Noise processing method and system for electromechanical equipment based on integration of Internet of Things

Abstract

The invention relates to the technical field of noise processing, in particular to an electromechanical equipment noise processing method and system based on Internet of Things integration, and the method comprises the steps: collecting and preprocessing an original noise signal of electromechanical equipment; judging whether the preprocessed noise signal is a periodic signal or not; performing time domain analysis and frequency domain analysis on the preprocessed noise signal for feature extraction; wavelet transformation and feature extraction are carried out on the preprocessed noise signals; the extracted features are classified and recognized based on a machine learning algorithm, and noise processing is performed after a recognition model is constructed in a cloud platform to recognize the features. According to the invention, data with small calculation amount is processed through a local embedded system or a DSP system; the construction of the identification model with a large calculation amount is processed through the cloud platform; the local end communicates with the cloud platform, so that the calculation amount of the cloud platform is simplified, and the calculation efficiency is improved; the calculation efficiency is improved; the system power consumption is reduced; and meanwhile, a contribution is made to miniaturization of a noise processing system.

Description

Technical Field

[0001] The present invention relates to the technical field of noise processing, and in particular to a method and system for processing electromechanical equipment noise based on Internet of Things integration. Background Art

[0002] Electromechanical equipment is widely used in all walks of life and is closely related to people's production and daily lives. The emergence and upgrading of electromechanical equipment is a significant symbol of the evolution of human civilization and a fundamental material foundation for social modernization. Monitoring and real-time fault diagnosis of key electromechanical equipment to ensure safe operation is one of the most popular research topics.

[0003] Electromechanical equipment is accompanied by various noise signals during operation. On the other hand, noise detection of electromechanical equipment can also be used to obtain the operating status of the equipment and conduct fault diagnosis. For example, when machining bearings, the status and impact of the spindle will be reflected in the form of noise. The conventional method is for experienced workers to judge the fault location by listening to the location of the noise and abnormal sounds. With the development of artificial intelligence and Internet of Things technologies, feature extraction algorithms based on noise signals can quickly identify and locate fault points. However, large-scale data collection and computational processing for embedded systems or DSP systems will bring a large burden to the system. Therefore, how to reduce the amount of computing and efficiently process data has become a technical problem that needs to be solved urgently. Summary of the Invention

[0004] The present invention is proposed in view of the above-mentioned background problems.

[0005] Therefore, the problem to be solved by the present invention is how to optimize the calculation method, reduce the amount of calculation to improve the miniaturization and distribution of the system, reduce costs so that it can be applied to a wider range of equipment, provide real-time monitoring and optimized processing for equipment operation, and reduce noise and failure rate.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] A first aspect of the present invention provides a method for processing electromechanical equipment noise based on Internet of Things integration, comprising the following specific steps:

[0008] S1, collect the original noise signal of electromechanical equipment and preprocess it;

[0009] The pre-processing steps in step S1 include:

[0010] Filtering: Use a bandpass filter to filter the original signal and remove environmental noise;

[0011] Segment processing: Segment the waveform according to the waveform characteristics of the noise signal.

[0012] S2, determining whether the pre-processed noise signal is a periodic signal, if so, proceeding to step S3; if not, proceeding to step S6;

[0013] S3, performing time domain analysis on the pre-processed noise signal;

[0014] In step S3, the pre-processed noise signal is subjected to time domain analysis, specifically including: maximum value, minimum value, mean value, root mean square value, root mean square amplitude, standard deviation, kurtosis, skewness, and margin;

[0015] Maximum value: T1 = max(X(i))

[0016] Minimum value: T2 = min(X(i))

[0017] Mean:

[0018] RMS value:

[0019] Square root amplitude:

[0020] Standard Deviation:

[0021] Kurtosis:

[0022] Skewness:

[0023] Margin:

[0024] Where Xi represents the signal to be analyzed; N represents the number of signals to be analyzed;

[0025] The maximum and minimum values ​​are used to reflect the range of equipment operation noise signals;

[0026] The mean value is used to reflect the vibration signal generated by the change of the axis position of the equipment during operation;

[0027] The root mean square value is used to reflect the intensity and stability of the vibration signal;

[0028] Kurtosis is used to reflect the impact characteristics of vibration signals;

[0029] Skewness is used to reflect the asymmetry of the vibration signal;

[0030] Margin is used to reflect the wear condition of mechanical equipment.

