An air main shaft monitoring and early warning method, a monitoring and early warning system, a computer device, and a storage medium

Abstract

The application relates to an air main shaft monitoring and early warning method, a monitoring and early warning system, computer equipment and a storage medium. The air main shaft monitoring and early warning method comprises the following steps: collecting state data of an air main shaft in real time through a sensor group arranged on the air main shaft; pre-processing and feature extraction are performed on the collected state data to generate a feature parameter set representing the running state of the equipment; based on the feature parameter set, a preset fault diagnosis model is used for judgment to identify whether the air main shaft has a potential fault, and the fault type is determined when the fault is identified. Through air main shaft data collection comparison, the fault is identified and the fault type is determined, so that the technical effect of reducing the maintenance cost of the air main shaft is achieved.

Description

Technical Field

[0001] This application relates to the field of precision machining technology, and in particular to an air spindle monitoring and early warning method, a monitoring and early warning system, a computer device, and a storage medium. Background Technology

[0002] In the semiconductor wafer manufacturing process, the air spindle is a core and critical piece of equipment. Its operational stability directly determines the wafer surface processing accuracy, processing efficiency, manufacturing cost, and final product yield. It is one of the core links affecting the quality of high-end semiconductor manufacturing.

[0003] In existing technologies, the operation and maintenance mode of air spindles generally relies on post-maintenance: lacking effective real-time monitoring methods and fault prediction mechanisms, it is impossible to identify potential faults in advance. Repairs are only carried out when the fault becomes apparent and the equipment exhibits obvious abnormalities (such as exceeding machining accuracy or spindle jamming), which leads to increased maintenance costs.

[0004] Therefore, the technical problem with the existing technology is that the air spindle has high maintenance costs. Summary of the Invention

[0005] This application provides an air spindle monitoring and early warning method, a monitoring and early warning system, a computer device, and a storage medium. By collecting and comparing air spindle data, faults are identified and the fault type is determined, thereby achieving the technical effect of reducing the maintenance cost of air spindles.

[0006] Firstly, this application provides an air spindle monitoring and early warning method, which adopts the following technical solution: An air spindle monitoring and early warning method, comprising: The sensor array deployed on the air spindle collects real-time status data during the operation of the air spindle. The collected state data is preprocessed and features are extracted to generate a set of feature parameters characterizing the operating state of the equipment. Based on the set of feature parameters, a preset fault diagnosis model is used to identify whether there is a potential fault in the air spindle, and to determine the fault type when a fault is identified.

[0007] Preferably, the real-time acquisition of the air spindle's operational status data via a sensor array deployed on the air spindle includes: The status data includes vibration data, temperature data, air film pressure data, and humidity data; A microelectromechanical system vibration sensor is installed at the bearing end of the air spindle to collect multidimensional vibration data; A temperature sensor and an air film pressure sensor are installed at the air bearing of the air spindle to collect the temperature data and the air film pressure data. A humidity sensor is installed at the air inlet of the air shaft to collect the humidity data.

[0008] Preferably, the preprocessing and feature extraction of the collected state data to generate a set of feature parameters characterizing the device's operating state includes: The vibration data collected by the vibration sensor of the microelectromechanical system is filtered and denoised, and frequency domain features are extracted by fast Fourier transform. The temperature data and air film pressure data collected by the temperature sensor and air film pressure sensor are smoothed to extract trend change characteristics. The duration and instantaneous rate of change of humidity exceeding the standard are calculated from the humidity data collected by the humidity sensor. Based on the processed data, feature indicators that are strongly correlated with preset fault modes are selected to form the feature parameter set.

[0009] Preferably, the feature parameter set includes: Amplitude values ​​of specific frequency components related to bearing wear extracted from vibration data; The mean, trend, and fluctuation indicators extracted from temperature and film pressure data; and The duration of exceeding the standard and the rate of change were extracted from humidity data.

[0010] Preferably, the judgment based on the feature parameter set and using a preset fault diagnosis model includes: The extracted feature parameter set is matched one by one with the fault feature mapping library to form a fault feature mapping library, and the feature parameter combination corresponding to the potential fault is selected. Among them, the characteristic parameter combination of the continuous decrease in air film pressure, the rate of temperature rise, and the low-frequency change of vibration amplitude is corresponding to the air film attenuation. The characteristic parameter combination of vibration frequency domain characteristic peak value and local abnormal temperature rise value corresponding to bearing wear; The characteristic parameter combination of spindle water inlet corresponding to the duration of excessive humidity at the air inlet, sudden change value of humidity at the bearing end, and air film pressure fluctuation coefficient; Based on the matching results, potential fault types that match the current combination of feature parameters are initially screened out.

[0011] Preferably, the preset fault diagnosis model is a fault tree-based diagnosis model, and the method for constructing the fault tree includes: The fault tree uses at least one of the following as top-level events: air film decay, bearing wear, and water ingress into the spindle; the direct causes of each top-level event are intermediate events; and the judgment conditions of each indicator in the feature parameter set are bottom-level events. The diagnostic model associates events at each layer through logic gates, and is used to determine the fault type by executing logic from the bottom layer events to the top layer events based on the input of the feature parameter set.

[0012] Preferably, the intermediate events include: Intermediate events corresponding to film degradation include insufficient gas path pressure and increased bearing clearance. Intermediate events corresponding to bearing wear include insufficient lubrication, wear caused by impurities; and Intermediate events corresponding to water ingress into the spindle include excessive humidity in the air source and aging of seals.

[0013] As a preferred option, it also includes: In response to the judgment result of the fault diagnosis model, if the identified fault is of the adjustable type, the corresponding adaptive control strategy is automatically triggered, including adjusting the gas path pressure or triggering electromagnetic damping compensation. If the identified fault is of the maintenance-required type, or if the fault characteristics are not eliminated after adaptive control is performed, a solution containing the specific fault location and maintenance steps will be generated and output.

[0014] Secondly, the air spindle monitoring and early warning system provided in this application adopts the following technical solution: An air spindle monitoring and early warning system, comprising: The parameter acquisition module is used to collect multi-dimensional status data of the air spindle in real time through the deployed sensor group; A data processing module, communicatively connected to the parameter acquisition module, is used to preprocess and extract features from the state data to generate a feature parameter set; and The fault diagnosis module is communicatively connected to the data processing module. The fault diagnosis module has a preset fault diagnosis model and is used to perform logical judgments based on the feature parameter set to identify and confirm potential faults and their types.

[0015] Thirdly, the computer device provided in this application adopts the following technical solution: A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the method.

[0016] Fourthly, the computer-readable storage medium provided in this application adopts the following technical solution: A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method.

[0017] In summary, this application includes at least one of the following beneficial technical effects: This application provides an air spindle monitoring and early warning method. By constructing a complete technical system for multi-dimensional data acquisition, processing, diagnosis, and control, it enables early identification, accurate location, and proactive control of potential air spindle faults, reducing downtime probability and maintenance costs, and ensuring continuous and stable operation of the production line. It solves the technical problems of existing air spindles lacking effective monitoring and fault prediction methods, relying solely on post-failure maintenance, resulting in significant downtime losses and reduced processing yield. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the air spindle described in this application; Figure 2 This is a flowchart illustrating the air spindle monitoring and early warning method described in this application; Figure 3 This is a schematic diagram of the data monitoring process of the air spindle monitoring and early warning method described in this application; Figure 4 This is a schematic diagram of the feature extraction process of the air spindle monitoring and early warning method described in this application; Figure 5 This is a schematic diagram of the fault matching process of the air spindle monitoring and early warning method described in this application; Figure 6 This is a schematic diagram of the fault tree construction process of the air spindle monitoring and early warning method described in this application; Figure 7 This is a schematic diagram of the computer device structure described in this application.

