A spindle fault detection system, method, and storage medium

By comparing the vibration characteristic curve of the spindle with historical curves, the problem of static and fixed detection standards in existing technologies is solved, thereby improving the accuracy and efficiency of spindle fault detection. It has self-learning and self-optimization capabilities, enabling early warning and dynamic optimization of the detection process.

CN123065749APending Publication Date: 2026-11-10GUANGZHOU HAOZHI ELECTROMECHANICAL
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611270784.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-20
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

Existing spindle vibration detection solutions suffer from problems such as static and fixed detection standards, inability to adapt to the manufacturing tolerances of batch spindles, easy omission of minor defects, low detection efficiency, and inability to improve judgment accuracy.

Method used

By comparing the vibration characteristic curve of the spindle with a large number of similar historical curves accumulated in mass production, early anomaly warnings can be achieved. By utilizing real-time database updates and multiple decision results, intelligent detection with self-learning and self-optimization can be realized.

Benefits of technology

It improves the accuracy and efficiency of spindle fault detection, can capture subtle vibration characteristics caused by minute manufacturing defects, avoids missed detection, and the detection process is dynamically optimized as data accumulates, continuously improving detection accuracy and speed.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123065749A_ABST
    Figure CN123065749A_ABST
Patent Text Reader

Abstract

This application belongs to the field of spindle fault detection technology, specifically relating to a spindle fault detection system, method, and storage medium. The detection method includes recording the spindle's identification information; measuring the vibration of the spindle and generating vibration data; generating a vibration characteristic curve based on the vibration data; retrieving historical curves from a database based on the identification information; comparing the vibration characteristic curve with each historical curve and obtaining a decision result: if the decision result is that further testing is required, a testing process is executed on the spindle to obtain a fault detection judgment result and store it in the database; if the decision result is that subsequent testing is omitted, the fault detection judgment result of the spindle is directly output and stored in the database. This application achieves early anomaly warning and avoids missed detection by comparing the vibration characteristic curve with a large number of similar historical curves; by storing the fault detection judgment result in the database for real-time updates and obtaining multiple decision results through comparison, it has the advantages of becoming more accurate and faster with each test.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application belongs to the field of spindle fault detection technology, specifically relating to a spindle fault detection system, method and storage medium. Background Technology

[0002] Spindles, especially high-precision, high-stability high-end electric spindles, are core functional components of CNC machine tools, machining centers, and other mechanical equipment. Their performance directly determines the machining accuracy, surface quality, and long-term operational reliability of the equipment. Even minor flaws during spindle manufacturing and assembly can lead to a decrease in accuracy and reliability, placing extremely high demands on machining precision and assembly processes. To ensure high precision and reliability of products leaving the factory, spindles must undergo vibration testing, break-in testing, and cutting machining testing in sequence before leaving the factory.

[0003] Existing spindle vibration detection solutions vary in their specific methods of acquiring and processing vibration signals, but they all share the inherent technical approach of single-testing, static judgment, and isolated data, resulting in the following drawbacks: 1. Static and fixed testing standards, with fixed thresholds that cannot adapt to the manufacturing tolerances of batch spindles, making it easy to miss minor defects; 2. Curve comparison is only used for manual visualization and no reusable quantitative evaluation and record management system has been established; 3. The test data of each spindle are isolated from each other, making it impossible to use batch historical data for horizontal comparison, completely wasting the value of batch statistics; 4. The testing process is fixed and rigid, requiring all spindles to undergo a complete set of tests, resulting in extremely low batch testing efficiency; 5. The judgment rules are based on preset static logic, making it impossible to iterate and optimize after misjudgment, and the recognition accuracy cannot be improved with the increase of test samples.

[0004] In other words, the existing technology is essentially still in the rudimentary stage of "instrument measurement + manual threshold setting", lacking an intelligent closed-loop mechanism for self-learning, self-optimization and self-improvement. This easily leads to problems such as missed detections, low detection efficiency and inability to improve the accuracy of system judgment, which urgently need to be improved. Summary of the Invention

[0005] The purpose of this application is to at least solve one of the technical problems existing in the prior art, and to provide a spindle fault detection system, method and storage medium. By comparing the vibration characteristic curve with a large number of similar historical curves accumulated in mass production, it can capture subtle vibration characteristic changes caused by minor manufacturing defects, realize early anomaly warning and avoid missed detection. By storing the fault detection judgment result in the database for real-time updates and obtaining multiple decision results through comparison, it has the outstanding advantages of becoming more accurate and faster with each test.

