Explosion-proof motor fault diagnosis method and system based on big data

Abstract

The invention relates to the technical field of motor fault diagnosis, and discloses an explosion-proof motor fault diagnosis method and system based on big data, and the method comprises the steps: obtaining real-time multi-source signal data, and generating an initial data set; performing data preprocessing according to the initial data set to obtain a standardized data set; according to the standardized data set, performing dimension reduction processing by adopting a principal component analysis method to generate a simplified feature data set; constructing a fault feature library according to the simplified feature data set, and generating a structured feature mapping table; performing fault type feature matching according to the structured feature mapping table and the real-time multi-source signal data to obtain a fault type identifier; and according to the fault type identifier and the real-time multi-source signal data, in combination with historical feature distribution data acquired in advance, evaluating the severity of the fault, and generating a fault mode identification report. According to the method, systematic overall analysis can be carried out on multi-dimensional motor operation data.

Description

Technical Field

[0001] The present invention relates to the technical field of motor fault diagnosis, and in particular to an explosion-proof motor fault diagnosis method and system based on big data. Background Art

[0002] The stable operation of motors is crucial to industrial manufacturing and equipment maintenance. Monitoring their health and diagnosing faults are key to ensuring the smooth functioning of industrial systems. With the advancement of industrial intelligence, timely detection of motor faults through data analysis and effective action have become a key focus of the industry. Research in this area not only extends equipment life but also directly impacts production cost control and the prevention of safety incidents.

[0003] In one existing technology, the fault diagnosis of explosion-proof motors mainly relies on a simple threshold judgment method, which monitors a single parameter such as vibration, current or temperature and combines it with a preset normal range threshold to determine whether the motor has an abnormality. For example, technicians collect the vibration frequency data of the motor through a vibration sensor and compare it with a fixed threshold. If it exceeds the threshold, it is considered that a fault exists. In addition, some methods attempt to introduce data analysis technology, such as statistical-based anomaly detection, which identifies abnormal points by calculating the mean and variance of vibration or current. However, these methods usually only focus on a single or a few parameters, and the correlation between multiple source signals such as vibration frequency, current intensity, temperature changes and load status is often ignored, resulting in the diagnosis process being unable to fully capture the complex characteristics of the motor's operating status. In addition, when dealing with different operating conditions (such as load status and speed changes), the existing technology has difficulty in dynamically adjusting the diagnostic mode, and is prone to misjudgment or omission due to the diversity of data features.

[0004] In summary, the existing technology has the problem of difficulty in achieving systematic and comprehensive analysis of multi-dimensional motor operation data. Summary of the Invention

[0005] The present invention provides an explosion-proof motor fault diagnosis method and system based on big data, so as to realize systematic and comprehensive analysis of multi-dimensional motor operation data.

[0006] In the first aspect, in order to solve the above technical problems, the present invention provides an explosion-proof motor fault diagnosis method based on big data, comprising:

[0007] Acquire real-time multi-source signal data during motor operation and generate an initial data set containing timestamps and operating condition identifiers;

[0008] Performing data preprocessing on the initial data set to obtain a standardized data set;

[0009] Based on the standardized data set, a principal component analysis method is used to perform dimensionality reduction processing to generate a simplified feature data set;

[0010] Building a fault feature library based on the simplified feature data set and generating a structured feature mapping table;

[0011] Perform fault type feature matching based on the structured feature mapping table and the real-time multi-source signal data to obtain a fault type identifier;

[0012] According to the fault type identification and the real-time multi-source signal data, combined with the pre-acquired historical feature distribution data, a support vector machine algorithm is used to evaluate the fault severity and generate a fault pattern recognition report including the fault type, severity and working condition description.

[0013] In an optional embodiment, the step of acquiring real-time multi-source signal data during motor operation and generating an initial data set including a timestamp and an operating condition identifier includes:

[0014] Acquire real-time multi-source signal data during motor operation;

[0015] The real-time multi-source signal data includes the speed parameters, vibration frequency, amplitude, current intensity, fluctuation frequency, surface temperature, temperature rise rate and load status of the motor during operation;

[0016] According to the real-time multi-source signal data, identification division is performed to obtain an initial data set including a timestamp and an operating condition identification.

[0017] In an optional embodiment, performing data preprocessing on the initial data set to obtain a standardized data set includes:

[0018] According to the initial data set, a wavelet transform filtering algorithm is used to perform denoising on the data to obtain a denoised data set;

[0019] Based on the denoised data set, the threshold ranges of vibration frequency and current intensity are calculated using the interquartile range statistical method;

[0020] Checking whether the data in the denoised data set is within the threshold range, and when the data exceeds the threshold range, removing the corresponding data; when the data is within the threshold range, retaining the corresponding data, and obtaining a data set without outliers;

[0021] Based on the data without outliers, a k-means clustering algorithm is used to perform classification to generate a segmented data set with working condition identification;

[0022] According to the segmented data set, a z-score normalization method is used to perform normalization processing to obtain a standardized data set.