[0019] Explanation of reference numerals in the attached drawings: 100, housing; 200, main shaft; 300, air bearing; 410, stator; 420, mover. Detailed Implementation

[0020] The serial numbers assigned to components in this document, such as "first" and "second," are used solely to distinguish the described objects and have no sequential or technical meaning. The terms "connection" and "linkage" used in this application, unless otherwise specified, include both direct and indirect connections (linkages). It should be understood that the terms "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings and are used solely for the convenience of describing this application and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this application.

[0021] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "on top of," and "over" the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0022] To better understand the above technical solutions, a detailed description of the technical solutions will be provided below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit the scope of this application.

[0023] As semiconductor technology advances to smaller process nodes, the requirements for wafer surface flatness and roughness are becoming increasingly stringent. This necessitates that air spindles maintain extremely high operational accuracy and reliability. However, existing air spindle maintenance technologies have significant shortcomings, making it difficult to meet the continuous production demands of high-end manufacturing.

[0024] The operation and maintenance mode of air spindles generally relies on post-failure repair, lacking effective real-time monitoring methods and fault prediction mechanisms. It is impossible to identify potential faults in advance, and maintenance is only carried out when the fault becomes explicit and the equipment exhibits obvious abnormalities (such as exceeding machining accuracy or spindle jamming). This operation and maintenance mode has significant drawbacks. On the one hand, sudden failures force the production line to stop, resulting in huge downtime losses and wasted production capacity. On the other hand, post-failure repairs cannot trace the cause of the fault, have long repair cycles, lack specificity, and frequent downtime can easily aggravate the wear and tear of spindle components, further affecting the service life and machining stability of the equipment. It is difficult to adapt to the core requirements of continuous production and precise control in high-end semiconductor manufacturing.

[0025] Furthermore, some existing technologies attempt to monitor air spindles using a single parameter, such as vibration or temperature data. However, a single parameter cannot comprehensively reflect the operating status of the air spindle, easily leading to misjudgments or omissions. At the same time, these technologies lack effective methods for processing and analyzing monitoring data, failing to extract key fault-related features from raw data, and making it even more difficult to accurately identify fault types based on feature data.

[0026] Therefore, there is an urgent need for a technology that can monitor and warn of air spindle failures, in order to solve the problems of lack of fault prediction capability, passive maintenance mode, and large downtime losses in existing technologies.

[0027] This application proposes a method for monitoring and early warning of air spindles, such as... Figure 1As shown, this is applicable to air spindles. An air spindle includes a housing, a spindle, and air bearings. The spindle is rotatably connected to the inside of the housing via the air bearings. The spindle is driven to rotate by a mover located on the spindle and a stator located on the housing. Figure 2 As shown, the air spindle monitoring and early warning methods include: S1: Real-time status data of the air spindle during operation is collected through a sensor array deployed on the air spindle; S2: Preprocess and extract features from the collected status data to generate a set of feature parameters characterizing the operating status of the equipment; S3: Based on the feature parameter set, a preset fault diagnosis model is used to identify whether there is a potential fault in the air spindle, and to determine the fault type when a fault is identified.

[0028] The air spindle monitoring and early warning method of this application achieves accurate identification and type determination of potential air spindle faults through the core logic of data acquisition, data processing, and fault judgment; replacing the existing single-parameter monitoring and passive maintenance mode, it constructs a full-process, intelligent monitoring and early warning system to ensure that each step can provide reliable support for subsequent links, and ultimately achieve early warning of faults.

[0029] Specifically, firstly, a sensor array deployed on the air spindle collects real-time status data during its operation. The sensor array employs a multi-dimensional deployment strategy, covering key status indicators throughout the air spindle's operation to ensure the collected data comprehensively reflects the equipment's operating status. The data collection process must accompany the air spindle's full operating conditions, including no-load start-up, load processing, high-speed operation, and low-speed shutdown, to avoid data loss due to incomplete coverage. The core purpose of real-time data acquisition is to capture the dynamic changes in the air spindle's operating status, providing continuous and complete raw data support for subsequent fault diagnosis and preventing missed fault detections due to data lag or incompleteness.

[0030] Secondly, the collected status data undergoes preprocessing and feature extraction to generate a set of feature parameters characterizing the equipment's operating status. The raw status data contains a large amount of interference signals from various factors such as environmental vibrations, sensor noise, and power fluctuations. Directly using this data for fault diagnosis would severely impact accuracy. Therefore, preprocessing is essential to remove interference signals and retain only the valid signals relevant to the equipment's operating status and faults. After preprocessing, feature parameters are further extracted, transforming the raw data into core indicators that accurately characterize the equipment's health status, forming a feature parameter set. This process purifies and transforms the raw data, providing high-quality input data for subsequent fault diagnosis models and ensuring the accuracy of fault diagnosis.

[0031] Next, based on the feature parameter set, a pre-set fault diagnosis model is used to identify potential faults in the air spindle and determine the fault type when a fault is identified. This pre-set fault diagnosis model, trained and optimized with extensive experimental data, can accurately determine the presence of potential faults in the air spindle based on core indicators in the feature parameter set and the inherent logic of fault occurrence. Compared to existing single-threshold judgment methods, this model can perform multi-parameter correlation analysis, effectively avoiding misjudgments caused by fluctuations in a single parameter. When a potential fault is identified, the model can further determine the fault type, providing a clear basis for subsequent early warning, control, and maintenance, achieving a leap from anomaly detection to precise fault location.

[0032] Based on data acquisition, supported by preprocessing and feature extraction, and with fault diagnosis as the goal, this system enables full-process monitoring of the air spindle's operating status and accurate fault prediction, completely changing the existing passive maintenance model and ensuring the continuous and stable operation of the production line.

[0033] Furthermore, such as Figure 3 As shown, the sensor array deployed on the air spindle collects real-time status data during air spindle operation, including: S11: Status data includes vibration data, temperature data, air film pressure data, and humidity data; S12: The bearing end of the air spindle is equipped with a microelectromechanical system vibration sensor to collect multidimensional vibration data; Temperature sensors and air film pressure sensors are installed at the air bearing of the air spindle to collect temperature and air film pressure data. A humidity sensor is installed at the air inlet of the air shaft to collect humidity data.

[0034] The aforementioned status data acquisition steps utilize a specific sensor deployment strategy to achieve comprehensive and accurate acquisition of multi-dimensional data. The acquired status data specifically includes vibration data, temperature data, air film pressure data, and humidity data. These four data types correspond to different key states during the operation of the air spindle, collectively forming a complete status monitoring system. The acquisition of each data type is achieved by deploying corresponding sensors at specific locations on the air spindle, ensuring the relevance and effectiveness of the collected data.