[0006] The technical solution adopted by this application to solve its technical problem is: Firstly, a spindle fault detection method includes: Enter the spindle identification information; Vibration measurements were performed on the spindle, and vibration data was generated. A vibration characteristic curve is generated based on the vibration data; Historical curves are retrieved from the database based on the identification information: If no matching historical curve is found, a test procedure is performed on the spindle to obtain the fault detection and judgment result of the spindle and store it in the database; If a matching historical curve is found, the vibration characteristic curve is compared with each of the historical curves, and a decision result is obtained: The decision result is that if further testing is required, a test procedure is performed on the spindle to obtain the fault detection and judgment result of the spindle, and the result is stored in the database; The decision result is to omit subsequent tests, directly output the fault detection judgment result of the spindle, and store it in the database.

[0007] In conjunction with the first aspect, in some implementations of the first aspect, the vibration data includes vibration acceleration, vibration velocity, and vibration displacement data of the spindle at several sets of rotational speeds. When generating vibration characteristic curves based on the vibration data, the vibration data is preprocessed, time-domain features are extracted, and frequency-domain features are extracted to obtain characteristic value groups corresponding to each set of rotational speeds.

[0008] In combination with the first aspect and the above implementation methods, in some implementation methods of the first aspect, the comparison includes calculating the deviation value between the vibration characteristic curve and the historical curve at each set of rotational speeds corresponding to the characteristic value group, obtaining multiple deviation values, summing the deviation values ​​and taking the average value to obtain the difference index, wherein the deviation value = |a - b| / max(a, b) / 100, a represents a certain characteristic value in the characteristic value group corresponding to the vibration characteristic curve at a specific rotational speed, and b represents a characteristic value of the same type as a in the characteristic value group corresponding to the historical curve at the same rotational speed as a; The difference index is compared with a preset similarity threshold, and the vibration characteristic curve is determined to be the same as or similar to the historical curve currently being compared, based on the comparison result, so as to obtain a decision result.

[0009] In combination with the first aspect and the above implementation methods, in some implementation methods of the first aspect, after the comparison is completed, if there is a historical curve that is the same as or similar to the vibration characteristic curve, the confidence level of the fault detection result of the corresponding historical curve is determined. If the confidence level is high, the decision result is to omit subsequent testing; if the confidence level is low, the decision result is to require further testing. If no historical curve is found that is the same as or similar to the vibration characteristic curve, the decision result is determined to require further testing.

[0010] Combining the first aspect and the above implementation methods, in some implementation methods of the first aspect, when the fault detection result of the corresponding historical curve has a unique result branch and the number of result branches reaches a preset number threshold, the confidence level is determined to be high. If the fault detection result of the corresponding historical curve has multiple result branches, and the proportion of each result branch does not reach the preset concentration threshold, the confidence level is determined to be low, and a comprehensive test is performed on the spindle. The fault detection result of the corresponding historical curve has multiple result branches, and when the proportion of the number of a certain result branch reaches the preset concentration threshold, it is determined that the confidence level is not high, and partial testing is performed on the spindle.

[0011] In combination with the first aspect and the above implementation methods, in some implementation methods of the first aspect, the identification information includes the spindle serial number and characteristic information. The characteristic information includes at least one or more of the following: spindle model, maximum speed, bearing type, and rated power. When retrieving the historical curve, the database is searched based on the characteristic information.

[0012] Combining the first aspect and the above implementation methods, in some implementation methods of the first aspect, when the decision result is to omit subsequent tests, if the spindle produces an abnormal fault, the abnormal fault is stored as the result branch in the database, and when the decision result is to omit subsequent tests in the same or similar vibration characteristic curves, the abnormal fault detection is added to the spindle.

[0013] In a second aspect, a storage medium storing a computer program that, when executed by a processor, implements the spindle fault detection method described in any implementation of the first aspect.