[0031] S4. Use Fourier transform to convert the time domain signal into frequency domain and perform frequency domain analysis;

[0032] In step S4, the frequency domain is analyzed after Fourier transformation, specifically including: center of gravity frequency, average frequency and frequency variance;

[0033] Center of gravity frequency:

[0034] Average frequency:

[0035] Frequency variance:

[0036] where f k represents the frequency component; S(k) represents the power spectrum value of the corresponding frequency component; K represents the number of frequency components;

[0037] The center of gravity frequency reflects the main frequency component of the vibration signal. Identifying the vibration source is helpful to indicate the fault type and fault severity.

[0038] The average frequency reflects the energy or power difference of different frequency components, which is used to evaluate the intensity and stability of vibration and its impact on the equipment structure;

[0039] Frequency variance reflects the energy difference between different frequency components in the vibration signal and is used to identify abnormal frequency components in the vibration signal and monitor the changing trend of the vibration signal.

[0040] S5, extract features from the time domain signal and the frequency domain signal, and execute step S7;

[0041] S6, performing wavelet transform on the pre-processed noise signal and performing feature extraction, and executing step S7;

[0042] In step S6, a multi-scale wavelet transform is performed on the pre-processed data; the signal is first decomposed into a series of sub-band signals by performing multi-level high-pass and low-pass filtering on the signal to form components of different scales; and then feature extraction is performed according to the following formula:

[0043]

[0044] Where W(a,b) represents the wavelet coefficient, which is used to describe the characteristics of the signal X(i) at scale a and position b; X(i) is the signal to be analyzed; ψ is the mother wavelet function; a is the scaling factor; b is the translation factor;

[0045] The wavelet coefficients of the noise signal at different scales are obtained and the feature vectors are constructed based on them.

[0046] S7. Classify and identify the extracted features based on machine learning algorithms and build a recognition model in the cloud platform;

[0047] In step S7, the feature points obtained in step S5 and step S6 are normalized to construct a feature vector set A and a feature vector set B respectively; wherein the feature vector set A is a set of feature points in the time domain and the frequency domain; the feature vector set B is a set of wavelet coefficients at different scales;

[0048] S8. After identifying features using the recognition model, perform noise processing according to a predetermined scheme.

[0049] The obtained feature vector set is divided into a test set and a validation set in proportion, and a recognition model is constructed based on a neural network; the obtained feature vector is input into the recognition model, the noise type is identified through a classifier, and noise processing is performed according to a predetermined procedure.

[0050] A second aspect of the present invention provides an electromechanical equipment noise processing system based on Internet of Things integration, which is applied to the above method and includes a noise sensor, an embedded processing unit, a communication unit, a cloud platform, and an execution unit;

[0051] The noise sensor is used to collect noise signals when the equipment is running;

[0052] The embedded processing unit is used to process the collected noise signal and generate feature points and feature vectors;

[0053] The communication unit is used for communicating between the embedded processing unit and the cloud platform;

[0054] The cloud platform builds a recognition model and uses a classifier to identify the noise type based on the feature vector;

[0055] The execution unit performs processing according to a predetermined noise type processing procedure.

[0056] A third aspect of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned method and system for electromechanical equipment noise processing based on Internet of Things integration when executing the computer program.

[0057] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned electromechanical equipment noise processing method and system based on Internet of Things integration.