[0035] Specifically, for vibration data acquisition, a microelectromechanical system (MEMS) vibration sensor is installed at the bearing end of the air spindle to collect multi-dimensional vibration data. The bearing end is the primary source of vibration for the air spindle, and its vibration state directly reflects core indicators such as bearing operation and air film stability. Therefore, the MEMS vibration sensor is deployed on the surface of the bearing end housing, and bolts or adhesive are used to ensure a tight fit between the sensor and the housing, preventing vibration signal attenuation or distortion due to loose installation. The MEMS vibration sensor has multi-dimensional acquisition capabilities, simultaneously acquiring vibration data in the axial, radial, and tangential directions, comprehensively capturing the vibration characteristics during air spindle operation. During acquisition, the sensor maintains a stable sampling frequency to ensure the capture of vibration changes at different speeds, providing comprehensive vibration data support for subsequent fault diagnosis. Accurate vibration data acquisition can promptly detect vibration anomalies caused by bearing wear, air film instability, and other faults, facilitating the identification of potential faults.

[0036] For acquiring temperature and film pressure data, temperature sensors and film pressure sensors are installed at the air bearing of the air spindle. The air bearing is the core component of the air spindle, and its operating state directly determines the accuracy and stability of the air spindle. Film pressure is a key parameter for maintaining the normal operation of the air bearing, while temperature directly reflects the friction state and film stability of the air bearing. Therefore, the temperature and film pressure sensors are deployed on the end face of the air bearing, using miniaturized sensors to avoid interfering with the structure of the air bearing and film formation. The temperature sensor, by being in contact with the surface of the air bearing, senses changes in bearing temperature in real time, enabling it to detect abnormalities such as increased friction due to bearing wear and localized overheating due to film decay. The film pressure sensor, located near the film formation area, collects film pressure values ​​in real time, accurately reflecting film thickness and stability, and promptly detecting problems such as insufficient film pressure and excessive pressure fluctuations. The two sensors collect data simultaneously, ensuring the temporal consistency of temperature and film pressure data, providing conditions for subsequent correlation analysis.

[0037] For humidity data acquisition, a humidity sensor is installed at the air inlet of the air spindle. The air spindle's air path system relies on dry compressed air to form a stable air film. If the air source humidity exceeds the standard, moisture will enter the air spindle with the compressed air, leading to decreased air film stability, bearing corrosion, damage to internal components, and ultimately causing water ingress failure in the spindle. Therefore, the humidity sensor is deployed near the air source access point at the air inlet to ensure real-time monitoring of the external air source humidity, detecting humidity anomalies before moisture enters the air path system and the air spindle. The humidity sensor continuously collects air source humidity data, enabling timely detection of excessive or sudden humidity changes, providing early warning for preventing water ingress failure in the spindle and avoiding irreversible damage to the air spindle.

[0038] The aforementioned sensor deployment strategy enables the synchronous and real-time acquisition of four key state data points. These four data points reflect the operating status of the air spindle from different dimensions, complementing and verifying each other to form a complete multi-dimensional data acquisition system. This provides comprehensive and reliable raw data for subsequent data processing and fault diagnosis, effectively avoiding the problem of missed fault diagnoses caused by single-parameter acquisition.

[0039] The aforementioned sensor deployment strategy enables the synchronous and real-time acquisition of multiple data sources. These four types of data reflect the operating status of the air spindle from different dimensions, forming a complete multi-dimensional data acquisition system. This provides comprehensive and reliable raw data for subsequent data processing and fault diagnosis, effectively avoiding the problem of missed fault diagnosis caused by single parameter acquisition.

[0040] Furthermore, after collecting the four types of status data, preprocessing and feature extraction are required to transform the raw data into a set of feature parameters that characterize the equipment's operating status. This process employs differentiated processing strategies based on the characteristics of different data types to ensure that the processed data accurately reflects the equipment's health status and provides effective support for subsequent fault diagnosis. For example, Figure 4 As shown, the collected status data undergoes preprocessing and feature extraction to generate a set of feature parameters characterizing the equipment's operating status, including: S21: The vibration data collected by the vibration sensor of the microelectromechanical system is filtered and denoised, and the frequency domain features are extracted by fast Fourier transform. S22: Smooth the temperature data and air film pressure data collected by the temperature sensor and air film pressure sensor to extract trend change features; S23: Calculate the duration and instantaneous rate of change of humidity exceeding the standard based on the humidity data collected by the humidity sensor; S24: Based on the processed data, select characteristic indicators that are strongly correlated with the preset fault modes to form a set of characteristic parameters.

[0041] For vibration data acquired by vibration sensors in microelectromechanical systems (MEMS), filtering and denoising are performed first. The vibration data contains a large number of interference signals, mainly including external environmental vibration interference, sensor electronic noise, and vibration transmission interference from other components of the equipment. These interference signals can mask the effective vibration features related to the fault, leading to inaccurate feature extraction. Therefore, a low-pass filtering algorithm is used to process the vibration data, setting a reasonable filtering threshold to remove high-frequency interference signals and retain low-frequency and mid-frequency vibration signals related to the operation of the air spindle. After filtering and denoising, the waveform of the vibration data is smoother, and the effective signals are more prominent. Subsequently, the time-domain vibration signal is converted into a frequency-domain signal using Fast Fourier Transform (FFT). The time-domain signal can only reflect the change of vibration amplitude over time and cannot accurately identify the characteristic frequencies corresponding to different faults. However, the frequency-domain signal can decompose the vibration into the superposition of different frequency components, with each frequency component corresponding to a specific operating state of the air spindle. Through FFT processing, core indicators such as vibration amplitude and characteristic frequency are extracted. These indicators can accurately reflect the characteristics of faults such as bearing wear and air film attenuation. For example, bearing wear will cause a significant increase in the amplitude of a specific frequency component, and air film attenuation will cause continuous fluctuations in the vibration amplitude in the low-frequency range.

[0042] Temperature and film pressure data collected by temperature and film pressure sensors are smoothed to extract trend characteristics. These data are affected by instantaneous fluctuations during acquisition, often caused by sensor errors, air source pressure fluctuations, and other accidental factors. These fluctuations do not accurately reflect the equipment's operating status and could lead to misjudgments if used directly for analysis. Therefore, a moving average method is used to smooth the temperature and film pressure data. A reasonable sliding window is selected, and the average value within the window is taken to eliminate the influence of instantaneous fluctuations and retain the data's trend. After smoothing, the temperature data clearly reflects the temperature change trend of the air bearing, such as the slow temperature rise due to bearing wear and the continuous local temperature increase caused by film attenuation. The film pressure data shows the stable state and variation pattern of the film pressure, such as the continuous decrease in film pressure due to insufficient air path pressure and the periodic pressure fluctuations caused by film instability. These trend characteristics are directly related to potential faults in the air spindle and are important bases for fault diagnosis.

[0043] For humidity data collected by humidity sensors, the duration and instantaneous rate of humidity exceeding the standard are calculated. First, a reasonable humidity threshold is set, based on the humidity requirements of the air source for the air spindle operation, ensuring that the compressed air dryness meets the needs of air film formation and equipment operation. During data processing, the collected humidity data is compared with the preset threshold in real time, and the duration of humidity exceeding the threshold is statistically analyzed. If the duration of humidity exceeding the standard is short, it may be a temporary fluctuation in the air source and will not have a significant impact on the equipment. If the duration of humidity exceeding the standard is long, it indicates a stable humidity anomaly in the air source, posing a risk of water ingress into the spindle. Simultaneously, the instantaneous rate of change of humidity data is calculated, i.e., the amount of humidity change per unit time. If the instantaneous rate of change of humidity is too large, it indicates a sudden change in air source humidity, which may be a malfunction in the air source system, requiring timely warning to prevent moisture from entering the air spindle. By extracting the duration of humidity exceeding the standard and the instantaneous rate of change indicators, abnormal characteristics of humidity data can be accurately captured, providing a reliable basis for preventing water ingress into the spindle.