[0014] Thirdly, a spindle fault detection system includes: The information input module is used to input the identification information of the spindle; A vibration measurement module is used to measure the vibration of the spindle and generate vibration data; The curve generation and archiving module is connected to the information input module and the vibration measurement module, and is used to generate vibration characteristic curves based on the vibration data and store them in the database. The classification and retrieval module is connected to the curve generation and archiving module and is used to retrieve historical curves in the database based on the identification information. The curve comparison module is connected to the classification and retrieval module and is used to compare the vibration characteristic curve with the retrieved historical curve. The statistical decision module is used to make decisions based on the distribution of the result branches of the historical curve and the statistical quantities, and output the decision results; The tracking feedback module is used to receive the decision result and perform corresponding operations based on the decision result to obtain the fault detection and judgment result of the spindle; The file collection module is used to collect the identification information, vibration data, vibration characteristic curves and fault detection judgment results of the spindle and store them in the database. When the classification and retrieval module fails to retrieve the historical curve, the tracking and feedback module sends a test control signal to the external testing equipment and receives the test results fed back by the external testing equipment, which are used as the fault detection and judgment results of the spindle and stored in the archive collection module.

[0015] In conjunction with the third aspect, some implementations of the third aspect also include a result archive management module, which is used to manage all the result branches and corresponding statistical quantities of each of the historical curves and store them in the database; When the tracking feedback module receives the decision result of omitting subsequent tests, it determines the result branch specified by the statistical decision module as the fault detection result of the spindle and stores the fault detection result of the spindle in the result file management module; if the spindle produces an abnormal fault, the tracking feedback module feeds back the abnormal fault to the result file management module. When the tracking feedback module receives the decision result that further testing is required, it sends a test control signal to the external testing equipment and receives the test results fed back by the external testing equipment, which are used as the fault detection judgment result of the spindle and stored in the result file management module.

[0016] One of the above technical solutions has at least one of the following advantages or beneficial effects: The spindle fault detection system, method and storage medium of this application compares the vibration characteristic curve with the matching historical curve one by one, so that the spindle fault detection no longer makes a static comparison with a pre-set fixed threshold, but makes a dynamic horizontal comparison with a large number of similar historical curves accumulated in mass production. This overcomes the static judgment defect of the single threshold comparison in the prior art, can capture the subtle vibration characteristic changes caused by minor manufacturing defects, realize early abnormality warning, and avoid the risk of missed detection.

[0017] After obtaining the fault detection and judgment results of the spindle, the relevant spindle information is stored in the database, so that each fault detection and judgment result can be used as a statistical sample for subsequent decision-making. This realizes the continuous accumulation and reuse of detection knowledge, which is conducive to iterative optimization and has the ability to learn and optimize itself. The accuracy of the fault detection and judgment results can be improved as the number of detection samples increases, with the outstanding advantage of becoming more accurate with each test.

[0018] Furthermore, by using curve comparison to determine further testing and omitting tests, the testing process can be transformed from a "fixed and rigid process" to an "intelligent and dynamic tailoring," significantly improving testing efficiency. Simultaneously, this invention enables the testing process to be dynamically optimized as data accumulates, achieving multiple outstanding advantages such as increasing accuracy and speed with each test. Attached Figure Description

[0019] The following description, in conjunction with the accompanying drawings, further illustrates this application: Figure 1 This is a schematic diagram of a spindle fault detection system according to an embodiment of this application; Figure 2 This is a schematic diagram of the collection and archiving of fault detection for two spindles in one embodiment of this application; Figure 3 This is a schematic diagram of vibration characteristic curve generation and archiving according to an embodiment of this application; Figure 4 This is a schematic diagram comparing the vibration characteristic curve with historical curves according to one embodiment of this application; Figure 5 This is a schematic diagram illustrating the calculated deviation values ​​of the vibration characteristic curve and the historical curve for each group of characteristic values ​​under different speed groups in an embodiment of the application. Detailed Implementation

[0020] This section will describe in detail the specific embodiments of this application. Preferred embodiments of this application are shown in the accompanying drawings. The purpose of the drawings is to supplement the textual description with graphics, so that people can intuitively and vividly understand each technical feature and the overall technical solution of this application, but they should not be construed as limiting the scope of protection of this application.

[0021] In this application, when directions (up, down, left, right, front, and back) are described, it is only for the purpose of describing the technical solution of this application, and does not indicate or imply that the technical features referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation on this application.

[0022] In this application, "several" means one or more, "multiple" means two or more, "greater than," "less than," "exceeding," etc. are understood to exclude the stated number; "above," "below," "within," etc. are understood to include the stated number. In the description of this application, the terms "first" and "second" are used only for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance or implicitly indicating the number of indicated technical features or the order of the indicated technical features.

