Wireless telemetering and edge diagnosis method and system based on compressed lightweight sensing

Abstract

The invention belongs to the technical field of mechanical fault diagnosis, and particularly discloses a wireless telemetering and edge diagnosis method and system based on compressed lightweight sensing, and the method comprises the following steps: collecting vibration and temperature signals of equipment, and carrying out the preprocessing of original signals; envelope detection is carried out on the preprocessed vibration data at the edge end, kurtosis values of the temperature data and the vibration data are calculated, anomaly detection is carried out by using an empirically preset kurtosis value, and fault information is output if anomaly exists; discrete cosine Huffman dual-mode coding is used, compressed coding is carried out on the preprocessed vibration data, and a telemetering node sends the coded data; and a data receiving end decodes and reconstructs coded data through Huffman anti-coding and discrete cosine inverse transformation, and performs fault identification by using a neural network. By adopting the technical scheme, the transmission data volume is remarkably reduced through the data compression technology, so that the bandwidth requirement and the transmission delay are reduced, and fault diagnosis can be more efficiently carried out.

Description

Technical Field

[0001] The present invention belongs to the technical field of mechanical fault diagnosis, and relates to a wireless telemetry and edge diagnosis method and system based on compressed lightweight perception. Background Art

[0002] Mechanical fault diagnosis is of great significance in modern industrial production. With the increasing complexity and automation of industrial equipment and production lines, the reliability and operating efficiency of equipment directly affect the safety and economic benefits of production.

[0003] During the operation of mechanical equipment, a large number of physical signals, such as vibration and temperature, are generated. Real-time monitoring, analysis, and diagnosis of these signals are crucial to ensuring proper operation and preventing failures. However, traditional detection methods face challenges such as slow data transmission due to large amounts of data and poor timeliness. Summary of the Invention

[0004] The purpose of the present invention is to provide a wireless telemetry and edge diagnosis method and system based on compressed lightweight perception to solve the problems of large data volume, slow transmission, poor timeliness and high power consumption in traditional fault diagnosis.

[0005] To achieve the above objectives, the basic solution of the present invention is: a wireless telemetry and edge diagnosis method based on compressed lightweight sensing, comprising the following steps:

[0006] Collect vibration and temperature signals of the equipment and pre-process the original signals;

[0007] The pre-processed vibration data is subjected to envelope detection at the edge, the kurtosis values ​​of the temperature data and vibration data are calculated, and anomaly detection is performed using the empirically preset kurtosis values. If an anomaly exists, fault information is output;

[0008] Use discrete cosine Huffman dual-mode coding to compress and encode the pre-processed vibration data, and the telemetry node sends the encoded data;

[0009] The data receiving end decodes and reconstructs the encoded data through Huffman inverse coding and inverse discrete cosine transform, and uses neural networks to identify equipment faults. A convolutional neural network (CNN) is constructed and trained to effectively identify and classify different types of faults, thereby achieving real-time monitoring of equipment status and predictive maintenance.

[0010] The working principle and beneficial effects of this basic solution are as follows: Based on compressed lightweight sensing, this technical solution significantly reduces the amount of transmitted data through data compression technology, thereby reducing bandwidth requirements and transmission latency. Simultaneously, lightweight sensing technology effectively reduces data processing complexity and energy consumption without compromising diagnostic accuracy. Edge computing nodes can more efficiently perform fault diagnosis, extend battery life, and reduce maintenance frequency, thereby improving the reliability and economic benefits of the entire system.

[0011] Furthermore, the method for preprocessing the original signal is:

[0012] Calculate and subtract the mean of the original signal to remove the DC component. The mean is:

[0013]

[0014] in, is the mean of the signal, N is the number of samples of the signal, x i is the i-th sample value;

[0015] The weighted moving average method is used to smooth and filter the original signal. Different weights are assigned to data points to calculate the average value, retaining the characteristics of the data and achieving better smoothing effect:

[0016]

[0017] Among them, WMH t is the weighted moving average at time point t, x t-i is the value of the i-th data point in the window, w i is the weight of the corresponding data point, and A is the window size;

[0018] Based on the block threshold strategy denoising, visual shrinkage Visushrink calculates the threshold λ used by wavelet block threshold denoising:

[0019]

[0020] Where σ represents the standard deviation of the noise;

[0021] Use Haar wavelet to perform multi-layer decomposition of the signal and decompose the original signal into multi-scale detail components A i ,AA i ,ADA i With the approximate component D i ,DA i ,DDA i , use block thresholding to perform block thresholding on the different components obtained by decomposition:

[0022]

[0023] Among them, (w j,k ) is the original wavelet packet decomposition coefficient, (w j,k ) ⊥ is the coefficient after threshold processing;

[0024] The coefficients after block threshold processing (w j,k ) ⊥ By performing inverse wavelet packet transform, the denoised original signal can be restored.