[0023] In an optional embodiment, the step of performing dimensionality reduction processing using a principal component analysis method based on the standardized data set to generate a simplified feature data set includes:

[0024] Analyzing the statistical relationship between the load state and the speed parameter through covariance calculation based on the standardized data set to generate a covariance matrix;

[0025] Performing eigendecomposition according to the covariance matrix, calculating eigenvalues ​​and eigenvectors, and obtaining an eigenvalue set and an eigenvector set;

[0026] According to the eigenvalue set and the eigenvector set, the cumulative contribution rate is calculated. When the cumulative contribution rate is greater than the preset contribution rate threshold, the corresponding eigenvector is determined to be the principal component and retained. When the cumulative contribution rate is less than the preset contribution rate threshold, the corresponding eigenvector is eliminated to generate a streamlined feature data set.

[0027] In an optional implementation, constructing a fault feature library based on the reduced feature data set and generating a structured feature mapping table includes:

[0028] Performing normalization processing on the simplified feature data set to obtain a normalized feature data set;

[0029] According to the normalized feature data set, the vibration frequency and current intensity are classified using a k-means clustering algorithm to generate a fault feature data set;

[0030] The fault feature data set is associated with a preset fault mode to obtain a structured feature mapping table.

[0031] In an optional embodiment, performing fault type feature matching on the structured feature mapping table and the real-time multi-source signal data to obtain a fault type identifier includes:

[0032] Performing normalization processing on the real-time multi-source signal data to obtain normalized real-time signal data;

[0033] Calculating the Euclidean distance between the normalized real-time signal data and the feature vector of the fault mode in the structured feature mapping table to obtain a Euclidean distance set;

[0034] When the Euclidean distance in the Euclidean distance set is less than the preset similarity threshold, the data corresponding to the Euclidean distance is marked with the corresponding type of fault. When the Euclidean distance in the Euclidean distance set is greater than the preset similarity threshold, it is determined that the data corresponding to the Euclidean distance does not have the corresponding type of fault, and the fault type identification is obtained.

[0035] In an optional embodiment, the fault severity is assessed using a support vector machine algorithm based on the fault type identifier and the real-time multi-source signal data in combination with pre-acquired historical feature distribution data, and a fault pattern recognition report including the fault type, severity, and operating condition description is generated, including:

[0036] According to the fault type identification, combined with the previously acquired historical feature distribution data, feature distribution statistics are calculated to obtain a fault feature distribution set;

[0037] Based on the fault feature distribution set, a support vector machine algorithm is used to perform classification calculation, and a classification hyperplane is constructed to distinguish different fault severity categories to obtain a fault severity value;

[0038] Data integration is performed based on the fault severity value and the real-time multi-source signal data to generate a fault mode recognition report including the fault type, severity and working condition description.

[0039] In a second aspect, the present invention provides an explosion-proof motor fault diagnosis device based on big data, comprising:

[0040] The data acquisition module is used to obtain real-time multi-source signal data during motor operation and generate an initial data set containing a timestamp and operating condition identification;

[0041] A standard processing module, configured to perform data preprocessing based on the initial data set to obtain a standardized data set;

[0042] A dimensionality reduction and simplification module is used to perform dimensionality reduction processing based on the standardized data set using a principal component analysis method to generate a simplified feature data set;

[0043] A feature mapping module is used to construct a fault feature library based on the simplified feature data set and generate a structured feature mapping table;

[0044] A fault identification module is used to match fault type features with the real-time multi-source signal data according to the structured feature mapping table to obtain a fault type identification;

[0045] The result output module is used to evaluate the severity of the fault using a support vector machine algorithm based on the fault type identification and the real-time multi-source signal data, combined with the pre-acquired historical feature distribution data, and generate a fault pattern recognition report including the fault type, severity and working condition description.

[0046] In a third aspect, the present invention also provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements any one of the above-mentioned explosion-proof motor fault diagnosis methods based on big data.

[0047] In a fourth aspect, the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned explosion-proof motor fault diagnosis methods based on big data.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] (1) The present invention obtains real-time multi-source signal data during motor operation, including speed parameters, vibration frequency, amplitude, current intensity, fluctuation frequency, surface temperature, temperature rise rate and load status, and performs identification division to generate an initial data set containing a timestamp and working condition identification. This process can effectively organize the data structure and improve the efficiency of data processing through the synchronous acquisition of multi-source signals and classification of working conditions. The comprehensive acquisition of multi-source signals avoids the time delay caused by multiple independent measurements, and the division of working condition identification enables subsequent analysis to quickly locate specific operating conditions, thereby reducing the time for data screening and improving computing efficiency. In addition, by extracting features from abnormal working condition data, the pertinence of the data is enhanced, further improving the accuracy of diagnosis.

[0050] (2) Based on the initial data set, the present invention removes noise through a wavelet transform filtering algorithm, uses the interquartile range statistical method to eliminate outliers, and generates a standardized data set based on physical attribute classification. This process can significantly improve computational efficiency by optimizing data quality step by step. The wavelet transform filtering algorithm reduces interference items in the data by separating high-frequency noise and low-frequency effective signals, speeding up subsequent processing. The interquartile range statistical method simplifies the data scale by determining the normal range and eliminating outliers. K-means clustering and z-score standardization make the data structure clearer and reduce unnecessary computational complexity through classification and normalization. Standardization enhances the consistency of the data, thereby improving the accuracy of fault feature extraction.

