Device fault diagnosis method, system and device based on data feature analysis
By using an edge-cloud collaborative equipment fault diagnosis method, vibration signal features are extracted in real time and combined with cloud analysis, solving the problems of low efficiency and high bandwidth pressure in existing technologies, and achieving efficient and accurate fault diagnosis.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA RESOURCES POWER TECH RES INST CO LTD
- Filing Date
- 2026-02-25
- Publication Date
- 2026-06-26
AI Technical Summary
Existing equipment fault diagnosis methods suffer from low efficiency and poor automation in vibration signal analysis, and data transmission puts a heavy burden on network bandwidth, making it difficult to meet timeliness requirements.
By adopting an edge-cloud collaborative approach, fault diagnosis is performed through real-time feature extraction and preliminary diagnosis at the edge, combined with dynamic time planning and machine learning classification methods in the cloud.
It significantly improves the real-time performance and accuracy of fault diagnosis, reduces the data transmission burden, and enables precise and real-time equipment status monitoring and fault diagnosis.
Smart Images
Figure CN122286243A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of equipment monitoring technology, and in particular to a method, system and equipment for equipment fault diagnosis based on data feature analysis. Background Technology
[0002] Whether in traditional thermal power generation or new energy wind power generation, malfunctions of mechanical equipment such as wind turbines, bearings, and gearboxes are common. Vibration signal analysis always plays a crucial role in the fault diagnosis process, whether it is the appearance of fault symptoms or after a fault has already occurred.
[0003] Currently, equipment fault diagnosis based on vibration signals mainly employs two typical implementation schemes: one is on-site data collection and manual analysis, relying on professionals to collect data using specialized instruments and make judgments based on experience. This approach is inefficient, lacks automation, and lacks correlation analysis with historical and similar equipment data, resulting in limited diagnostic accuracy. The other approach involves uploading all data to the cloud for centralized processing. While this leverages the powerful computing capabilities of the cloud, the large volume of raw vibration data, high sampling frequency, and continuous transmission place enormous pressure on network bandwidth. Furthermore, the system response latency is high, making it difficult to meet timeliness requirements. Therefore, how to reduce data transmission load and improve real-time performance while ensuring diagnostic accuracy has become a pressing technical problem in the field of fault diagnosis. Summary of the Invention
[0004] The technical problem to be solved by the present invention is to provide a method, system and equipment for equipment fault diagnosis based on data feature analysis, so as to solve the above-mentioned technical problem.
[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: a device fault diagnosis method based on data feature analysis, comprising: acquiring vibration signals of the device to be diagnosed; extracting features from the vibration signals to obtain feature data, wherein the feature data includes statistical distribution features, time-series trend features, and machine learning pattern features; comparing the statistical distribution features with preset benchmark distribution features to generate preliminary anomaly judgment information based on the obtained comparison results; when the preliminary anomaly judgment information indicates an anomaly, uploading the feature data and the vibration signals to the cloud, and using dynamic time planning and machine learning classification methods on the cloud to match and analyze the feature data with a pre-constructed historical fault feature sample library to determine the fault diagnosis result.
[0006] The beneficial effects of this invention are as follows: This method, through edge-cloud collaborative processing, decentralizes the real-time feature extraction and preliminary diagnosis of vibration signals to the edge, effectively alleviating the pressure on transmission bandwidth from massive amounts of raw data and significantly improving response speed and real-time performance. Simultaneously, the cloud, based on a historical fault sample database, utilizes dynamic time programming and machine learning classification methods for comprehensive matching analysis, overcoming the limitations of traditional methods that rely on human experience and lack historical and similar data for reference. This significantly improves the accuracy and automation level of fault diagnosis, achieving precise, real-time, and intelligent equipment status monitoring and fault diagnosis under limited bandwidth conditions.
[0007] Based on the above technical solution, the present invention can be further improved as follows.
[0008] Furthermore, the step of comparing the statistical distribution features with a preset benchmark distribution feature to generate preliminary anomaly judgment information based on the comparison result includes: calculating the deviation between the statistical distribution features and the preset benchmark distribution features to obtain a distribution deviation value; and generating preliminary anomaly judgment information indicating an anomaly when the distribution deviation value is greater than a preset deviation threshold.
