Bearing processing equipment operation quality detection system and method based on data analysis

Abstract

The present invention discloses a bearing processing equipment operation quality detection system and method based on data analysis, which specifically relates to the technical field of industrial equipment, and includes the following steps: collecting and preprocessing the operation data of the bearing processing equipment through a fixed time window to generate a data set; extracting statistical features and change features, performing stability analysis, and generating a production stability index; combining historical fault records to perform fault analysis and generating a fault risk index; inputting the two indexes into a pre-trained machine learning model to classify the processing accuracy of the equipment as high-precision or low-precision processing; if it is low-precision processing, triggering an early warning mechanism, and dynamically optimizing the sampling frequency according to the equipment status to improve the monitoring efficiency and the accuracy of fault prediction; the present invention can comprehensively monitor the operation status of the equipment, reflect its stability, and can dynamically adjust the data sampling frequency, and can provide more frequent monitoring in case of high risk or instability to ensure the acquisition of key data.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial equipment, and more specifically, to a system and method for detecting the operating quality of bearing processing equipment based on data analysis. Background Art

[0002] Bearings are key components in mechanical equipment, and their processing quality directly affects the performance and service life of bearings. In modern industry, bearing processing equipment is widely used in high-precision industries such as automobiles, aerospace, and machine tools, so the requirements for processing quality are extremely strict. However, due to the long-term operation of processing equipment, problems such as component wear, vibration fluctuations, and mechanical deviations may lead to a decrease in processing accuracy, thereby affecting product quality.

[0003] Traditional equipment quality detection is usually based on changes in a single index, but it cannot comprehensively capture the operating state and potential faults of the equipment, and the fixed sampling frequency cannot adapt to the changes in the equipment operating state, which may lead to the omission of abnormal information of high-risk equipment, or waste of sampling resources in a low-risk state. Moreover, traditional methods can only identify after obvious faults occur in the equipment, making it difficult to predict and prevent the occurrence of faults in a timely manner, resulting in a decrease in production efficiency or downtime losses. Therefore, a system and method for detecting the operating quality of bearing processing equipment based on data analysis are proposed here to solve the above problems. Summary of the Invention

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A method for detecting the operating quality of bearing processing equipment based on data analysis, comprising the following steps:

[0006] Collect the operating data of the bearing processing equipment through a fixed time window and perform preprocessing to obtain a bearing processing equipment data set;

[0007] Based on the bearing processing equipment data set, extract statistical features and change features describing the equipment working state, and perform stability analysis;

[0008] Based on the bearing processing equipment data set and the historical fault records of the bearing processing equipment, perform fault analysis;

[0009] Input the stability analysis result and the fault analysis result into a pre-trained machine learning model together to classify the equipment processing accuracy into high-precision processing and low-precision processing;

[0010] According to the equipment processing accuracy classification result, decide whether to issue a warning signal and optimize and adjust the sampling frequency of the operating data.

[0011] In a preferred embodiment, the stability analysis generates a production stability index, and the fault analysis generates a fault risk index.

[0012] In a preferred embodiment, the preprocessing includes data denoising, standardization, and filling in missing values to generate structured time series data.

[0013] In a preferred embodiment, statistical features and change features describing the working state of the device are obtained. The statistical features refer to: obtaining the actual machining coordinates and the preset standard coordinates at each sampling moment within a fixed time window, then calculating the Euclidean distance value between the actual machining coordinates and the preset standard coordinates, and then respectively obtaining the average value one, standard deviation one, and skewness value one of all the Euclidean distance values;

[0014] The change features refer to: respectively obtaining the vibration signals of each machining equipment component at each sampling moment within a fixed time window, then performing Fourier transform to obtain the frequency spectrum signal, and obtaining the corresponding frequency spectrum slope of the machining equipment component, summing up all the frequency spectrum slopes at the same sampling moment to obtain the time series data of the fluctuation value, and then respectively obtaining the average value two, standard deviation two, and skewness value two corresponding to the time series data of the fluctuation value.

[0015] In a preferred embodiment, the acquisition logic of the production stability index is as follows:

[0016] Calculate the stability influence coefficient one based on the average value one, standard deviation one, and skewness value one: ; represents the average value one, represents the standard deviation one, represents the skewness value one, , , are all preset proportional coefficients, represents the stability influence coefficient one; ; represents the average value two, represents the standard deviation two, represents the skewness value two, represents the stability influence coefficient two;

[0017] The calculation formula for the production stability index is:

[0018] ; , represent the preset mapping coefficients, represents the production stability index.

[0019] In a preferred embodiment, the acquisition logic of the fault risk index is as follows:

[0020] From the bearing processing equipment dataset in the current time window, obtain various preset processing parameters to get the first set of processing parameters. From the historical fault records of the bearing processing equipment, obtain multiple second sets of processing parameters. Calculate the cosine similarity between the first set of processing parameters and each second set of processing parameters respectively, and then take the maximum value of the cosine similarity as the risk value. Then substitute it into the fault risk index calculation formula:

[0021] ; represents the risk value, represents the risk trend function, represents the fault risk index.