[0051] (3) Based on the standardized data set, the present invention adopts the principal component analysis method to analyze statistical relationships through covariance calculation, extract eigenvalues ​​and eigenvectors through eigendecomposition, and generate a simplified feature data set based on the cumulative contribution rate. This process effectively improves computational efficiency through dimensionality reduction technology. Covariance calculation and eigendecomposition identify the main direction of change. The screening of cumulative contribution rate retains key information and eliminates redundant data, reducing the data dimension and thus reducing the subsequent computational workload. K-means clustering combines working conditions and physical properties to further optimize data organization, improve the pertinence of feature extraction, and thus improve the accuracy of fault diagnosis.

[0052] (4) Based on a streamlined feature dataset, the present invention generates a fault feature dataset through normalization and k-means clustering, and then associates it with preset fault patterns to generate a structured feature mapping table. This process improves computational efficiency through data structuring. Normalization unifies the data scale and reduces computational differences between different parameters; k-means clustering accelerates the processing speed of fault pattern identification by classifying and clustering similar features; and the generation of the structured feature mapping table reduces the search range during real-time matching by pre-associating fault patterns, thereby improving computational efficiency, enhancing feature differentiation, and improving the accuracy of fault diagnosis.

[0053] (5) The present invention matches fault type features based on structured feature mapping tables and real-time multi-source signal data through normalization processing and Euclidean distance calculation to obtain fault type identification. This process improves computational efficiency through predefined mapping. Normalization processing unifies the data format and reduces dimensional differences during matching; Euclidean distance calculation quickly identifies the fault type by comparing the similarity between real-time data and the feature library, reducing unnecessary data comparison steps; the application of preset similarity thresholds further optimizes the matching process and improves computational efficiency. At the same time, threshold judgment improves the accuracy of fault type identification, thereby improving the accuracy of diagnosis.

[0054] (6) The present invention uses a support vector machine algorithm to assess fault severity and generate a fault pattern recognition report based on the fault type identification and pre-acquired historical feature distribution data. This process improves computational efficiency by utilizing historical data. The statistical parameter calculation of historical feature distribution data provides a preprocessing basis for the support vector machine, reducing the complexity of real-time data analysis. The support vector machine quickly distinguishes fault severity categories by constructing a classification hyperplane, shortening the evaluation time. The integration of historical data and the application of the classification hyperplane enhance the accuracy of severity assessment and improve the accuracy of diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 This is a flowchart of a method for diagnosing explosion-proof motor faults based on big data provided by the first embodiment of the present invention;

[0056] Figure 2 This is a schematic diagram of the structure of an explosion-proof motor fault diagnosis system / device based on big data provided by the second embodiment of the present invention. DETAILED DESCRIPTION

[0057] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0058] Reference Figure 1 The first embodiment of the present invention provides a method for diagnosing explosion-proof motor faults based on big data, comprising the following steps:

[0059] S11, acquiring real-time multi-source signal data during motor operation and generating an initial data set including a timestamp and an operating condition identifier;

[0060] S12, performing data preprocessing based on the initial data set to obtain a standardized data set;

[0061] S13, performing dimensionality reduction processing using a principal component analysis method based on the standardized data set to generate a simplified feature data set;

[0062] S14, constructing a fault feature library based on the simplified feature data set and generating a structured feature mapping table;

[0063] S15, performing fault type feature matching based on the structured feature mapping table and the real-time multi-source signal data to obtain a fault type identifier;

[0064] S16, based on the fault type identification and the real-time multi-source signal data, combined with the pre-acquired historical feature distribution data, a support vector machine algorithm is used to evaluate the fault severity and generate a fault pattern recognition report including the fault type, severity and working condition description.

[0065] In step S11 , it is necessary to obtain real-time multi-source signal data during the operation of the motor and generate an initial data set including a timestamp and an operating condition identifier.

[0066] In one implementation, real-time multi-source signal data during motor operation is acquired to generate an initial data set containing a timestamp and an operating condition identifier, including:

[0067] Real-time multi-source signal data during motor operation is acquired; the real-time multi-source signal data includes speed parameters, vibration frequency, amplitude, current intensity, fluctuation frequency, surface temperature, temperature rise rate, and load status during motor operation; and identification division is performed based on the real-time multi-source signal data to obtain an initial data set including a timestamp and an operating condition identification.