[0044] After completing the aforementioned differentiated preprocessing, based on the processed data, feature indicators strongly correlated with preset fault modes are selected to form a feature parameter set. The preset fault modes include three core fault types: air film attenuation, bearing wear, and spindle water ingress. Feature indicators corresponding to each fault mode are summarized through a large amount of experimental data. During the selection process, indicators with weak correlation to the fault modes are eliminated, while indicators with strong correlation and accurate characterization of fault features are retained. For example, characteristic frequency amplitudes related to bearing wear are selected from vibration data, trend indicators related to air film attenuation are selected from temperature and air film pressure data, and the duration of exceeding the standard related to spindle water ingress is selected from humidity data. The feature parameter set formed through this selection can accurately reflect the operating status and potential fault characteristics of the air spindle, providing high-quality input data for subsequent fault diagnosis model judgment and ensuring the accuracy and reliability of fault judgment.

[0045] The feature parameter set includes: amplitude of specific frequency components related to bearing wear extracted from vibration data; mean, trend and fluctuation indicators extracted from temperature and film pressure data; and indicators of duration of exceedance and rate of change extracted from humidity data. Specifically, the amplitude of specific frequency components related to bearing wear, extracted from vibration data, is a core indicator characterizing the bearing wear state. During normal operation, the bearings of an air spindle generate stable vibration frequencies. When bearing wear occurs, the clearance between internal components changes, friction intensifies, and the vibration frequency changes, resulting in specific frequency vibration components. The amplitude of these components gradually increases with the degree of wear. Through Fast Fourier Transform (FFT) processing, these specific frequency components can be accurately identified, and their amplitude directly reflects the degree of bearing wear. For example, when the bearing is slightly worn, the amplitude of the specific frequency component is at a low level; as wear intensifies, clearance increases, and friction worsens, the amplitude of this frequency component rises significantly. By monitoring and analyzing this indicator, the bearing wear state can be identified in advance, preventing further wear deterioration that could lead to bearing damage and equipment failure.

[0046] The mean, trend, and fluctuation indicators extracted from temperature and film pressure data are important indicators characterizing film decay and bearing operating status. The mean temperature reflects the overall temperature level of the air bearing; under normal operating conditions, the mean temperature remains within a stable range. When film decay or bearing wear occurs, friction intensifies, and the mean temperature gradually increases. The temperature trend reflects the temperature change over time; the temperature rise caused by film decay shows a continuous and slow upward trend, while the temperature rise caused by bearing wear may show a phased upward trend. The mean film pressure reflects the overall stability of the film; under normal operating conditions, the mean film pressure remains within a set range, while during film decay, the mean film pressure continuously decreases. The fluctuation of film pressure reflects the stability of the film pressure; when the film is unstable, the pressure fluctuation will increase significantly, which may lead to a decrease in the operating accuracy of the air spindle. These indicators reflect the operating status of the film and bearing from different perspectives, complementing each other and accurately capturing the early characteristics of film decay and bearing wear.

[0047] The duration of exceeding the limit and the rate of change of humidity, extracted from humidity data, are core indicators characterizing the risk of water ingress into the spindle. The duration of exceeding the limit refers to the duration for which humidity data exceeds a preset threshold. When the air source humidity exceeds the limit for a long period, moisture will gradually accumulate in the air path system and eventually enter the air spindle, causing water ingress failure. If the duration of exceeding the limit is short, it is mostly a temporary fluctuation and will not have a significant impact on the equipment. The rate of change of humidity refers to the amount of change in humidity per unit time. When the rate of change of humidity is too large, it indicates that the air source humidity has changed drastically, which may be a failure of the air source drying system. Timely warning is required to avoid a large amount of moisture entering the air spindle. By analyzing these two indicators, the risk of water ingress into the spindle can be detected in advance, and timely intervention measures can be taken to prevent moisture from causing corrosion and damage to the internal components of the air spindle, ensuring the normal operation of the equipment.

[0048] Furthermore, such as Figure 5 As shown, the judgment based on the feature parameter set and the preset fault diagnosis model includes: S31: Match the extracted feature parameter set with the fault feature mapping library one by one to form the fault feature mapping library; S32: Filter out the combination of characteristic parameters corresponding to potential faults; Among them, the characteristic parameter combination of air film attenuation corresponds to the continuous decrease of air film pressure, the rate of temperature rise, and the low-frequency change of vibration amplitude; the characteristic parameter combination of bearing wear corresponds to the characteristic peak value of vibration frequency domain and the local abnormal increase value of temperature; the characteristic parameter combination of spindle water ingress corresponds to the duration of excessive humidity at the air inlet, the sudden change value of humidity at the bearing end, and the air film pressure fluctuation coefficient. S33: Based on the matching results, initially screen out potential fault types that match the current combination of feature parameters.

[0049] The process of judging based on a set of feature parameters using a pre-set fault diagnosis model hinges on the matching analysis between the feature parameter set and the fault feature mapping library. This, combined with multi-parameter combination judgment, enables the initial screening of potential fault types. In other words, the judgment based on the feature parameter set using the pre-set fault diagnosis model includes: matching the extracted feature parameter set with the fault feature mapping library one by one to form the library, and then selecting feature parameter combinations corresponding to potential faults. Specifically, for example, air film attenuation corresponds to the feature parameter combination of the continuous decrease in air film pressure, the rate of temperature rise, and low-frequency changes in vibration amplitude; bearing wear corresponds to the feature parameter combination of vibration frequency domain characteristic peak values ​​and local abnormal temperature increases; and spindle water ingress corresponds to the feature parameter combination of the duration of excessive humidity at the air inlet, sudden changes in humidity at the bearing end, and the air film pressure fluctuation coefficient. Based on the matching results, potential fault types matching the current feature parameter combinations are initially screened.

[0050] First, a pre-defined fault feature mapping library is invoked. This library, built upon extensive experimental data and real-world maintenance cases, stores the relationships between different fault types and their corresponding characteristic parameter combinations. During its construction, the library is constructed by simulating the operation of the air spindle under different fault conditions, collecting corresponding characteristic parameters, and summarizing typical characteristic parameter combinations for each fault type. Simultaneously, fault cases encountered in actual maintenance are used to supplement and improve the mapping library, ensuring it covers all typical characteristics of the three core fault types: air film decay, bearing wear, and water ingress into the spindle. Each relationship in the mapping library includes a fault type and its corresponding characteristic parameter combination. This combination consists of multiple indicators from a set of characteristic parameters, accurately characterizing the features of a specific fault type.

[0051] Subsequently, the extracted feature parameter set is matched one by one with the fault feature mapping library to filter out the feature parameter combinations corresponding to potential faults. The matching process adopts a multi-parameter association matching strategy, rather than independent matching of single parameters. That is, by comparing multiple indicators in the feature parameter set with the feature parameter combinations corresponding to a certain fault type in the mapping library, it is determined whether the combination conditions are met. This multi-parameter combination matching method can effectively avoid misjudgments caused by fluctuations in a single parameter and improve the accuracy of fault diagnosis. For example, if only one feature parameter exceeds the standard, while other associated parameters are normal, it may be an occasional fluctuation and not caused by a fault; if multiple associated parameters exceed the standard at the same time and meet the feature parameter combinations of a certain fault type in the mapping library, then a corresponding potential fault is determined to exist.