[0009] Furthermore, the process of matching and analyzing the feature data with a pre-built historical fault feature sample library using dynamic time planning and machine learning classification methods based on the cloud to determine the fault diagnosis result includes: calculating the dynamic time warping distance between the feature data and each historical fault case feature in the pre-built historical fault feature sample library; determining the fault type corresponding to the historical fault case feature with the smallest dynamic time warping distance with the feature data as the first candidate fault type; judging the fault based on the feature data using a pre-trained machine learning classification model to obtain a second candidate fault type; comparing the scores of the first candidate fault type and the second candidate fault type, and generating a fault diagnosis result based on the fault type with the higher score.
[0010] Furthermore, the historical fault feature sample library is obtained through the following methods: acquiring historical fault vibration data, wherein the historical fault vibration data represents the historical fault data of multiple devices associated with the device to be diagnosed; performing preprocessing and feature extraction on the historical fault vibration data in sequence to obtain historical fault feature data; wherein, the feature extraction includes statistical distribution feature extraction, time series trend feature extraction and machine learning pattern feature extraction; constructing a historical fault feature sample library based on the historical fault feature data, and storing the historical fault feature sample library in the cloud.
[0011] Furthermore, the step of extracting features from the vibration signal to obtain feature data includes: extracting features from the vibration signal using statistical analysis, model analysis, and machine learning analysis methods, respectively, to obtain the feature data.
[0012] Furthermore, before feature extraction from the vibration signal, the method further includes: performing data cleaning and filtering processing on the vibration signal.
[0013] Furthermore, after determining the fault diagnosis result, the method also includes: generating fault repair guidance information corresponding to the fault diagnosis result based on a preset case knowledge base.
[0014] To address the aforementioned technical problems, the present invention also provides a device fault diagnosis system based on data feature analysis, comprising: The signal acquisition module is used to acquire the vibration signal of the device under diagnosis. The feature extraction module is used to extract features from the vibration signal to obtain feature data, which includes statistical distribution features, time-series trend features, and machine learning pattern features. The preliminary anomaly identification module is used to compare the statistical distribution features with the preset benchmark distribution features, so as to generate preliminary anomaly judgment information based on the comparison results. The fault diagnosis module is used to upload the feature data and the vibration signal to the cloud when the preliminary anomaly judgment information indicates an anomaly. Based on the cloud, the feature data is matched and analyzed with a pre-built historical fault feature sample library using dynamic time planning and machine learning classification methods to determine the fault diagnosis result.
[0015] To address the aforementioned technical problems, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the device fault diagnosis method based on data feature analysis as described above.
[0016] To address the aforementioned technical problems, the present invention also provides a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the device fault diagnosis method based on data feature analysis as described above. Attached Figure Description
[0017] Figure 1 This is a flowchart of the equipment fault diagnosis method based on data feature analysis according to the present invention; Figure 2 This is a schematic diagram of the equipment fault diagnosis system based on data feature analysis according to the present invention; Figure 3This is a schematic diagram of the electronic device of the present invention. Detailed Implementation
[0018] The principles and features of the present invention are described below. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.
[0019] Currently, online vibration monitoring mainly relies on dedicated sensors and analysis software to collect and display data within a localized area, followed by manual judgment by professionals based on the data curves. For suspected faulty equipment, the common practice is to collect high-frequency vibration data on-site and connect it to professional analysis instruments. This method has significant limitations: first, it is inflexible, requiring point-by-point analysis on-site, making it difficult to achieve comprehensive comparison of multiple data points; second, it relies solely on real-time data, lacking correlation analysis with the equipment's historical status and similar fault cases; third, the diagnostic process is overly dependent on personnel experience and cannot automatically retrieve historical data for accurate judgment; and fourth, the lack of maintenance case studies and knowledge bases can easily lead to diagnostic omissions or wasted resources.
