε. There is no need to pre-set the number of decomposition modes during the decomposition process. Therefore, a good decomposition effect can be achieved by only setting the penalty coefficient α.

[0038] Furthermore, the method of performing spectrum analysis and envelope spectrum analysis on the sub-signals after the signal is decomposed by SVMD is as follows:

[0039] Obtain the pitch diameter, roller diameter, number of rollers, rotation speed and contact angle parameters of the rolling element, and calculate the failure frequency of the rolling element;

[0040] The characteristic frequency is a multiple of the fault frequency. If the characteristic frequency of the rolling element exists in the spectrum and envelope spectrum of the sub-signal, it means that there is a real fault.

[0041] Simple operation and easy to use.

[0042] The present invention also provides an intelligent rolling body distributed detection system based on end-edge collaborative diagnosis, comprising a device layer, a wireless data acquisition node layer, a relay module layer and a host computer layer connected in sequence, wherein the wireless data acquisition node layer is used to collect vibration and temperature data of multiple rolling bodies on a belt conveyor device in the device layer, and the host computer layer executes the method of the present invention to perform fault analysis on the intelligent rolling body.

[0043] This system uses the coordination of various levels to perform fault analysis on intelligent rolling elements.

[0044] Furthermore, the wireless data acquisition node layer includes a plurality of wireless vibration and temperature sensors, and the wireless vibration and temperature sensors are arranged at corresponding rolling bodies of the belt conveyor device of the equipment layer.

[0045] Simple structure and easy to use. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the process of obtaining a characteristic distribution vector of the intelligent rolling element distributed detection method based on end-edge collaborative diagnosis of the present invention;

[0047] Figure 2 It is a schematic diagram of the process of finding outliers in the intelligent rolling element distributed detection method based on end-edge collaborative diagnosis of the present invention;

[0048] Figure 3 It is a flow chart of setting the optimal penalty coefficient α of SVMD in the intelligent rolling element distributed detection method based on end-edge collaborative diagnosis of the present invention;

[0049] Figure 4 It is a flow chart of the intelligent rolling element distributed detection method based on end-edge collaborative diagnosis of the present invention;

[0050] Figure 5 It is a structural schematic diagram of the intelligent rolling element distributed detection system based on end-edge collaborative diagnosis of the present invention;

[0051] Figure 6 It is a structural schematic diagram of the intelligent rolling element distributed detection method based on end-edge collaborative diagnosis of the present invention, which uses WPT to decompose each vibration signal. DETAILED DESCRIPTION

[0052] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0053] In the description of the present invention, it is necessary to understand that the terms "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0054] In the description of the present invention, unless otherwise specified and limited, it should be noted that the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a mechanical connection or an electrical connection, or it can be the internal connection between two components. It can be a direct connection or an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to the specific circumstances.

[0055] There are many rollers in long-distance belt conveyors, and because the service conditions are consistent, the signals collected from each roller are similar. These rollers rotate smoothly when working normally, and if a fault occurs, a periodic vibration pulse is generated at the fault point. Based on the above characteristics of the rolling element signal, the present invention uses a distributed outlier detection method to locate the faulty rolling element based on the difference between the normal rolling element and the faulty rolling element signal.

[0056] The present invention discloses a distributed detection method for intelligent rolling elements based on end-edge collaborative diagnosis, such as Figure 4 As shown, the following steps are included:

[0057] S1, collecting vibration and temperature data of multiple rolling elements on a belt conveyor device;

[0058] S2, based on Wavelet Packet Transformation (WPT), the vibration signals of each rolling element are decomposed into a unified scale, the kurtosis value of each scale signal after decomposition of each rolling element is calculated, the characteristic distribution vector is obtained and uploaded to the relay module layer (i.e., relay layer);

[0059] S3, in the relay module layer, the characteristic distribution vectors and temperature data of each rolling body (the temperature data collected by different rolling bodies are their temperature characteristic values) are mutually checked, and the K nearest neighbor algorithm is used to find outliers as candidate fault points, and send them to the upper computer layer; since each position works under the same working conditions, theoretically the sensitive modal characteristic values ​​of its vibration signal should be similar, and the position corresponding to the outlier indicates that it may have a fault.