[0022] In a preferred embodiment, the risk trend function refers to:

[0023] ; represents the usage duration since the last fault repair, represents a preset non-zero adjustment factor.

[0024] In a preferred embodiment, classifying the equipment processing precision into high-precision processing and low-precision processing refers to:

[0025] Take both the fault risk index and the production stability index in the current time window as the input variables of a pre-trained fuzzy logic controller, and the equipment processing precision type as the output variable. Perform fuzzy processing on the input variables to convert the values of the input variables into fuzzy sets, perform fuzzy processing on the output variable to convert the output variable into a fuzzy set, formulate fuzzy rules to describe the fitness of each equipment processing precision type under different combinations of data types, and perform reasoning on the fuzzy input variables through the fuzzy rules to obtain the classification type of the bearing processing equipment. And when the classification type is low-precision processing, activate the warning mechanism to send a warning signal.

[0026] In a preferred embodiment, optimizing the sampling frequency of the operation data refers to:

[0027] ; represents the preset basic sampling frequency, 、 are both preset non-zero adjustment coefficients, and 、 the sum of them is one, represents the optimized sampling frequency of the operation data, which is applied to obtain data in the next time window.

[0028] In a preferred embodiment, a bearing processing equipment operation quality detection system based on data analysis includes:

[0029] A data acquisition module that collects the operation data of the bearing processing equipment through a fixed time window;

[0030] A data preprocessing module that preprocesses the data collected by the data acquisition module to obtain a bearing processing equipment dataset;

[0031] A stability analysis module that extracts statistical features and change features describing the working state of the equipment based on the bearing processing equipment dataset and conducts stability analysis;

[0032] A fault analysis module that conducts fault analysis based on the bearing processing equipment dataset and the historical fault records of the bearing processing equipment;

[0033] An accuracy classification module that inputs the stability analysis results and the fault analysis results into a pre-trained machine learning model to classify the equipment processing accuracy into high-precision processing and low-precision processing;

[0034] A fault warning module that activates the warning mechanism and issues a warning signal when the classification type is low-precision processing;

[0035] An optimization adjustment module that optimizes and adjusts the operation data sampling frequency of the next time window.

[0036] The technical effects and advantages of the present invention:

[0037] Through the combination of statistical features (such as mean, standard deviation, skewness) and change features (such as spectral slope), the present invention can comprehensively monitor the operation state of the equipment and reflect its stability. The calculation of the Production Stability Index (PSI) and the Fault Risk Index (FRI) further enhances the comprehensiveness and accuracy of equipment state monitoring.

[0038] Based on the real-time state of the equipment and combined with the Production Stability Index and the Fault Risk Index, the present invention can dynamically adjust the data sampling frequency. The sampling frequency of high-risk or unstable equipment will be automatically increased to obtain more monitoring data, which helps to detect potential faults in a timely manner; while for low-risk or stable equipment, the sampling frequency will be reduced to save resources and avoid unnecessary data processing.

[0039] By combining the stability and risk level of the equipment with the processing accuracy classification, the present invention can adjust the production process according to the processing accuracy of the equipment to ensure that the processing accuracy of each equipment always meets the production requirements. When the equipment is classified as low-precision processing, the system will automatically start the warning mechanism to prompt the operator to make adjustments or stop for inspection to prevent production quality problems caused by equipment failures.

[0040] By dynamically adjusting the sampling frequency, the present invention can provide more frequent monitoring during high-risk or unstable situations to ensure the acquisition of critical data. When the device is operating stably, it reduces resource consumption, improves the efficiency of data collection, and reduces the storage and computing burdens. Combining fuzzy logic reasoning and data analysis, the present invention provides more intelligent and scientific decision-making support for device management. It can adjust the device monitoring strategy in real time to help managers optimize the maintenance plan and production arrangement according to the device health status. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] For the convenience of those skilled in the art to understand, the present invention will be further described below in conjunction with the accompanying drawings;

[0042] Figure 1 It is a schematic diagram of the method for detecting the operation quality of bearing processing equipment based on data analysis in the present invention.

[0043] Figure 2 It is a schematic diagram of the system for detecting the operation quality of bearing processing equipment based on data analysis in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0045] Refer to Figure 1 - Figure 2 The following embodiments are obtained:

[0046] Embodiment 1: As the importance of industrial equipment in the production line gradually increases, the stability of equipment operation directly affects product quality and production efficiency. Bearings are core components in mechanical equipment, and the operation status of their processing equipment is crucial for the accuracy and performance of the final product.

[0047] Limitations of traditional monitoring methods: Single monitoring index: Traditional monitoring is usually based on a single index (such as vibration) and cannot comprehensively capture the operation status of the equipment. Lack of dynamic adjustment: The fixed sampling frequency cannot adapt to the dynamic changes of equipment risks, easily leading to resource waste or data omission. Early warning lag: Traditional methods are difficult to identify fault risks in the early stage, and usually measures can only be taken after obvious faults occur in the equipment.