[0068] It should be noted that, first, a sensor array collects multi-source signal data during motor operation in real time. Specifically, this involves utilizing vibration sensors, current sensors, temperature sensors, and other devices to synchronously record parameters such as motor speed, vibration frequency, amplitude, current intensity, fluctuation frequency, surface temperature, temperature rise rate, and load status, ensuring real-time and accurate data. The collected data is stored in a time series format, with one record generated per second, containing all parameter values ​​and corresponding timestamps. Next, the content of the real-time multi-source signal data is clarified. Speed ​​parameters are measured by speed sensors, reflecting the number of revolutions per minute of the motor. Vibration frequency and amplitude are acquired by vibration sensors, representing the periodicity and intensity of vibration, respectively. Current intensity and fluctuation frequency are measured by current sensors, reflecting the motor's energy consumption and current stability. Surface temperature and temperature rise rate are recorded by temperature sensors, reflecting the motor's thermal state during operation. Load status is determined by a power meter or related sensors, representing the motor's actual workload. Together, these parameters constitute multi-source signal data, covering multiple dimensions of motor operation. Then, based on the real-time multi-source signal data, identification and segmentation are performed. Specifically, the data is segmented according to load conditions and speed parameters. For example, load is divided into light load, medium load, and heavy load, and speed is divided into low speed, medium speed, and high speed. Each data segment is attached with a working condition identifier, such as "light load-low speed", while retaining the timestamp of each data item, ultimately generating an initial data set. The initial data set is a structured data set containing timestamps, working condition identifiers, and multi-source signal parameters. It is obtained through sensor array acquisition and working condition classification. It is used for subsequent data preprocessing and feature extraction, providing basic data support for fault diagnosis.

[0069] In step S12, data preprocessing needs to be performed based on the initial data set to obtain a standardized data set.

[0070] In one implementation, data preprocessing is performed on the initial data set to obtain a standardized data set, including:

[0071] According to the initial data set, a wavelet transform filtering algorithm is used to denoise the data to obtain a denoised data set; according to the denoised data set, an interquartile range statistical method is used to calculate the threshold range of vibration frequency and current intensity; it is checked whether the data in the denoised data set is within the threshold range, when the data exceeds the threshold range, the corresponding data is eliminated, and when the data is within the threshold range, the corresponding data is retained to obtain a data set without outliers; according to the data without outliers, a k-means clustering algorithm is used for classification to generate a segmented data set with working condition identification; according to the segmented data set, a z-score normalization method is used for normalization to obtain a standardized data set.

[0072] It should be noted that data containing vibration frequency, current fluctuation, and timestamps are first extracted from the initial dataset. The data is then denoised using a wavelet transform filtering algorithm. The wavelet transform splits the data into different frequency components through multi-scale decomposition, retaining low-frequency signals related to motor operation and removing high-frequency noise interference, thereby generating a denoised dataset. Next, the interquartile range statistical method is used to calculate the threshold range for the vibration frequency and current intensity in the denoised dataset. Specifically, the upper and lower quartiles of the data are determined based on the distribution characteristics of the statistical data, the interquartile range is calculated, and the normal range is set to a reasonable multiple of the interquartile range, for example, 1.5 times the interquartile range, to form a threshold range. Then, the vibration frequency and current intensity data in the denoised dataset are checked, and each data point is compared one by one to see if it is within the calculated threshold range. If the data point exceeds the range, it is considered an outlier and eliminated. If it is within the range, it is retained, ultimately resulting in a dataset without outliers. Afterwards, the k-means clustering algorithm is used to classify the data set without outliers. K-means clustering uses iterative optimization to assign data points to the most similar operating condition groups according to load conditions and speed parameters, such as light load and low speed, heavy load and high speed, etc. Each set of data is attached with an operating condition identifier to generate a segmented data set. Finally, the vibration frequency and current fluctuations in the segmented data set are normalized using the z-score normalization method. Z-score normalization calculates the deviation of each data point from the mean and divides it by the standard deviation, converting the data into a standard distribution with a mean of zero and a standard deviation of one, resulting in a standardized data set. The standardized data set is a set of structured data that has been denoised, outlier removed, classified, and normalized. The data dimensions are consistent and the distribution is uniform. It is used for subsequent dimensionality reduction and feature extraction, providing a high-quality data foundation for fault diagnosis.

[0073] In step S13, it is necessary to perform dimensionality reduction processing using a principal component analysis method based on the standardized data set to generate a simplified feature data set.

[0074] In one implementation, a principal component analysis method is used to perform dimensionality reduction processing based on the standardized data set to generate a simplified feature data set, including:

[0075] According to the standardized data set, the statistical relationship between the load state and the speed parameter is analyzed through covariance calculation to generate a covariance matrix; according to the covariance matrix, eigendecomposition is performed, the eigenvalues ​​and eigenvectors are calculated, and the eigenvalue set and the eigenvector set are obtained; according to the eigenvalue set and the eigenvector set, the cumulative contribution rate is calculated. When the cumulative contribution rate is greater than the preset contribution rate threshold, the corresponding eigenvector is determined to be the principal component and retained. When the cumulative contribution rate is less than the preset contribution rate threshold, the corresponding eigenvector is eliminated to generate a streamlined feature data set.