[0060] S4, collects vibration data of candidate fault points of the rolling element, and decomposes them (such as SVMD decomposition, which can continuously extract sub-signals without setting the number of IMFs and has low computational complexity) to obtain sub-signals, performs spectrum analysis and envelope spectrum analysis on the sub-signals, and determines whether the fault frequency of the rolling element actually exists. If so, outputs the corresponding fault information and performs fault cause analysis (performs cause analysis based on the fault phenomenon).

[0061] In a preferred embodiment of the present invention, Figure 1As shown, based on wavelet packet transform, the vibration signals of each rolling body are decomposed into a unified scale, and the kurtosis value of each scale signal after decomposition of each rolling body is calculated to obtain the characteristic distribution vector. The specific steps are as follows:

[0062] At a certain moment i , assuming that the vibration signal collected by the wireless vibration and temperature sensor nodes of N rolling elements is f k (t), k = 1, 2, ..., N; use WPT to decompose each vibration signal (such as Figure 6 As shown, it is divided into low-frequency components and high-frequency components), and m sub-signals of each vibration signal are obtained;

[0063] Based on the kurtosis value, the characteristic distribution vector M of the sub-signal after each vibration signal is decomposed is calculated. k =(ξ1,ξ2,…,ξ m ),ξ m is the kurtosis value of sub-signal m, and the kurtosis value is:

[0064]

[0065] Among them, kurt is the kurtosis value, E means expectation, μ means the sequence mean, σ means the sequence standard deviation, and x means the signal sequence.

[0066] In a preferred embodiment of the present invention, Figure 2 As shown, the characteristic distribution vectors and temperature characteristic values ​​of each rolling element are mutually checked, and the method of finding outliers using the K-Nearest Neighbor (KNN) algorithm is as follows:

[0067] For a sample set X = [x1, x2, ..., x m ];

[0068] In each sample set x i With x j The Euclidean distance d between i,j for:

[0069] d i,j =||x i -x j || 2

[0070] Where, i = 1, 2, ... m, j = 1, 2, ... m;

[0071] According to the Euclidean distance difference, find each sample point x i The first K nearest neighbors of x oj , j = 1, 2, ... K, and then calculate the average distance D between each sample point and its K nearest neighbor samples i :

[0072]

[0073] D i It is x i The K nearest neighbor distance of the feature distribution vector and the temperature feature value is set, and the sample points exceeding the K nearest neighbor distance threshold are judged as outlier sample points;

[0074] The outlier sample points of the feature distribution vector sample set and the temperature feature sample set are calculated respectively by the K-nearest neighbor algorithm, and these outlier sample points are the fault candidate points.

[0075] In a preferred embodiment of the present invention, in step S4, vibration and temperature data of the rolling body at L consecutive moments are collected, and steps S2 and S3 are executed to find outliers of the wireless vibration and temperature sensor transmission nodes at each moment. If a wireless vibration and temperature sensor transmission node is continuously an abnormal outlier, the vibration data signal of the wireless vibration and temperature sensor transmission node is obtained, the optimal penalty coefficient α of SVMD is set, and SVMD is used to perform adaptive decomposition diagnosis on the vibration data signal;

[0076] The sub-signals after SVMD decomposition are subjected to spectrum analysis and envelope spectrum analysis to determine whether a fault actually exists (if the characteristic frequency of the rolling element exists in the spectrum analysis and envelope spectrum of the sub-signal, it means that a fault actually exists) and to perform fault diagnosis analysis. Sub-signals are obtained by SVMD decomposition of the signal. These sub-signals have different frequency components. By analyzing the sub-signals with specific frequency components (such as the characteristic frequency of the rolling element), it is determined whether the rolling element has a fault.