[0023] In this application, unless otherwise explicitly defined, the terms "setup," "installation," and "connection" should be interpreted broadly. For example, they can refer to direct connection or indirect connection through an intermediate medium; they can refer to fixed connection, detachable connection, or integral molding; they can refer to mechanical connection, electrical connection, or connection capable of mutual communication; they can refer to the internal connection of two components or the interaction between two components. Those skilled in the art can reasonably determine the specific meaning of the above terms in this application based on the specific content of the technical solution.

[0024] See Figure 1 , Figure 3 and Figure 4 The embodiments of this application provide a spindle fault detection method, including: Enter the spindle identification information; Vibration measurements are performed on the spindle, and vibration data is generated. Generate vibration characteristic curves based on vibration data; Based on the identification information, historical curves are searched in the database to determine if there is a matching historical curve for the spindle under test. If no matching historical curve is found, a comprehensive testing process is performed on the spindle to obtain the spindle fault detection and judgment results, which are then stored in the database. If a matching historical curve is found, the vibration characteristic curve is compared with the matching historical curve. That is, the vibration characteristic curve of the current spindle to be tested is compared with each of the retrieved historical curves one by one, and a decision is obtained based on the comparison results. If the decision is that further testing is required, a testing process will be performed on the spindle to obtain the spindle fault detection and judgment results, which will then be stored in the database. The decision is to omit subsequent tests, directly output the spindle fault detection and judgment results, and store them in the database.

[0025] The spindle fault detection method of this application compares the vibration characteristic curve with the matching historical curve one by one. This makes the spindle fault detection no longer a static comparison with a pre-set fixed threshold, but a dynamic horizontal comparison with a large number of similar historical curves accumulated in mass production. This overcomes the static judgment defect of the single threshold comparison in the prior art, and can capture subtle vibration characteristic changes caused by minor manufacturing defects, realize early anomaly warning, and avoid the risk of missed detection.

[0026] After obtaining the spindle fault detection judgment result, the spindle fault detection method stores the relevant spindle information in the database, so that each fault detection judgment result can be used as a statistical sample for subsequent decision-making. This realizes the continuous accumulation and reuse of detection knowledge, which is conducive to iterative optimization. It has the ability to learn and optimize itself, and the accuracy of the fault detection judgment result can be improved as the number of detection samples increases. It has the outstanding advantage of becoming more accurate with each test.

[0027] Furthermore, when the decision result obtained through curve comparison is to omit subsequent tests, the inspection process can be transformed from a "fixed and rigid process" to an "intelligent and dynamic trimming" process, significantly improving inspection efficiency. The inspection time for a single spindle can be reduced from an average of 4 hours to less than 40 minutes. The proportion of spindles that omit subsequent tests using this spindle fault detection method can exceed 95%, greatly shortening the spindle fault detection time. At the same time, the inspection method of this invention allows the inspection process to be dynamically optimized as data accumulates, achieving the outstanding advantages of becoming more accurate and faster with each test.

[0028] In some embodiments, the vibration data includes vibration acceleration, vibration velocity, and vibration displacement data of the spindle under several sets of rotational speeds. When generating vibration characteristic curves based on the vibration data, the vibration data is preprocessed, time-domain features are extracted, and frequency-domain features are extracted to obtain characteristic value sets corresponding to each set of rotational speeds, so as to generate standardized vibration characteristic curves.

[0029] Each rotational speed corresponds to a set of characteristic values, and each set of characteristic values ​​contains 6 characteristic values. Preprocessing includes trend elimination (least squares method), smoothing (five-point cubic smoothing algorithm), and filtering (10Hz~5000Hz bandpass filter). After preprocessing, the original time-domain features of the vibration acceleration are obtained. The acceleration waveform is integrated once to obtain the vibration velocity waveform, and integrated again to obtain the vibration displacement waveform. The original time-domain features of velocity and displacement are extracted respectively. Fourier transforms are performed on the acceleration, velocity, and displacement waveforms to obtain their spectra, and the proportions of the 3rd, 5th, and 7th harmonics are extracted as the original frequency-domain features. After extracting all original time-domain and frequency-domain features at each test speed, each original feature is subjected to min-max standardization to obtain a dimensionless single characteristic value. All standardized time-domain and frequency-domain values ​​at the same speed are set into a set of characteristic values. The characteristic values ​​corresponding to each set of speeds are arranged and connected with speed as the horizontal axis to generate a standardized vibration characteristic curve, which is then bound to the spindle identification information to establish a curve file and stored in the database.

