Precise numerical control machine tool operation state intelligent control method and system

Abstract

The invention discloses a precise numerical control machine tool operation state intelligent control method and system, and particularly relates to the technical field of numerical control machine tools. Multi-dimensional operation data, including working temperature, load, current, pressure, cutting force and tool wear data, of a machine tool are collected in real time, a Z-score algorithm is adopted for data preprocessing, abnormal values are extracted, a prediction model is constructed and trained based on a machine learning algorithm, abnormal data are analyzed in real time, and the abnormal values of the machine tool are analyzed in real time. And accurately identifying the operation state of the machine tool and evaluating the health condition. According to the health state evaluation result, an intelligent control strategy is automatically generated, and the feeding speed, the spindle rotating speed and the cutting depth of the machine tool are dynamically adjusted, so that the operation state of the machine tool is optimized, potential faults are found and dealt with in time, the machining quality is improved, the faults are reduced, the production efficiency and the equipment stability are improved, and the maintenance cost is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of numerically controlled machine tools, and in particular to an intelligent control method and system for the operating status of a precision numerically controlled machine tool. Background Art

[0002] With the continuous improvement of the level of industrial automation, the requirements for precision CNC machine tools in terms of processing accuracy and efficiency are increasing. Precision CNC machine tools often require high-precision control and monitoring during the production process to ensure product quality and production stability. However, traditional CNC machine tool control methods usually rely on empirical parameters or simple control algorithms. These methods are difficult to respond to complex working state changes in real time, and cannot effectively make accurate predictions and adjustments to abnormal conditions during machine tool operation. Therefore, how to perform intelligent and precise control based on the operating status of the machine tool has become a key technical problem to improve the performance of CNC machine tools, extend their service life, and improve processing quality.

[0003] The prior art has the following deficiencies: In the existing technology, the real-time abnormal warning and automatic adjustment capabilities of machine tools under complex working conditions are seriously insufficient. Traditional control systems often rely only on simple data collection of machine hardware, such as basic parameters such as temperature, vibration or speed, and lack in-depth analysis and prediction of multi-dimensional and multi-level dynamic working condition changes. This leads to hidden faults or operating errors that cannot be identified in time in certain extreme working conditions, which in turn affects the processing quality and even causes equipment damage. This problem not only affects the stability of production, but also increases maintenance costs, and may not respond in time under extreme conditions, seriously affecting production efficiency. Summary of the invention

[0004] The purpose of the present invention is to provide a method and system for intelligently controlling the operating status of a precision CNC machine tool, so as to solve the deficiencies in the background technology.

[0005] In order to achieve the above object, the present invention provides the following technical solution: a method for intelligently controlling the operating state of a precision CNC machine tool, comprising the following steps: S1: real-time collection of actual operation data of precision CNC machine tools in several time periods, wherein the actual operation data includes the working temperature, load, current and pressure data of the machine tools, and real-time status data during the machining process, including cutting force data and tool wear data; S2: Preprocess the collected actual operation data and extract abnormal values ​​of temperature, load, current and pressure in the actual operation data through the Z-score algorithm; S3: Based on the machine learning algorithm, a prediction model is constructed for real-time identification and prediction of the machine tool operation status, wherein the prediction model is trained based on abnormal values ​​of temperature, load, current and pressure to identify the machine tool operation status; S4: Analyze the real-time status data using the trained prediction model, detect abnormal behavior of the machine tool in real time, and evaluate the health status of the precision CNC machine tool; S5: Automatically generate an intelligent control strategy based on the operating status evaluation result, wherein the intelligent control strategy includes dynamically adjusting the machine tool feed speed, spindle speed and cutting depth.

[0006] Preferably, in S2, the actual operation data collected from each sensor of the machine tool is preprocessed, and each collected operation data is standardized using a Z-score algorithm; the calculation formula of the Z-score algorithm is: ; Where X is the actual value of a data point, μ is the mean of the parameter, and σ is the standard deviation of the parameter; the standardized data will have a distribution with a mean of 0 and a standard deviation of 1. After standardization, outlier detection is performed on each data point; according to the principle of the Z-score algorithm, when the Z-score value exceeds the set threshold, the data point is considered to be an outlier.

[0007] Preferably, in S3, a data set for training a machine learning model is constructed by extracting outliers and standardized data, wherein the data set includes the operating temperature, load, current and pressure of the machine tool in the input data, and whether the output data is in an abnormal state; Divide the data set into a training set and a test set. For real-time data, a validation set needs to be divided to adjust model parameters. Perform feature selection based on the temperature, load, current, and pressure characteristics in the actual operating data. Use the random forest algorithm to train the training set data; During the training process, random forests automatically learn the complex relationship between features and abnormal states through the ensemble learning method of decision trees; During the training process, each tree will randomly select a feature subset and a sample subset, gradually build multiple decision trees, and finally get the final prediction result through voting; A binary classification model is used, where normal state and abnormal state are the target categories. During the training process, the model is used to distinguish between the two states; The hyperparameters of the model are optimized through cross-validation, and the training set is retrained using the adjusted model to ensure the generalization ability of the model.