[0025] The original signal is preprocessed and the wavelet packet block threshold method is used to denoise the data and optimize the data.

[0026] Furthermore, the steps for performing anomaly detection on the preprocessed temperature data are as follows:

[0027] Compute the kurtosis value for temperature data:

[0028]

[0029] Where E[] represents the expected value, X is the original signal, and μ is the mean of the original signal X;

[0030] The kurtosis value of the temperature data is compared with a preset value. If the kurtosis value of the temperature data exceeds the preset value, a temperature fault message is output.

[0031] Perform anomaly detection on the preprocessed temperature data to obtain corresponding temperature fault information.

[0032] Furthermore, the steps for performing anomaly detection on the pre-processed vibration data are as follows:

[0033] After preprocessing, the vibration signal is subjected to envelope detection to extract the envelope of the original signal. The envelope spectrum Envelope(t) is obtained using the Hilbert function:

[0034]

[0035] Where x(t) is the original signal, is the Hilbert transform;

[0036] Calculate the kurtosis value based on the envelope spectrum value:

[0037]

[0038] Where m is the number of samples, x i is the i-th sample value, is the sample mean, s is the sample standard deviation; an empirical threshold is used to compare with the kurtosis value, and a vibration fault message is sent if it exceeds the empirical threshold.

[0039] Perform anomaly detection on the preprocessed vibration data to obtain corresponding vibration fault information.

[0040] Furthermore, the steps of compressing and encoding the pre-processed vibration data using discrete cosine Huffman dual-mode coding are as follows:

[0041] Combine the discrete cosine transform with Huffman coding to compress the original signal into a coded bit stream and an element dictionary:

[0042]

[0043] Where X(k) is the kth frequency domain coefficient, x(n) is the nth sample value of the original time domain signal, N is the number of samples of the input signal, n is the index of the signal, and k is the index of the transformed frequency domain coefficient.

[0044] The transmission efficiency of the encoded bit stream after efficient compression is greatly improved.

[0045] Furthermore, the data receiving end decodes and reconstructs the encoded data through Huffman inverse coding and inverse discrete cosine transform, which is:

[0046]

[0047] Where X(k) is the kth frequency domain coefficient, x(n) is the nth sample value of the original time domain signal, k is the index of the transformed frequency domain coefficient; N is the number of samples of the signal, and n is the index of the signal.

[0048] The data is reconstructed with high precision through Huffman inverse coding and inverse discrete cosine transform to facilitate fault identification.

[0049] The present invention also provides a wireless telemetry and edge diagnosis system based on compressed and lightweight perception, including a vibration sensor, a temperature sensor, a dual-channel acquisition module, a transmission module and an edge processing module;

[0050] The vibration sensor and the temperature sensor are simultaneously connected to a dual-channel acquisition module, which is used to synchronously acquire the vibration and temperature signals of the device;

[0051] The edge processing module receives the signal collected by the dual-channel collection module through the transmission module, and executes the method of the present invention to identify equipment faults.

[0052] This system uses sensors and modules to realize rolling element edge diagnosis.

[0053] Furthermore, it also includes a remote terminal, which is connected to the output end of the edge processing module through the wireless communication module.

[0054] Receive data remotely for easy viewing.

[0055] Furthermore, it also includes a power supply module, which is respectively connected to the power supply ends of the vibration sensor, temperature sensor, dual-channel acquisition module, transmission module and edge processing module.

[0056] Set up a power supply module to ensure power supply to each device. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a flow chart of the wireless telemetry and edge diagnosis method based on compressed lightweight sensing of the present invention;

[0058] Figure 2 It is a schematic diagram of the process of preprocessing the original signal in the wireless telemetry and edge diagnosis method based on compressed lightweight perception of the present invention;

[0059] Figure 3 It is a schematic diagram of a flow chart of abnormality detection on pre-processed temperature data and vibration data according to the wireless telemetry and edge diagnosis method based on compressed lightweight perception of the present invention;

[0060] Figure 4 It is a schematic diagram of a compressed sensing process based on discrete cosine Huffman dual-mode coding of a wireless telemetry and edge diagnosis method based on compressed lightweight perception of the present invention. DETAILED DESCRIPTION

[0061] The following describes embodiments of the present invention in detail. Examples of the embodiments 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 intended only to explain the present invention and are not to be construed as limiting the present invention.