[0030] In some embodiments, the comparison includes calculating the deviation value between the vibration characteristic curve and the retrieved historical curve at each set of rotational speeds. Each set of characteristic values ​​yields 6 deviation values, and N sets of rotational speeds will yield 6*N deviation values. All deviation values ​​are summed and averaged to obtain the difference index, where the deviation value = |a - b| / max(a, b) / 100. (See [reference]). Figure 5 'a' represents a characteristic value in the characteristic value group corresponding to the vibration characteristic curve at a specific speed, and 'b' represents a characteristic value of the same type as 'a' in the characteristic value group corresponding to the historical curve at the same speed.

[0031] The difference index is compared with a preset similarity threshold. Based on the comparison results, it is determined whether the vibration characteristic curve is the same as or similar to the historical curve being compared, so as to obtain a decision result.

[0032] Furthermore, after the comparison is completed, if there is a historical curve that is the same as or similar to the vibration characteristic curve, the confidence level of the fault detection result of the corresponding historical curve is judged. If the confidence level is high, the decision is to omit subsequent testing; if the confidence level is low, the decision is to require further testing. By judging the decision based on the confidence level, the spindle fault detection judgment result is more accurate.

[0033] If no historical curve is found that is identical or similar to the vibration characteristic curve, the decision is that further testing is required. This testing is comprehensive and may include break-in testing, no-load operation testing, and machining testing.

[0034] Furthermore, when the fault detection result of the corresponding historical curve has a unique result branch, and the number of result branches reaches a preset number threshold, the confidence level is high.

[0035] Furthermore, if the fault detection results of the corresponding historical curve have multiple result branches, and the proportion of each result branch does not reach the preset concentration threshold, the confidence level is deemed low, and a comprehensive test is performed on the spindle. The corresponding historical curve fault detection results have multiple result branches, and when the proportion of a certain result branch reaches the preset concentration threshold, the confidence level is judged to be low, and partial testing is performed on the spindle.

[0036] When the fault detection result has multiple result branches, the concentration threshold is used to help determine the confidence level and to obtain the execution results of the full test and the partial test. When the confidence level is low, the full test will not be executed uniformly, but will be flexibly adjusted according to the actual situation of the result branches, which greatly improves the efficiency and accuracy of spindle fault detection.

[0037] In some embodiments, the identification information includes a spindle serial number and characteristic information. The spindle serial number serves as a unique identifier for the spindle within the system and is used for subsequent tracking and management throughout its entire lifecycle. The characteristic information includes at least one or more of the following: spindle model, maximum speed, bearing type, and rated power. When retrieving historical curves, historical curves are retrieved from the database based on the characteristic information to obtain historical curves that are identical to one or more of the spindle model, maximum speed, bearing type, or rated power of the spindle under test.

[0038] In some embodiments, when the decision result is to omit subsequent tests, if the spindle produces an abnormal fault, the abnormal fault is stored in the database as a result branch. When the decision result is to omit subsequent tests for the same or similar vibration characteristic curves, abnormal fault detection is added to the spindle to achieve feedback correction and system self-optimization. This ensures that subsequent spindles with the same or similar vibration characteristic curves can be detected more comprehensively during fault detection, further improving the efficiency of the detection process and the accuracy of the detection results.

[0039] This application also provides a storage medium storing a computer program. When the computer program is executed by a processor, it implements the spindle fault detection method described in any of the above embodiments. The storage medium can be understood as any computer-readable tangible carrier capable of persistently storing program code. Common types include hard disks, flash drives, ROMs, optical discs, and server storage disks. After the program corresponding to this method is written into the storage medium, the processor of the industrial computer or controller can read the program and execute all the spindle fault detection steps of this invention.

[0040] See Figure 1 This application also provides a spindle fault detection system, comprising: The information input module is used to input the identification information of the spindle. The information input module can be one or more of the following: touch screen, keyboard, mouse or barcode scanner. It can be installed or electrically connected to the industrial control computer of the testing platform to realize flexible information input.

