Machine tool running state monitoring and fault early warning system

Abstract

The invention provides a machine tool running state monitoring and fault early warning system, belongs to the technical field of machine tool state monitoring, and aims at solving the problems that a traditional machine tool is single in monitoring parameter, ungraded in early warning and low in fault diagnosis accuracy. The system comprises a data acquisition module, a data processing module, an analysis module, an early warning module, a historical data storage module and a parameter configuration module which are connected according to a preset relation; the data acquisition module acquires multi-dimensional state data through multiple types of sensors; the data processing module preprocesses the data; the analysis module diagnoses a fault risk by using a trained deep learning model; the early warning module realizes graded early warning; the system can comprehensively monitor the state of the machine tool, improve fault diagnosis accuracy and early warning effectiveness, enhance adaptability, reduce fault shutdown and guarantee stable operation of the machine tool.

Description

Technical Field

[0001] This invention relates to the field of machine tool condition monitoring technology, specifically to a machine tool operating condition monitoring and fault early warning system. Background Technology

[0002] Traditional machine tool operation status monitoring relies heavily on manual inspections or single-parameter monitoring, which has significant limitations: manual inspections lack real-time performance and struggle to capture instantaneous abnormalities; single parameters (such as monitoring only temperature or vibration) cannot comprehensively reflect the complex operating conditions of the machine tool, easily leading to missed or misdiagnosed faults; fault warnings are often simple alarms lacking a tiered mechanism, making it difficult to differentiate risk levels, frequently resulting in wasted maintenance resources or overlooking critical faults. Furthermore, traditional systems lack historical data accumulation and model optimization mechanisms, leading to a decline in diagnostic accuracy over long-term use and difficulty adapting to changes in machine tool conditions during long-term operation, resulting in high downtime due to faults and impacted production efficiency.

[0003] To address this, a machine tool operation status monitoring and fault early warning system is proposed. Summary of the Invention

[0004] The present invention aims to solve the problems mentioned in the background art by providing a machine tool operation status monitoring and fault early warning system.

[0005] The specific technical solution is as follows: A machine tool operation status monitoring and fault early warning system includes a data acquisition module, a data processing module, an analysis module, and an early warning module. The data acquisition module is connected to the data processing module, the data processing module is connected to the analysis module, and the analysis module is connected to the early warning module. The data acquisition module is used to collect various status data during machine tool operation. The data processing module is used to preprocess the collected status data. The analysis module is used to analyze the preprocessed status data to determine whether there is a fault risk in the machine tool. The early warning module is used to issue an early warning message when the analysis module determines that there is a fault risk.

[0006] The aforementioned machine tool operation status monitoring and fault early warning system includes a data acquisition module comprising a vibration sensor, a temperature sensor, a current sensor, and a voltage sensor. The vibration sensor is used to collect vibration data of the machine tool, the temperature sensor is used to collect temperature data of key components of the machine tool, the current sensor is used to collect current data of the machine tool during operation, and the voltage sensor is used to collect voltage data of the machine tool during operation.

[0007] The aforementioned machine tool operation status monitoring and fault early warning system includes a data processing module comprising a data cleaning unit, a data denoising unit, and a feature extraction unit. The data cleaning unit is used to remove outliers and missing values ​​from the status data. The data denoising unit is used to denoise the cleaned status data. The feature extraction unit is used to extract feature parameters from the denoised status data.

[0008] The aforementioned machine tool operation status monitoring and fault early warning system includes an analysis module comprising a model training unit and a fault diagnosis unit. The model training unit is used to train a fault diagnosis model using historical status data and corresponding fault records. The fault diagnosis unit is used to input preprocessed status data into the trained fault diagnosis model to determine the fault risk.

[0009] The aforementioned machine tool operation status monitoring and fault early warning system includes an early warning module comprising an early warning level judgment unit and an early warning execution unit. The early warning level judgment unit is used to determine the early warning level based on the judgment result of the analysis module, and the early warning execution unit is used to issue corresponding early warning signals according to different early warning levels.

[0010] The aforementioned machine tool operation status monitoring and fault early warning system includes an early warning execution unit comprising an audible and visual alarm and a communication module. The audible and visual alarm is used to issue audible and visual early warning signals, and the communication module is used to send early warning information to a designated terminal.

[0011] The aforementioned machine tool operation status monitoring and fault early warning system also includes a historical data storage module, which is connected to the data processing module and is used to store preprocessed status data, corresponding analysis results, and early warning records.

[0012] In the aforementioned machine tool operation status monitoring and fault early warning system, the historical data storage module is also connected to the analysis module to provide historical data to the analysis module to support the updating and optimization of the fault diagnosis model.

[0013] The aforementioned machine tool operation status monitoring and fault early warning system further includes a parameter configuration module, which is connected to the data acquisition module, the analysis module, and the early warning module, and is used to configure data acquisition parameters, analysis and judgment thresholds, and early warning parameters.

[0014] The aforementioned machine tool operation status monitoring and fault early warning system includes a fault diagnosis model that includes a deep learning-based neural network model. This neural network model is capable of identifying various fault types of the machine tool and determining the probability of fault occurrence.

[0015] The present invention has the following beneficial effects: 1. Comprehensive monitoring: The combination of multi-dimensional sensor acquisition and multi-module data processing avoids the limitations of single parameter monitoring, and can fully reflect the machine tool's operating status, providing sufficient basis for fault diagnosis.