[0062] In the description of the present invention, it should be understood that the terms "longitudinal", "transverse", "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 limiting the present invention.

[0063] 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 communication 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.

[0064] The present invention discloses a wireless telemetry and edge diagnosis method based on compressed lightweight perception. It addresses the problem that traditional manual inspections cannot understand the operating status of equipment in real time and issue early warnings and alarms. It monitors the vibration and temperature data of mechanical equipment during operation in real time and uses envelope detection anomaly detection technology to achieve real-time monitoring of mechanical equipment. Data compression technology significantly reduces the amount of transmitted data, thereby reducing bandwidth requirements and transmission delays. Figure 1 As shown, the wireless telemetry and edge diagnosis method includes the following steps:

[0065] Collect vibration and temperature signals of the equipment and pre-process the original signals;

[0066] Envelope detection is performed on the pre-processed vibration data at the edge, the kurtosis value of the temperature data and vibration data is calculated, and anomaly detection is performed using the empirically preset kurtosis value. If an anomaly exists, fault information is output;

[0067] Use discrete cosine Hoffman dual-mode coding (DCT-Hoffman-Dual-mode Coding, DHDMC) to compress and encode the pre-processed vibration data, and the telemetry node sends the encoded data;

[0068] The data receiving end decodes and reconstructs the encoded data through Huffman inverse coding and inverse discrete cosine transform, and uses neural networks (such as CNN, etc.) to identify equipment faults (identifying bearing and gear faults, such as bearing inner ring, outer ring, rolling element, cage faults, gear tooth breakage, wear, misalignment, etc.). Convolutional neural networks (CNNs) are constructed and trained to effectively identify and classify different types of faults, thereby realizing real-time monitoring of equipment status and predictive maintenance.

[0069] In a preferred embodiment of the present invention, Figure 2 As shown in the figure, the original signal often contains noise caused by the limitations of the sensor and circuit. Based on the traditional preprocessing method and wavelet packet block threshold, the method for preprocessing the original signal is as follows:

[0070] Calculate and subtract the mean of the original signal to remove the DC component. The mean is:

[0071]

[0072] in, is the mean of the signal, N is the number of samples of the signal, x i is the i-th sample value;

[0073] The weighted moving average method is used to smooth and filter the original signal. Different weights are assigned to data points to calculate the average value, retaining the characteristics of the data and achieving better smoothing effect:

[0074]

[0075] Among them, WMH t is the weighted moving average at time point t, x t-i is the value of the i-th data point in the window, w i is the weight of the corresponding data point, A is the window size, and the preferred window size is 5;

[0076] Perform wavelet packet decomposition on the original data after removing the DC component and filtering and smoothing, and use the block threshold strategy to denoise the decomposition result. Visual shrinkage Visushrink calculates the threshold λ used by wavelet block threshold denoising:

[0077]

[0078] Where σ represents the standard deviation of the noise;

[0079] Use Haar wavelet to perform multi-layer (such as 4-layer) decomposition of the signal, and decompose the original signal into multi-scale detail components A i ,AA i ,ADA i With the approximate component D i ,DA i ,DDA i , use block thresholding to perform block thresholding on the different components obtained by decomposition:

[0080]

[0081] Among them, (w j,k ) is the original wavelet packet decomposition coefficient, (w j,k ) ⊥ is the coefficient after threshold processing;

[0082] The coefficients after block threshold processing (w j,k ) ⊥ By performing inverse wavelet packet transform, the denoised original signal can be restored.

[0083] In a preferred embodiment of the present invention, Figure 3 As shown in Figure 2, the steps for anomaly detection on preprocessed temperature data are as follows:

[0084] Compute the kurtosis value for temperature data:

[0085]

[0086] Where E[] represents the expected value, X is the original signal, and μ is the mean of the original signal X;

[0087] The kurtosis value of the temperature data is compared with a preset value. If the kurtosis value of the temperature data exceeds the preset value, a temperature fault message is output.

[0088] In a preferred embodiment of the present invention, the steps of performing abnormality detection on the pre-processed vibration data are as follows:

[0089] After preprocessing, the vibration signal is subjected to envelope detection to extract the envelope of the original signal. The envelope spectrum Envelope(t) is obtained using the Hilbert function:

[0090]

[0091] Where x(t) is the original signal, is the Hilbert transform;

[0092] Calculate the kurtosis value based on the envelope spectrum value:

[0093]

[0094] Where m is the number of samples, x i is the i-th sample value, is the sample mean, s is the sample standard deviation; an empirical threshold is used to compare with the kurtosis value, and a vibration fault message is sent if it exceeds the empirical threshold.