[0048] Potential of data-driven methods: With the progress of industrial equipment monitoring technology, more and more operation data can be collected and stored in real time. Data-driven analysis methods can combine historical fault records and use machine learning models for fault prediction and processing accuracy classification.

[0049] Purpose of the present invention: The operating quality of bearing processing equipment is directly related to the precision and service life of products. The present invention aims to comprehensively monitor the equipment status through data analysis, and improve the stability and reliability of equipment operation. By combining the current operating data with historical fault records, the present invention calculates the fault risk index, realizes the early prediction of faults, and thus reduces the risk of unplanned equipment shutdown. Dynamically adjusting the sampling frequency can effectively reduce the consumption of monitoring resources during equipment operation, and ensure the density of data collection in high-risk states. By combining the production stability index (PSI) and the fault risk index (FRI), the present invention comprehensively quantifies the equipment operating status, and uses a machine learning model to classify high-precision processing and low-precision processing, optimizing the decision-making process. In the industrial context, intelligent monitoring and predictive maintenance have become the core requirements of modern industry. The present invention is proposed to meet this demand and improve the intelligent level of equipment management through a data-driven monitoring method.

[0050] The present invention provides a method for detecting the operating quality of bearing processing equipment based on data analysis, comprising the following steps:

[0051] Collect the operating data of the bearing processing equipment through a fixed time window and perform preprocessing to obtain a bearing processing equipment data set; the operating data is the core basis for detecting and analyzing the equipment working status. Data preprocessing (such as denoising, standardization, and filling in missing values) ensures the integrity, accuracy, and consistency of the data, providing a high-quality data set for subsequent analysis. Dividing the data by time window can capture the state characteristics of the equipment within a specific time period, which is helpful for dynamic monitoring.

[0052] Based on the bearing processing equipment data set, extract the statistical characteristics and change characteristics describing the equipment working status, and perform stability analysis; evaluate the stability of equipment operation through statistical characteristics (such as mean, standard deviation, skewness value). The change characteristics (such as the spectral slope of the vibration signal) can capture the dynamic changes of the equipment state. Combine the statistical characteristics and the change characteristics to generate a production stability index (PSI), which intuitively quantifies the operating stability of the equipment and helps to identify potential problems.

[0053] Based on the bearing processing equipment data set and the historical fault records of the bearing processing equipment, perform fault analysis; the fault analysis combines the current state with the historical fault records, and identifies whether the current operating state is close to the historical fault state by calculating the similarity. Generate a fault risk index (FRI) to evaluate the current fault risk level of the equipment. Through the historical fault trend function, consider the cumulative effect of the usage time on the fault risk to ensure more accurate risk prediction.

[0054] Input the stability analysis results and fault analysis results into a pre-trained machine learning model to classify the processing accuracy of the equipment into high-precision processing and low-precision processing; by combining PSI and FRI, the pre-trained model can classify the processing accuracy of the equipment, providing a more scientific decision-making basis. High-precision processing equipment can continue to run, while low-precision processing equipment will trigger an early warning mechanism to help detect problems in a timely manner and prevent further damage.

[0055] Based on the classification results of the equipment processing accuracy, decide whether to issue an early warning signal and optimize the sampling frequency of the operation data. By reminding the operation and maintenance personnel through the early warning signal, early prevention of faults can be achieved, reducing production downtime and maintenance costs. The sampling frequency optimization adjusts the sampling frequency dynamically according to the real-time risk status of the equipment (combining PSI and FRI), ensuring denser data collection in high-risk situations and reducing the sampling frequency in low-risk situations to reduce resource waste.

[0056] The stability analysis generates a production stability index, and the fault analysis generates a fault risk index. The production stability index (PSI) is calculated by analyzing the statistical characteristics and change characteristics during the operation of the equipment and is used to quantitatively measure the stability of the equipment operation status. The following are the specific meanings of generating PSI: Quantify the equipment operation stability: Through statistical characteristics such as mean, standard deviation, and skewness values, PSI can reflect the overall volatility and regularity of the equipment operation status. A high PSI value represents large fluctuations and instability in the equipment operation status; a low PSI value indicates relatively stable equipment operation. Early identification of potential anomalies: When the equipment operation status begins to fluctuate but has not fully reached the fault level, PSI can provide an early warning to help maintenance personnel take measures before the problem spreads. For example, the change trend of PSI can indicate that the equipment may gradually enter an unstable stage. Improve processing quality: The stability of equipment operation is the basis for high-precision processing. By generating PSI through stability analysis, it can provide an important basis for classifying the processing accuracy of the equipment, ensuring that the equipment operates in a stable state to meet the processing quality requirements. Assist dynamic monitoring and decision-making: As a dynamic monitoring indicator, PSI can help factory managers understand the operation status of the equipment in real time.