[0052] Different fault types correspond to specific combinations of characteristic parameters, enabling precise differentiation between them. Among these, film attenuation corresponds to a combination of characteristic parameters including the sustained decrease in film pressure, the rate of temperature rise, and changes in vibration amplitude in the low-frequency range. The core mechanism of film attenuation is a reduction in film thickness and stability, leading to a sustained decrease in film pressure. Simultaneously, increased friction between the film and the bearing causes a temperature rise, resulting in significant changes in vibration amplitude in the low-frequency range. Therefore, when the sustained decrease in film pressure exceeds a set threshold, the rate of temperature rise matches the characteristics of film attenuation, and abnormal fluctuations occur in the low-frequency range of vibration amplitude, a film attenuation fault can be preliminarily identified through matching.

[0053] Bearing wear corresponds to a combination of characteristic parameters, including the peak value of vibration in the frequency domain and the abnormal increase in local temperature. Bearing wear leads to changes in the clearance of internal components and increased friction, causing both an increase in the peak value of vibration at specific frequencies and an abnormal increase in local bearing temperature. Therefore, when the amplitude of a specific frequency component in the vibration frequency domain of the characteristic parameter set exceeds a set threshold, and the local temperature increase matches the characteristics of bearing wear, a bearing wear fault can be preliminarily identified. As the wear intensifies, the abnormality of these two indicators will gradually become more apparent, reflecting the severity of the fault.

[0054] The spindle water ingress is characterized by a combination of characteristic parameters: the duration of excessive humidity at the air inlet, sudden changes in humidity at the bearing end, and the air film pressure fluctuation coefficient. The core cause of spindle water ingress is excessive humidity in the air source. Moisture enters the air path system and the interior of the air spindle, leading to an increased duration of excessive humidity at the air inlet. The humidity at the bearing end also experiences sudden changes due to moisture infiltration. Simultaneously, the moisture affects the stability of the air film, resulting in an increased air film pressure fluctuation coefficient. Therefore, when these three indicators simultaneously meet the characteristic parameter combination conditions for spindle water ingress in the mapping library, it can be preliminarily determined that there is a risk of spindle water ingress or a minor spindle water ingress fault has already occurred.

[0055] Based on the matching results above, potential fault types that match the current combination of feature parameters are initially identified. During the screening process, if the combination of feature parameters matches some features of multiple fault types simultaneously, it is necessary to further compare the degree of abnormality and correlation of each indicator to eliminate interference and determine the most likely potential fault type. For example, if the film pressure continues to decrease while the temperature rises, but the vibration amplitude shows no significant change in the low-frequency range, it is necessary to combine other indicators for further judgment to avoid misjudging it as film attenuation. This matching and screening process enables the initial identification of potential fault types, which is beneficial for the fault tree-based diagnostic model to further accurately confirm the fault.

[0056] like Figure 6 As shown, the preset fault diagnosis model is a fault tree-based diagnostic model. The fault tree construction method includes: S301: The fault tree uses at least one of the following as top-level events: air film decay, bearing wear, and spindle water ingress; the direct cause of each top-level event is the intermediate event; and the judgment conditions of each indicator in the feature parameter set are the bottom-level events. S302: The diagnostic model associates events at each level through logic gates, and is used to determine the fault type by executing logic from the bottom-level events to the top-level events based on the input of the feature parameter set.

[0057] The preset fault diagnosis model is a fault tree-based model. This model constructs a hierarchical event structure and combines logic gates to associate events at each level, enabling accurate deduction from feature parameters to fault types and improving the logic and accuracy of fault diagnosis. The construction of the fault tree model follows a top-down deductive logic, starting from the final fault type and gradually breaking it down into intermediate causes and bottom-level feature parameter judgment conditions, forming a complete fault diagnosis logic chain.

[0058] The top-level event in the fault tree is set as a potential fault type of the air spindle, specifically at least one of the following: air film decay, bearing wear, and water ingress into the spindle. These three faults are the most common and damaging types during the operation of the air spindle, covering the core risk points in air spindle maintenance. Using them as top-level events can specifically address the lack of fault prediction capabilities in existing technologies. The setting of the top-level event directly corresponds to the core objective of the invention: to accurately identify these three potential faults and achieve early warning and control.

[0059] Intermediate events in the fault tree are defined as the direct causes of each top-level event. These intermediate events serve as transitions between top-level faults and bottom-level characteristic parameters, reflecting the underlying problems that lead to the fault. Each top-level event corresponds to multiple intermediate events, all of which are direct triggers for the top-level fault. These intermediate events are derived from extensive experimental data and fault case studies to ensure comprehensive coverage of all possible causes of the fault. For example, the top-level event for film degradation corresponds to intermediate events such as insufficient air pressure and increased bearing clearance; the top-level event for bearing wear corresponds to intermediate events such as insufficient lubrication and wear due to impurities; and the top-level event for water ingress into the spindle corresponds to intermediate events such as excessive air source humidity and aging seals. The definition of intermediate events must have a clear causal relationship to ensure accurate deduction of top-level faults and to provide a basis for defining the bottom-level events.

[0060] The bottom-level events of the fault tree are set as the judgment conditions for each indicator in the feature parameter set. Each bottom-level event corresponds one-to-one with the extracted feature parameters, serving as the basic input for the fault tree diagnostic model. Each bottom-level event is a quantifiable and monitorable judgment condition. By comparing the indicators in the feature parameter set with preset thresholds, it is determined whether a bottom-level event has occurred. For example, an intermediate event of insufficient air pressure corresponds to a bottom-level event where the continuous drop in air film pressure exceeds a set threshold; an intermediate event of excessive air source humidity corresponds to a bottom-level event where the duration of excessive humidity exceeds a set threshold. The judgment conditions for bottom-level events are based on the normal range settings of the feature parameters, ensuring accurate reflection of abnormal equipment operating conditions and providing reliable support for the judgment of intermediate and top-level events.

[0061] The diagnostic model uses logic gates to connect events at each level. These gates primarily consist of AND and OR gates, used to establish causal relationships between events. An AND gate indicates that the output event will only occur if all input events occur simultaneously; an OR gate indicates that the output event will occur as long as at least one input event occurs. By combining these logic gates, a complete fault diagnosis logic chain is constructed. For example, the top-level event of film gas decay is associated with two intermediate events—insufficient air pressure and increased bearing clearance—through an OR gate. This means that the occurrence of either insufficient air pressure or increased bearing clearance can trigger the top-level event of film gas decay. Similarly, the intermediate event of insufficient air pressure is associated with two bottom-level events—excessive decrease in film pressure and excessive fluctuation in film pressure—through an AND gate. This means that the occurrence of the intermediate event of insufficient air pressure can only be determined when both bottom-level events occur simultaneously.

[0062] Fault tree diagnostic models are used to determine fault types by performing logical deduction from bottom-level events to top-level events based on the input of a feature parameter set. Specifically, it first determines whether each bottom-level event has occurred based on the indicators of the feature parameter set; based on the occurrence of the bottom-level events, it deduces whether intermediate events have occurred through logic gates; finally, based on the occurrence of the intermediate events, it deduces whether top-level events have occurred through logic gates, thereby determining whether there is a potential fault in the air spindle and its specific fault type. This top-down logical deduction method can accurately locate the root cause of the fault, not only identifying the fault type but also clarifying the direct cause of the fault.

[0063] Specifically, intermediate events include: intermediate events corresponding to air film decay include insufficient air pressure and increased bearing clearance; intermediate events corresponding to bearing wear include insufficient lubrication and wear caused by impurities; and intermediate events corresponding to water ingress into the spindle include excessive air source humidity and aging of seals.