[0041] The vibration measurement module is used to measure the vibration of the spindle and generate vibration data. The vibration measurement module includes a speed control unit, a sensor unit, and a data acquisition unit. The speed control unit divides the spindle into multiple discrete speed groups according to its highest speed, such as low-speed group, medium-speed group, and high-speed group, or according to fixed speed intervals, forming multi-dimensional speed detection points. The sensor unit uses a piezoelectric accelerometer sensor, which is arranged radially on the spindle bearing housing. The data acquisition unit can amplify the signal and perform analog-to-digital conversion by setting the sampling frequency to output standardized vibration data.

[0042] The curve generation and archiving module is connected to the information input module and the vibration measurement module, and is used to generate vibration characteristic curves based on vibration data and store them in the database. The classification and retrieval module is connected to the curve generation and archiving module, and is used to retrieve historical curves in the database based on the identification information. The curve comparison module is connected to the classification and retrieval module and is used to compare the vibration characteristic curve with the retrieved historical curves. The statistical decision module is used to make decisions based on the distribution and statistical quantity of the result branches of the historical curve, and output the decision results. The tracking feedback module is used to receive decision results and perform corresponding operations based on the decision results to obtain the fault detection and judgment results of the spindle; The file collection module is used to collect the spindle's identification information, vibration data, vibration characteristic curves, and spindle fault detection and judgment results, and store them in the database; When the classification and retrieval module fails to find a matching historical curve, the tracking and feedback module sends a test control signal to the external test equipment and receives the test results fed back by the external test equipment. These results are used as the main axis fault detection and judgment results and stored in the archive collection module for storage in the database.

[0043] Furthermore, the spindle fault detection system also includes a result archive management module. The result archive management module is used to manage all result branches and corresponding statistical quantities of each historical curve and store them in the database. The result archive management module independently stores all fault result branches (qualified, poor dynamic balance, bearing wear, bearing preload deviation, etc.) corresponding to each type of vibration curve, and updates the cumulative statistical sample quantity of each branch in real time to provide a basis for confidence judgment for the statistical decision module.

[0044] Specifically, when the tracking feedback module receives a decision to omit subsequent testing, it determines the fault detection result of the spindle as specified by the statistical decision module and stores the fault detection result of the spindle in the result file management module. If the spindle experiences an abnormal fault, the tracking feedback module will report the abnormal fault to the result file management module. At the same time, after the completion of all testing processes for a single spindle, the file collection module integrates the identification information, original vibration data, standardized curves, fault determination, and tracking correction records to generate a full life cycle health file for the spindle, which is then stored in the database for permanent retention and used for batch defect tracing.

[0045] When the tracking feedback module receives a decision that further testing is required, it sends a test control signal to the external testing equipment and receives the test results fed back by the external testing equipment. These results are used as the main fault detection judgment results and stored in the result file management module.

[0046] To more clearly illustrate the specific implementation of the spindle fault detection system, method, and storage medium of this application, four different embodiments are listed below for further explanation: Example 1: If the model is a newly launched product and there are no historical curve files for the same model in the database, the curve comparison module determines that there is no matching curve and sends a new curve file creation command to the curve generation and archiving module (this module creates a new curve file based on the current vibration characteristic curve), and simultaneously sends a new result file creation command to the result file management module. The tracking feedback module sends a complete test command to the external testing equipment, controlling the external testing equipment to sequentially perform the break-in test (2 hours), operation test, and processing test. The external testing equipment feeds back the test results to the tracking feedback module. The tracking feedback module determines the final judgment conclusion ("qualified" in this embodiment) based on the test results, stores this conclusion as the first result branch in the result file management module, and the file collection module completes the collection (refer to...). Figure 2 (File 1). With this, the system's first collected file was written into the system, providing the first set of historical data for subsequent detection of similar spindles.

[0047] Example 2 If this model is a mass-produced product, after obtaining the curve file using the above method (refer to...). Figure 2 (File 2) The classification and retrieval module searches for historical curve files using the model "GDZ-120-15" as the search condition. The newly measured vibration characteristic curve is compared with the historical curves in the file. The comparison method is average percentage error, calculated using the formula: Deviation = |a - b| / max(a, b) / 100. That is, at a specific speed, the deviation value is obtained by subtracting all corresponding characteristic values ​​in each characteristic value group of the vibration characteristic curve from those of the matched historical curve. Figure 2By comparing the curves of the two archives and calculating the deviation, please refer to [reference needed]. Figure 5 .