[0016] 2. Diagnostic accuracy: The high-quality preprocessed data combined with the fault diagnosis model trained based on historical experience improves the fault identification capability and reduces misjudgments and omissions, especially for the identification of complex faults.

[0017] 3. Effectiveness of early warning: The combination of a tiered early warning mechanism and multi-channel early warning execution avoids the waste of resources from indiscriminate early warnings and ensures that relevant personnel are informed of risks in a timely manner, thereby improving the timeliness and pertinence of early warning response.

[0018] 4. Adaptability and longevity: The parameter configuration module enables the system to adapt to different machine tools and working conditions, while the model update mechanism supported by historical data ensures diagnostic accuracy during long-term use and extends the effective life cycle of the system.

[0019] 5. Maintenance and Optimization Support: Historical data storage provides a data foundation for tracing operational history and analyzing fault patterns, helping to optimize maintenance strategies, reduce downtime due to faults, ensure stable machine tool operation, and improve production efficiency. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the architecture of the machine tool operation status monitoring and fault early warning system provided in an embodiment of the present invention; Figure 2 The graph shows the change in kurtosis of the machine tool spindle vibration, which indicates that the vibration kurtosis value gradually increases from 2.5g to 8.5g over time, reflecting the changes in the spindle's operating state. Figure 3 The temperature change curve of the spindle bearing in the machine tool shows the process of the temperature rising uniformly from 45℃ to 65℃, reflecting the heat accumulation effect. Figure 4 The spindle current fluctuation curve of the machine tool shows the sinusoidal fluctuation characteristics in the range of 5-15A, reflecting the load change characteristics. Detailed Implementation

[0021] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0022] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual images. They should not be construed as limiting the scope of this application. To better illustrate the embodiments of the present invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual dimensions of the product. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0023] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "inner," and "outer" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and 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. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present application. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0024] In the description of this invention, unless otherwise explicitly specified and limited, the term "connection" or similar designation indicating a connection between components should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication between two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0025] Example: The machine tool operation status monitoring and fault early warning system provided in this example, such as Figures 1-4 As shown, it includes: a data acquisition module, a data processing module, an analysis module, and an early warning module. The data acquisition module is connected to the data processing module, the data processing module is connected to the analysis module, and the analysis module is connected to the early warning module. The data acquisition module is used to collect various status data during the machine tool's operation. The data processing module is used to preprocess the collected status data. The analysis module is used to analyze the preprocessed status data to determine whether the machine tool has any fault risks. The early warning module is used to issue early warning information when the analysis module determines that there are fault risks.

[0026] By setting up and connecting data acquisition, data processing, analysis, and early warning modules, the modules can work together to achieve full-process monitoring of the machine tool's operating status. Through the complete process of data acquisition, processing, analysis, and early warning, potential fault risks of the machine tool can be detected in a timely manner and early warnings can be issued, thereby reducing the probability of fault occurrence and ensuring the stable operation of the machine tool.

[0027] Specifically, in this embodiment, the data acquisition module includes a vibration sensor, a temperature sensor, a current sensor, and a voltage sensor. The vibration sensor is used to collect vibration data of the machine tool, the temperature sensor is used to collect temperature data of key components of the machine tool, the current sensor is used to collect current data of the machine tool during operation, and the voltage sensor is used to collect voltage data of the machine tool during operation.

[0028] The data acquisition module in this solution includes vibration, temperature, current, and voltage sensors, which can collect machine tool status data from multiple dimensions such as vibration, temperature of key components, operating current, and voltage. This can comprehensively reflect the operating status of the machine tool, avoid the limitations of single data monitoring, improve the comprehensiveness of the understanding of the machine tool status, and provide richer evidence for subsequent fault diagnosis.

[0029] Specifically, in this embodiment, the data processing module includes a data cleaning unit, a data denoising unit, and a feature extraction unit. The data cleaning unit is used to remove outliers and missing values ​​from the state data, the data denoising unit is used to denoise the cleaned state data, and the feature extraction unit is used to extract feature parameters from the denoised state data.

[0030] The data processing module of this solution removes outliers and missing values ​​through a data cleaning unit, reduces interference through a data noise reduction unit, and extracts key parameters through a feature extraction unit. This improves the quality of the original data, reduces the impact of invalid data and interference factors on subsequent analysis, and makes the extracted feature parameters more reflective of the machine tool's true state. This provides a reliable data foundation for the analysis module and improves the accuracy of fault diagnosis.

[0031] Specifically, in this embodiment, the analysis module includes a model training unit and a fault diagnosis unit. The model training unit is used to train a fault diagnosis model using historical state data and corresponding fault records. The fault diagnosis unit is used to input preprocessed state data into the trained fault diagnosis model to determine the fault risk.

[0032] The analysis module of this solution trains the model using historical data and fault records through the model training unit, and then the fault diagnosis unit applies the model to judge the fault risk. This can integrate historical operating experience and fault cases into the diagnosis process, making the fault judgment more consistent with the actual operating conditions, improving the ability to identify potential faults and the efficiency of judgment, and reducing misjudgment or omission.

[0033] Specifically, in this embodiment, the early warning module includes an early warning level judgment unit and an early warning execution unit. The early warning level judgment unit is used to determine the early warning level based on the judgment result of the analysis module, and the early warning execution unit is used to issue corresponding early warning signals according to different early warning levels.

[0034] The early warning module of this solution determines the early warning level through the early warning level judgment unit, and then the early warning execution unit sends out the corresponding signal. Differentiated early warnings can be taken according to the severity of the fault risk, avoiding the waste of resources or the neglect of important risks caused by indiscriminate early warnings, making the early warning more targeted and improving the actual effect of the early warning.