[0064] Regarding the top-level event of film vapor deposition (PVD), the corresponding intermediate events include insufficient air pressure and increased bearing clearance. Insufficient air pressure is one of the main causes of PVD. The formation of the PV film in an air spindle depends on a stable air source pressure. When the air pressure is insufficient, the PV film thickness decreases, and the PV film's load-bearing capacity and stability decline, leading to PVD and affecting the operating accuracy of the air spindle. Insufficient air pressure may be caused by factors such as air source system failure, air leakage, or pressure regulating valve failure, ultimately manifesting as a continuous decrease in PV film pressure and increased fluctuations. Increased bearing clearance is also an important cause of PVD. After long-term operation, bearings will wear, leading to increased internal clearance. Increased clearance changes the space for PV film formation, reducing PV film stability and thus causing PVD. Increased bearing clearance manifests as changes in vibration amplitude, temperature rise, and PV film pressure fluctuations. Both of these intermediate events can directly trigger PVD. By linking them to the top-level event using an OR gate, we can ensure comprehensive coverage of the direct causes of PVD.

[0065] For the top-level event of bearing wear, the corresponding intermediate events include insufficient lubrication and impurity wear. Insufficient lubrication leads to increased friction between internal bearing components, accelerating component wear. Long-term insufficient lubrication can cause increased bearing clearance, decreased precision, and even bearing seizure. Insufficient lubrication may be caused by factors such as lubrication system failure, insufficient lubricating oil, or deteriorated lubricating oil, manifesting through underlying events such as increased temperature and abnormal vibration. Impurity wear refers to impurities from the air source or environment entering the bearing and causing wear on the surface of components during bearing operation. The presence of impurities exacerbates friction and causes abnormal vibration, accelerating bearing wear. Impurity wear is manifested through underlying events such as changes in vibration frequency domain characteristics, localized temperature increases, and fluctuations in air film pressure. These two intermediate events, from the perspectives of lubrication status and the influence of impurities, cover the main direct causes of bearing wear. By correlating them with the top-level event using OR gates, the inducing factors of bearing wear can be accurately identified.

[0066] Regarding the top-level event of water ingress into the spindle, the corresponding intermediate events include excessive air source humidity and aging of seals. Excessive air source humidity is the primary cause of spindle water ingress. When the air source humidity exceeds a set threshold, moisture enters the air path system with the compressed air and seeps into the air spindle, leading to corrosion of internal components and decreased air film stability. Excessive air source humidity is manifested through underlying events such as the duration of humidity exceeding the limit and the rate of humidity change. Aging of seals is another important cause of spindle water ingress. Air spindles have seals at the air path interfaces and shaft ends to prevent moisture and impurities from entering the equipment. When these seals age, their sealing performance deteriorates, allowing external moisture to seep into the spindle through the seal gaps, causing water ingress. Aging of seals is manifested through underlying events such as sudden changes in humidity at the bearing end and fluctuations in air film pressure. These two intermediate events, from the perspectives of air source and equipment sealing, cover the main direct causes of spindle water ingress. By linking them to the top-level event using OR gates, the causes of spindle water ingress can be comprehensively identified.

[0067] Furthermore, it also includes: responding to the judgment results of the fault diagnosis model, if the identified fault is of the adjustable type, the corresponding adaptive control strategy is automatically triggered, including adjusting the gas path pressure or triggering electromagnetic damping compensation; if the identified fault is of the maintenance type, or if the fault characteristics are not eliminated after executing adaptive control, a solution containing the specific fault location and maintenance steps is generated and output.

[0068] This application also provides an air spindle monitoring and early warning system for implementing the above-mentioned air spindle monitoring and early warning method. The system includes a parameter acquisition module, a data processing module, and a fault diagnosis module, specifically including: The parameter acquisition module is used to collect multi-dimensional status data of the air spindle in real time through the deployed sensor group. The data processing module communicates with the parameter acquisition module. The data processing module is used to preprocess and extract features from the state data, generating a feature parameter set; and The fault diagnosis module communicates with the data processing module. The fault diagnosis module has a preset fault diagnosis model and is used to make logical judgments based on the feature parameter set to identify and confirm potential faults and their types.

[0069] The parameter acquisition module is used to collect multi-dimensional status data of the air spindle in real time through a deployed sensor array. This module is the core of the system's data source, and its performance directly determines the accuracy of subsequent data processing and fault diagnosis. The parameter acquisition module consists of a sensor array, a data acquisition unit, and a data transmission unit. The sensor array includes a microelectromechanical system vibration sensor deployed at the bearing end, a temperature sensor and an air film pressure sensor deployed at the air bearing, and a humidity sensor deployed at the air inlet. The data acquisition unit connects to each sensor, receiving the raw data collected by the sensors, converting analog signals into digital signals, and performing preliminary format standardization processing to ensure data format uniformity. The data transmission unit adopts industrial-grade communication technologies, including 5G industrial modules and Ethernet, to achieve real-time and stable transmission of collected data to the data processing module, ensuring the timeliness and integrity of data transmission and avoiding missed fault diagnosis due to data loss or delay. During operation, the parameter acquisition module maintains synchronous operation with the air spindle, covering all operating conditions of the equipment, ensuring that the collected data comprehensively reflects the equipment's operating status.

[0070] The data processing module communicates with the parameter acquisition module to preprocess and extract features from the status data, generating a feature parameter set. This module employs an edge computing architecture, deployed near the air spindle at the field end, enabling real-time data processing and avoiding bandwidth consumption and latency issues caused by transmitting large amounts of raw data to a remote platform. The data processing module incorporates differentiated processing algorithms, including filtering and denoising algorithms, fast Fourier transform algorithms, and moving average algorithms, each addressing the preprocessing needs of different data types. The filtering and denoising algorithm removes interference signals from vibration, temperature, film pressure, and humidity data; the fast Fourier transform algorithm converts the vibration time-domain signal to the frequency domain signal to extract vibration features; the moving average algorithm smooths the temperature and film pressure data to extract trend features; and threshold determination and rate calculation algorithms extract the duration of humidity exceedances and the rate of change. After processing, the data processing module uses a feature selection algorithm to filter out indicators strongly correlated with the fault mode, generating a standardized feature parameter set, which is then transmitted to the fault diagnosis module.

[0071] The fault diagnosis module communicates with the data processing module. This module has a preset fault diagnosis model used for logical judgment based on a set of feature parameters to identify and confirm potential faults and their types. The fault diagnosis module has a built-in fault feature mapping library and a fault tree-based diagnostic model, storing the correlation between fault types and combinations of feature parameters. The fault tree-based diagnostic model has top-level, intermediate, and bottom-level event structures and logic gate association capabilities. After receiving the feature parameter set transmitted by the data processing module, the fault diagnosis module first matches it with the fault feature mapping library to initially screen potential fault types. Then, it uses the fault tree diagnostic model to perform top-down logical deduction to accurately confirm the fault type and root cause. Finally, the diagnostic results are transmitted to the early warning module and the control execution module, providing a basis for subsequent early warning and control. Furthermore, the fault diagnosis module has self-optimization capabilities, continuously updating the fault feature mapping library and fault tree model based on actual operating data and fault handling results to improve diagnostic accuracy.