[0058] The beneficial effects of the present invention are:

[0059] The present invention uses a method of separate calculations on the local and cloud platforms to process the parts of data collection, preprocessing and feature point recognition with low computational complexity through a local embedded system or a DSP system; and the parts of recognition model construction iteration with high computational complexity are processed through the cloud platform; the local end and the cloud platform can communicate; at the same time, different local ends collect and process different data and send them to the cloud platform, and the cloud platform can perform more rounds of training and verification on the large amount of acquired data, thereby further improving the accuracy of the recognition model, and a large amount of basic and tedious data preprocessing content does not need to be uploaded to the cloud platform, which also simplifies the computational complexity of the cloud platform and improves computational efficiency; finally, the present application reduces the amount of data that needs to be calculated through a variety of preprocessing methods, and adopts different feature extraction methods according to periodic and non-periodic data, thereby solving the shortcomings of time domain-frequency domain analysis in extracting non-periodic waveform features, further optimizing the computational complexity, and helping to improve computational efficiency and reduce system power consumption; at the same time, it contributes to the miniaturization of the noise processing system. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0061] Figure 1 Flowchart of a method for processing electromechanical equipment noise based on Internet of Things integration in an embodiment of the present invention.

[0062] Figure 2 This is a structural diagram of an electromechanical equipment noise processing system based on Internet of Things integration in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0064] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0065] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0066] Example 1

[0067] A method for processing electromechanical equipment noise based on Internet of Things integration includes the following specific steps:

[0068] S1. Collect the original noise signal from the electromechanical equipment and perform preprocessing. The preprocessing steps in step S1 include filtering and segmentation. Filtering refers to filtering the original signal using a bandpass filter to remove environmental noise. The bandpass filter can be set to different frequency ranges based on the actual processing equipment, with a common range of 300Hz-2000Hz being adjustable. At this frequency, most noise from human voices and the environment can be filtered out, reducing the amount of data and improving the system's computational efficiency.

[0069] Segmentation processing: Segment the waveform according to the waveform characteristics of the noise signal. For example, select a 30-second noise signal and divide the waveform into 30 segments in 1-second units and number them in sequence.

[0070] In addition to the preprocessing process mentioned above, the original noise signal can also be processed using conventional denoising and waveform trimming methods to improve the quality of the waveform and reduce the subsequent processing workload.

[0071] S2. Determine whether the pre-processed noise signal is a periodic signal. If so, proceed to step S3; if not, proceed to step S6. Determine whether the noise signal is a periodic signal in step S2 by randomly extracting a segment of the noise signal and comparing it one by one in numerical order to determine whether the same noise signal appears and the number of times it appears. This determines whether the extracted noise signal is a periodic signal. For periodic signals, step S3 can be added for processing.

[0072] S3, performing time domain analysis on the pre-processed noise signal;

[0073] In step S3, the pre-processed noise signal is subjected to time domain analysis, specifically including: maximum value, minimum value, mean value, root mean square value, root mean square amplitude, standard deviation, kurtosis, skewness, and margin;

[0074] Maximum value: T1 = max(X(i))

[0075] Minimum value: T2 = min(X(i))

[0076] Mean:

[0077] RMS value:

[0078] Square root amplitude:

[0079] Standard Deviation:

[0080] Kurtosis:

[0081] Skewness:

[0082] Margin:

[0083] Where Xi represents the signal to be analyzed; N represents the number of signals to be analyzed;

[0084] The maximum and minimum values ​​are used to reflect the range of equipment operation noise signals;

[0085] The mean value is used to reflect the vibration signal generated by the change of the axis position of the equipment during operation;

[0086] The root mean square value is used to reflect the intensity and stability of the vibration signal;

[0087] Kurtosis is used to reflect the impact characteristics of vibration signals;

[0088] Skewness is used to reflect the asymmetry of the vibration signal;

[0089] Margin is used to reflect the wear condition of mechanical equipment.

[0090] S4. Use Fourier transform to convert the time domain signal into frequency domain and perform frequency domain analysis;

[0091] In step S4, the frequency domain is analyzed after Fourier transformation, specifically including: center of gravity frequency, average frequency and frequency variance;

[0092] Center of gravity frequency:

[0093] Average frequency:

[0094] Frequency variance:

[0095] where f k represents the frequency component; S(k) represents the power spectrum value of the corresponding frequency component; K represents the number of frequency components;

[0096] The center of gravity frequency reflects the main frequency component of the vibration signal. Identifying the vibration source is helpful to indicate the fault type and fault severity.