[0035] Specifically, in this embodiment, the early warning execution unit includes an audible and visual alarm and a communication module. The audible and visual alarm is used to issue audible and visual early warning signals, and the communication module is used to send early warning information to a designated terminal.

[0036] The early warning execution unit of this solution includes an audible and visual alarm and a communication module. The audible and visual alarm can promptly alert nearby personnel on-site, while the communication module can send early warning information to designated terminals, ensuring that relevant personnel can receive early warning information in a timely manner whether on-site or remotely, thereby expanding the coverage of the early warning and improving the timeliness of the early warning response.

[0037] Specifically, in this embodiment, a historical data storage module is also included. The historical data storage module is connected to the data processing module and is used to store preprocessed status data as well as corresponding analysis results and early warning records.

[0038] The historical data storage module of this solution stores preprocessed status data, analysis results, and early warning records. It can retain key data during machine tool operation, providing data support for tracing the machine tool's operating history and analyzing fault patterns, and facilitating the summarization of experience to optimize machine tool maintenance strategies.

[0039] Specifically, in this embodiment, the historical data storage module is also connected to the analysis module to provide historical data to the analysis module to support the updating and optimization of the fault diagnosis model.

[0040] In this solution, the historical data storage module is connected to the analysis module and provides it with historical data. This enables the fault diagnosis model of the analysis module to be continuously updated and optimized based on newly accumulated data, adapting to the changes in the machine tool's state during long-term operation, maintaining the model's diagnostic accuracy, and extending the effective service life of the system.

[0041] Specifically, in this embodiment, a parameter configuration module is also included. The parameter configuration module is connected to the data acquisition module, the analysis module, and the early warning module, respectively, and is used to configure data acquisition parameters, analysis and judgment thresholds, and early warning parameters.

[0042] The parameter configuration module of this solution can configure data acquisition parameters, analysis and judgment thresholds, and early warning parameters. It can adjust system parameters according to the characteristics of different types of machine tools or the needs of different operating scenarios, so that the system is applicable to a variety of machine tools and working conditions, thereby improving the system's flexibility and applicability.

[0043] Specifically, in this embodiment, the fault diagnosis model includes a deep learning-based neural network model, which can identify various fault types of the machine tool and determine the probability of fault occurrence.

[0044] This solution uses a deep learning-based neural network model as the fault diagnosis model, which can process complex state data, identify multiple fault types and determine the probability of fault occurrence, thereby improving the ability to identify complex faults. At the same time, it quantifies fault risks through probability judgment, making fault judgment more accurate and providing a more detailed basis for maintenance decisions.

[0045] It is worth noting that the specific meanings of the three graphs in the machine tool operation status monitoring and fault early warning system are as follows: 1. Spindle vibration kurtosis variation curve Significance: Quantifying the risk of spindle impact failure, such as wear, loosening, or imbalance, through kurtosis values.

[0046] Application Scenario: In this example, when the kurtosis value increases abnormally, the LSTM model identifies it as a risk of "spindle wear" and triggers a Level 1 warning (yellow light + intermittent beeping). If it continues to rise to the threshold, it is upgraded to a Level 2 warning (red light + continuous beeping + 4G remote push), supporting accurate fault location and preventative maintenance.

[0047] Technical effect: Combining 3σ criterion noise reduction and time domain feature extraction improves the sensitivity of impact fault detection and avoids missed detection.

[0048] 2. Spindle bearing temperature variation curve Significance: To monitor the heat accumulation effect of bearings and provide early warning of lubrication failure or overload risks.

[0049] Application scenario: When the temperature exceeds 60℃, a first-level warning is triggered. In the example, the process of the spindle bearing temperature rising from 45℃ to 65℃ can be correlated with the servo motor load change to diagnose "poor bearing lubrication" or "cooling system failure".

[0050] Technical effect: Combining the temperature rise rate characteristics with the steady-state temperature threshold enables early warning of overheating faults and extends bearing life.

[0051] 3. Spindle current fluctuation curve Significance: Reflects the stability of the electrical system and identifies voltage fluctuations or circuit faults.

[0052] Application scenario: When current fluctuations exceed ±15%, the system identifies it as a "circuit fault" risk and verifies whether it is a power supply problem or motor overload by combining voltage sensor data. In the embodiment, sinusoidal fluctuations in current from 5A to 15A can be correlated with changes in spindle load to assist in diagnosing "motor efficiency decline" or "circuit short circuit".

[0053] Technical effect: By analyzing the fluctuation amplitude and mean characteristics, the accuracy of electrical fault diagnosis is improved and misjudgments are reduced.

[0054] Comprehensive technical value Upgraded monitoring dimensions: Three-curve multi-parameter collaborative monitoring covers three major systems: mechanical, thermal, and electrical, avoiding missed reports caused by single parameter failures.

[0055] Early warning classification optimization: By combining risk probability (30%-70% Level 1 / >70% Level 2) with curve trends, a closed-loop management system from "status monitoring" to "fault prediction" is achieved.

[0056] Data-driven decision-making: Historical data storage supports model iteration. For example, in the embodiment, LSTM is trained with 500 normal data and 180 fault data to continuously improve diagnostic accuracy.

[0057] These graphs, by dynamically visualizing changes in key parameters, transform abstract technical parameters into intuitive fault warning signals, and are the core tool for this system to achieve "early detection, early diagnosis, and early treatment".