[0020] Example 1 Based on this, such as Figure 1 As shown, this embodiment provides a device fault diagnosis method based on data feature analysis, including: S101. Obtain the vibration signal of the device to be diagnosed.
[0021] At the edge, high-frequency vibration signals at the target location (bearing, rotor, etc.) of the equipment are collected in real time by sensors.
[0022] S102. Extract features from the vibration signal to obtain feature data, which includes statistical distribution features, time-series trend features, and machine learning pattern features.
[0023] S103. Compare the statistical distribution characteristics with the preset benchmark distribution characteristics to generate preliminary anomaly judgment information based on the comparison results.
[0024] S104. When the preliminary anomaly judgment information indicates an anomaly, the feature data and vibration signal are uploaded to the cloud. Based on the cloud, the feature data is matched and analyzed with a pre-built historical fault feature sample library using dynamic time planning and machine learning classification methods to determine the fault diagnosis result.
[0025] High-performance data processing algorithms are deployed at the edge to collect vibration signals in real time. Based on the obtained vibration signals, characteristic indicators are calculated, and preliminary feature comparisons and anomaly judgments are performed. Simultaneously, the raw data is downsampled and compressed before the characteristic data and processed signals are uploaded to the cloud. Upon receiving the data, the cloud uses more precise analysis algorithms to match and verify it against a built-in fault feature library, thereby achieving accurate fault diagnosis and classification. This method, by decentralizing real-time feature extraction to the edge and centralizing complex analysis in the cloud, reduces the transmission bandwidth burden while improving the real-time performance and accuracy of diagnosis, effectively realizing collaborative operation between the edge and cloud.
[0026] Optionally, in an embodiment, before feature extraction from the vibration signal, the method further includes: performing data cleaning and filtering processing on the vibration signal.
[0027] Specifically, the vibration signal is cleaned and filtered in real time to remove abnormal data such as those that are obviously out of range or dead values.
[0028] Optionally, in the embodiments, feature extraction is performed on the vibration signal to obtain feature data, including: extracting features from the vibration signal using statistical analysis, model analysis, and machine learning analysis methods respectively to obtain feature data.
[0029] Specifically, statistical analysis methods (empirical probability distribution estimation and theoretical probability distribution fitting methods) are used to obtain the probability distribution characteristics of vibration signals, such as normal distribution, Weibull distribution, or sharp distribution, to achieve quantitative characterization of fault features. The consistency between the empirical and theoretical distributions is verified through goodness-of-fit tests.
[0030] Using common model analysis methods (ARIMA or SARIMA models), we obtain data trends and periodic characteristics, including the variation cycle of stationary data or the fitting residual data of non-stationary data.
[0031] Machine learning analysis methods (XGBoost, LightGBM, CatBoost) are used to obtain data patterns and features, including residual extraction of model predictions and periodic features, and important periodic features (hourly, seasonal) are selected.
[0032] Optionally, in an embodiment, the historical fault feature sample library is obtained by: acquiring historical fault vibration data, which represents the historical fault data of multiple devices associated with the device to be diagnosed; performing preprocessing and feature extraction on the historical fault vibration data in sequence to obtain historical fault feature data; wherein, feature extraction includes statistical distribution feature extraction, time-series trend feature extraction, and machine learning pattern feature extraction; constructing a historical fault feature sample library based on the historical fault feature data, and storing the historical fault feature sample library in the cloud.
[0033] By collecting data such as equipment maintenance records, historical fault correlation data from the monitoring system, and equipment fault vibration signal files, historical fault vibration data of the equipment is obtained. After obtaining the above data, the reasonableness of the data is judged according to the dimensions of manufacturer, model, equipment, component, fault code, fault time, and data duration. Duplicate and invalid data are removed to obtain preprocessed data.
[0034] Based on the preprocessed data, feature extraction is performed using statistical analysis, model analysis, and machine learning analysis to obtain historical fault feature data. A historical fault feature sample library is then constructed based on this data. This library contains various fault types and their corresponding data features (statistical distribution features, time-series trend features, and machine learning pattern features), and is stored in the cloud.