[0076] It should be noted that first, the vibration frequency, current fluctuation, load state, and speed parameter data are extracted from the standardized data set, and the covariance calculation method is used to analyze the statistical relationship between the load state and the speed parameters. Specifically, by calculating the covariance between each pair of parameters, the linear correlation between the parameters is reflected, such as the positive correlation between the load state and the vibration frequency, and a covariance matrix is ​​generated. The covariance matrix provides the mathematical basis for subsequent eigendecomposition. Next, the eigendecomposition method is used to calculate the covariance matrix. The eigendecomposition decomposes the covariance matrix into eigenvalues ​​and eigenvectors through mathematical transformation. The eigenvalue represents the variance of the data in the direction of the corresponding eigenvector, and the eigenvector represents the main direction of change of the data, thereby obtaining a set of eigenvalues ​​and a set of eigenvectors. Then, the cumulative contribution rate is calculated based on the set of eigenvalues. The specific method is to add all the eigenvalues ​​to obtain the total variance, and then calculate the proportion of the sum of the first several eigenvalues ​​to the total variance to obtain the cumulative contribution rate. If the cumulative contribution rate is greater than the preset contribution rate threshold, the preset contribution rate threshold is determined by analyzing the characteristic distribution and diagnostic accuracy requirements of the historical motor operation data. Combined with multiple principal component analysis tests, the impact of feature retention on fault identification under different thresholds is evaluated. Finally, according to industry standards, 85% is selected as a reasonable value that balances data representativeness and computational efficiency, and then the corresponding eigenvector is selected as the principal component and retained. If it is less than the threshold, the corresponding eigenvector is eliminated. Through this screening process, a simplified feature data set is generated. The simplified feature data set is a set of eigenvectors that have been processed by dimensionality reduction, which retains the main information of the data and eliminates redundant dimensions. It is used for the subsequent construction of the fault feature library and fault pattern matching, which can effectively reduce computational complexity and improve diagnostic efficiency.

[0077] In step S14, it is necessary to construct a fault feature library based on the simplified feature data set and generate a structured feature mapping table.

[0078] In one implementation, a fault feature library is constructed based on the simplified feature data set, and a structured feature mapping table is generated, including:

[0079] According to the simplified feature data set, normalization processing is performed to obtain a normalized feature data set; according to the normalized feature data set, the vibration frequency and current intensity are classified using the k-means clustering algorithm to generate a fault feature data set; the fault feature data set is associated with a preset fault mode to obtain a structured feature mapping table.

[0080] It should be noted that feature data such as vibration frequency and current intensity are first extracted from the streamlined feature dataset and normalized using the z-score normalization method. Specifically, the deviation of each data point from the mean is calculated and divided by the standard deviation. This transforms the data into a standard distribution with a mean of zero and a standard deviation of one, generating a normalized feature dataset. Next, the k-means clustering algorithm is used to classify the vibration frequency and current intensity within the normalized feature dataset. K-means clustering sets several cluster centers and iteratively adjusts the assignment of each data point to the nearest center. Based on load conditions and speed parameters, the data is classified into different categories, such as normal operation and minor faults, generating a fault feature dataset. Then, the fault feature dataset is associated with the preset fault mode. The specific method is to label each type of feature data as a specific fault mode, such as bearing wear or circuit fault, based on the clustering results. A structured feature mapping table is generated through mapping relationships. The mapping relationship is a rule established by corresponding each type of feature vector to the preset fault mode. It is formulated based on the clustering results. For example, the class with high vibration frequency and current intensity is mapped to "bearing wear", and the class with abnormal current is mapped to "circuit fault". The process of generating the structured feature mapping table is to extract the representative value of each type of feature vector and combine it with the operating condition identifier based on the mapping relationship to form a table containing "operating condition identifier", "vibration feature value", "current feature value", and "fault mode". It is used for subsequent fault type matching. The structured feature mapping table is a set of tabular data containing operating condition identifier, vibration feature value, current feature value and corresponding fault mode. It is obtained from the simplified feature dataset through the above normalization, clustering and association steps. It is used for subsequent fault type feature matching and can provide a basis for rapid query and diagnosis.

[0081] In step S15 , it is necessary to perform fault type feature matching based on the structured feature mapping table and the real-time multi-source signal data to obtain a fault type identifier.

[0082] In one implementation, performing fault type feature matching on the structured feature mapping table and the real-time multi-source signal data to obtain a fault type identifier includes:

[0083] Normalization processing is performed on the real-time multi-source signal data to obtain normalized real-time signal data; the Euclidean distance between the normalized real-time signal data and the characteristic vector of the fault mode in the structured feature mapping table is calculated to obtain a Euclidean distance set; when the Euclidean distance in the Euclidean distance set is less than a preset similarity threshold, the data corresponding to the Euclidean distance is labeled with a corresponding type of fault; when the Euclidean distance in the Euclidean distance set is greater than the preset similarity threshold, it is determined that the data corresponding to the Euclidean distance does not have a corresponding type of fault, and a fault type identifier is obtained.