[0072] To achieve complete early warning and control functions, the system can be expanded to include an early warning module and a control execution module. The early warning module communicates with the fault diagnosis module and is used to execute tiered early warnings based on fault diagnosis results. Depending on the severity and type of the fault, different levels of early warning information are pushed, including primary and advanced warnings. Early warning information is output through audible and visual alarms, remote platform push notifications, and terminal SMS notifications to ensure that maintenance personnel are promptly informed of the fault situation. The control execution module communicates with the fault diagnosis and early warning modules and is used to execute adaptive control strategies. This includes actuators such as gas pressure regulators and electromagnetic damping devices. It can automatically trigger corresponding control operations based on fault diagnosis results and also has a fault handling result feedback function, transmitting the control effect to the fault diagnosis module in real time to form a closed-loop control system.

[0073] This application also proposes a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method. The provided computer device can be a terminal, and its internal structure diagram can be as follows: Figure 7As shown. The computer device includes a processor, memory, communication interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements the method of this application. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0074] This application also proposes a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the methods described above. Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0075] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0076] Furthermore, this application also proposes a monitoring method for an air spindle, applicable to an air spindle structure comprising a housing, a spindle, and an air bearing, wherein the spindle is rotatably connected inside the housing via the air bearing. The monitoring method for the air spindle includes: Constructing bearing wear risk factors air film attenuation risk factors Main shaft water ingress risk factors Risk of failure The mapping model; parameters strongly correlated with three types of core faults are collected by multi-dimensional sensors, and after edge computing processing, bearing wear risk factors are obtained respectively. air film attenuation risk factors Main shaft water ingress risk factors The bearing wear risk factor was calculated. air film attenuation risk factors Main shaft water ingress risk factors Substitute into the established functional relationship model to output the failure risk. This enables a global assessment of the operating status of the air spindle.

[0077] Specifically, establish bearing wear risk factors. air film attenuation risk factors Main shaft water ingress risk factors Risk of failure The functional relationship model; Bearing wear risk factors based on data collection or calculation air film attenuation risk factors Main shaft water ingress risk factors And the aforementioned functional relationship model, to determine the risk of failure. .

[0078] The functional relationship model is as follows: Failure risk

[0079] in, Bearing wear risk factor Weighting coefficients; Risk factors for air film attenuation Weighting coefficients; Main shaft water ingress risk factor Weighting coefficients; This is the trend correction coefficient.

[0080] Weighting coefficient , , The allocation based on the degree of impact of faults on the spindle (downtime losses, maintenance costs, and accuracy impact) was obtained by fitting data from over 500 sets of fault experiment data. In one embodiment, the value is: =0.4 (Bearing wear directly affects machining accuracy, is difficult to repair, and has the highest weight). =0.35 (The air film is the core of the main shaft support; its attenuation can easily lead to a chain reaction of vibration failures). =0.25 (Water ingress is mostly caused by external factors, which can be intervened early, so the weight is relatively low). The weight satisfies... + + =1, ensuring that the value range of the weighted summation part is [0,1].

[0081] Trend correction coefficient This is used to amplify the impact of the failure degradation rate on the overall risk. Experimental verification showed that the value was [value missing]. When the value is 0.1, it can effectively reflect trend changes without excessively amplifying the impact of the rate, thus ensuring overall risk. The value range is [0, 1.2].

[0082] Fault deterioration rate This rate is the time derivative of each risk factor, reflecting how quickly the fault worsens. For example, if the air film attenuation risk factor R2 rises from 0.3 to 0.5 in only 10 minutes, then... This indicates that the fault is deteriorating rapidly and the risk is higher; take the maximum value among the three. Focus on the most dangerous deteriorating trends.

[0083] Establish vibration kurtosis Effective value of vibration velocity Bearing end temperature Bearing wear risk factors The first functional relationship model; Vibration kurtosis based on acquisition or calculation Effective value of vibration velocity Bearing end temperature And the first functional relationship model, to determine the bearing wear risk factors. .

[0084] The first functional relationship model is: Bearing wear risk factors

[0085] in, For vibration kurtosis Weighting coefficients; Effective value of vibration velocity Weighting coefficients; Bearing end temperature Weighting coefficients; This is the reference value for the normal operating temperature of the bearing; This is the first error correction term.

[0086] Vibration kurtosis A dimensionless parameter characterizing the deviation between the peak and effective values ​​of the vibration signal. During bearing wear, the impact between metal particles significantly increases the peak value of the vibration signal, and consequently, the kurtosis value. The effective value of the vibration velocity is calculated by acquiring time-domain acceleration signals of the X / Y / Z axes (sampling frequency ≥ 1kHz) using a MEMS vibration sensor at the bearing end. The root mean square (RMS) value of vibration velocity reflects the magnitude of vibration energy. Increased bearing wear leads to a continuous rise in vibration energy. The velocity time-domain signal is obtained by integrating the acceleration signal acquired by the MEMS sensor, and then the RMS value is calculated; bearing end temperature... Real-time temperature of the bearing outer ring; bearing wear increases frictional heat, and the temperature rises with the degree of wear. This temperature is directly collected by a high-precision temperature sensor embedded in the bearing end; weighting coefficient. , , The correlation strength between parameters and bearing wear is obtained through fitting; in one embodiment, =0.4 (Vibration kurtosis is most sensitive to wear impact). =0.35 (The effective value of vibration velocity reflects the accumulation of vibration energy); =0.25 (Temperature correction factor, taking into account the hysteresis of frictional heat). Reference value for normal bearing operating temperature. The normal operating temperature reference value for the bearing was calibrated experimentally. First error correction term Used to reduce the impact of data noise on results, with a value range of [value range missing]. ≤0.03, obtained after noise reduction of vibration and temperature data using the Kalman filter algorithm.

[0087] Establishing air film pressure deviation Pressure attenuation slope Peak vibration frequency With air film attenuation risk factor The second functional relationship model; Based on the collected or calculated air film pressure deviation Pressure attenuation slope Peak vibration frequency And the second functional relationship model, to determine the risk factors of air film attenuation. .

[0088] The second functional relationship model is as follows: air film attenuation risk factor

[0089] in, For film pressure deviation Weighting coefficients; Pressure decay slope Weighting coefficients; Peak vibration frequency Weighting coefficients; This is the reference value for normal air film frequency; This is the second error correction term.

[0090] air film pressure deviation This is a dimensionless parameter characterizing the percentage deviation between the actual pressure and the rated pressure of the gas film. Gas film attenuation or leakage will lead to an increase in pressure deviation; the pressure attenuation slope... The slope represents the rate of decrease in air film pressure per unit time, reflecting the degree of air leakage; a steeper slope indicates a more severe leakage. The slope is obtained by linearly fitting continuously collected pressure data (sampling frequency ≥ 500 Hz) from the pressure sensor; the peak vibration frequency is also measured. The peak frequency in the power spectrum of the vibration signal is the characteristic frequency. Air film decay induces principal axis resonance, and the peak frequency increases in a specific frequency band (500-1500Hz). Acceleration signals are acquired using a MEMS vibration sensor, converted to a frequency domain signal via Fast Fourier Transform (FFT), and the peak frequency in this band is extracted. Weighting coefficients are used. , , Based on the correlation strength fitting between the parameters and gas film attenuation, the recommended values ​​are: =0.5 (Air film pressure deviation directly reflects the support capacity and has the highest weight). =0.3 (The pressure decay slope reflects the leakage rate). =0.2 (peak vibration frequency to aid in verifying air film stability); normal air film frequency reference value The characteristic frequency reference value corresponding to a normal air film can be calibrated experimentally; the second error correction term... Used to improve the stability of model output, with a value range of [value missing]. ≤0.02, obtained by smoothing the pressure data using a moving average filter.