[0035] Optionally, in an embodiment, the statistical distribution features are compared with preset benchmark distribution features to generate preliminary anomaly judgment information based on the comparison results, including: calculating the deviation between the statistical distribution features and the preset benchmark distribution features to obtain a distribution deviation value; when the distribution deviation value is greater than a preset deviation threshold, generating preliminary anomaly judgment information indicating an anomaly.
[0036] Specifically, normal vibration signals are collected over a period of time, and their amplitude probability distribution is fitted (e.g., using Weibull distribution or kernel density estimation) to obtain distribution parameters (e.g., shape parameters, scale parameters), which serve as the baseline distribution features.
[0037] Calculate the difference between the statistical distribution characteristics (empirical distribution of signal amplitude) of the vibration signal within the current window and the baseline distribution characteristics. The deviation value can be calculated using Euclidean distance. Set a deviation threshold; when the distribution deviation value exceeds the preset threshold, generate preliminary anomaly assessment information indicating an anomaly.
[0038] After generating preliminary anomaly assessment information indicating abnormality, the feature data and vibration signals are compressed and uploaded to the cloud.
[0039] Optionally, in this embodiment, based on the cloud, the feature data is matched and analyzed with a pre-built historical fault feature sample library using dynamic time planning and machine learning classification methods to determine the fault diagnosis result. This includes: calculating the dynamic time warping distance between the feature data and each historical fault case feature in the pre-built historical fault feature sample library; determining the fault type corresponding to the historical fault case feature with the smallest dynamic time warping distance with the feature data as the first candidate fault type; judging the fault based on the feature data using a pre-trained machine learning classification model to obtain the second candidate fault type; comparing the scores of the first candidate fault type and the second candidate fault type, and generating the fault diagnosis result based on the fault type with the higher score.
[0040] Dynamic time programming is a classic algorithm used to measure the similarity between two time series of different lengths or with local deformation.
[0041] Specifically, the feature data (such as a sequence of root mean square values within a fault event window or a sequence of amplitude values at a specific fault frequency) is converted into a one-dimensional time-series vector: X=(x1,x2,…,x m ).
[0042] Retrieve the feature sequence Y=(y1,y2,…,y) of a historical fault case to be compared from the historical fault feature sample database. n ).
[0043] To calculate the local distance between each pair of points in sequences X and Y, Euclidean distance is typically used. d(i,j)=(x i -y j ) 2 ; This forms an m×n distance matrix D.
[0044] Define the cumulative distance matrix C, where C(i,j) represents the minimum cumulative distance from the starting point (1,1) to the point (i,j).
[0045] The calculation is performed using a recursive formula; the classic recursive formula is as follows: ; The formula allows for three operations: matching, stretching X, or stretching Y.
[0046] Backtrack from the endpoint (m,n) to the starting point (1,1), find the path P that minimizes the cumulative distance.
[0047] The total cumulative distance DTW(X,Y)=C(m,n) along this path is the minimum cumulative distance between the two sequences after optimal elastic alignment. The smaller this value, the more similar the two sequences are in shape.
[0048] Calculate the DTW distance between the test sequence X and all fault case sequences in the historical sample database. The fault type corresponding to the historical fault case feature with the smallest dynamic time warped distance to the feature data is identified as the first candidate fault type.
[0049] XGBoost was chosen as the classifier. The XGBoost classifier was trained using historical fault feature data from a historical fault feature sample library. The training process iteratively constructs multiple decision trees, with each tree learning and correcting the residuals of the previous tree. Key hyperparameters include: maximum tree depth, learning rate, number of trees, and regularization parameter, which can be tuned through cross-validation. The training objective is to maximize classification accuracy and minimize the multi-class log loss function.
[0050] The extracted feature data is converted into feature vectors, which are then input into a machine learning classification model deployed in the cloud. The model performs voting or weighted calculations using multiple internal trees, outputting a probability distribution vector, and selecting the category with the highest probability as the second candidate fault type.