[0057] The Fault Risk Index (FRI) is calculated based on the similarity between the device operation parameters in the current time window and the historical fault records, as well as the trend function of the device operation time, and is used to quantify the fault risk of the device. The following are the specific significances of generating FRI: Evaluating device fault risk: FRI can measure the proximity of the current device state to the historical fault state. The higher the similarity, the greater the risk value. The larger the FRI, the higher the fault risk of the current operation state of the device. Predicting the trend of device faults: By combining the influence of the device usage time (such as the risk trend function), FRI can dynamically reflect the risk accumulation of the device after long-term operation. The longer the device operates, the higher FRI will be with the increase of usage time, providing a scientific basis for device maintenance. Supporting preventive maintenance: Through the real-time monitoring of FRI, managers can identify potential faults in advance and arrange preventive maintenance to avoid downtime or product scrapping caused by sudden device failures. This significantly reduces the device maintenance cost and production losses. Ensuring the continuity of the production line: In manufacturing, device failures usually lead to the shutdown of the entire production line. By real-time monitoring FRI, managers can formulate maintenance plans to ensure the stable operation of the production line. Data-driven intelligent decision-making: By comparing the current operation data with the historical fault records, FRI provides an intelligent fault prediction method based on data, providing a scientific basis for the traditional experience-based maintenance mode.

[0058] The collaborative significance of the Production Stability Index (PSI) and the Fault Risk Index (FRI): Comprehensively evaluating the device operation state: PSI evaluates the stability of the device, and FRI evaluates the fault risk of the device. The combination of the two can comprehensively reflect the operation status of the device. For example: when the device operates stably but the FRI is high, it may imply that the device is about to fail; when the device operates unstably and the FRI is high, it indicates a higher fault risk. Intelligent classification and maintenance: Using PSI and FRI as input variables into a machine learning model can intelligently classify the processing accuracy of the device (high-precision processing or low-precision processing). According to the classification results, managers can conduct targeted maintenance or adjustment of the device. Optimizing production and device management: By combining PSI and FRI, the optimization of production plans and device management can be achieved. For example, reducing the load of unstable devices or arranging preventive maintenance to avoid greater losses caused by faults. Dynamically adjusting the sampling frequency: Adjust the data sampling frequency according to the changing trends of PSI and FRI. High-risk devices are sampled more intensively, and low-risk devices are sampled less, achieving efficient utilization of resources.

[0059] Preprocessing includes data denoising, standardization, and filling in missing values to generate structured time-series data. Data denoising is to eliminate random noise or outliers in the original data and retain the key information reflecting the true operating state of the device. The collected original data may be affected by sensor errors, electromagnetic interference, or environmental changes, and there is noise. Denoising can remove these irrelevant interferences and make the data more accurate. Undenoised data may lead to misjudgments, wrongly regarding noise as abnormal device states. After denoising, misjudgments can be reduced, and the reliability of fault analysis and stability analysis can be improved. Noise will mask the actual change trends in the data. After denoising, the true characteristics of the device operation can be more clearly extracted, providing support for subsequent feature analysis.

[0060] Data standardization is to uniformly process data with different dimensions or ranges so that they have the same scale, thereby eliminating the dimensional differences between different features. The data may contain features with different dimensions (such as the acceleration of vibration signals, etc.). Standardization can eliminate dimensional differences, making these features comparable in analysis.

[0061] Filling in missing values is to fill in the missing data caused by sensor failures, communication interruptions, or other reasons during data collection. Missing values in time-series data will cause discontinuities in analysis. Filling in missing values can ensure the integrity of time-series data and avoid information loss. Missing values will cause biases when calculating feature values (such as mean, standard deviation, etc.), affecting the accuracy of analysis results. Filling in missing values can reduce these effects. The device operation data has the characteristic of continuity. Filling in missing values can retain the dynamic characteristics of the data, making the subsequent state analysis more in line with reality. Time-series data is structured data arranged in chronological order and can reflect the changes in the device operation state over time. After generating structured time-series data, statistical features (mean, standard deviation, skewness) and dynamic change features (vibration spectrum, fluctuation value, etc.) of the device operation can be extracted from it. Time-series data can help identify the trends and patterns of the device operation state, support real-time monitoring and fault trend prediction. Structured time-series data is convenient for comparison with historical fault records, identifying the similarities between the current state and historical fault patterns, and supporting fault analysis. Structured time-series data has consistency and standardization and can directly be used as the input of machine learning models, improving the training efficiency and prediction accuracy of the models.

[0062] Obtain the statistical features and change features describing the working state of the device. The statistical features refer to: obtaining the actual machining coordinates and the preset standard coordinates at each sampling moment within a fixed time window, then calculating the Euclidean distance value between the actual machining coordinates and the preset standard coordinates, and then respectively obtaining the average value one, standard deviation one, and skewness value one of all Euclidean distance values; The statistical features are eigenvalue calculated by analyzing the deviation degree between the actual machining coordinates of the device and the preset standard coordinates within a fixed time window, including: average value one (μ1): representing the average level of deviation during the machining process. Standard deviation one (σ1): reflecting the volatility of the machining process. Skewness value one (γ1): reflecting the symmetry of the machining deviation distribution.