[0097] The average frequency reflects the energy or power difference of different frequency components, which is used to evaluate the intensity and stability of vibration and its impact on the equipment structure;

[0098] Frequency variance reflects the energy difference between different frequency components in the vibration signal and is used to identify abnormal frequency components in the vibration signal and monitor the changing trend of the vibration signal.

[0099] S5, extract features from the time domain signal and the frequency domain signal, and execute step S7;

[0100] S6, performing wavelet transform on the pre-processed noise signal and performing feature extraction, and executing step S7;

[0101] In step S6, a multi-scale wavelet transform is performed on the pre-processed data; the signal is first decomposed into a series of sub-band signals by performing multi-level high-pass and low-pass filtering on the signal to form components of different scales; and then feature extraction is performed according to the following formula:

[0102]

[0103] Where W(a,b) represents the wavelet coefficient, which is used to describe the characteristics of the signal X(i) at scale a and position b; X(i) is the signal to be analyzed; ψ is the mother wavelet function; a is the scaling factor; b is the translation factor;

[0104] The wavelet coefficients of the noise signal at different scales are obtained and the feature vectors are constructed based on them.

[0105] S7. Classify and identify the extracted features based on machine learning algorithms and build a recognition model in the cloud platform;

[0106] In step S7, the feature points obtained in step S5 and step S6 are normalized to construct a feature vector set A and a feature vector set B respectively; wherein the feature vector set A is a set of feature points in the time domain and the frequency domain; the feature vector set B is a set of wavelet coefficients at different scales;

[0107] S8. After identifying features using the recognition model, perform noise processing according to a predetermined scheme.

[0108] The obtained feature vector set is divided into a test set and a validation set in proportion, and a recognition model is constructed based on a neural network; the obtained feature vector is input into the recognition model, the noise type is identified through a classifier, and noise processing is performed according to a predetermined procedure.

[0109] In this embodiment, after constructing feature vector set A and feature vector set B, the two feature vector sets are separated and divided into test set and verification set respectively; based on the RNN recurrent neural network, the test set and verification set are used alternately for training, and a recognition model for feature vector set A and feature vector set B is constructed with a smaller amount of data. Then, the convergence parameters are adjusted and the recognition model is verified. When the F1-score (F1-score is the harmonic mean of accuracy and recall) index on the verification set reaches 0.85 or above, the model is considered to have been trained; the recognition model is placed in the server of the cloud platform; feature vector set A and feature vector set B are both set with a check code to distinguish the two feature vector sets, which can be checked with 0 or 1; 0 represents feature vector set A, and 1 represents feature vector set B. Only one feature vector set needs to be sent to the cloud platform each time, and the recognition result is fed back to the local computer. In this embodiment, noise data is collected and preprocessed to first eliminate environmental noise and human voice noise to improve the cleanliness of the data; then, different schemes are used to construct different feature vector sets according to the type of noise signal, which greatly reduces the amount of data calculation on the local side; after the feature vector set communicates with the cloud platform, the result is recognized by the recognition model and returned to the local side, and then optimized according to the preset program.

[0110] Example 2

[0111] The second aspect of the present invention provides an electromechanical equipment noise processing system based on Internet of Things integration, which is applied to the method in Example 1, including a noise sensor, an embedded processing unit, a communication unit, a cloud platform, and an execution unit; the noise sensor is used to collect noise signals during equipment operation; the embedded processing unit is used to process the collected noise signals and generate feature points and feature vectors; the communication unit is used to communicate between the embedded processing unit and the cloud platform; the cloud platform constructs a recognition model and identifies the noise type through a classifier based on the feature vector; the execution unit performs processing according to a predetermined noise type processing procedure.