[0095] In a preferred embodiment of the present invention, Figure 4 As shown in FIG, the steps of compressing and encoding the pre-processed vibration data using discrete cosine Hoffman dual-mode coding (DCT-Hoffman-Dual-mode Coding, DHDMC) are as follows:

[0096] Combine the discrete cosine transform with Huffman coding to compress the original signal into a coded bit stream and an element dictionary:

[0097]

[0098] Where X(k) is the kth frequency domain coefficient, x(n) is the nth sample value of the original time domain signal, N is the number of samples of the input signal, n is the index of the signal, and k is the index of the transformed frequency domain coefficient.

[0099] Discrete cosine transform can convert the original signal into a sparse representation in the frequency domain, greatly reducing data complexity; then, Huffman coding is a lossless coding method that can encode long data into a bit stream and an element dictionary. The memory occupied by the bit stream and the element dictionary is much smaller than the memory occupied by the original signal.

[0100] In the present invention, discrete cosine transform is combined with Huffman coding, and discrete cosine Huffman dual-mode coding can effectively compress the space occupied by original data.

[0101] Based on this, the data uploaded by telemetry is no longer a raw signal that takes up a huge amount of space, but a coded bit stream after efficient compression, which greatly improves the transmission efficiency.

[0102] In a preferred embodiment of the present invention, after the transmission is completed, the data receiving end decodes and reconstructs the encoded data through Huffman inverse coding and inverse discrete cosine transform, which is:

[0103]

[0104] Where X(k) is the kth frequency domain coefficient, x(n) is the nth sample value of the original time domain signal, k is the index of the transformed frequency domain coefficient, and n is the index of the signal.

[0105] The present invention also provides a wireless telemetry and edge diagnosis system based on compressed lightweight perception, including a vibration sensor (such as a MEMS vibration sensor, which captures the vibration amplitude and frequency of mechanical equipment with high precision), a temperature sensor (such as a DS18B20 temperature sensor, which accurately measures temperature changes during equipment operation), a dual-channel acquisition module (synchronously acquires vibration and temperature signals to ensure data integrity and consistency), a transmission module, and an edge processing module. The vibration sensor and the temperature sensor are electrically connected to the dual-channel acquisition module at the same time, and the dual-channel acquisition module is used to synchronously acquire the vibration and temperature signals of the equipment. The vibration sensor, temperature sensor, and dual-channel acquisition module are installed at the edge of the equipment to capture key data such as vibration and temperature of the mechanical equipment in real time.

[0106] The edge processing module receives the signals collected by the dual-channel acquisition module through the transmission module and executes the method of the present invention to identify equipment faults. This system uses sensors and modules to achieve rolling element edge diagnosis.

[0107] Preferably, components such as the vibration sensor, temperature sensor, dual-channel acquisition module, transmission module, and edge processing module are all housed within the housing, facilitating centralized installation and providing protection. The exterior of the housing can be provided with a raised threaded structure (e.g., similar to a screw structure) for connection to the device (i.e., screw connection), ensuring a secure connection.

[0108] In a preferred embodiment of the present invention, the wireless telemetry and edge diagnostic system further includes a remote terminal electrically connected to the output of the edge processing module via a wireless communication module (e.g., WiFi, Bluetooth, 5G network, etc.). The remote terminal can be a smart display, a computer, etc., to remotely receive data for easy viewing.

[0109] In a preferred embodiment of the present invention, the wireless telemetry and edge diagnostic system further includes a power supply module (e.g., a battery). The power supply module is electrically connected to the power supply terminals of the vibration sensor, temperature sensor, dual-channel acquisition module, transmission module, and edge processing module. The power supply module ensures power supply to each component.

[0110] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that a specific feature, structure, material, or characteristic described in conjunction with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0111] While 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 invention, and that the scope of the invention is defined by the claims and their equivalents.

Claims

1. A wireless telemetry and edge diagnosis method based on compressed lightweight perception, characterized in that: The steps include: Collect vibration and temperature signals of the equipment and pre-process the original signals; Envelope detection is performed on the pre-processed vibration data at the edge, the kurtosis values ​​of the temperature data and vibration data are calculated, and the empirically preset kurtosis values ​​are used for anomaly detection. If an anomaly exists, fault information is output; Discrete cosine Huffman dual-mode coding is used to compress and encode the pre-processed vibration data, and the telemetry node sends the encoded data; The data receiving end decodes and reconstructs the encoded data through Huffman inverse coding and discrete cosine inverse transform, and uses neural networks to identify equipment faults.