[0008] Preferably, in S4, the obtained cutting force data is analyzed to generate a cutting force abnormality index, and the method for obtaining the cutting force abnormality index is: The cutting force data collected within a period of time is set as , where n is the total number of data points, It's in time Cutting force values ​​on the cutting surface; prepare a set of reference data representing normal cutting force , where m is the length of the normal data sequence; Construct a distance matrix D(i,j), where D(i,j) represents the distance between the i-th value of the sequence F(t) and the sequence The distance between the jth values ​​of : ; Define the cumulative distance matrix C(i,j): ; where C(i,j) represents the value from the 1st to the i-th value of F(t), and The minimum cumulative distance between the first value and the jth value of , and the DTW distance is calculated: ;in, is the final cumulative distance value, indicating the difference between F(t) and The overall similarity between them; generate the cutting force abnormality index, the expression is: ;in, is the maximum threshold of DTW distance, It is the cutting force abnormality index.

[0009] Preferably, in S4, the acquired tool wear data is analyzed to generate a tool wear deviation index, and the tool wear deviation index is obtained by: The tool wear data obtained is a time series , where v is the number of data points, t represents the sampling time point, Represents at time The tool wear measurement value at the time of , the tool wear data is normalized, and the tool wear data is trained using the LSTM neural network. LSTM predicts future wear values ​​by learning historical wear patterns; the LSTM model is used to calculate the tool wear value at Tool wear prediction at a point in time , and then calculate the prediction error: ; where e(t+1) represents the error at the t+1th moment, for The tool wear measurement value at a time point; by calculating the prediction error e(t+1), the tool wear deviation index SF is obtained, which is expressed as: ;in, is the maximum prediction error under normal wear mode.

[0010] Preferably, the cutting force anomaly index and the tool wear deviation index are converted into a comprehensive feature vector, and the comprehensive feature vector is used as the input of the machine learning model. The machine learning model predicts the health status assessment value label of the precision CNC machine tool with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the health status assessment value labels of all precision CNC machine tools as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The health status assessment value of the precision CNC machine tool is determined according to the model output results, wherein the machine learning model is a pre-trained random forest model.

[0011] Preferably, the acquired health status assessment value of the precision CNC machine tool is compared with a health status assessment threshold value pre-set based on historical data. If the health status assessment value of the precision CNC machine tool is greater than or equal to the pre-set health status assessment threshold value, it indicates that the precision CNC machine tool is in a healthy state, and no warning signal is generated at this time; if the health status assessment value of the precision CNC machine tool is less than the pre-set health status assessment threshold value, it indicates that the precision CNC machine tool is in an unhealthy state, and a warning signal is generated at this time.

[0012] Preferably, in S5, when the health status assessment value of the precision CNC machine tool If the value is less than the preset health status assessment threshold, it means that the machine tool is in an unhealthy state and an intelligent control strategy needs to be generated immediately to dynamically adjust the operating parameters of the machine tool. The core of the intelligent control strategy is to dynamically adjust the feed speed. , spindle speed and cutting depth To reduce the workload of the machine tool.

[0013] Preferably, the feed speed It directly affects the cutting force and cutting temperature of the machine tool. When the machine tool is in an unhealthy state, the feed speed should be reduced to reduce the burden on the machine tool and the temperature rise during the cutting process. The formula for adjusting the feed speed is: ;in, is the current feed speed, is the standard feed speed when the machine tool is working normally, γ is the index of the adjustment factor, which is used to control the change ratio of the feed speed. assessing thresholds for health states; Spindle speed It will affect the cutting force and temperature during the cutting process. The spindle speed is adjusted dynamically according to the health status assessment value. The formula for adjusting the spindle speed is: ;in: is the spindle speed at the current moment, is the standard spindle speed when the machine tool is working normally, and δ is the index of the adjustment factor, which is used to control the change ratio of the spindle speed; Cutting depth It has a direct impact on the cutting force and the load of the machine tool. The cutting depth is dynamically adjusted according to the health status assessment value. The formula for adjusting the cutting depth is: ;in, is the cutting depth at the current moment, is the standard cutting depth when the machine tool is working normally, and β is the exponent of the adjustment factor, which is used to control the change ratio of the cutting depth.