[0051] For score comparison, a comprehensive scoring function can be designed to quantitatively evaluate the first and second candidate fault types. For example, the minimum dynamic time warp distance and the probability (or confidence level) corresponding to the second candidate fault type can be normalized to obtain their respective scores. Alternatively, a weighted fusion calculation can be performed on the dynamic time warp distance and probability (or confidence level) corresponding to the first candidate fault type to obtain the corresponding score; and the same weighted fusion calculation can be performed on the dynamic time warp distance and probability (or confidence level) corresponding to the second candidate fault type to obtain the corresponding score.
[0052] The scores corresponding to the first candidate fault type are compared with those corresponding to the second candidate fault type. The candidate fault type with the higher score is selected as the final fault type, and the fault diagnosis result is obtained.
[0053] If the above methods fail to match the fault type, the fault currently occurring on the device is considered a newly emerging fault. This diagnosis will be recorded as a new fault case for future use.
[0054] Optionally, in this embodiment, after determining the fault diagnosis result, the method further includes: generating fault repair guidance information corresponding to the fault diagnosis result based on a preset case knowledge base.
[0055] The pre-set case knowledge base stores relevant troubleshooting solutions and knowledge for each type of fault. By matching the fault diagnosis results with the case knowledge base, corresponding fault repair cases and knowledge are provided to assist in fault diagnosis and repair.
[0056] Compared to existing equipment fault diagnosis methods, this method achieves multiple optimizations through edge-cloud collaboration: First, by delegating the extraction of key vibration signal features to the edge of the equipment, the timeliness of fault identification and response is significantly improved. Second, by fully utilizing the powerful computing resources and unified operation and maintenance capabilities of the cloud, the stable operation of complex diagnostic methods is ensured, greatly expanding the scope of application. Furthermore, by integrating multi-dimensional feature extraction methods and combining dynamic time planning and machine learning classification strategies, the accuracy of diagnostic results is further improved. Simultaneously, by deeply linking real-time diagnosis with historical fault case libraries and maintenance guidance knowledge, the closed-loop efficiency and operational convenience from diagnosis to repair are effectively improved. Finally, the divide-and-conquer design of this edge-cloud collaboration aligns with the hierarchical management concept of IoT systems, reducing overall complexity while achieving functional decoupling.
[0057] Example 2 like Figure 2 As shown, this embodiment provides a device fault diagnosis system 200 based on data feature analysis, including: Signal acquisition module 201 is used to acquire vibration signals of the device under diagnosis; Feature extraction module 202 is used to extract features from vibration signals to obtain feature data, which includes statistical distribution features, time-series trend features and machine learning pattern features. The preliminary anomaly identification module 203 is used to compare statistical distribution characteristics with preset benchmark distribution characteristics to generate preliminary anomaly judgment information based on the comparison results. The fault diagnosis module 204 is used to upload feature data and vibration signals to the cloud when the preliminary abnormality judgment information indicates an abnormality. Based on the cloud, the feature data is matched and analyzed with a pre-built historical fault feature sample library through dynamic time planning and machine learning classification methods to determine the fault diagnosis result.
[0058] Optionally, in this embodiment, the preliminary anomaly identification module 203 includes: The deviation calculation unit is used to calculate the deviation between the statistical distribution characteristics and the preset benchmark distribution characteristics to obtain the distribution deviation value; The comparison unit is used to generate preliminary anomaly judgment information indicating an anomaly when the distribution deviation value is greater than a preset deviation threshold.
[0059] Optionally, in an embodiment, the fault diagnosis module 204 includes: The distance calculation unit is used to calculate the dynamic time-normalized distance between the feature data and each historical fault case feature in the pre-built historical fault feature sample library; The first fault determination unit is used to determine the fault type corresponding to the historical fault case feature with the smallest dynamic time warp distance from the feature data as the first candidate fault type. The second fault determination unit is used to determine the fault based on the feature data and through a pre-trained machine learning classification model to obtain the second candidate fault type. The fault diagnosis unit is used to score and compare the first candidate fault type and the second candidate fault type, and generate a fault diagnosis result based on the fault type with the higher score.