[0091] Establish humidity mutation rate Temperature difference between shaft and housing relative humidity Risk factors for water ingress into the main shaft The third functional relationship model; Humidity abrupt change rate based on data collection or calculation Temperature difference between shaft and housing relative humidity And the third functional relationship model, to determine the risk factors of water ingress into the main shaft. .

[0092] The third functional relationship model is as follows: Main shaft water ingress risk factors

[0093] in, Humidity abrupt change rate Weighting coefficients; Temperature difference between shaft and housing Weighting coefficients; relative humidity Weighting coefficients; The humidity reference value for normal spindle operation; This is the third error correction term.

[0094] Humidity mutation rate The change in relative humidity per unit time is measured, and the internal humidity will experience a momentary jump when water enters the spindle. The relative humidity is collected by a humidity sensor (accuracy ±2%RH) inside the spindle cavity, and the result is calculated. The temperature difference between the spindle body and the housing is also considered. The temperature difference between the spindle surface and the spindle housing is measured by the temperature difference between the two. Water ingress into the spindle accelerates heat dissipation from the spindle, increasing the temperature difference. This temperature difference is obtained by measuring the temperature difference between the corresponding temperature sensors located at the mid-section of the spindle and the housing. Relative humidity is also considered. The real-time relative humidity value inside the spindle cavity directly reflects the internal moisture content. It is collected directly by a humidity sensor, with units of %RH; weighting coefficients are used. , , Based on the correlation strength fitting between the parameters and the influent, the recommended values ​​are: =0.45 (humidity abrupt change rate is most sensitive to ingress water); =0.3 (Temperature difference reflects the change in heat dissipation after water enters the system); =0.25 (relative humidity correction factor, considering the influence of steady-state humidity); normal operating humidity reference value for the spindle. The normal operating humidity reference value for the spindle, calibrated experimentally, is H0 = 45%RH. The humidity is normalized in this form to adapt to different environmental conditions; the third error correction term Used to adapt to complex scenarios with coupled temperature and humidity, the value range is: ≤0.03, obtained after completing the humidity data through linear interpolation.

[0095] This application also includes an air spindle monitoring system for performing the above-described monitoring method, comprising the following modules: Function creation module: Used to create bearing wear risk factors. air film attenuation risk factors Main shaft water ingress risk factors Overall failure risk The functional relationship model includes the algorithmic implementation of the three sub-fault risk models and the overall risk model mentioned above.

[0096] Data acquisition module: used to collect or calculate bearing wear risk factors. air film attenuation risk factors Main shaft water ingress risk factors It includes multi-dimensional sensors (MEMS vibration sensor, high-precision temperature sensor, air film pressure sensor, humidity sensor) and edge computing units to complete data acquisition, preprocessing (noise reduction, completion, standardization) and feature extraction.

[0097] Risk Confirmation Module: Used for identifying bearing wear risk factors based on collected or calculated data. air film attenuation risk factors Main shaft water ingress risk factors And functional relationship models to determine overall failure risk This module is deployed on a cloud server and communicates with the edge computing unit via a 5G industrial module to run the risk quantification model in real time and output the risk level.

[0098] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0099] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for monitoring and early warning of air spindles, characterized in that, include: The sensor array deployed on the air spindle collects real-time status data during the operation of the air spindle. The collected state data is preprocessed and features are extracted to generate a set of feature parameters characterizing the operating state of the equipment. Based on the set of feature parameters, a preset fault diagnosis model is used to identify whether there is a potential fault in the air spindle, and to determine the fault type when a fault is identified.

2. The air spindle monitoring and early warning method according to claim 1, characterized in that, The real-time acquisition of the air spindle's operational status data through a sensor array deployed on the air spindle includes: The status data includes vibration data, temperature data, air film pressure data, and humidity data; A microelectromechanical system vibration sensor is installed at the bearing end of the air spindle to collect multidimensional vibration data; A temperature sensor and an air film pressure sensor are installed at the air bearing of the air spindle to collect the temperature data and the air film pressure data. A humidity sensor is installed at the air inlet of the air shaft to collect the humidity data.

3. The air spindle monitoring and early warning method according to claim 2, characterized in that, The preprocessing and feature extraction of the collected state data to generate a set of feature parameters characterizing the device's operating state includes: The vibration data collected by the vibration sensor of the microelectromechanical system is filtered and denoised, and frequency domain features are extracted by fast Fourier transform. The temperature data and air film pressure data collected by the temperature sensor and air film pressure sensor are smoothed to extract trend change characteristics. The duration and instantaneous rate of change of humidity exceeding the standard are calculated from the humidity data collected by the humidity sensor. Based on the processed data, feature indicators that are strongly correlated with preset fault modes are selected to form the feature parameter set.

4. The air spindle monitoring and early warning method according to claim 3, characterized in that, The feature parameter set includes: Amplitude values ​​of specific frequency components related to bearing wear extracted from vibration data; The mean, trend, and fluctuation indicators extracted from temperature and film pressure data; and The duration of exceeding the standard and the rate of change were extracted from humidity data.

5. The air spindle monitoring and early warning method according to claim 3 or 4, characterized in that, The judgment based on the feature parameter set and using a preset fault diagnosis model includes: The extracted feature parameter set is matched one by one with the fault feature mapping library to form a fault feature mapping library, and the feature parameter combination corresponding to the potential fault is selected. Among them, the characteristic parameter combination of the continuous decrease in air film pressure, the rate of temperature rise, and the low-frequency change of vibration amplitude is corresponding to the air film attenuation. The characteristic parameter combination of vibration frequency domain characteristic peak value and local abnormal temperature rise value corresponding to bearing wear; The characteristic parameter combination of spindle water inlet corresponding to the duration of excessive humidity at the air inlet, sudden change value of humidity at the bearing end, and air film pressure fluctuation coefficient; Based on the matching results, potential fault types that match the current combination of feature parameters are initially screened out.

6. The air spindle monitoring and early warning method according to claim 5, characterized in that, The preset fault diagnosis model is a fault tree-based diagnosis model, and the method for constructing the fault tree includes: The fault tree uses at least one of the following as top-level events: air film decay, bearing wear, and water ingress into the spindle; the direct causes of each top-level event are intermediate events; and the judgment conditions of each indicator in the feature parameter set are bottom-level events. The diagnostic model associates events at each layer through logic gates, and is used to determine the fault type by executing logic from the bottom layer events to the top layer events based on the input of the feature parameter set.

7. The air spindle monitoring and early warning method according to claim 6, characterized in that, The intermediate events include: Intermediate events corresponding to film degradation include insufficient gas path pressure and increased bearing clearance. Intermediate events corresponding to bearing wear include insufficient lubrication, wear caused by impurities; and Intermediate events corresponding to water ingress into the spindle include excessive humidity in the air source and aging of seals.

8. An air spindle monitoring and early warning system, characterized in that, include: The parameter acquisition module is used to collect multi-dimensional status data of the air spindle in real time through the deployed sensor group; A data processing module is communicatively connected to the parameter acquisition module. The data processing module is used to preprocess and extract features from the state data to generate a feature parameter set. as well as The fault diagnosis module is communicatively connected to the data processing module. The fault diagnosis module has a preset fault diagnosis model and is used to perform logical judgments based on the feature parameter set to identify and confirm potential faults and their types.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 7.

Images