[0084] It should be noted that the real-time multi-source signal data is first normalized using the z-score normalization method. Specifically, by calculating the deviation of each data point from the mean and dividing it by the standard deviation, the data is converted to a standard distribution with a mean of zero and a standard deviation of one, generating normalized real-time signal data. Next, the Euclidean distance between the normalized real-time signal data and the fault mode feature vectors in the structured feature map is calculated. The Euclidean distance is calculated by taking the square root of the sum of the squared differences between the corresponding components of the two feature vectors. The real-time data is then compared with the feature vectors of each fault mode in the map one by one to generate a set of Euclidean distances. Then, each distance value in the Euclidean distance set is checked. If a distance value is less than a preset similarity threshold, the real-time data is considered to be highly matched with the corresponding fault mode and is marked as the fault type. If the distance value is greater than the preset threshold, the real-time data is considered to not match any known fault mode and is judged to have no corresponding type of fault. Finally, a fault type identifier is generated. The preset similarity threshold is determined by analyzing the accuracy of the matching between historical data and fault modes. Multiple experiments are conducted to statistically analyze the accuracy at different distances, and 0.2 is determined as the threshold that balances the false positive rate and the missed positive rate. This is applicable to different operating conditions and ensures the reliability and consistency of identification. The fault type identifier is a fault category label that represents the current operating status of the motor, such as bearing wear or circuit fault. It is obtained from real-time multi-source signal data through the above-mentioned normalization, distance calculation, and threshold comparison steps. It is used for subsequent fault severity assessment and generation of fault mode recognition reports, providing a basis for maintenance decisions.

[0085] In step S16, it is necessary to use a support vector machine algorithm to evaluate the severity of the fault based on the fault type identification and the real-time multi-source signal data, combined with the pre-acquired historical feature distribution data, and generate a fault pattern recognition report containing the fault type, severity and working condition description.

[0086] In one implementation, based on the fault type identifier and the real-time multi-source signal data, combined with pre-acquired historical feature distribution data, a support vector machine algorithm is used to assess the fault severity and generate a fault pattern recognition report containing the fault type, severity, and operating condition description, including:

[0087] Based on the fault type identification, combined with pre-acquired historical feature distribution data, feature distribution statistical calculations are performed to obtain a fault feature distribution set; based on the fault feature distribution set, a support vector machine algorithm is used to perform classification calculations, and different fault severity categories are distinguished by constructing a classification hyperplane to obtain a fault severity value; based on the fault severity value and the real-time multi-source signal data, data integration is performed to generate a fault pattern recognition report that includes a description of the fault type, severity, and operating condition.

[0088] It should be noted that, first, based on the fault type identifier and the real-time multi-source signal data, historical feature distribution data related to the corresponding fault type and operating parameters are extracted from a pre-established historical database. Statistical analysis methods are used to calculate parameters such as the mean and standard deviation of these historical data. Statistical calculations are then performed in conjunction with real-time data to generate a fault feature distribution set reflecting the fault feature distribution. Specifically, parameters corresponding to the fault type identifier, such as vibration frequency and current intensity, are first extracted from the real-time multi-source signal data. Sample data matching the current operating parameters is then obtained from the historical database. The sample data includes feature records from the past under the same fault type and operating conditions. Next, a statistical analysis method is used to merge the real-time data with the historical data, and the joint mean of the two is calculated. The specific method is to find the average of all data points, calculate the standard deviation, and then calculate the sum of the squares of the deviations of each data point from the mean and divide it by the total number of data points minus one to obtain an indicator reflecting the degree of data dispersion. Furthermore, the distribution differences between real-time data and historical data are analyzed, and the statistical results are adjusted to reflect the dynamic changes in the current fault characteristics. Ultimately, a fault feature distribution set is generated. This set includes updated statistical parameters such as the mean and standard deviation, comprehensively describing the characteristic distribution of specific fault types under the current operating conditions, and providing accurate data support for the subsequent severity classification of the support vector machine algorithm. Next, a support vector machine algorithm is used to perform classification calculations on the fault feature distribution set. The support vector machine optimizes the calculation to find the optimal classification hyperplane, classifying the data into different severity categories such as normal, mild fault, and moderate fault. The fault severity value of the real-time data is determined based on the distance of the hyperplane. Specifically, the support vector machine algorithm is first optimized to determine the optimal classification hyperplane, which divides the data points in the fault feature distribution set into different severity categories. The hyperplane is trained based on historical data and defines multiple threshold intervals. The specific judgment process is to compare the feature vector of the real-time data, such as the normalized value of the vibration frequency or current intensity, with the corresponding distance of the hyperplane. If the index is less than the first threshold, for example, the distance from the hyperplane is less than 0.5 normalized units, it is judged as normal, indicating that the motor operating parameters are within the safe range. If the indicator is greater than the first threshold but less than the second threshold, for example, the distance to the hyperplane is between 0.5 and 1.0, it is determined to be a mild fault, reflecting that the parameter has begun to deviate from the normal range but has not yet reached a serious level. If the indicator is greater than the second threshold but less than the third threshold, for example, the distance to the hyperplane is between 1.0 and 1.5, it is determined to be a moderate fault, indicating that the parameter anomaly has intensified and there is a moderate risk. If the indicator is greater than the third threshold, for example, the distance to the hyperplane exceeds 1.5, it is determined to be a severe fault, reflecting that the parameter has seriously deviated from the normal range and requires immediate attention. Among them, the first threshold is less than the second threshold, and the second threshold is less than the third threshold, which are 0.5, 1.0 and 1.5 respectively. These thresholds are set by historical data distribution and expert experience. Through the step-by-step comparison of the hyperplane distance, the health status of real-time data can be clearly distinguished.Next, a structured table is created based on the fault severity value and real-time multi-source signal data. The table contains fields such as "fault type," "severity," "operating condition description," and "key parameters." The specific fault category, such as "bearing wear," is extracted from the fault type identifier, and the severity value is filled in the corresponding column, such as "minor fault." A description is generated based on the operating condition parameters of the real-time multi-source signal data, such as "heavy load-medium speed-2025-06-0719:56." Representative parameters, such as vibration frequency and current intensity, are selected from the signal data and filled in the key parameter column, such as "vibration frequency 50 Hz, current intensity 10 A." A programming script is then used to sequentially integrate these scattered data fields into the table, ensuring that each row of data corresponds to a complete report entry. Finally, the integrated table is converted into a readable report format, such as PDF or database record, and stored in the monitoring system database to generate a fault pattern recognition report. This process ensures that the report content is clear and easy for maintenance personnel to review and analyze the fault status by filling and formatting the fields in the structured table, guiding subsequent maintenance decisions. The fault pattern identification report is a document containing the fault type, severity level and operating condition description. It is obtained from the fault type identification and real-time data through the above statistical calculation, classification and data integration steps. It is used to store in the monitoring system database and provide maintenance personnel with fault analysis and decision support.