2. The wireless telemetry and edge diagnosis method based on compressed lightweight perception according to claim 1, characterized in that: The method for preprocessing the original signal is: Calculate and subtract the mean of the original signal to remove the DC component. The mean is: in, is the mean of the signal, N is the number of samples of the signal, x i is the i-th sample value; The weighted moving average method is used to smooth and filter the original signal, and different weights are assigned to data points to calculate the average value, retain the characteristics of the data, and achieve a better smoothing effect: Among them, WMH t is the weighted moving average at time point t, x t-i is the value of the i-th data point in the window, w i is the weight of the corresponding data point, and A is the window size; Based on the block threshold strategy denoising, visual shrinkage Visushrink calculates the threshold λ used by wavelet block threshold denoising: Where σ represents the standard deviation of the noise; Use Haar wavelet to perform multi-layer decomposition of the signal and decompose the original signal into multi-scale detail components A i ,AA i ,ADA i With the approximate component D i ,DA i ,DDA i , use block thresholding to perform block thresholding on the different components obtained by decomposition: Among them, (w j,k ) is the original wavelet packet decomposition coefficient, (w j,k ) ⊥ is the coefficient after threshold processing; The coefficients after block threshold processing (w j,k ) ⊥ By performing inverse wavelet packet transform, the denoised original signal can be restored.

3. The wireless telemetry and edge diagnosis method based on compressed lightweight perception according to claim 1, characterized in that: The steps for anomaly detection on preprocessed temperature data are: Compute the kurtosis value for temperature data: Where E[] represents the expected value, X is the original signal, and μ is the mean of the original signal X; The kurtosis value of the temperature data is compared with a preset value. If the kurtosis value of the temperature data exceeds the preset value, a temperature fault message is output.

4. The wireless telemetry and edge diagnosis method based on compressed lightweight perception according to claim 2, characterized in that: The steps for anomaly detection on preprocessed vibration data are: For the vibration signal after preprocessing, envelope detection is performed to extract the envelope of the original signal, and the envelope spectrum Envelope(t) is obtained using the Hilbert function: Where x(t) is the original signal, is the Hilbert transform; Calculate the kurtosis value based on the envelope spectrum value: Where m is the number of samples, x i is the i-th sample value, is the sample mean, s is the sample standard deviation; the empirical threshold is used to compare with the kurtosis value, and a vibration fault message is sent if it exceeds the empirical threshold.

5. The wireless telemetry and edge diagnosis method based on compressed lightweight perception according to claim 1, characterized in that: The steps of compressing and encoding the preprocessed vibration data using discrete cosine Huffman dual-mode coding are as follows: Combine the discrete cosine transform with Huffman coding to compress the original signal into a coded bit stream and a dictionary of elements: Among them, X(k) is the kth frequency domain coefficient, x(n) is the nth sampling value of the original time domain signal, N is the number of samples of the input signal, n is the index of the signal, and k is the index of the transformed frequency domain coefficient.

6. The wireless telemetry and edge diagnosis method based on compressed lightweight perception according to claim 5, characterized in that: The data receiving end decodes and reconstructs the encoded data through Huffman inverse coding and inverse discrete cosine transform, which is: Among them, X(k) is the kth frequency domain coefficient, x(n) is the nth sampling value of the original time domain signal, k is the index of the transformed frequency domain coefficient; and n is the index of the signal.

7. A wireless telemetry and edge diagnosis system based on compressed lightweight perception, characterized in that: It includes vibration sensor, temperature sensor, dual-channel acquisition module, transmission module and edge processing module; The vibration sensor and the temperature sensor are simultaneously connected to a dual-channel acquisition module, and the dual-channel acquisition module is used to synchronously acquire the vibration and temperature signals of the device; The edge processing module receives the signal collected by the dual-channel acquisition module through the transmission module, and executes the method described in any one of claims 1-6 to identify equipment faults.

8. The wireless telemetry and edge diagnosis system based on compressed lightweight perception as claimed in claim 7, characterized in that: It also includes a remote terminal, which is connected to the output end of the edge processing module through the wireless communication module.

9. The wireless telemetry and edge diagnosis system based on compressed lightweight perception according to claim 7, characterized in that: It also includes a power supply module, which is respectively connected to the power supply ends of the vibration sensor, the temperature sensor, the dual-channel acquisition module, the transmission module and the edge processing module.

Images