[0014] The present invention also provides an intelligent control system for the operation status of a precision CNC machine tool, including a data acquisition module, a data processing module, a machine learning training module, a detection module and an intelligent control module: Data acquisition module: real-time acquisition of actual operation data of precision CNC machine tools within a certain period of time, the actual operation data including the working temperature, load, current and pressure data of the machine tools, and real-time status data during the machining process, including cutting force data and tool wear data; Data processing module: pre-processes the collected actual operation data and extracts abnormal values ​​of temperature, load, current and pressure in the actual operation data through the Z-score algorithm; Machine learning training module: Based on the machine learning algorithm, a prediction model is constructed for real-time identification and prediction of the machine tool operation status. The prediction model is trained based on abnormal values ​​of temperature, load, current and pressure to identify the machine tool operation status; Detection module: using the trained prediction model to analyze real-time status data, detect abnormal behavior of machine tools in real time, and evaluate the health status of precision CNC machine tools; Intelligent control module: According to the operation status evaluation results, the intelligent control strategy is automatically generated, and the intelligent control strategy includes dynamic adjustment of the machine tool feed speed, spindle speed and cutting depth.

[0015] In the above technical solution, the technical effects and advantages provided by the present invention are: 1. The present invention not only collects the basic operating data of the machine tool, but also integrates the state data of the machining process such as cutting force and tool wear. The Z-score algorithm is used to extract outliers in the data, and the prediction model built by the machine learning algorithm can accurately identify the health status of the machine tool and detect potential faults in real time. This analysis method based on multi-level data greatly improves the adaptability of the machine tool to extreme working conditions and the fault prevention ability.

[0016] 2. The present invention also introduces an intelligent control strategy, which effectively reduces the workload and temperature rise of the machine tool by dynamically adjusting the feed speed, spindle speed and cutting depth, ensuring that the machine tool can still maintain stable operation in an unhealthy state. Based on the output of the machine learning model, the intelligent control strategy can adaptively adjust various working parameters to avoid potential failures or operating errors affecting the processing quality. Ultimately, the present invention not only improves production efficiency and reduces maintenance costs, but also optimizes the working stability of precision CNC machine tools, significantly improving the reliability and service life of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0018] Figure 1 The figure is a flow chart of the method of the present invention.

[0019] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Example 1, please refer to Figure 1 As shown, the intelligent control method for the operation status of a precision CNC machine tool described in this embodiment includes the following steps: S1: real-time collection of actual operation data of precision CNC machine tools in several time periods, wherein the actual operation data includes the working temperature, load, current and pressure data of the machine tools, and real-time status data during the machining process, including cutting force data and tool wear data; S2: Preprocess the collected actual operation data and extract abnormal values ​​of temperature, load, current and pressure in the actual operation data through the Z-score algorithm; S3: Based on the machine learning algorithm, a prediction model is constructed for real-time identification and prediction of the machine tool operation status, wherein the prediction model is trained based on abnormal values ​​of temperature, load, current and pressure to identify the machine tool operation status; S4: Analyze the real-time status data using the trained prediction model, detect abnormal behavior of the machine tool in real time, and evaluate the health status of the precision CNC machine tool; S5: Automatically generate an intelligent control strategy based on the operating status evaluation result, wherein the intelligent control strategy includes dynamically adjusting the machine tool feed speed, spindle speed and cutting depth.

[0022] In S1, the specific steps include: Step 1: Set the data collection time period According to the specific requirements of the processing task and the working characteristics of the machine tool, several collection time periods are set. The data collection interval in each time period can be optimized and adjusted according to the dynamic changes of the processing process, such as time units such as per second, per minute or per hour.

[0023] Step 2: Working temperature data collection a) In each time period, the temperature sensor is used to monitor the working temperature of the main components of the machine tool in real time.

[0024] b) Focus on monitoring the temperature of the machine tool's spindle, motor, guide rail, cooling system and other parts.

[0025] c) Record the temperature data of each component to facilitate subsequent analysis of whether the machine tool has problems such as overheating or temperature abnormalities.

[0026] Step 3: Load data collection a) In each time period, monitor the load of the machine tool in real time, especially the load of the main motor and servo motor.

[0027] b) Obtain the workload data of the machine tool drive system through the load sensor.

[0028] c) Compare the machine tool load data with the set normal load range to promptly detect any abnormal conditions such as overload or underload.

[0029] Step 4: Current data collection a) Use current sensors to monitor the working current of each electrical component of the machine tool in real time, especially the current data of the main motor, servo motor, cooling system and other equipment.

[0030] b) The collection of current data can help identify whether there are abnormalities in electrical equipment, such as excessive or insufficient current.

[0031] c) Record the current change trend and analyze whether there are any abnormal fluctuations or early warning signs of equipment failure.

[0032] Step 5: Pressure data collection a) Install pressure sensors at key components of the machine tool such as the hydraulic system and air pressure system to monitor the pressure values ​​of the hydraulic oil and air pressure in real time.

[0033] b) Collect pressure data of each working part of the machine tool, especially in complex cutting processes, to ensure the normal operation of the hydraulic and pneumatic systems.

[0034] c) Compare the pressure data with the predetermined range to promptly detect abnormal pressure and avoid the impact of hydraulic or pneumatic system failures on processing quality.

[0035] Step 6: Cutting force data collection a) During the machining process, a cutting force sensor is installed to monitor the cutting force in real time.