[0089] In summary, the present invention realizes a systematic and comprehensive analysis of multi-dimensional motor operation data through comprehensive acquisition of multi-source signals, optimization of data preprocessing, simplification of dimensionality reduction processing, feature library constructed by clustering, precise identification of feature matching, and severity assessment of support vector machine.

[0090] Reference Figure 2 The second embodiment of the present invention provides an explosion-proof motor fault diagnosis device based on big data, comprising:

[0091] The data acquisition module is used to obtain real-time multi-source signal data during motor operation and generate an initial data set containing a timestamp and operating condition identification;

[0092] A standard processing module, configured to perform data preprocessing based on the initial data set to obtain a standardized data set;

[0093] A dimensionality reduction and simplification module is used to perform dimensionality reduction processing based on the standardized data set using a principal component analysis method to generate a simplified feature data set;

[0094] A feature mapping module is used to construct a fault feature library based on the simplified feature data set and generate a structured feature mapping table;

[0095] A fault identification module is used to match fault type features with the real-time multi-source signal data according to the structured feature mapping table to obtain a fault type identification;

[0096] The result output module is used to evaluate the severity of the fault using a support vector machine algorithm based on the fault type identification and the real-time multi-source signal data, combined with the pre-acquired historical feature distribution data, and generate a fault pattern recognition report including the fault type, severity and working condition description.

[0097] It should be noted that the explosion-proof motor fault diagnosis device based on big data provided in an embodiment of the present invention is used to execute all the process steps of the explosion-proof motor fault diagnosis method based on big data in the above embodiment. The working principles and beneficial effects of the two correspond one to one, so they will not be repeated here.

[0098] An embodiment of the present invention further provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as an explosion-proof motor fault diagnosis program based on big data. When the processor executes the computer program, the steps in each of the above-mentioned explosion-proof motor fault diagnosis method embodiments based on big data are implemented, such as Figure 1 Alternatively, when the processor executes the computer program, the functions of the modules / units in the above-mentioned device embodiments are realized, such as the fault identification module.

[0099] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device.

[0100] The electronic device may be a computing device such as a desktop computer, notebook, PDA, or smart tablet. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will appreciate that the aforementioned components are merely examples of electronic devices and do not constitute a limitation of the electronic device. The electronic device may include more or fewer components than those described above, or a combination of certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, and the like.

[0101] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the electronic device, connecting various parts of the entire electronic device using various interfaces and lines.

[0102] The memory can be used to store the computer programs and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created based on the use of the mobile phone (such as audio data, a phone book, etc.). In addition, the memory can include a high-speed random access memory and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0103] Wherein, if the module / unit integrated in the electronic device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the process in the above-mentioned embodiment method, and can also be completed by a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, it can implement the steps of each of the above-mentioned method embodiments. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0104] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement the present invention without inventive effort.

[0105] The specific embodiments described above further illustrate the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A method for diagnosing explosion-proof motor faults based on big data, characterized in that: include: Acquire real-time multi-source signal data during motor operation and generate an initial data set containing timestamps and operating condition identifiers; Performing data preprocessing on the initial data set to obtain a standardized data set; Based on the standardized data set, a principal component analysis method is used to perform dimensionality reduction processing to generate a simplified feature data set; Building a fault feature library based on the simplified feature data set and generating a structured feature mapping table; Perform fault type feature matching based on the structured feature mapping table and the real-time multi-source signal data to obtain a fault type identifier; According to the fault type identification and the real-time multi-source signal data, combined with the pre-acquired historical feature distribution data, a support vector machine algorithm is used to evaluate the fault severity and generate a fault pattern recognition report including the fault type, severity and working condition description.