[0063] Significance: Quantify machining accuracy: The Euclidean distance calculation reflects the degree of deviation of the machining coordinates from the standard coordinates. The average value one can intuitively evaluate the overall level of machining accuracy. Capture the stability of the device operation: The standard deviation one reflects the stability of the device operation state by evaluating the volatility of the machining coordinates. Larger fluctuations may indicate abnormalities in the device, and the device operates more stably when the fluctuations are small. Diagnose deviation patterns: The skewness value one can reveal the tendency of the deviation direction of the machining coordinates. For example, positive skewness may indicate that the machining results are systematically deviated in the positive direction, while negative skewness indicates deviation in the negative direction, providing a basis for fault mode analysis. Evaluate long-term stability: The time series of statistical features can reveal the stability changes of the device during long-term operation, providing support for predictive maintenance and process improvement.

[0064] The change features refer to: respectively obtaining the vibration signals of each machining device component at each sampling moment within a fixed time window, then performing Fourier transform to obtain the spectral signal, and obtaining the spectral slope corresponding to the machining device component, summing all spectral slopes at the same sampling moment to obtain the time series data of the fluctuation value, and then respectively obtaining the average value two, standard deviation two, and skewness value two corresponding to the time series data of the fluctuation value. Average value two (μ2): reflecting the overall level of vibration fluctuation. Standard deviation two (σ2): measuring the change amplitude of vibration fluctuation. Skewness value two (γ2): describing the symmetry of the vibration fluctuation distribution.

[0065] Significance: Revealing dynamic operating status: Vibration signals can reflect the dynamic characteristics of equipment components during operation. By extracting the slope through spectral analysis, the characteristics that change rapidly during equipment operation can be captured. Detecting potential faults: Significant changes in vibration fluctuations are often early signals of equipment anomalies or faults. By observing the changing trends of the mean two and standard deviation two, the potential risks of equipment faults can be identified. Evaluating equipment operation stability: The mean two and standard deviation two can comprehensively reflect the stability of equipment operation. For example, when the vibration fluctuations are small, the equipment operates more stably; conversely, it indicates a poor operating state of the equipment. Analyzing the collaborative effect of mechanical components: The slope of vibration signals from different components comprehensively reflects the collaborative relationship between mechanical components. If the spectral slope of certain components changes significantly, it may indicate that the operating state of these components is inconsistent with that of other components, suggesting potential problems.

[0066] The acquisition logic of the production stability index is as follows:

[0067] Calculate the stability influence coefficient one based on the mean one, standard deviation one, and skewness value one:

[0068] ; represents the mean one, which represents the average level of the machining coordinates deviating from the standard coordinates within a fixed time window and reflects the overall machining accuracy. The larger the mean value, the more serious the deviation. represents the standard deviation one, which represents the fluctuation range of the machining coordinates and reflects the stability of the equipment operating state. The larger the standard deviation, the greater the equipment operation fluctuations. represents the skewness value one, which represents the asymmetry of the machining coordinate deviation distribution. A larger skewness value indicates that the deviation may have a systematic tendency (such as continuously being positive or negative). , , are all preset proportionality coefficients. represents the stability influence coefficient one; Limit the skewness value within the range of -1 to 1 to ensure the smoothness of the influence of skewness on stability.

[0069] ; represents the mean two, which represents the average spectral slope of the vibration signal and reflects the overall vibration level of the equipment. The larger the mean value, the more unstable the equipment operating state. represents the standard deviation two, which represents the fluctuation range of the spectral slope of the vibration signal and reflects the degree of change in the equipment vibration state. A larger standard deviation indicates the instability of equipment operation. represents the skewness value two, which represents the asymmetry of the spectral slope distribution and describes whether the vibration signal has a bias. An abnormal change in the skewness value may indicate a fault trend in a specific direction. represents the stability influence coefficient two;

[0070] The calculation formula for the production stability index is as follows:

[0071] ; 、 represents a preset mapping coefficient, which is used to map the parameter value into the standardized value interval. represents the production stability index. The stability influence coefficients one (WD1) and two (WD2) start from the statistical characteristics and change characteristics respectively, and balance the contributions between characteristics through non-linear transformations (such as logarithm, hyperbolic function, etc.) to avoid the excessive influence of a certain characteristic. Through the non-linear processing of the sum of squares of the spectral characteristics in WD2, the dynamic change characteristics of the equipment operation can be captured more sensitively. The final PSI formula uses the tangent function (tan) to further amplify the combined effect of the stability influence coefficients WD1 and WD2, making the instability characteristics more prominent.