[0058] Specifically, in this embodiment, the fault diagnosis unit in the machine tool operation status monitoring and fault early warning system uses a fault risk quantification equation based on multi-dimensional feature fusion when judging fault risk. The expression of this equation is: ; Where: R is the fault risk quantification value, the core indicator for comprehensively assessing machine tool fault risk. The larger the value, the higher the risk. The value range is [0,10]. R<3 is judged as no fault risk, 3≤R<7 is judged as first-level warning risk, and R≥7 is judged as second-level warning risk. The feature weighting coefficients are obtained through training on historical fault data, and their values ​​range from [0.1, 3]. Specifically, α is the vibration feature weighting coefficient, which adjusts the importance of vibration features in risk assessment and is determined based on historical data training; β is the temperature feature weighting coefficient, which adjusts the importance of temperature features in risk assessment and is determined based on historical data training; γ is the temperature influence coefficient, which controls the exponential growth rate of temperature risk values; and δ is the electrical parameter weighting coefficient, which adjusts the importance of current and voltage parameters in risk assessment and is determined based on historical data training. F i The actual value of the i-th vibration characteristic parameter, such as peak value, root mean square, kurtosis, etc., reflects the vibration state of the machine tool's mechanical structure. It is determined according to the sensor type, such as the vibration peak value being in g. F i0 The normal threshold for the i-th vibration characteristic parameter is the reference value of this vibration characteristic during normal machine tool operation, determined by factory standards and historical normal data, and compared with F. i Units are consistent; ΔF imaxThe maximum allowable deviation value for the i-th vibration characteristic parameter represents the maximum allowable fluctuation range of this vibration characteristic during normal machine tool operation. This range is determined by the equipment manual and experimental data, and is related to F. i Units are consistent; n represents the number of vibration characteristic parameters, which is the number of vibration characteristic parameters selected according to monitoring requirements, such as 3 (peak value, root mean square, kurtosis), and is a positive integer; T represents the actual temperature value of key components, such as the real-time temperature of key components like spindle bearings and servo motors, reflecting the thermal state of the components, and is expressed in °C. T0 is the normal temperature threshold of the critical component, which is the reference temperature for the critical component during normal operation. It is determined by the equipment design standard and is in °C. ΔT max This refers to the maximum allowable temperature deviation of critical components; the maximum allowable temperature fluctuation range of critical components during normal operation is determined by the heat resistance performance of the equipment, and the unit is °C. I j This is the actual value of the j-th current parameter, such as the main circuit current or the servo motor current, reflecting the load status of the electrical system, and the unit is A; I j0 The j-th current parameter is the normal threshold value. It is the reference value of the current parameter when the electrical system is running normally. It is determined by the electrical parameters of the equipment and the unit is A. ΔI jmax The j-th current parameter is the maximum allowable deviation value. It represents the maximum allowable fluctuation range of the current parameter during normal operation of the electrical system, determined by electrical design standards, and is expressed in amperes (A). m represents the number of current parameters, which is the number of current parameters selected according to the monitoring requirements, such as 2 (main circuit current, servo motor current), and is a positive integer; U l This is the actual value of the l-th voltage parameter, such as the main circuit voltage or the control circuit voltage, reflecting the power supply status of the electrical system, and the unit is V; U l0 The normal threshold value for the l-th voltage parameter is the reference value of this voltage parameter when the electrical system is operating normally. It is determined by the power supply standard and the unit is V. ΔU lmax The maximum allowable deviation value of the l-th voltage parameter is the maximum allowable fluctuation range of this voltage parameter during normal operation of the electrical system, which is determined by the power supply standard and is in V. k represents the number of voltage parameters, which is the number of voltage parameters selected according to the monitoring requirements, such as 1 (main circuit voltage), and is a positive integer.

[0059] The derivation of the equation is as follows: 1. Construction of a basic feature standardization model: For the four core monitoring parameters—vibration, temperature, current, and voltage—a standardized processing model is first constructed to eliminate the dimensional differences between the various parameters. For any parameter X (such as vibration characteristic F...), a standardized processing model is then constructed. i Temperature T, Current I j Voltage U l Its standardized deviation expression is: Where X0 is the normal threshold for the parameter, ΔX max Given the maximum allowable deviation value for the parameter, this formula can convert the actual deviation of the parameter into a standardized value within the range of [0,1], which facilitates cross-dimensional feature comparison.

[0060] 2. Derivation of the vibration characteristic fusion model: Vibration signals contain multiple characteristic parameters such as peak value, root mean square (RMS), and kurtosis. Each feature has different sensitivities to faults but exhibits synergy. A method using the mean square error (RMS) approach is employed to fuse these multiple vibration features. By balancing the weights of each feature through mean calculation, and amplifying the influence of abnormal features through squared terms, the algorithm better reflects the nonlinear changes in vibration characteristics during fault occurrence. A weighting coefficient α is introduced to adjust the importance of vibration characteristics in the overall risk assessment. This coefficient is obtained through correlation analysis between vibration characteristics and fault occurrence in historical fault data.

[0061] 3. Derivation of the nonlinear model for temperature characteristics: The effect of temperature on failure has a cumulative effect; when the temperature approaches or exceeds the allowable threshold, the failure risk increases exponentially. A temperature risk model is constructed based on an exponential function: Where γ is the temperature influence coefficient, obtained by fitting the relationship between the duration of temperature exceeding the limit and the failure rate in historical data. When the temperature is within the normal range... The exponential term is approximately 1, contributing little to the risk value; however, when the temperature exceeds the limit, the exponential term increases rapidly, highlighting the risk warning effect of abnormal temperature. A weighting coefficient β is introduced to adjust the overall influence weight of temperature characteristics.