[0036] b) The collected data include the size, direction and torque changes of the cutting force, reflecting the interaction force between the tool and the workpiece during the machining process.

[0037] c) Evaluate machining stability based on changes in cutting force data, and promptly identify machining errors or tool wear that may be caused by excessive cutting force.

[0038] Step 7: Tool wear data collection a) Use tool wear sensors to monitor tool wear during machining and collect tool wear data in real time.

[0039] b) Indirect detection of tool wear changes through auxiliary tools such as force sensors, vision sensors or vibration sensors.

[0040] c) Record tool wear trend data to provide a basis for subsequent fault warning and tool replacement.

[0041] Step 8: Data storage and transmission a) Real-time data storage is performed at each collection point, and all collected data is transmitted to the central control system or cloud platform through the industrial data acquisition system.

[0042] b) When storing data, it is necessary to ensure efficient data compression and denoising to ensure data accuracy and real-time performance.

[0043] c) Data transmission should have high reliability and low latency to facilitate real-time monitoring and subsequent analysis.

[0044] Step 9: Data synchronization and fusion a) Synchronize the multi-dimensional data (temperature, load, current, pressure, cutting force, tool wear, etc.) collected above and merge them into a complete machine tool operation status data set.

[0045] b) Ensure the timeliness and consistency of data to facilitate subsequent intelligent analysis and fault prediction.

[0046] S2: The specific steps include: Data preprocessing a) Pre-process the actual operation data collected from various sensors of the machine tool to ensure the accuracy and consistency of the data.

[0047] b) Preprocessing steps include denoising, missing value filling, outlier correction and standardization to eliminate the impact of environmental noise, sensor errors and data missing on subsequent analysis.

[0048] c) Data denoising uses appropriate filtering algorithms, such as low-pass filtering or median filtering, to remove irrelevant fluctuations from the external environment or the sensor itself.

[0049] Z-score normalization a) For each collected operating parameter (such as temperature, load, current, pressure, etc.), the Z-score algorithm is used for standardization.

[0050] b) The calculation formula of the Z-score algorithm is: ; Where X is the actual value of a data point, μ is the mean of the parameter, and σ is the standard deviation of the parameter.

[0051] c) The standardized data will have a distribution with a mean of 0 and a standard deviation of 1, which can eliminate the dimensional differences of different data dimensions and make the data comparable.

[0052] Outlier Detection a) After normalization, perform outlier detection on each data point. The specific method is as follows: b) According to the principle of the Z-score algorithm, when the Z-score value exceeds a certain set threshold (such as 3 or -3), the data point is considered an outlier. That is: >3 This means that the data point deviates from the mean by more than 3 standard deviations and has a high probability of being an abnormal data point.

[0053] c) For example, if the Z-score of a certain temperature data is 4, the temperature data is considered abnormal, which may indicate an abnormal state of a malfunctioning temperature sensor or overheating of the machine tool.

[0054] Outlier correction and labeling a) Mark detected outliers and make corrections as needed.

[0055] b) Methods for outlier correction include: Use interpolation methods (such as linear interpolation or polynomial interpolation) to fill in abnormal values ​​of parameters such as temperature and current; For relatively stable parameters such as load and pressure, smoothing algorithms (such as moving average method) are used to remove sudden fluctuations.

[0056] c) Mark abnormal data points to facilitate special processing in subsequent analysis.

[0057] Output cleaned data a) The data after preprocessing and outlier detection will be used in subsequent steps, such as intelligent model training, operation status analysis, etc.

[0058] b) These cleaned data can effectively improve the accuracy of subsequent algorithm (such as fault prediction, status assessment) analysis and avoid analysis errors caused by outliers.

[0059] S3: Construct a dataset for training machine learning models by extracting outliers and normalizing data. The dataset should include the following fields: Characteristic variables (input data): including real-time data collected from machine tools such as operating temperature, load, current, and pressure; Label variable (output data): includes marking whether there is an abnormality (i.e., whether a fault or abnormal state occurs). It can use binary classification labels (normal / abnormal) or multi-classification labels (such as temperature abnormality, current abnormality, etc.).

[0060] Dataset Partitioning a) Divide the dataset into a training set and a test set. The common ratio is 70% for training and 30% for testing (or use methods such as cross-validation).

[0061] b) For real-time data, it may be necessary to divide the validation set to adjust the model parameters and ensure the generalization ability of the model.

[0062] Feature Selection and Engineering a) Perform feature selection based on the temperature, load, current, pressure and other characteristics in the actual operation data.

[0063] b) Select features that have a significant impact on outlier prediction, such as selecting features with a high correlation with abnormal states through correlation analysis.

[0064] c) If necessary, feature engineering, such as polynomial feature expansion, feature normalization, etc., can be performed to improve the predictive performance of the model.

[0065] Model Training a) Use the random forest algorithm to train the training set data.