[0048] Then, the deviation values ​​are summed and averaged to obtain the difference index. In this embodiment, the difference index is approximately 1.89%, which is less than the preset similarity threshold of 10%, indicating that it is identical to the historical records. A retrospective check of the historical records reveals that out of 300 accumulated detection data points, 298 are classified as "qualified" and only 2 are classified as "poor dynamic balance." (See [link to relevant documentation]). Figure 4 If the statistical decision module determines that the result has high confidence, it outputs an instruction to omit subsequent tests to the tracking feedback module. The tracking feedback module directly determines the result as "qualified" and transmits the result to the result file management module. The number of "qualified" branches is updated to 299, and the file retains the tracking attribute. Subsequent tests can be omitted based on usage to ensure accurate results. Simultaneously, the statistical data for this curve type is richer, and the statistical confidence is further improved, making the decision to omit tests in subsequent tests more reliable—fully demonstrating the system's characteristic of "becoming more accurate and faster with each test."

[0049] Example 3 The difference between this embodiment and Embodiment 2 is that, after determining that the difference index in this embodiment is less than the preset similarity threshold, it is considered to be the same as the historical archive. However, the result archive management module stores three result branches: "Qualified" (45), "Slight Bearing Wear" (30), and "Poor Dynamic Balance" (25). The branches are relatively scattered, with no significantly dominant branch (the highest branch accounts for only 45%). The statistical decision module determines that the confidence level is insufficient to support omitting the test, outputs a further test instruction to the tracking feedback module, and determines the order of tracking analysis and processing test to be performed based on the test item requirements corresponding to each branch. The tracking feedback module sends the corresponding test control signal to the external test equipment. After the external test equipment performs the test, it feeds back the results to the tracking feedback module, confirming the existence of slight dynamic imbalance. The tracking feedback module stores the final judgment result "Poor Dynamic Balance Requires Rework" in the result archive management module, and the statistical count of this branch is updated to 26. The conclusion of this test became the 26th statistical sample in the "poor dynamic balance requires rework" branch under this curve type, which improved the statistical confidence of this branch and made the subsequent testing decisions for similar spindles more accurate—the system completed another self-optimization in this process.

[0050] Example 4 This example demonstrates the application of a tracking and correction mechanism. A batch of spindles had a large number of "qualified" branches (over 500), leading the statistical decision module to determine a high confidence level. The tracking feedback module skipped the testing instructions, directly classifying them as qualified and adding them to the archive. One spindle (serial number "SP-2026-01234") developed abnormal vibration after 100 hours of operation at the customer's site. Return testing confirmed the issue was caused by a bearing preload deviation. After-sales personnel entered this anomaly feedback into the system. Upon receiving this feedback, the tracking feedback module created a new "Bearing Preload Deviation" result branch in the result archive management module and triggered a re-statistical update. The system's judgment rule was updated to: for subsequent tests of this curve type, a rapid preload check is added as an auxiliary confirmation to the statistical judgment. Thus, the system completed a full closed loop of "misjudgment → correction → rule iteration," further improving its judgment accuracy and verifying the effectiveness of the system's self-improvement mechanism.

[0051] In the description of this specification, references to terms such as "example," "embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0052] Of course, the present invention is not limited to the above-described embodiments. Those skilled in the art can make equivalent modifications or substitutions without departing from the spirit of this application, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A spindle fault detection method, characterized in that, include: Enter the spindle identification information; Vibration measurements are performed on the spindle, and vibration data is generated. A vibration characteristic curve is generated based on the vibration data; Historical curves are retrieved from the database based on the identification information: If no matching historical curve is found, a test procedure is performed on the spindle to obtain the fault detection and judgment result of the spindle and store it in the database; If a matching historical curve is found, the vibration characteristic curve is compared with the historical curve, and a decision result is obtained: The decision result is that if further testing is required, a test procedure is performed on the spindle to obtain the fault detection and judgment result of the spindle, and the result is stored in the database; The decision result is to omit subsequent tests, directly output the fault detection judgment result of the spindle, and store it in the database.

2. The spindle fault detection method according to claim 1, characterized in that, The vibration data includes the vibration acceleration, vibration velocity, and vibration displacement data of the spindle at several sets of rotational speeds. When generating vibration characteristic curves based on the vibration data, the vibration data is preprocessed, time-domain features are extracted, and frequency-domain features are extracted to obtain characteristic value groups corresponding to each set of rotational speeds.