[0062] 4. Derivation of the electrical parameter fusion model: Current and voltage parameters reflect the operating status of the machine tool's electrical system; they are correlated and require joint evaluation. The root mean square (RMS) fusion of current and voltage parameters is performed separately. Square root calculation can reduce the excessive influence of extreme outliers while incorporating overall deviations in current and voltage. A weighting coefficient δ is introduced to adjust the weight of electrical parameters in risk assessment.

[0063] 5. Integration of overall risk quantification equations: By weighting and superimposing the fusion model of vibration, temperature, and electrical parameters, the overall failure risk quantification equation is obtained: , Determined through training with historical fault data The optimal value of R is obtained so that R can accurately quantify the risk level under different failure scenarios.

[0064] Example: Taking the spindle monitoring of a certain type of CNC lathe as an example, the specific parameter settings and calculation process are as follows: 1. Parameter settings: Vibration characteristic parameters (n=3): Peak value F1 (F 10 =5g, ΔF 1max =3g), root mean square F2 (F 20 =2g, ΔF 2max =1.5g), kurtosis F3 (F 30 =3, ΔF 3max =2); Temperature parameters: T0 = 45℃, ΔT max =20℃; Current parameters (m=2): Main circuit current I1 (I 10 =10A, ΔI 1max =5A), servo motor current I2 (I 20 =8A, ΔI 2max =3A); Voltage parameters (k=1): Main circuit voltage U1 (U 10 =380V, ΔU 1max =20V); Weight coefficient: α=2.5, β=1.8, γ=1.2, δ=2.2.

[0065] 2. Actual monitoring data: Vibration characteristics: F1=6.5g, F2=2.8g, F3=4.2g; Temperature: T=58℃; Current: I1 = 12.3A, I2 = 9.5A; Voltage: U1 = 372V.

[0066] 3. Calculation process Vibration characteristic fusion value: ; Temperature risk value: exp(1.2× )=exp(1.2×0.65)≈exp(0.78)≈2.182; Combined electrical parameter values: ≈0.625; Fault risk quantification value: R=2.5×0.298+1.8×2.182+2.2×0.625≈0.745+3.928+1.375=6.048.

[0067] 4. Risk Assessment: Since 3≤6.048<7, it is determined to be a Level 1 warning risk. The warning module emits a yellow light and intermittent buzzer according to the Level 1 warning standard, and pushes the warning information to the workshop management terminal.

[0068] Technical effect 1. Achieving Precise Fusion of Multi-Dimensional Features: Traditional monitoring systems often rely on single-parameter threshold judgments or simple weighted summations, which cannot effectively handle the nonlinear relationships between different parameters. This equation, by piecewise fusing vibration, temperature, and electrical parameters and combining various mathematical forms such as mean square error, exponential functions, and square root operations, accurately characterizes the influence of each parameter on the fault, solving the problem of difficult multi-dimensional feature fusion and making risk assessment more closely aligned with actual fault occurrence mechanisms.

[0069] 2. Improved Accuracy of Risk Quantification: The weight coefficients in the equation are obtained through training on historical fault data, rather than being subjectively set, ensuring that the weights of each feature match the correlation with the fault. Simultaneously, an exponential function is introduced to amplify the cumulative effect of temperature, better reflecting the actual situation where excessive temperature leads to a surge in fault risk. Experimental verification shows that using this equation improves the accuracy of fault risk assessment by more than 35% compared to traditional methods, while reducing the false positive rate by 28%.

[0070] 3. Enhance the rationality of early warning classification: The fault risk quantification value R is divided into three intervals: [0,3), [3,7), and [7,10], corresponding to no risk, Level 1 warning, and Level 2 warning, avoiding the limitations of the traditional "black and white" warning system. By quantifying the degree of risk with specific values, maintenance personnel can more accurately judge the urgency of the fault and rationally allocate maintenance resources. For example, a Level 1 warning allows for planned shutdown for maintenance, while a Level 2 warning requires immediate shutdown, reducing the waste of maintenance resources and the risk of fault escalation.

[0071] 4. Adapt to different machine tools and operating conditions: Threshold values ​​for each parameter in the equation (such as F) i0 T0, I j0 The parameters and weighting coefficients can be adjusted through the parameter configuration module, allowing the system to adapt to different types of machine tools such as lathes, milling machines, and machining centers, as well as different processing materials (such as 45# steel and aluminum alloy) and processing accuracy requirements without modifying the equation structure, thus enhancing the system's versatility and adaptability.

[0072] Working principle and process 1. Parameter initialization phase After the system starts, the parameter configuration module calls the fault data of similar machine tools in the historical data storage module and obtains the data through the gradient descent algorithm. Equal weighting coefficients are set based on the current machine tool model and operating conditions. Normal threshold and Once the maximum allowable deviation value is reached, the equation parameters are initialized.

[0073] 2. Data Acquisition and Preprocessing Stage The data acquisition module collects machine tool operation data in real time through vibration, temperature, current, and voltage sensors; the data processing module cleans the collected data (removing outliers and filling in missing values), reduces noise (e.g., wavelet thresholding), and extracts features (e.g., extracting peak values, root mean square values, and kurtosis from vibration data) to obtain the data required for the equations. Actual monitoring values, etc.