[0066] b) During the training process, the random forest will automatically learn the complex relationship between features and abnormal states through the ensemble learning method of decision trees. Each tree will randomly select a subset of features and a subset of samples during the training process, gradually build multiple decision trees, and finally obtain the final prediction result through voting.

[0067] c) For the outlier detection problem, a binary classification model is usually used, where normal state and abnormal state are the target categories. During the training process, the model will learn how to distinguish between the two states.

[0068] Model optimization and parameter adjustment a) Optimize the model's hyperparameters, such as the number of trees, tree depth, maximum number of features, etc., through methods such as cross-validation or grid search.

[0069] b) Common optimization methods include adjusting: n_estimators: the number of trees in the random forest; max_depth: the maximum depth of each tree; min_samples_split: the minimum number of samples required to split a node; max_features: Maximum number of features considered by each tree, etc.

[0070] c) Use the adjusted model to retrain the training set and ensure the model's generalization ability and prediction effect on new data.

[0071] Model Evaluation a) Use the test set to evaluate the trained random forest model and check its ability to identify abnormal values ​​(such as abnormal fluctuations in temperature, load, current, pressure, etc.).

[0072] b) Common evaluation indicators include: Accuracy: The proportion of correct predictions; Precision: The accuracy of predicting abnormal conditions; Recall: The ratio of the identified abnormal states to all abnormal states; F1-score: the harmonic mean of precision and recall; ROC curve and AUC value: judge the classification ability and discrimination of the model.

[0073] Model real-time prediction a) After the model has passed the test, it can be integrated into the machine tool control system for real-time prediction.

[0074] b) For machine tool data collected in real time, the trained model is used to predict new data and identify whether there are outliers.

[0075] c) If the model predicts an abnormal state (such as overtemperature, current overload, etc.), an alarm mechanism can be activated to remind the operator to intervene or adjust parameters.

[0076] Continuous learning and model updating a) During actual use, the working environment of the machine tool may change, so the model needs to be continuously learned and updated.

[0077] b) Regularly collect new data and use it to retrain the model so that the model can adapt to new working conditions and improve the accuracy of identifying abnormal conditions.

[0078] c) Incremental learning can be used to enable the model to be updated quickly when new data is received without the need to retrain the entire model.

[0079] S4: using the trained prediction model to analyze the real-time status data, detect abnormal behavior of the machine tool in real time, and evaluate the health status of the precision CNC machine tool, wherein the real-time status data during the machining process includes cutting force data and tool wear data; After analyzing the obtained cutting force data, the cutting force abnormality index is generated. The method for obtaining the cutting force abnormality index is as follows: The cutting force data collected within a period of time is set as , where n is the total number of data points, It's in time Cutting force values ​​on the cutting surface; prepare a set of reference data representing normal cutting force , where m is the length of the normal data sequence.

[0080] Construct a distance matrix D(i,j), where D(i,j) represents the distance between the i-th value of the sequence F(t) and the sequence The distance between the jth values ​​of is usually calculated using the Euclidean distance: ; Define the cumulative distance matrix C(i,j): ; where C(i,j) represents the value from the 1st to the i-th value of F(t), and The minimum cumulative distance between the first value and the jth value of , and the DTW distance is calculated: ;in, is the final cumulative distance value, indicating the difference between F(t) and The overall similarity between them; generate the cutting force abnormality index, the expression is: ;in, It is the maximum threshold of DTW distance, which can be calculated by taking the maximum value of DTW of multiple normal data or set based on experience. Usually, the maximum distance of DTW is set based on factors such as machine operating parameters and cutting conditions. It is the cutting force abnormality index.

[0081] Normal state: If the abnormal index is close to 0, it means that the current cutting force data is highly similar to the normal mode and the cutting process is in a normal state.

[0082] Abnormal state: If the abnormal index is close to 1, it means that the current cutting force data has a low similarity with the normal mode and there is an obvious abnormality.

[0083] Critical state: If the abnormal index is between 0 and 1 and gradually increases, it may indicate that the cutting process is approaching an abnormal state and deserves attention.

[0084] After analyzing the acquired tool wear data, a tool wear deviation index is generated. The method for obtaining the tool wear deviation index is as follows: The tool wear data obtained is a time series , where v is the number of data points, Represents at time The tool wear data is normalized so that the data has the same range (e.g. [0, 1]). The tool wear data is trained using an LSTM neural network, which predicts future wear values ​​by learning historical wear patterns.

[0085] LSTM model calculation Tool wear prediction at a given point in time , and then calculate the prediction error: ; where e(t+1) represents the error at the t+1th moment, for The measurement value of tool wear at a point in time; by calculating the prediction error e(t+1), the tool wear deviation index SF is obtained, which is expressed as: ;in, is the maximum prediction error in the normal wear mode. Normal wear: If the deviation index is close to 0, it means that the tool wear is consistent with the normal mode. Abnormal wear: If the deviation index is close to 1, it means that the tool wear deviates greatly from the normal mode.