[0060] Optionally, in an embodiment, the historical fault feature sample library is obtained by: acquiring historical fault vibration data, which represents the historical fault data of multiple devices associated with the device to be diagnosed; performing preprocessing and feature extraction on the historical fault vibration data in sequence to obtain historical fault feature data; wherein, feature extraction includes statistical distribution feature extraction, time-series trend feature extraction, and machine learning pattern feature extraction; constructing a historical fault feature sample library based on the historical fault feature data, and storing the historical fault feature sample library in the cloud.
[0061] Optionally, in an embodiment, the feature extraction module 202 includes: The feature extraction unit is used to extract features from the vibration signal using statistical analysis, model analysis, and machine learning analysis methods, respectively, to obtain feature data.
[0062] Optionally, in an embodiment, before the feature extraction module 202, the following further step is taken: The preprocessing module is used to clean and filter the vibration signals.
[0063] Optionally, in an embodiment, after the fault diagnosis module 204, the following is also included: The guidance information generation module is used to generate fault repair guidance information corresponding to the fault diagnosis results based on a preset case knowledge base.
[0064] In some embodiments, the equipment fault diagnosis system 200 based on data feature analysis of the present invention can be implemented in a combination of hardware and software. As an example, the equipment fault diagnosis system 200 based on data feature analysis of the present invention can be a processor in the form of a hardware decoding processor, which is programmed to execute the equipment fault diagnosis method based on data feature analysis of the present invention. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.
[0065] The modules described in the embodiments of this invention can be implemented in software or hardware. The names of the modules are not, in some cases, limiting the scope of the module itself.
[0066] Example 3 like Figure 3 As shown, this embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the device fault diagnosis method based on data feature analysis as described in Embodiment 1.
[0067] In other words, an electronic device according to an embodiment of the present invention may include, but is not limited to: a processor and a memory; the memory is used to store computer programs; the processor is used to execute the device fault diagnosis method based on data feature analysis shown in any embodiment of the present invention by calling the computer program.
[0068] In one alternative embodiment, an electronic device is provided. Figure 3 The illustrated electronic device 300 includes a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302. Optionally, the electronic device 300 may further include a transceiver 304, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 304 is not limited to one type, and the structure of the electronic device 300 does not constitute a limitation on the embodiments of the present invention.
[0069] Processor 301 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. Processor 301 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.
[0070] Bus 302 may include a path for transmitting information between the aforementioned components. Bus 302 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 302 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 3 The bus 302 is represented by only one thick line, but this does not mean that there is only one bus or one type of bus.
[0071] The memory 303 may be a ROM (Read Only Memory) or other type of static storage device capable of storing static information and instructions, RAM (Random Access Memory) or other type of dynamic storage device capable of storing information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.
[0072] The memory 303 is used to store application code (computer program) for executing the present invention, and its execution is controlled by the processor 301. The processor 301 is used to execute the application code stored in the memory 303 to implement the content shown in the foregoing method embodiments.
[0073] Among them, electronic devices can also be terminal devices, which can be any device that can install applications, including at least one of smartphones, tablets, laptops, desktop computers, smart speakers, smartwatches, smart TVs, and smart in-vehicle devices.
[0074] It should be noted that, Figure 3 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.
[0075] Example 4 This embodiment provides a non-transitory computer-readable storage medium that stores computer instructions for causing a computer to execute a device fault diagnosis method based on data feature analysis as described in Embodiment 1.
[0076] Alternatively, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, a floppy disk, and an optical data storage device, etc.
[0077] In an exemplary embodiment, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the aforementioned device fault diagnosis method based on data feature analysis.
[0078] Computer program code for performing the operations of this invention can be written in one or more programming languages or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, and C++, and conventional procedural programming languages such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).
[0079] It should be understood that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0080] The computer-readable storage medium provided in this invention can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EEPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.
[0081] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the method shown in the above embodiments.
[0082] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this invention.
[0083] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and represent a limitation on a specific order or sequence. Where appropriate, the order of use for similar objects can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that shown or described.