[0074] 3. Risk Quantification and Calculation Stage The fault diagnosis unit of the analysis module substitutes the pre-processed actual monitoring values ​​into the fault risk quantification equation, and calculates the vibration characteristic fusion value, temperature risk value, and electrical parameter fusion value in sequence. These are then weighted and summed to obtain the fault risk quantification value R. During the calculation process, the system calls historical parameters from the historical data storage module in real time to compare the deviation between the current calculation result and the historical fault risk value, ensuring calculation accuracy.

[0075] 4. Risk Assessment and Early Warning Stage The early warning module's early warning level judgment unit determines the risk level based on the range of R values: R < 3 indicates no fault risk, and no early warning is issued; 3 ≤ R < 7 indicates a Level 1 early warning risk; and R ≥ 7 indicates a Level 2 early warning risk. The early warning execution unit, based on the risk level, issues on-site early warning signals via audible and visual alarms (Level 1 early warning: yellow light + intermittent buzzer; Level 2 early warning: red light + continuous buzzer), and simultaneously pushes the early warning information (including R value, abnormal parameters, and risk level) to the designated terminal via the communication module.

[0076] 5. Parameter Iterative Optimization Stage The historical data storage module stores the fault risk quantification value R, actual monitoring data, early warning records, and subsequent fault handling results in real time for each fault. After the system has run for a full cycle (e.g., one month), the model training unit of the analysis module calls up the historical data and uses the least squares method for iterative optimization. Equal weighting coefficients make the equations more adaptable to changes in the current operating status of the machine tool, ensuring the accuracy of risk assessment during long-term use.

[0077] In summary, the machine tool operation status monitoring and fault early warning system provided in this embodiment has the following advantages: This system achieves a comprehensive upgrade in machine tool operating status monitoring and fault early warning through multi-module collaboration and end-to-end design: 1. Comprehensive monitoring: The combination of multi-dimensional sensor acquisition and multi-module data processing avoids the limitations of single parameter monitoring, and can fully reflect the machine tool's operating status, providing sufficient basis for fault diagnosis.

[0078] 2. Diagnostic accuracy: The high-quality preprocessed data combined with the fault diagnosis model trained based on historical experience improves the fault identification capability and reduces misjudgments and omissions, especially for the identification of complex faults.

[0079] 3. Effectiveness of early warning: The combination of a tiered early warning mechanism and multi-channel early warning execution avoids the waste of resources from indiscriminate early warnings and ensures that relevant personnel are informed of risks in a timely manner, thereby improving the timeliness and pertinence of early warning response.

[0080] 4. Adaptability and longevity: The parameter configuration module enables the system to adapt to different machine tools and working conditions, while the model update mechanism supported by historical data ensures diagnostic accuracy during long-term use and extends the effective life cycle of the system.

[0081] 5. Maintenance and Optimization Support: Historical data storage provides a data foundation for tracing operational history and analyzing fault patterns, helping to optimize maintenance strategies, reduce downtime due to faults, ensure stable machine tool operation, and improve production efficiency.

[0082] Working principle: This machine tool operation status monitoring and fault early warning system achieves full-process monitoring and early warning through the coordinated work of various modules. The specific principle is as follows: 1. Data Acquisition: The data acquisition module uses vibration sensors, temperature sensors, current sensors, and voltage sensors to collect machine tool operating status data in real time from four dimensions: vibration, temperature of key components, operating current, and voltage, ensuring comprehensive capture of machine tool operating information.

[0083] 2. Data Processing: The collected raw data is transmitted to the data processing module. The data cleaning unit removes outliers and missing values, the data noise reduction unit reduces environmental interference, and finally the feature extraction unit extracts key feature parameters that reflect the true state of the machine tool, providing a high-quality data foundation for subsequent analysis.

[0084] 3. Fault Analysis: The preprocessed feature parameters are input into the analysis module. The model training unit uses historical state data and corresponding fault records to train a fault diagnosis model (such as a neural network model based on deep learning). The fault diagnosis unit then inputs the real-time feature parameters into the trained model to determine whether the machine tool has fault risks and the degree of risk.

[0085] 4. Tiered early warning: The analysis results are transmitted to the early warning module. The early warning level judgment unit determines the early warning level according to the degree of risk. The early warning execution unit issues on-site reminders through the sound and light alarm and sends the information to the designated terminal through the communication module, realizing multi-channel tiered early warning.

[0086] 5. Data support and parameter adaptation: The historical data storage module stores preprocessed data, analysis results, and early warning records, providing a basis for tracing historical operating status and data support for the analysis module to update and optimize the fault diagnosis model; the parameter configuration module can adjust the acquisition parameters, analysis thresholds, and early warning parameters, enabling the system to adapt to different machine tool types and operating conditions.

[0087] How to use 1. System Deployment: Install vibration sensors and temperature sensors on key parts of the machine tool, and install current sensors and voltage sensors on the power supply circuit to ensure reliable connection between each sensor and the data acquisition module; deploy the data acquisition module, data processing module, analysis module, and early warning module according to their connection relationships, and connect them to the historical data storage module and parameter configuration module.

[0088] 2. Parameter Configuration: The parameter configuration module allows you to set data acquisition parameters such as data acquisition frequency and sensor range, as well as set fault judgment thresholds for the analysis module and grading standards for the early warning module (such as early warning methods corresponding to minor risk, medium risk, and severe risk).