[0086] The cutting force anomaly index and tool wear deviation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts the health status assessment value label of the precision CNC machine tool with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the health status assessment value labels of all precision CNC machine tools as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The health status assessment value of the precision CNC machine tool is determined according to the model output results, wherein the machine learning model is a pre-trained random forest model.

[0087] The obtained health status assessment value of the precision CNC machine tool is compared with the health status assessment threshold value pre-set according to historical data. If the health status assessment value of the precision CNC machine tool is greater than or equal to the pre-set health status assessment threshold value, it means that the precision CNC machine tool is in a healthy state, and no warning signal is generated at this time; if the health status assessment value of the precision CNC machine tool is less than the pre-set health status assessment threshold value, it means that the precision CNC machine tool is in an unhealthy state, and a warning signal is generated at this time.

[0088] S5: Automatically generate an intelligent control strategy based on the operating status evaluation result, wherein the intelligent control strategy includes dynamically adjusting the machine tool feed speed, spindle speed and cutting depth.

[0089] When the health status assessment value of precision CNC machine tools If the value is less than the preset health status assessment threshold, it means that the machine tool is in an unhealthy state and an intelligent control strategy needs to be generated immediately to dynamically adjust the operating parameters of the machine tool to reduce the machine tool load, extend its service life, and avoid further damage.

[0090] The core of intelligent control strategy is to dynamically adjust the feed speed , spindle speed and cutting depth To reduce the workload of the machine tool and avoid increasing the wear of the machine tool as much as possible. The specific control goal is to balance the load and reduce the risk of failure by adjusting these key parameters until the machine tool is restored to a healthy state.

[0091] Feed speed It directly affects the cutting force and cutting temperature of the machine tool. When the machine tool is in an unhealthy state, the feed speed should be reduced to reduce the burden on the machine tool and the temperature rise during the cutting process. The formula for adjusting the feed speed is: ;in, is the current feed speed, It is the standard feed rate when the machine tool is working normally (unit: mm / min). γ (γ>0) is the index of the adjustment factor, which is used to control the change ratio of the feed rate. It is usually taken as 0.5≤γ≤1.0, which means that the feed rate will be significantly reduced when the machine tool is in poor health. is the health status assessment threshold, and its value can be determined based on experimental data or the recommended value of the machine tool manufacturer.

[0092] When the health status assessment value of precision CNC machine tools When the feed speed is less than the preset health status assessment threshold, that is, when the machine tool is in an unhealthy state, the feed speed The feed rate is adjusted in proportion to the health status assessment value to reduce the load on the machine tool. The sensitivity of the adjustment speed can be determined by controlling the index γ. A large γ value will result in a significant drop in the feed rate.

[0093] Spindle speed It will affect the cutting force and temperature during the cutting process. In order to reduce the load and slow down the wear, the spindle speed can be dynamically adjusted according to the health status assessment value. The formula for adjusting the spindle speed is: ;in: is the spindle speed at the current moment. It is the standard spindle speed (unit: rpm) when the machine tool is working normally. δ is the index of the adjustment factor, which is used to control the change ratio of the spindle speed. It is usually taken as 0.5≤δ≤1.0. The control index δ determines the sensitivity of the adjustment. A larger δ will lead to a significant reduction in the spindle speed.

[0094] Cutting depth It has a direct impact on the cutting force and the load of the machine tool. In order to reduce the wear of the machine tool, the cutting depth can be dynamically adjusted according to the health status assessment value. The formula for adjusting the cutting depth is: ;in, is the cutting depth at the current moment, It is the standard cutting depth when the machine tool is working normally (unit: mm). β (β>0) is the index of the adjustment factor, which is used to control the change ratio of the cutting depth. It is usually taken as 0.5≤β≤1.0. The control index β determines the degree of adjustment of the cutting depth. A larger β value will lead to a significant reduction in the cutting depth.

[0095] In this embodiment, first, the actual operation data of the precision CNC machine tool in several time periods are collected in real time, including the working temperature, load, current and pressure data of the machine tool, as well as the real-time status data during the processing, such as cutting force and tool wear data. Then, the Z-score algorithm is used to pre-process the actual operation data collected to extract the abnormal values ​​of temperature, load, current and pressure. Next, a prediction model is built based on the machine learning algorithm, and the model is trained to identify and predict these abnormal values ​​to achieve real-time identification of the operating status of the machine tool. The real-time status data is analyzed using the trained prediction model to detect the abnormal behavior of the machine tool in real time and evaluate its health status. When the evaluation results show that the health status of the machine tool is abnormal, the system automatically generates an intelligent control strategy to dynamically adjust the feed speed, spindle speed and cutting depth of the machine tool, thereby optimizing the operating status of the machine tool, extending the service life of the equipment, and reducing the risk of failure.