[0112] Specifically, in this embodiment, the No. 1 equipment in the bearing processing workshop of a factory in the south is used as an example for introduction;

[0113] First, multiple acquisition units are grouped around the electromechanical equipment according to component location. Each acquisition unit consists of multiple evenly spaced noise sensors. For example, the first acquisition unit is located at the spindle, and the second acquisition unit is located at the motor. The waveforms of the diverse noise signals collected by the noise sensors are compared using a comparator. Data from different noise sensors can further eliminate interference from environmental and human noise, thereby improving data cleanliness. Taking the second acquisition unit located at the motor as an example, the noise sensors should be evenly spaced around the motor housing, including the top, bottom, and sides. Environmental or human noise signals are far from the noise sensors and are overwhelmed by the noise generated by the motor's operation. Therefore, the noise signals from the same noise source obtained by different noise sensors will vary. However, the noise generated by the motor's operation or fault noise tends to vary little within a small range. Therefore, the noise waveforms collected by different noise sensors during the same period can be screened. Generally, deviations greater than 20% are considered noise. This noise can be removed, thereby filtering out interference from environmental and human noise. A bandpass filter can significantly reduce environmental interference and reduce the amount of subsequent computation.

[0114] Secondly, the embedded unit processes the noise signal after the preprocessing step, and first determines whether the noise signal is a periodic signal; specifically, a segment of the segmented noise signal is randomly extracted, and compared one by one in the order of numbering to determine whether the same noise signal appears and the number of times it appears; thereby determining whether the extracted noise signal is a periodic signal; time domain and frequency domain analysis are used to extract features of the periodic signal; for non-periodic signals, wavelet transform is used to extract features; for periodic data, conventional time domain-frequency domain analysis can quickly and accurately identify feature points, and based on the feature points, the fault location and cause can be quickly determined, and targeted processing methods can be made; for non-periodic signals, such as occasional faults, the waveform changes due to the superposition of fault noise and operating noise, and the time domain-frequency domain analysis cannot quickly identify feature points, resulting in slow processing speed and abnormal state; the present invention uses wavelet transform to quickly extract local feature information in the time domain and frequency domain; with the multi-threading technology of the processor, time domain-frequency domain analysis and wavelet transform can be processed simultaneously, further improving the data processing speed.

[0115] Specifically: By performing time domain analysis on the pre-processed noise signal, characteristic information including maximum value, minimum value, mean value, root mean square value, root square amplitude, standard deviation, kurtosis, skewness, and margin can be obtained; among them, the maximum value and minimum value are used to reflect the range of the equipment operation noise signal; the mean value is used to reflect the vibration signal generated by the change of the equipment operation axis position; the root mean square value is used to reflect the intensity and stability of the vibration signal; the kurtosis is used to reflect the impact characteristics of the vibration signal; the skewness is used to reflect the asymmetry of the vibration signal; and the margin is used to reflect the wear condition of the mechanical equipment.

[0116] The time domain is then converted to the frequency domain based on Fourier transform for analysis, which can obtain characteristic information including: center of gravity frequency, average frequency and frequency variance. The center of gravity frequency reflects the main frequency component of the vibration signal, identifying the vibration source and facilitating the indication of the fault type and fault severity. The average frequency reflects the energy or power difference between different frequency components, which is used to evaluate the intensity and stability of the vibration and its impact on the equipment structure. The frequency variance reflects the energy difference between different frequency components in the vibration signal, which is used to identify abnormal frequency components in the vibration signal and monitor the changing trend of the vibration signal. The time domain and frequency domain feature points are superimposed to construct the feature vector set A.

[0117] The signal is decomposed into a series of sub-band signals by performing multi-level high-pass and low-pass filtering on the signal to form components of different scales; then feature extraction is performed according to the following formula:

[0118]

[0119] Where W(a,b) represents the wavelet coefficient, which is used to describe the characteristics of the signal X(i) at scale a and position b; X(i) is the signal to be analyzed; ψ is the mother wavelet function; a is the scaling factor; b is the translation factor;

[0120] The wavelet coefficients of the noise signal at different scales are obtained and the feature vector set B is constructed based on them.