[0089] 3. Operation monitoring: After the system starts, the data acquisition module collects various status data in real time. After preprocessing by the data processing module, the data is transmitted to the analysis module. The analysis module continuously analyzes the data to determine the risk of failure.

[0090] 4. Early Warning Response: When the analysis module determines that there is a risk of failure, the early warning module issues an early warning according to the set level: the audible and visual alarm will remind nearby personnel on site, and the communication module will send the information to the designated terminal (such as the mobile phone of the management personnel or the monitoring platform); relevant personnel will take corresponding measures according to the warning level (such as strengthening observation for minor risks and shutting down for maintenance for serious risks).

[0091] 5. Maintenance and optimization: The system periodically retrieves operation records through the historical data storage module to analyze fault patterns; the system automatically uses newly accumulated historical data to update and optimize the fault diagnosis model through the analysis module, ensuring diagnostic accuracy during long-term use.

[0092] Specific example: This embodiment takes the operation status monitoring and fault early warning of a certain type of CNC lathe as an example. The system includes a data acquisition module, a data processing module, an analysis module, an early warning module, a historical data storage module, and a parameter configuration module. The connection relationship of each module is as follows: the data acquisition module is connected to the data processing module, the data processing module is connected to the analysis module and the historical data storage module respectively, the analysis module is connected to the early warning module and the historical data storage module respectively, and the parameter configuration module is connected to the data acquisition module, the analysis module, and the early warning module respectively.

[0093] Specific configuration of each module 1. Data acquisition module: It includes 4 vibration sensors (installed on the lathe spindle, X-axis feed mechanism, Y-axis feed mechanism and Z-axis feed mechanism respectively), 3 temperature sensors (monitoring the spindle bearing, servo motor and cooling system temperature respectively), 1 current sensor (connected in series with the main circuit) and 1 voltage sensor (connected in parallel with the main circuit). The sampling frequency is set to 1kHz through the parameter configuration module.

[0094] 2. Data Processing Module: The data cleaning unit uses the 3σ criterion to remove outliers from vibration, temperature, and other data, and fills in missing values ​​through linear interpolation; the data denoising unit uses a wavelet threshold denoising algorithm to process mechanical noise in vibration signals; the feature extraction unit extracts time-domain features such as peak value, root mean square, and kurtosis from vibration data, features such as temperature rise rate and steady-state temperature from temperature data, and features such as fluctuation amplitude and mean value from current and voltage data.

[0095] 3. Analysis Module: The model training unit trains an LSTM neural network model based on historical data (including 500 normal operation records, 80 spindle failure records, 60 feed mechanism failure records, and 40 circuit failure records of this model of lathe over the past 3 years); the fault diagnosis unit inputs real-time feature parameters into the model and outputs the fault type (such as spindle wear, feed mechanism jamming, and abnormal voltage fluctuation) and the corresponding risk probability.

[0096] 4. Early Warning Module: The early warning level judgment unit sets the risk probability as follows: <30% for no warning, 30%-70% for Level 1 warning, and >70% for Level 2 warning. In the early warning execution unit, the audible and visual alarm emits a yellow light and intermittent buzzer during Level 1 warning and emits a red light and continuous buzzer during Level 2 warning. The communication module pushes the early warning information (including fault type and risk probability) to the workshop management terminal and maintenance personnel's mobile phones via the 4G network.

[0097] 5. Historical data storage module: An industrial-grade database is used to store the pre-processed feature parameters, analysis results and early warning records every day, with a storage period of 5 years.

[0098] 6. Parameter configuration module: The sensor sampling frequency (500Hz-2kHz), fault judgment threshold (such as adjusting the first-level warning risk probability threshold to 25%-65%), and warning information push targets can be adjusted through the touch screen interface.

[0099] Working principle of the embodiment 1. Data Acquisition: Each sensor collects vibration, temperature, current, and voltage data during the operation of the CNC lathe in real time and transmits them to the data processing module at a frequency of 1kHz.

[0100] 2. Data Processing: The data cleaning unit removes outliers caused by momentary interference from sensors and fills in missing values ​​caused by sensor communication interruptions; the data noise reduction unit filters environmental noise from mechanical vibrations; the feature extraction unit extracts key feature parameters that reflect the equipment status from the processed data, such as the kurtosis value of the spindle vibration (which can characterize impact failures) and the temperature rise rate of the servo motor (which can characterize overload risk).

[0101] 3. Fault Analysis: The analysis module inputs real-time feature parameters into the pre-trained LSTM model. The model compares the feature patterns of historical normal and fault data to determine whether there is a fault risk and the specific fault type (e.g., when the spindle vibration kurtosis value is abnormal, the model identifies it as a "spindle wear" risk) and outputs the risk probability.

[0102] 4. Tiered early warning: The early warning level judgment unit determines the early warning level based on the risk probability. When the early warning is at level one, the audible and visual alarm emits a yellow signal and the communication module pushes a prompt message. When the early warning is at level two, a red signal is emitted and an emergency maintenance notice is pushed.

[0103] 5. Data Iteration and Parameter Adaptation: The historical data storage module updates the operation records daily, and the analysis module optimizes the LSTM model monthly using the newly added data; if the lathe changes tools or adjusts machining parameters, the sampling frequency can be increased through the parameter configuration module to ensure monitoring adaptability.

[0104] Experimental data 1. To verify the system's effectiveness, two CNC lathes of the same model were selected for a comparative experiment: one lathe was equipped with this system (experimental group), and the other lathe used traditional single vibration monitoring + manual inspection (control group). The experiment lasted for 6 months, during which the lathes were used to process 45# steel parts and ran for 8 hours a day.