[0072] The acquisition logic of the fault risk index is as follows:

[0073] From the bearing processing equipment dataset in the current time window, obtain each preset processing parameter to get the first set of processing parameters. From the historical fault records of the bearing processing equipment, obtain multiple second sets of processing parameters, calculate the cosine similarity between the first set of processing parameters and each second set of processing parameters respectively, and then take the maximum value of the cosine similarity as the risk value. Then substitute it into the calculation formula for the fault risk index:

[0074] ; represents the risk value, which represents the similarity between the current set of processing parameters and the historical fault records, and is calculated through the cosine similarity. The higher the similarity, the closer the current equipment state is to the historical fault state. represents the risk trend function, which represents the influence of the equipment operation time on the fault risk. The longer the time, the higher the fault risk. Represents the Fault Risk Index. FRI combines the dynamic characteristics of equipment operation data and historical fault records, and can quantitatively measure the current fault risk of the equipment in a numerical way. By calculating the similarity between the current processing parameters and historical fault parameters, it can accurately evaluate whether the equipment state is close to the known fault mode. Through the maximum similarity of the risk values in historical data, it can judge whether there are potential faults in the equipment. Even if there are no obvious fault signs in the equipment, but if the risk value is high, FRI can provide early warnings in time to help operation and maintenance personnel take measures in advance. The risk trend function takes into account the cumulative effect of equipment usage time, indicating that the longer the equipment runs, the greater the possibility of failure. This design conforms to the actual law of equipment failure and provides a scientific basis for equipment maintenance. FRI can dynamically monitor the equipment fault risk. As an important indicator of predictive maintenance, it helps managers repair the equipment in time before a failure occurs and reduce the downtime. The higher the FRI value, the more necessary it is to focus on monitoring the equipment or arranging a maintenance plan.

[0075] The risk trend function refers to:

[0076] ; Represents the usage duration since the last fault repair, Represents a preset non-zero adjustment factor. T represents the time that the equipment has been running continuously since the last repair. As the equipment operation time T increases, problems such as aging and wear of each component of the equipment will gradually accumulate, and the fault risk will also increase. As a time variable, T directly reflects the cumulative state of equipment operation and is the core driving factor of the risk trend function. λ is a constant used to adjust the influence degree of equipment operation time on the fault risk. The size of λ determines the risk growth rate: when λ is larger, the risk increases exponentially with time; when λ is smaller, the risk increases more slowly with time. Under different equipment or working conditions, the value of λ can be adjusted according to the actual situation to ensure the adaptability of the function. The exponential function simulates the non-linear cumulative effect of risk over time, that is, the risk growth rate will accelerate over time. This design conforms to the actual law of equipment failure: the risk is relatively low during the early operation of the equipment, but as the operation time increases, the risk shows an accelerating upward trend. Equipment in long-term operation usually experiences problems such as component wear and fatigue aging, which will accumulate over time, leading to a gradual increase in the fault risk. By associating the operation time T with the risk index, the contribution of the time cumulative effect to the equipment fault risk is quantified. The risk trend function provides the ability to dynamically adjust the calculation of the Fault Risk Index (FRI). Even if the similarity between the current equipment state and historical fault records is not high, but if the operation time T is long, the risk trend function will amplify the risk index, thus indicating potential risks.

[0077] The classification of high-precision machining and low-precision machining for equipment machining accuracy refers to:

[0078] Both the fault risk index and the production stability index of the current time window are used as input variables of a pre-trained fuzzy logic controller. The equipment machining accuracy type is used as the output variable. The input variables are fuzzified, converting the values of the input variables into fuzzy sets. The output variable is fuzzified, converting the output variable into a fuzzy set. Fuzzy rules are formulated to describe the fitness of each equipment machining accuracy type under different combinations of data types. The fuzzified input variables are inferred through the fuzzy rules to obtain the classification type of the bearing processing equipment. And when the classification type is low-precision machining, the warning mechanism is activated to send a warning signal.

[0079] When using the pre-trained fuzzy logic controller, the input variables are defined. The input variables include: Fault risk index: Quantifies the similarity between the current operating state of the equipment and the historical fault records and the time cumulative risk. Production stability index: Quantifies the stability of the equipment during operation, reflecting the coordinate deviation and dynamic vibration characteristics of the equipment machining.

[0080] The specific values of the input variables are converted into fuzzy sets. The fuzzy sets are represented in the form of linguistic variables, such as "low", "medium", etc.: Fault risk index: Low risk, Medium risk, High risk; Production stability index: Stable, Medium stable, Unstable. Through the membership function, the values of the input variables are mapped to the membership degrees in the fuzzy sets. The membership degree ranges from 0 to 1, indicating the degree of belonging of the input value to each fuzzy set.

[0081] The output variable is defined: The output variable is the equipment machining accuracy type, represented by a linguistic variable: High-precision machining: The equipment is in good operating condition, and the machining result meets the high-precision requirements. Low-precision machining: The equipment is in a poor operating condition, and the machining result may not meet the precision requirements. Fuzzification process: Similar to the input variables, the classification result of the output variable is also represented by a fuzzy set: High-precision, Medium-precision, Low-precision. The fuzzification result of the output variable is used for comprehensive inference, and finally provides the membership degree of the equipment machining accuracy classification.