[0096] Example 2, please refer to Figure 2 As shown, the intelligent control system for the operation status of a precision CNC machine tool described in this embodiment includes a data acquisition module, a data processing module, a machine learning training module, a detection module and an intelligent control module: Data acquisition module: real-time acquisition of actual operation data of precision CNC machine tools within a certain period of time, the actual operation data including the working temperature, load, current and pressure data of the machine tools, and real-time status data during the machining process, including cutting force data and tool wear data; Data processing module: pre-processes the collected actual operation data and extracts abnormal values ​​of temperature, load, current and pressure in the actual operation data through the Z-score algorithm; Machine learning training module: Based on the machine learning algorithm, a prediction model is constructed for real-time identification and prediction of the machine tool operation status. The prediction model is trained based on abnormal values ​​of temperature, load, current and pressure to identify the machine tool operation status; Detection module: using the trained prediction model to analyze real-time status data, detect abnormal behavior of machine tools in real time, and evaluate the health status of precision CNC machine tools; Intelligent control module: According to the operation status evaluation results, the intelligent control strategy is automatically generated, and the intelligent control strategy includes dynamic adjustment of the machine tool feed speed, spindle speed and cutting depth.

[0097] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0098] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0099] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0100] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. An intelligent control method for the operation status of a precision CNC machine tool, characterized in that: The following steps are involved: S1: real-time collection of actual operation data of precision CNC machine tools in several time periods, wherein the actual operation data includes the working temperature, load, current and pressure data of the machine tools, and real-time status data during the machining process, including cutting force data and tool wear data; S2: Preprocess the collected actual operation data and extract abnormal values ​​of temperature, load, current and pressure in the actual operation data through the Z-score algorithm; S3: Based on the machine learning algorithm, a prediction model is constructed for real-time identification and prediction of the machine tool operation status, wherein the prediction model is trained based on abnormal values ​​of temperature, load, current and pressure to identify the machine tool operation status; S4: Analyze the real-time status data using the trained prediction model, detect abnormal behavior of the machine tool in real time, and evaluate the health status of the precision CNC machine tool; S5: Automatically generate an intelligent control strategy based on the operating status evaluation result, wherein the intelligent control strategy includes dynamically adjusting the machine tool feed speed, spindle speed and cutting depth.

2. According to claim 1, a method for intelligently controlling the operating status of a precision CNC machine tool is characterized in that: In S2, the actual operation data collected from each sensor of the machine tool is preprocessed, and each collected operation data is standardized using the Z-score algorithm; the calculation formula of the Z-score algorithm is: ; Where X is the actual value of a data point, μ is the mean of the parameter, and σ is the standard deviation of the parameter; the standardized data will have a distribution with a mean of 0 and a standard deviation of 1. After standardization, outlier detection is performed on each data point; according to the principle of the Z-score algorithm, when the Z-score value exceeds the set threshold, the data point is considered to be an outlier.

3. The intelligent control method for the operation status of a precision CNC machine tool according to claim 1 is characterized in that: In S3, a data set for training a machine learning model is constructed by extracting outliers and standardized data. The data set includes the operating temperature, load, current, and pressure of the machine tool in the input data, and whether the output data is in an abnormal state. Divide the data set into a training set and a test set. For real-time data, a validation set needs to be divided to adjust model parameters. Perform feature selection based on the temperature, load, current, and pressure characteristics in the actual operating data. Use the random forest algorithm to train the training set data; During the training process, random forests automatically learn the complex relationship between features and abnormal states through the ensemble learning method of decision trees; During the training process, each tree will randomly select a subset of features and a subset of samples, gradually build multiple decision trees, and finally get the final prediction result through voting; A binary classification model is used, where normal state and abnormal state are the target categories. During the training process, the model is used to distinguish between the two states; The hyperparameters of the model are optimized through cross-validation, and the training set is retrained using the adjusted model to ensure the generalization ability of the model.

4. The intelligent control method for the operation status of a precision CNC machine tool according to claim 3 is characterized in that: In S4, the obtained cutting force data is analyzed to generate a cutting force abnormality index. The method for obtaining the cutting force abnormality index is as follows: The cutting force data collected within a period of time is set as , where n is the total number of data points, It's in time Cutting force value on Prepare a set of reference data representing normal cutting forces , where m is the length of the normal data sequence; Construct a distance matrix D(i,j), where D(i,j) represents the distance between the i-th value of the sequence F(t) and the sequence The distance between the jth values ​​of : ; Define the cumulative distance matrix C(i,j): ; where C(i,j) represents the value from the 1st to the i-th value of F(t), and The minimum cumulative distance between the first value and the jth value of , and the DTW distance is calculated: ;in, is the final cumulative distance value, indicating the difference between F(t) and The overall similarity between them; generate the cutting force abnormality index, the expression is: ;in, is the maximum threshold of DTW distance, It is the cutting force abnormality index.