[0084] Those skilled in the art will recognize that this invention can be implemented as a system, method, or computer program product. Therefore, this invention can be specifically implemented in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, this invention can also be implemented as a computer program product contained in one or more computer-readable media, which includes computer-readable program code.
[0085] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.
Claims
1. A device failure diagnosis method based on data feature analysis, characterized by, The method comprises the following steps: obtaining a vibration signal of a device to be diagnosed; extracting features from the vibration signal to obtain feature data, the feature data comprising statistical distribution features, time series trend features and machine learning pattern features; comparing the statistical distribution features with preset reference distribution features to generate preliminary abnormality judgment information based on the comparison result obtained; when the preliminary abnormality judgment information indicates an abnormality, uploading the feature data and the vibration signal to a cloud server to match and analyze the feature data with a pre-constructed historical fault feature sample library based on dynamic time warping and machine learning classification methods to determine a fault diagnosis result.
2. The device failure diagnosis method based on data feature analysis according to claim 1, characterized by, The comparison of the statistical distribution features with the preset reference distribution features to generate preliminary abnormality judgment information based on the comparison result obtained comprises: calculating the deviation between the statistical distribution features and the preset reference distribution features to obtain a distribution deviation value; when the distribution deviation value is greater than a preset deviation threshold, generating preliminary abnormality judgment information indicating an abnormality.
3. The device failure diagnosis method based on data feature analysis according to claim 1, characterized by, The matching and analysis of the feature data with the pre-constructed historical fault feature sample library based on the dynamic time warping and machine learning classification methods to determine a fault diagnosis result comprises: calculating the dynamic time warping distance between the feature data and each historical fault case feature in the pre-constructed historical fault feature sample library; determining the fault type corresponding to the historical fault case feature with the minimum dynamic time warping distance from the feature data as a first candidate fault type; performing fault judgment on the feature data through a pre-trained machine learning classification model to obtain a second candidate fault type; scoring and comparing the first candidate fault type and the second candidate fault type to generate a fault diagnosis result based on the fault type with a higher score.
4. The device failure diagnosis method based on data feature analysis according to claim 1, characterized by, The historical fault feature sample library is obtained by the following method: obtaining historical fault vibration data representing historical fault data of a plurality of devices associated with the device to be diagnosed; sequentially pre-processing and extracting features from the historical fault vibration data to obtain historical fault feature data; wherein the feature extraction comprises statistical distribution feature extraction, time series trend feature extraction and machine learning pattern feature extraction; constructing a historical fault feature sample library based on the historical fault feature data and storing the historical fault feature sample library in the cloud server.
5. The method of claim 1, wherein, The feature extraction from the vibration signal to obtain feature data comprises: extracting features from the vibration signal by statistical analysis, model analysis and machine learning analysis respectively to obtain the feature data.
6. The method of claim 1, wherein, Before the feature extraction from the vibration signal, the method further comprises data cleaning and filtering processing on the vibration signal.
7. The method of claim 1, wherein, After the determination of the fault diagnosis result, the method further comprises generating fault repair guidance information corresponding to the fault diagnosis result based on a preset case knowledge base.
8. A device failure diagnosis system based on data feature analysis, characterized by, The method comprises the following steps: a signal acquisition module for obtaining a vibration signal of a device to be diagnosed; The feature extraction module is used to extract features from the vibration signal to obtain feature data, which includes statistical distribution features, time-series trend features, and machine learning pattern features. The preliminary anomaly identification module is used to compare the statistical distribution features with the preset benchmark distribution features, so as to generate preliminary anomaly judgment information based on the comparison results. The fault diagnosis module is used to upload the feature data and the vibration signal to the cloud when the preliminary anomaly judgment information indicates an anomaly. Based on the cloud, the feature data is matched and analyzed with a pre-built historical fault feature sample library using dynamic time planning and machine learning classification methods to determine the fault diagnosis result.
9. An electronic device, comprising: The device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the device fault diagnosis method based on data feature analysis as described in any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the device fault diagnosis method based on data feature analysis as described in any one of claims 1 to 7.