[0082] Formulation of fuzzy rules: The fuzzy rules describe the fitness of different input variable combinations to the output variable. For example: If the fault risk index is "Low risk" and the production stability index is "Stable", then the equipment machining accuracy is "High-precision". If the fault risk index is "High risk" and the production stability index is "Unstable", then the equipment machining accuracy is "Low-precision". Fuzzy rules are usually generated by expert experience or data training, covering all possible input combinations.

[0083] Fuzzy Inference: Rule Matching: Based on the fuzzy sets and membership degrees of the input variables, the fuzzy logic unit matches the corresponding rules. The contribution of each rule to the output variable is determined by the membership degrees of the input variables. Membership Degree Synthesis: The fuzzy logic unit synthesizes the results of multiple rules to obtain the membership degrees of the output variable in each fuzzy set. The commonly used synthesis methods are the "max-min method" or the "weighted average method".

[0084] Defuzzification: Defuzzification: Converts the result of fuzzy inference (the fuzzy set and membership degree of the output variable) into a specific numerical value. Defuzzification methods include the "centroid method" or the "maximum membership degree method", etc. The defuzzification result is used for the final classification. Output Classification: According to the defuzzification result, the processing accuracy of the equipment is divided into: High-precision processing: The equipment operates stably, with low risk, and the processing result meets the requirements. Low-precision processing: The equipment operates with large fluctuations, with high risk, and the processing result does not meet the requirements.

[0085] Optimally adjusting the sampling frequency of the operating data means that:

[0086] ; represents the preset basic sampling frequency, that is, the sampling frequency required for the equipment in the normal operating state, 、 are both preset non-zero adjustment coefficients, controlling the influence degrees of the production stability index (PSI) and the fault risk index (FRI) on the sampling frequency, and 、 the sum is one, Denote the sampling frequency of the operation data after optimization adjustment, which is applied to obtain data in the next time window. The production stability index reflects the stability of the equipment operation. The higher the PSI, the greater the fluctuations of the equipment and the stronger the instability. The fault risk index quantifies the fault risk level of the current operation state of the equipment. The higher the FRI, the closer the equipment is to the potential fault state. The formula incorporates the production stability index (PSI) and the fault risk index (FRI) into the calculation of the sampling frequency, enabling the sampling frequency to be dynamically adjusted according to the changes in the equipment operation state: When the equipment has high volatility (high PSI): the sampling frequency increases to capture subtle changes in the equipment operation with a higher sampling density. When the fault risk is high (high FRI): the sampling frequency increases to monitor the equipment state in real time and avoid the expansion of faults. When the equipment state is stable and the fault risk is low: the sampling frequency is appropriately reduced to reduce unnecessary data acquisition and processing costs. A high sampling frequency can provide more intensive operation data, helping to capture subtle anomalies in the equipment operation and improving the accuracy of fault detection and status monitoring. Through dynamic sampling, it is possible to capture the drastic fluctuations and potential risks of the equipment state more promptly. When the equipment state is stable, by reducing the sampling frequency, the amount of data collected is reduced, and the storage and processing costs are lowered. The dynamic allocation of resources improves the efficiency of the monitoring system, enabling the monitoring resources to be concentrated on high-risk equipment or high-volatility periods. Dynamically adjusting the sampling frequency can help the operation and maintenance personnel to grasp the changes in the equipment state in real time and provide accurate data support for predictive maintenance. When the sampling frequency increases, potential faults can be detected earlier, reducing the probability of sudden faults.

[0087] Embodiment 2: An operation quality detection system for bearing processing equipment based on data analysis, comprising:

[0088] A data acquisition module that collects the operation data of the bearing processing equipment through a fixed time window;

[0089] A data preprocessing module that preprocesses the data collected by the data acquisition module to obtain a bearing processing equipment data set;

[0090] A stability analysis module that extracts statistical features and change features describing the equipment working state based on the bearing processing equipment data set and conducts stability analysis;

[0091] A fault analysis module that conducts fault analysis based on the bearing processing equipment data set and the historical fault records of the bearing processing equipment;

[0092] An accuracy classification module that inputs the stability analysis results and the fault analysis results into a pre-trained machine learning model together to classify the equipment processing accuracy into high-precision processing and low-precision processing;

[0093] A fault warning module that activates the warning mechanism and issues a warning signal when the classified type is low-precision processing;

[0094] An optimization adjustment module optimizes and adjusts the sampling frequency of the operation data for the next time window.

[0095] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0096] It should be understood that in various embodiments of the present application, the magnitudes of the sequence numbers of the above processes do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0097] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0098] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0099] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present application, and all should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claimed rights.