2. The explosion-proof motor fault diagnosis method based on big data according to claim 1 is characterized in that: The method of acquiring real-time multi-source signal data during motor operation and generating an initial data set including a timestamp and an operating condition identifier includes: Acquire real-time multi-source signal data during motor operation; The real-time multi-source signal data includes the speed parameters, vibration frequency, amplitude, current intensity, fluctuation frequency, surface temperature, temperature rise rate and load status of the motor during operation; According to the real-time multi-source signal data, identification division is performed to obtain an initial data set including a timestamp and an operating condition identification.

3. The explosion-proof motor fault diagnosis method based on big data according to claim 1 is characterized in that: The step of performing data preprocessing on the initial data set to obtain a standardized data set includes: According to the initial data set, a wavelet transform filtering algorithm is used to perform denoising on the data to obtain a denoised data set; Based on the denoised data set, the threshold ranges of vibration frequency and current intensity are calculated using the interquartile range statistical method; Checking whether the data in the denoised data set is within the threshold range, and when the data exceeds the threshold range, removing the corresponding data; when the data is within the threshold range, retaining the corresponding data, and obtaining a data set without outliers; Based on the data without outliers, a k-means clustering algorithm is used to perform classification to generate a segmented data set with working condition identification; According to the segmented data set, a z-score normalization method is used to perform normalization processing to obtain a standardized data set.

4. The explosion-proof motor fault diagnosis method based on big data according to claim 1 is characterized in that: The method of performing dimensionality reduction processing based on the standardized data set using a principal component analysis method to generate a simplified feature data set includes: Analyzing the statistical relationship between the load state and the speed parameter through covariance calculation based on the standardized data set to generate a covariance matrix; Performing eigendecomposition according to the covariance matrix, calculating eigenvalues ​​and eigenvectors, and obtaining an eigenvalue set and an eigenvector set; According to the eigenvalue set and the eigenvector set, the cumulative contribution rate is calculated. When the cumulative contribution rate is greater than the preset contribution rate threshold, the corresponding eigenvector is determined to be the principal component and retained. When the cumulative contribution rate is less than the preset contribution rate threshold, the corresponding eigenvector is eliminated to generate a streamlined feature data set.

5. The explosion-proof motor fault diagnosis method based on big data according to claim 1 is characterized in that: The process of constructing a fault feature library and generating a structured feature mapping table based on the simplified feature data set includes: Performing normalization processing on the simplified feature data set to obtain a normalized feature data set; According to the normalized feature data set, the vibration frequency and current intensity are classified using a k-means clustering algorithm to generate a fault feature data set; The fault feature data set is associated with a preset fault mode to obtain a structured feature mapping table.

6. The explosion-proof motor fault diagnosis method based on big data according to claim 1 is characterized in that: The performing fault type feature matching based on the structured feature mapping table and the real-time multi-source signal data to obtain a fault type identifier includes: Performing normalization processing on the real-time multi-source signal data to obtain normalized real-time signal data; Calculating the Euclidean distance between the normalized real-time signal data and the feature vector of the fault mode in the structured feature mapping table to obtain a Euclidean distance set; When the Euclidean distance in the Euclidean distance set is less than a preset similarity threshold, the data corresponding to the Euclidean distance is marked with a corresponding type of fault. When the Euclidean distance in the Euclidean distance set is greater than the preset similarity threshold, it is determined that the data corresponding to the Euclidean distance does not have a corresponding type of fault, and a fault type identifier is obtained.

7. The explosion-proof motor fault diagnosis method based on big data according to claim 1 is characterized in that: The fault type identification and the real-time multi-source signal data are combined with pre-acquired historical feature distribution data to evaluate the fault severity using a support vector machine algorithm, and generate a fault pattern recognition report containing the fault type, severity, and operating condition description, including: According to the fault type identification, combined with the previously acquired historical feature distribution data, feature distribution statistics are calculated to obtain a fault feature distribution set; Based on the fault feature distribution set, a support vector machine algorithm is used to perform classification calculation, and a classification hyperplane is constructed to distinguish different fault severity categories to obtain a fault severity value; Data integration is performed based on the fault severity value and the real-time multi-source signal data to generate a fault mode recognition report including the fault type, severity and working condition description.

8. An explosion-proof motor fault diagnosis device / system based on big data, characterized in that: include: The data acquisition module is used to obtain real-time multi-source signal data during motor operation and generate an initial data set containing a timestamp and operating condition identification; A standard processing module, configured to perform data preprocessing based on the initial data set to obtain a standardized data set; A dimensionality reduction and simplification module is used to perform dimensionality reduction processing based on the standardized data set using a principal component analysis method to generate a simplified feature data set; A feature mapping module is used to construct a fault feature library based on the simplified feature data set and generate a structured feature mapping table; A fault identification module is used to match fault type features with the real-time multi-source signal data according to the structured feature mapping table to obtain a fault type identification; The result output module is used to evaluate the severity of the fault using a support vector machine algorithm based on the fault type identification and the real-time multi-source signal data, combined with the pre-acquired historical feature distribution data, and generate a fault pattern recognition report including the fault type, severity and working condition description.

9. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for diagnosing explosion-proof motor faults based on big data as claimed in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the explosion-proof motor fault diagnosis method based on big data according to any one of claims 1 to 7.

Images