[0077] In a preferred embodiment of the present invention, Figure 3 As shown, the method for setting the optimal penalty coefficient α of SVMD is:

[0078] When performing SVMD decomposition on the vibration signal f(t), the initial parameters α and ε are input, and each mode and modal center frequency are obtained through SVMD decomposition and iteratively updated until the convergence condition is met:

[0079]

[0080] Among them, α is the penalty coefficient, ε is the convergence threshold, is the nth mode, is the n+1th mode;

[0081] After the iteration stops, the kurt value of the signal reconstructed by SVMD decomposition under the α parameter is calculated, the α value is updated, and the next SVMD decomposition iteration is performed until the α group is traversed;

[0082] Compare the kurtosis values ​​of the decomposed and reconstructed signals under different α values, select the decomposition parameter α with the largest kurtosis value, and output the corresponding filter parameter group;

[0083] For the optimal decomposition parameters, a set of modal components IMF and corresponding center frequencies ω are obtained through SVMD decomposition. k , calculate the upper and lower frequency cutoff points corresponding to 0.1 times the center frequency amplitude of each modal component spectrum [f 下 ,f 上 ], the required bandwidth is B = f 上 -f 下 .

[0084] In a preferred embodiment of the present invention, the method of performing spectrum analysis and envelope spectrum analysis on the sub-signals after the signal is decomposed by SVMD is as follows:

[0085] Obtain the pitch diameter, roller diameter, number of rollers, rotation speed and contact angle parameters of the rolling element, and calculate the failure frequency of the rolling element;

[0086] The characteristic frequency is a multiple of the fault frequency. If the characteristic frequency of the rolling element exists in the spectrum and envelope spectrum of the sub-signal, it means that there is a real fault.

[0087] The present invention also provides an intelligent rolling element distributed detection system based on end-edge collaborative diagnosis, such as Figure 5 As shown, it includes a device layer, a wireless data acquisition node layer, a relay module layer and a host computer layer connected in sequence. The wireless data acquisition node layer is used to collect vibration and temperature data of multiple rolling bodies on the belt conveyor device in the device layer, and uploads the data to the host computer layer through the relay module layer. The host computer layer executes the method of the present invention to perform fault analysis on the intelligent rolling body.

[0088] In a preferred embodiment of the present invention, the wireless data acquisition node layer includes a plurality of wireless vibration and temperature sensors, and the wireless vibration and temperature sensors are arranged at corresponding rolling bodies of the belt conveyor device of the equipment layer.

[0089] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0090] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A distributed detection method for intelligent rolling elements based on end-edge collaborative diagnosis, characterized in that: The steps include: S1, collecting vibration and temperature data of multiple rolling elements on a belt conveyor device; S2, based on wavelet packet transform, the vibration signal of each rolling element is decomposed into a unified scale, and the kurtosis value of each scale signal after decomposition of each rolling element is calculated to obtain a characteristic distribution vector; S3, perform mutual inspection on the characteristic distribution vector and temperature data of each rolling element, and use the K nearest neighbor algorithm to find outliers as candidate fault points; S4, collecting vibration data of the candidate fault points of the rolling body, and decomposing them to obtain sub-signals, performing spectrum analysis and envelope spectrum analysis on the sub-signals to determine whether the fault frequency of the rolling body actually exists. If so, outputting corresponding fault information and performing fault cause analysis.

2. The intelligent rolling element distributed detection method based on end-edge collaborative diagnosis according to claim 1 is characterized in that: Based on wavelet packet transform, the vibration signals of each rolling element are decomposed into a unified scale, and the kurtosis value of each scale signal after decomposition of each rolling element is calculated to obtain the characteristic distribution vector. The specific steps are as follows: At a certain moment i , assuming that the vibration signal collected by the wireless vibration and temperature sensor nodes of N rolling elements is f k (t), k = 1, 2, ..., N; decompose each vibration signal using WPT to obtain m sub-signals of each vibration signal; Based on the kurtosis value, the characteristic distribution vector M of the sub-signal after each vibration signal is decomposed is calculated. k =(ξ1,ξ2,…,ξ m ),ξ m is the kurtosis value of sub-signal m, and the kurtosis value is: Among them, kurt is the kurtosis value, E means expectation, μ means the sequence mean, σ means the sequence standard deviation, and x means the signal sequence.