Claims

1. A method for detecting the operation quality of bearing processing equipment based on data analysis, characterized in that: The following steps are involved: The operation data of the bearing processing equipment is collected through a fixed time window and preprocessed to obtain a bearing processing equipment data set; Based on the data set of bearing processing equipment, the statistical characteristics and change characteristics describing the working status of the equipment are extracted to perform stability analysis; Conduct fault analysis based on the bearing processing equipment data set and historical fault records of the bearing processing equipment; The stability analysis results and the failure analysis results are input into the pre-trained machine learning model to classify the equipment processing accuracy into high-precision processing and low-precision processing; According to the classification results of equipment processing accuracy, decide whether to issue a warning signal and optimize the sampling frequency of operating data; Obtain statistical features and change features that describe the working status of the equipment. The statistical features refer to: obtaining the actual processing coordinates and the preset standard coordinates at each sampling moment in a fixed time window, then calculating the Euclidean distance value between the actual processing coordinates and the preset standard coordinates, and then respectively calculating the average value, standard deviation, and skewness value of all Euclidean distance values; The variation characteristics refer to: obtaining the vibration signal of each processing equipment component in a fixed time window at each sampling moment, then performing Fourier transform to obtain the spectrum signal, and obtaining the spectrum slope corresponding to the processing equipment component, summing up all spectrum slopes at the same sampling moment to obtain the time series data of the fluctuation value, and then respectively calculating the mean value 2, standard deviation 2, and skewness value 2 corresponding to the time series data of the fluctuation value; Stability analysis generates a production stability index, and failure analysis generates a failure risk index; The classification of equipment processing accuracy into high-precision processing and low-precision processing refers to: The fault risk index and production stability index of the current time window are used as input variables of the pre-trained fuzzy logic device, and the equipment processing precision type is used as the output variable. The input variables are fuzzified and the values ​​of the input variables are converted into fuzzy sets. The output variables are fuzzified and the output variables are converted into fuzzy sets. Fuzzy rules are formulated to describe the adaptability of the processing precision types of each equipment under different data type combinations. The fuzzified input variables are inferred through fuzzy rules to obtain the classification type of bearing processing equipment. When the classification type is low-precision processing, the early warning mechanism is activated and an early warning signal is issued. Optimizing the sampling frequency of operating data means: ; Indicates the preset basic sampling frequency, , are all preset non-zero adjustment coefficients, and , The sum of is one, Indicates the optimized and adjusted operating data sampling frequency, and the application acquires data in the next time window. represents the failure risk index, Represents the production stability index.

2. The method for detecting the operation quality of bearing processing equipment based on data analysis according to claim 1 is characterized in that: Preprocessing includes data denoising, standardization, and missing value filling to generate structured time series data.

3. The method for detecting the operation quality of bearing processing equipment based on data analysis according to claim 2 is characterized in that: The logic for obtaining the production stability index is: Calculate the stability influence coefficient based on the mean value, standard deviation and skewness value: ; represents the average value of one, represents a standard deviation of one, represents a skewness value of one, , , are preset proportional factors. represents the stability influence coefficient of one; ; represents the average value 2, represents the standard deviation of two, represents the skewness value two, represents the stability influence coefficient two; The calculation formula of production stability index is: ; , Represents the preset mapping coefficient, Represents the production stability index.

4. The method for detecting the running quality of bearing processing equipment based on data analysis according to claim 3 is characterized in that: The logic for obtaining the fault risk index is: From the data set of the bearing processing equipment in the current time window, various preset processing parameters are obtained to obtain processing parameter group 1. From the historical fault records of the bearing processing equipment, multiple processing parameter groups 2 are obtained. The cosine similarity between processing parameter group 1 and each processing parameter group 2 is calculated respectively, and then the maximum cosine similarity is used as the risk value, and then substituted into the fault risk index calculation formula: ; represents the risk value, represents the risk trend function, Indicates the failure risk index.

5. The method for detecting the running quality of bearing processing equipment based on data analysis according to claim 4 is characterized in that: The risk trend function is: ; Indicates the usage time since the last fault repair. Indicates a preset non-zero adjustment factor.

6. A bearing processing equipment operation quality detection system based on data analysis, used to implement the bearing processing equipment operation quality detection method based on data analysis according to any one of claims 1 to 5, characterized in that: include: Data acquisition module, which collects the operating data of bearing processing equipment through a fixed time window; A data preprocessing module preprocesses the data collected by the data acquisition module to obtain a bearing processing equipment data set; The stability analysis module extracts the statistical characteristics and change characteristics describing the working status of the equipment based on the bearing processing equipment data set, and performs stability analysis; Fault analysis module, which performs fault analysis based on the bearing processing equipment data set and the historical fault records of the bearing processing equipment; The precision classification module inputs the stability analysis results and the fault analysis results into the pre-trained machine learning model to classify the equipment processing accuracy into high-precision processing and low-precision processing; Fault warning module, when the classification type is low-precision processing, activates the warning mechanism and sends out a warning signal; The optimization and adjustment module optimizes and adjusts the sampling frequency of the operating data in the next time window.

Images