[0121] Based on the obtained feature vector set, a recognition model is constructed in the cloud platform; specifically: after constructing feature vector set A and feature vector set B, the two feature vector sets are separated and divided into test set and validation set respectively; based on the RNN recurrent neural network, the test set and validation set are used alternately for training, and a recognition model for feature vector set A and feature vector set B is constructed with a smaller amount of data, and then the convergence parameters are adjusted and the recognition model is verified. When the F1-score (F1-score is the harmonic mean of accuracy and recall) index on the validation set reaches 0.85 or above, the model is considered to have been trained; the recognition model is placed in the server of the cloud platform; feature vector set A and feature vector set B are both set with a check code to distinguish the two feature vector sets, which can be checked with 0 or 1; 0 represents feature vector set A, and 1 represents feature vector set B. Only one feature vector set needs to be sent to the cloud platform each time, and the recognition result is fed back to the local after it is recognized. The recognition model uses a classifier to output different results and develops a complete solution or procedure for each fault; for each feature vector set input into the recognition model, the type is first verified by the check bit; in the recognition model corresponding to the input, the recognition results are sorted according to the correlation degree of the feature point comparison, and the results are displayed in descending order; based on the feedback information, the user executes the corresponding action according to the preset program through the execution unit to eliminate the fault point automatically or manually.

[0122] In this embodiment, large-scale computing projects such as the construction and iteration of the recognition model are placed in the cloud platform. The local embedded system or DSP system only needs to pre-process the original data to construct a feature vector set, and cooperate with the communication unit to communicate with the recognition model in the cloud platform in real time; at the same time, for multiple discrete embedded systems, after the data is aggregated to the cloud platform, the data volume can be further increased, and the recognition model can improve the recognition accuracy through multiple iterative deep learning.

[0123] Example 3

[0124] A third aspect of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned method and system for electromechanical equipment noise processing based on Internet of Things integration when executing the computer program.

[0125] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0126] Example 4

[0127] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-mentioned electromechanical equipment noise processing method and system based on Internet of Things integration.

[0128] The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination of them, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0129] In summary, the present invention demonstrates the important value of the electromechanical equipment noise processing method and system based on Internet of Things integration in industrial production; by calculating separately locally and on the cloud platform, the parts of data collection, preprocessing and feature point recognition with small computational complexity are processed by the local embedded system or DSP system; and the parts such as construction and iteration of the recognition model with large computational complexity are processed by the cloud platform; the local end and the cloud platform can communicate; at the same time, different local ends collect and process different data and send them to the cloud platform, and the cloud platform can perform more rounds of training and verification on the large amount of data obtained, thereby further improving the accuracy of the recognition model, and a large amount of basic and tedious data preprocessing content does not need to be uploaded to the cloud platform, which also simplifies the computational complexity of the cloud platform and improves computational efficiency; finally, this application reduces the amount of data that needs to be calculated through a variety of preprocessing methods, and adopts different feature extraction methods according to periodic and non-periodic data, thereby solving the shortcomings of time domain-frequency domain analysis in extracting non-periodic waveform features, further optimizing the computational complexity, and helping to improve computational efficiency and reduce system power consumption; at the same time, it contributes to the miniaturization of the noise processing system.

[0130] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for processing electromechanical equipment noise based on Internet of Things integration, characterized by: The specific steps include: S1, collect the original noise signal of electromechanical equipment and preprocess it; S2, determining whether the pre-processed noise signal is a periodic signal, if so, proceeding to step S3; if not, proceeding to step S6; S3, performing time domain analysis on the pre-processed noise signal; S4. Use Fourier transform to convert the time domain signal into frequency domain and perform frequency domain analysis; S5, extract features from the time domain signal and the frequency domain signal, and execute step S7; S6, performing wavelet transform on the pre-processed noise signal and performing feature extraction, and executing step S7; S7. Classify and identify the extracted features based on machine learning algorithms and build a recognition model in the cloud platform; S8. After identifying features using the recognition model, perform noise processing according to a predetermined scheme.

2. The electromechanical equipment noise processing method based on Internet of Things integration according to claim 1 is characterized in that: The pre-processing steps in step S1 include: Filtering: Use a bandpass filter to filter the original signal and remove environmental noise; Segment processing: Segment the waveform according to the waveform characteristics of the noise signal.