5. The intelligent control method for the operation status of a precision CNC machine tool according to claim 4 is characterized in that: In S4, the acquired tool wear data is analyzed to generate a tool wear deviation index, and the tool wear deviation index is obtained by: The tool wear data obtained is a time series , where v is the number of data points, t represents the sampling time point, Represents at time The tool wear measurement value at the time of , the tool wear data is normalized, and the tool wear data is trained using the LSTM neural network. LSTM predicts future wear values ​​by learning historical wear patterns; the LSTM model is used to calculate the tool wear value at Tool wear prediction at a point in time , and then calculate the prediction error: ; where e(t+1) represents the error at the t+1th moment, for The tool wear measurement value at a time point; by calculating the prediction error e(t+1), the tool wear deviation index SF is obtained, and the expression is: ;in, is the maximum prediction error under normal wear mode.

6. The intelligent control method for the operation status of a precision CNC machine tool according to claim 5, characterized in that: The cutting force anomaly index and tool wear deviation index are converted into comprehensive feature vectors, and the comprehensive feature vectors are used as the input of the machine learning model. The machine learning model predicts the health status assessment value label of the precision CNC machine tool with each group of comprehensive feature vectors as the prediction target, and minimizes the sum of prediction errors of the health status assessment value labels of all precision CNC machine tools as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The health status assessment value of the precision CNC machine tool is determined according to the model output results, wherein the machine learning model is a pre-trained random forest model.

7. The intelligent control method for the operation status of a precision CNC machine tool according to claim 6 is characterized by: The obtained health status assessment value of the precision CNC machine tool is compared with the health status assessment threshold value pre-set according to historical data. If the health status assessment value of the precision CNC machine tool is greater than or equal to the pre-set health status assessment threshold value, it means that the precision CNC machine tool is in a healthy state, and no warning signal is generated at this time; if the health status assessment value of the precision CNC machine tool is less than the pre-set health status assessment threshold value, it means that the precision CNC machine tool is in an unhealthy state, and a warning signal is generated at this time.

8. The intelligent control method for the operation status of a precision CNC machine tool according to claim 1 is characterized by: In S5, when the health status assessment value of the precision CNC machine tool If the value is less than the preset health status assessment threshold, it means that the machine tool is in an unhealthy state and an intelligent control strategy needs to be generated immediately to dynamically adjust the operating parameters of the machine tool. The core of the intelligent control strategy is to dynamically adjust the feed speed. , spindle speed and cutting depth To reduce the workload of the machine tool.

9. The intelligent control method for the operation status of a precision CNC machine tool according to claim 8, characterized in that: Feed speed It directly affects the cutting force and cutting temperature of the machine tool. When the machine tool is in an unhealthy state, the feed speed should be reduced to reduce the burden on the machine tool and the temperature rise during the cutting process. The formula for adjusting the feed speed is: ;in, is the current feed speed, is the standard feed speed when the machine tool is working normally, γ is the index of the adjustment factor, which is used to control the change ratio of the feed speed. assessing thresholds for health states; Spindle speed It will affect the cutting force and temperature during the cutting process. The spindle speed is adjusted dynamically according to the health status assessment value. The formula for adjusting the spindle speed is: ;in: is the spindle speed at the current moment, is the standard spindle speed when the machine tool is working normally, and δ is the index of the adjustment factor, which is used to control the change ratio of the spindle speed; Cutting depth It has a direct impact on the cutting force and the load of the machine tool. The cutting depth is dynamically adjusted according to the health status assessment value. The formula for adjusting the cutting depth is: ;in, is the cutting depth at the current moment, is the standard cutting depth when the machine tool is working normally, and β is the exponent of the adjustment factor, which is used to control the change ratio of the cutting depth.

10. An intelligent control system for the operation status of a precision CNC machine tool, used to implement an intelligent control method for the operation status of a precision CNC machine tool according to any one of claims 1 to 9, characterized in that: Including data acquisition module, data processing module, machine learning training module, detection module and intelligent control module: Data acquisition module: real-time acquisition of actual operation data of precision CNC machine tools within a certain period of time, the actual operation data including the working temperature, load, current and pressure data of the machine tools, and real-time status data during the machining process, including cutting force data and tool wear data; Data processing module: pre-processes the collected actual operation data and extracts abnormal values ​​of temperature, load, current and pressure in the actual operation data through the Z-score algorithm; Machine learning training module: Based on the machine learning algorithm, a prediction model is constructed for real-time identification and prediction of the machine tool operation status. The prediction model is trained based on abnormal values ​​of temperature, load, current and pressure to identify the machine tool operation status; Detection module: using the trained prediction model to analyze real-time status data, detect abnormal behavior of machine tools in real time, and evaluate the health status of precision CNC machine tools; Intelligent control module: According to the operation status evaluation results, the intelligent control strategy is automatically generated, and the intelligent control strategy includes dynamic adjustment of the machine tool feed speed, spindle speed and cutting depth.

Images