3. The spindle fault detection method according to claim 2, characterized in that, The comparison includes calculating the deviation value between the vibration characteristic curve and the historical curve at each set of rotational speeds corresponding to the characteristic value group, obtaining multiple deviation values, summing the deviation values ​​and taking the average value to obtain the difference index, wherein the deviation value = |a - b| / max(a, b) / 100, a represents a certain characteristic value in the characteristic value group corresponding to the vibration characteristic curve at a specific rotational speed, and b represents a characteristic value of the same type as a in the characteristic value group corresponding to the historical curve at the same rotational speed as a; The difference index is compared with a preset similarity threshold, and the vibration characteristic curve is determined to be the same as or similar to the historical curve currently being compared, based on the comparison result, so as to obtain the decision result.

4. The spindle fault detection method according to claim 3, characterized in that, After the comparison is completed, if there is a historical curve that is the same as or similar to the vibration characteristic curve, the confidence level of the fault detection result of the corresponding historical curve is determined. If the confidence level is high, the decision result is to omit subsequent testing; if the confidence level is low, the decision result is to require further testing. If no historical curve is found that is the same as or similar to the vibration characteristic curve, the decision result is determined to require further testing.

5. The spindle fault detection method according to claim 4, characterized in that, When the fault detection result of the corresponding historical curve has a unique result branch, and the number of result branches reaches a preset number threshold, the confidence level is determined to be high. If the fault detection result of the corresponding historical curve has multiple result branches, and the proportion of each result branch does not reach the preset concentration threshold, the confidence level is determined to be low, and a comprehensive test is performed on the spindle. The fault detection result of the corresponding historical curve has multiple result branches, and when the proportion of the number of a certain result branch reaches the preset concentration threshold, it is determined that the confidence level is not high, and partial testing is performed on the spindle.

6. The spindle fault detection method according to claim 1, characterized in that, The identification information includes the spindle serial number and characteristic information. The characteristic information includes at least one or more of the following: spindle model, maximum speed, bearing type, and rated power. When retrieving the historical curve, the database is searched based on the characteristic information.

7. The spindle fault detection method according to claim 1, characterized in that, If the decision result is to omit subsequent tests, and the spindle produces an abnormal fault, the abnormal fault is stored as the result branch in the database. If the decision result is to omit subsequent tests when the vibration characteristic curve is the same or similar, the abnormal fault detection is added to the spindle.

8. A storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the spindle fault detection method according to any one of claims 1 to 7.

9. A spindle fault detection system, characterized in that, include: The information input module is used to input the identification information of the spindle; A vibration measurement module is used to measure the vibration of the spindle and generate vibration data; The curve generation and archiving module is connected to the information input module and the vibration measurement module, and is used to generate vibration characteristic curves based on the vibration data and store them in the database; The classification and retrieval module is connected to the curve generation and archiving module and is used to retrieve historical curves in the database based on the identification information. The curve comparison module is connected to the classification and retrieval module and is used to compare the vibration characteristic curve with the retrieved historical curve. The statistical decision module is used to make decisions based on the distribution and statistical quantity of the result branches of the historical curve, and output the decision results; The tracking feedback module is used to receive the decision result and perform corresponding operations based on the decision result to obtain the fault detection and judgment result of the spindle; The file collection module is used to collect the identification information, vibration data, vibration characteristic curves and fault detection judgment results of the spindle and store them in the database. When the classification and retrieval module fails to retrieve the historical curve, the tracking and feedback module sends a test control signal to the external testing equipment and receives the test results fed back by the external testing equipment, which are used as the fault detection and judgment results of the spindle and stored in the archive collection module.

10. The spindle fault detection system according to claim 9, characterized in that, It also includes a result archive management module, which is used to manage all the result branches and corresponding statistical quantities of each historical curve, and store them in the database; When the tracking feedback module receives the decision result of omitting subsequent tests, it determines the result branch specified by the statistical decision module as the fault detection result of the spindle and stores the fault detection result of the spindle in the result file management module; if the spindle produces an abnormal fault, the tracking feedback module feeds back the abnormal fault to the result file management module. When the tracking feedback module receives the decision result that further testing is required, it sends a test control signal to the external testing equipment and receives the test results fed back by the external testing equipment, which are used as the fault detection judgment result of the spindle and stored in the result file management module.