[0105] 2. Fault detection: The experimental group detected a total of 23 potential faults (including 10 signs of spindle wear, 8 instances of feed mechanism jamming, and 5 instances of circuit abnormalities), with no missed reports; the control group only detected 9 vibration abnormalities, with 14 missed reports (most of which were temperature and circuit-related faults).

[0106] 3. Timeliness of early warning: The experimental group received early warnings of spindle wear precursors 3-15 days earlier than the actual failure occurred; the control group, which relied on manual inspection, only discovered the problem after the failure occurred.

[0107] 4. Maintenance Response: After a Level 1 warning in the experimental group, maintenance personnel can arrange planned maintenance in advance by sending information; during a Level 2 warning, the on-site audible and visual alarms and remote notifications are triggered simultaneously, and the average response time is shorter than that of the control group.

[0108] Technical Effects of the Examples 1. Enhanced monitoring comprehensiveness: By collecting vibration, temperature, current, and voltage data through multiple sensors, it covers the key states of mechanical structure and electrical system, avoiding the omission problem of traditional single-parameter monitoring, and comprehensively reflecting the lathe's operating status.

[0109] 2. Data quality optimization: The feature parameters after cleaning and noise reduction are closer to the actual state of the equipment, providing a reliable basis for fault analysis and reducing misjudgments caused by data interference.

[0110] 3. Enhanced diagnostic accuracy: The LSTM model trained on historical fault data can identify multiple fault types and quantify risks, making it more objective than traditional manual judgment and improving the accuracy of fault identification.

[0111] 4. Improved early warning effectiveness: The tiered early warning mechanism, combined with on-site audible and visual alarms and remote information push, ensures that faults of different risk levels receive targeted responses, avoids waste of maintenance resources, and improves fault handling efficiency.

[0112] 5. Adaptability and Long-Term Effectiveness: The parameter configuration module can adapt to changes in processing conditions, and the model optimization mechanism supported by historical data ensures that the system can maintain diagnostic accuracy during long-term use and extend the effective service cycle.

[0113] 6. Maintenance strategy optimization: Historical data storage provides a basis for analyzing fault patterns, helps to formulate preventive maintenance plans, reduces downtime due to sudden failures, and ensures continuous and stable operation of the lathe.

[0114] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. A machine tool operating status monitoring and fault early warning system, characterized in that, The system includes a data acquisition module, a data processing module, an analysis module, and an early warning module. The data acquisition module is connected to the data processing module, the data processing module is connected to the analysis module, and the analysis module is connected to the early warning module. The data acquisition module is used to collect various status data during the machine tool's operation. The data processing module is used to preprocess the collected status data. The analysis module is used to analyze the preprocessed status data to determine whether the machine tool has any fault risks. The early warning module is used to issue early warning information when the analysis module determines that there is a fault risk.

2. The machine tool operation status monitoring and fault early warning system according to claim 1, characterized in that, The data acquisition module includes a vibration sensor, a temperature sensor, a current sensor, and a voltage sensor. The vibration sensor is used to collect vibration data of the machine tool, the temperature sensor is used to collect temperature data of key components of the machine tool, the current sensor is used to collect current data during machine tool operation, and the voltage sensor is used to collect voltage data during machine tool operation.

3. The machine tool operation status monitoring and fault early warning system according to claim 1, characterized in that, The data processing module includes a data cleaning unit, a data denoising unit, and a feature extraction unit. The data cleaning unit is used to remove outliers and missing values ​​from the state data. The data denoising unit is used to denoise the cleaned state data. The feature extraction unit is used to extract feature parameters from the denoised state data.

4. The machine tool operation status monitoring and fault early warning system according to claim 1, characterized in that, The analysis module includes a model training unit and a fault diagnosis unit. The model training unit is used to train a fault diagnosis model using historical state data and corresponding fault records. The fault diagnosis unit is used to input preprocessed state data into the trained fault diagnosis model to determine the fault risk.

5. The machine tool operation status monitoring and fault early warning system according to claim 1, characterized in that, The early warning module includes an early warning level judgment unit and an early warning execution unit. The early warning level judgment unit is used to determine the early warning level based on the judgment result of the analysis module, and the early warning execution unit is used to issue corresponding early warning signals according to different early warning levels.

6. The machine tool operation status monitoring and fault early warning system according to claim 5, characterized in that, The early warning execution unit includes an audible and visual alarm and a communication module. The audible and visual alarm is used to issue audible and visual early warning signals, and the communication module is used to send early warning information to a designated terminal.

7. The machine tool operation status monitoring and fault early warning system according to claim 1, characterized in that, It also includes a historical data storage module, which is connected to the data processing module and is used to store preprocessed status data as well as corresponding analysis results and early warning records.

8. The machine tool operation status monitoring and fault early warning system according to claim 7, characterized in that, The historical data storage module is also connected to the analysis module to provide historical data to support the updating and optimization of the fault diagnosis model.

9. The machine tool operation status monitoring and fault early warning system according to claim 1, characterized in that, It also includes a parameter configuration module, which is connected to the data acquisition module, the analysis module and the early warning module respectively, and is used to configure data acquisition parameters, analysis and judgment thresholds and early warning parameters.

10. The machine tool operation status monitoring and fault early warning system according to claim 4, characterized in that, The fault diagnosis model includes a deep learning-based neural network model, which can identify various fault types of machine tools and determine the probability of fault occurrence.

Images