3. The electromechanical equipment noise processing method based on Internet of Things integration according to claim 1 is characterized in that: In step S3, the pre-processed noise signal is subjected to time domain analysis, specifically including: maximum value, minimum value, mean value, root mean square value, root mean square amplitude, standard deviation, kurtosis, skewness, and margin; Maximum value: T1 = max(X(i)) Minimum value: T2 = min(X(i)) Mean: RMS value: Square root amplitude: Standard Deviation: Kurtosis: Skewness: Margin: Where Xi represents the signal to be analyzed; N represents the number of signals to be analyzed; The maximum and minimum values ​​are used to reflect the range of equipment operation noise signals; The mean value is used to reflect the vibration signal generated by the change of the axis position of the equipment during operation; The root mean square value is used to reflect the intensity and stability of the vibration signal; Kurtosis is used to reflect the impact characteristics of vibration signals; Skewness is used to reflect the asymmetry of the vibration signal; Margin is used to reflect the wear condition of mechanical equipment.

4. The method for processing electromechanical equipment noise based on Internet of Things integration according to claim 1, characterized in that: In step S4, the frequency domain is analyzed after Fourier transformation, specifically including: center of gravity frequency, average frequency and frequency variance; Center of gravity frequency: Average frequency: Frequency variance: where f k represents the frequency component; S(k) represents the power spectrum value of the corresponding frequency component; K represents the number of frequency components; The center of gravity frequency reflects the main frequency component of the vibration signal. Identifying the vibration source is helpful to indicate the fault type and fault severity. The average frequency reflects the energy or power difference of different frequency components, which is used to evaluate the intensity and stability of vibration and its impact on the equipment structure; Frequency variance reflects the energy difference between different frequency components in the vibration signal and is used to identify abnormal frequency components in the vibration signal and monitor the changing trend of the vibration signal.

5. The electromechanical equipment noise processing method based on Internet of Things integration according to claim 1 is characterized in that: In step S6, a multi-scale wavelet transform is performed on the pre-processed data; the signal is first decomposed into a series of sub-band signals by performing multi-level high-pass and low-pass filtering on the signal to form components of different scales; and then feature extraction is performed according to the following formula: Where W(a, b) represents the wavelet coefficient, which is used to describe the characteristics of the signal X(i) at scale a and position b; X(i) is the signal to be analyzed; ψ is the mother wavelet function; a is the scaling factor; b is the translation factor; The wavelet coefficients of the noise signal at different scales are obtained and the feature vectors are constructed based on them.

6. The electromechanical equipment noise processing method based on Internet of Things integration according to claim 1 is characterized in that: In step S7, the feature points obtained in step S5 and step S6 are normalized to construct feature vector set A and feature vector set B respectively; wherein feature vector set A is a set of feature points in the time domain and frequency domain; feature vector set B is a set of wavelet coefficients at different scales.

7. The electromechanical equipment noise processing method based on Internet of Things integration according to claim 1 is characterized in that: The obtained feature vector set is divided into a test set and a validation set in proportion, and a recognition model is constructed based on a neural network; the obtained feature vector is input into the recognition model, the noise type is identified through a classifier, and noise processing is performed according to a predetermined procedure.

8. An electromechanical equipment noise processing system based on Internet of Things integration, applied to the method according to any one of claims 1 to 7, characterized in that: Includes noise sensor, embedded processing unit, communication unit, cloud platform, and execution unit; The noise sensor is used to collect noise signals when the equipment is running; The embedded processing unit is used to process the collected noise signal and generate feature points and feature vectors; The communication unit is used for communicating between the embedded processing unit and the cloud platform; The cloud platform builds a recognition model and uses a classifier to identify the noise type based on the feature vector; The execution unit performs processing according to a predetermined noise type processing procedure.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the electromechanical equipment noise processing method and system based on Internet of Things integration according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the electromechanical equipment noise processing method and system based on Internet of Things integration according to any one of claims 1 to 7 are implemented.

Images