3. The intelligent rolling element distributed detection method based on end-edge collaborative diagnosis according to claim 2 is characterized in that: The characteristic distribution vectors and temperature characteristic values ​​of each rolling element are mutually checked, and the method of finding outliers using the K nearest neighbor algorithm is as follows: For a sample set X = [x1, x2, ..., x m ]; In each sample set x i With x j The Euclidean distance d between i,j for: d i,j =||x i -x j || 2 Where, i = 1, 2, ... m, j = 1, 2, ... m; According to the Euclidean distance difference, find each sample point x i The first K nearest neighbors of x oj , j = 1, 2, ... K, and then calculate the average distance D between each sample point and its K nearest neighbor samples i : Among them, D i is x i The K nearest neighbor distance of the feature distribution vector and the temperature feature value is set, and the sample points exceeding the K nearest neighbor distance threshold are judged as outlier sample points; The outlier sample points of the feature distribution vector sample set and the temperature feature sample set are calculated respectively by the K-nearest neighbor algorithm, and these outlier sample points are the fault candidate points.

4. The intelligent rolling element distributed detection method based on end-edge collaborative diagnosis according to claim 1, characterized in that: In step S4, the vibration and temperature data of the rolling body at L consecutive moments are collected, and steps S2 and S3 are executed to find the outliers of the wireless vibration and temperature sensor transmission nodes at each moment. If a wireless vibration and temperature sensor transmission node is continuously an abnormal outlier, the vibration data signal of the wireless vibration and temperature sensor transmission node is obtained, the optimal penalty coefficient α of SVMD is set, and the vibration data signal is adaptively decomposed and diagnosed by using SVMD; The sub-signals after SVMD decomposition are subjected to spectrum analysis and envelope spectrum analysis to determine whether a fault actually exists and to perform fault diagnosis analysis.

5. The intelligent rolling element distributed detection method based on end-edge collaborative diagnosis according to claim 4 is characterized in that: The method to set the optimal penalty coefficient α of SVMD is: When performing SVMD decomposition on the vibration signal f(t), the initial parameters α and ε are input, and each mode and modal center frequency are obtained through SVMD decomposition and iteratively updated until the convergence condition is met: Among them, α is the penalty coefficient, ε is the convergence threshold, is the nth mode, is the n+1th mode; After the iteration stops, the kurt value of the signal reconstructed by SVMD decomposition under the α parameter is calculated, the α value is updated, and the next SVMD decomposition iteration is performed until the α group is traversed; Compare the kurtosis values ​​of the decomposed and reconstructed signals under different α values, select the decomposition parameter α with the largest kurtosis value, and output the corresponding filter parameter group; For the optimal decomposition parameters, a set of modal components IMF and corresponding center frequencies ω are obtained through SVMD decomposition. k , calculate the upper and lower frequency cutoff points corresponding to 0.1 times the center frequency amplitude of each modal component spectrum [f 下 ,f 上 ], the required bandwidth is B = f 上 -f 下 .

6. The intelligent rolling element distributed detection method based on end-edge collaborative diagnosis according to claim 5 is characterized in that: The method of performing spectrum analysis and envelope spectrum analysis on the sub-signals after the signal is decomposed by SVMD is: Obtain the pitch diameter, roller diameter, number of rollers, rotational speed and contact angle parameters of the rolling element, and calculate the fault frequency of the rolling element; the characteristic frequency is the multiple of the fault frequency. If the characteristic frequency of the rolling element exists in the spectrum and envelope spectrum of the sub-signal, it means that a real fault exists.

7. An intelligent rolling element distributed detection system based on end-edge collaborative diagnosis, characterized in that: It includes a device layer, a wireless data acquisition node layer, a relay module layer and a host computer layer which are connected in sequence. The wireless data acquisition node layer is used to collect vibration and temperature data of multiple rolling bodies on the belt conveyor device in the device layer. The host computer layer executes the method described in one of claims 1-6 to perform fault analysis on the intelligent rolling body.

8. The intelligent rolling element distributed detection system based on end-edge collaborative diagnosis according to claim 7 is characterized in that: The wireless data acquisition node layer includes a plurality of wireless vibration and temperature sensors, and the wireless vibration and temperature sensors are arranged at corresponding rolling bodies of the belt conveyor device of the equipment layer.

Images