Numerical control machine tool cutter monitoring method, device, equipment, medium and program product

Abstract

The invention provides a numerical control machine tool cutter monitoring method, device and equipment, a medium and a program product, and relates to the technical field of numerical control machine tool machining. Inputting the machining load data into a tool monitoring model to obtain reference data of a tool load output by the tool monitoring model; the tool monitoring model is obtained based on processing load sample data training; and determining cutter failure information of the numerical control machine tool based on a comparison result of the processing load data and the reference data. The reference data of the tool load is determined through the machine learning algorithm, and then whether the tool fails is determined based on the machining load data and the reference data, so that the tool abrasion and fracture conditions in the self-adaptive control process of the feeding speed of the numerical control machine tool are monitored in real time, the tool monitoring accuracy and the machining efficiency are improved, and the machining cost is reduced. And machining interruption and workpiece scrapping caused by the cutter problem are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of numerical control machine tool processing, and in particular to a numerical control machine tool tool monitoring method, device, equipment, medium and program product. Background Art

[0002] At present, there are many methods for detecting tool wear and fracture of CNC machine tools, but they all have certain limitations. The direct measurement method usually involves regular shutdown during the processing process, using a measuring tool to directly measure the dimensional change of the tool to determine the wear of the tool. The experience-based prediction method is to establish a tool wear model based on previous processing experience and tool usage data to predict the service life of the tool, but the accuracy of this method depends on a large amount of accurate historical data and stable processing conditions. In the actual processing process, the processing conditions are often complex and changeable. For example, the material properties of different batches of workpieces may be different, and the cutting parameters may also be adjusted according to the processing requirements. These factors will lead to a decrease in the accuracy of the experience-based prediction method. Based on this, the low accuracy of CNC machine tool tool monitoring has become a problem that needs to be solved urgently. Summary of the invention

[0003] The present invention provides a CNC machine tool tool monitoring method, device, equipment, medium and program product, which are used to solve the defect of low accuracy of CNC machine tool tool monitoring in the prior art, realize real-time monitoring of tool wear and breakage during the CNC machine tool feed speed adaptive control process, improve the accuracy of tool monitoring and processing efficiency, and reduce processing interruptions and workpiece scrapping caused by tool problems.

[0004] The present invention provides a method for monitoring a tool of a numerically controlled machine tool, comprising the following steps: Obtain processing load data of CNC machine tools; Inputting the processing load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on the processing load sample data; Based on the comparison result between the machining load data and the reference data, tool failure information of the numerically controlled machine tool is determined.

[0005] According to a CNC machine tool tool monitoring method provided by the present invention, the reference data includes an upper limit value and an upper limit time of the tool load; the tool failure information includes tool wear; and the tool failure information of the CNC machine tool is determined based on a comparison result of the machining load data and the reference data, including: If the comparison result is that the machining load of the machining load data is greater than the upper limit value, and the duration for which the machining load is greater than the upper limit value is greater than the upper limit time, it is determined that the tool is worn.

[0006] According to a CNC machine tool tool monitoring method provided by the present invention, the reference data further includes a lower limit value and a lower limit time of the tool load; the tool failure information further includes tool breakage; and the tool failure information of the CNC machine tool is determined based on a comparison result of the machining load data and the reference data, including: If the comparison result is that the machining load of the machining load data is less than the lower limit value, and the duration for which the machining load is less than the lower limit value is greater than the lower limit time, it is determined that the tool is broken.

[0007] According to a CNC machine tool tool monitoring method provided by the present invention, the comparison result between the machining load data and the reference data is determined based on the following method: Converting the processing load data and the reference data from the time series domain to the processing length domain; After conversion into the processing length domain, aligning the processing load data with the reference data; The aligned machining load data and the reference data are compared to obtain the comparison result.

[0008] According to a CNC machine tool tool monitoring method provided by the present invention, the tool monitoring model is trained based on the following method: Extracting features from the processing load sample data, and constructing a feature index set based on the extracted features; Smoothing the feature index set, and using multiple evaluation indicators to evaluate the performance of features in the smoothed feature index set; Based on the performance evaluation results, features related to the prediction results are selected to construct a sensitive feature set; Model training is performed based on the sensitive feature set to obtain the tool monitoring model.

[0009] According to a CNC machine tool tool monitoring method provided by the present invention, the method further includes: When the machining load of the CNC machine tool is greater than the upper limit of the tool load, reducing the feed speed of the tool; When the machining load of the numerically controlled machine tool is less than a lower limit value of the tool load, the feed speed of the tool is increased.

[0010] The present invention also provides a CNC machine tool tool monitoring device, comprising the following modules: An acquisition module, used for acquiring processing load data of a CNC machine tool; A prediction module, used for inputting the processing load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on the processing load sample data; A determination module is used to determine tool failure information of the numerically controlled machine tool based on a comparison result of the machining load data and the reference data.

[0011] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, any one of the above-mentioned CNC machine tool tool monitoring methods is implemented.

[0012] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method for monitoring a tool for a CNC machine tool as described above is implemented.

[0013] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any one of the above-mentioned methods for monitoring a tool for a numerically controlled machine tool.

[0014] The CNC machine tool tool monitoring method, device, equipment, medium and program product provided by the present invention obtain the processing load data of the CNC machine tool; input the processing load data into the tool monitoring model to obtain the reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on the processing load sample data; based on the comparison result of the processing load data and the reference data, the tool failure information of the CNC machine tool is determined. The present invention determines the reference data of the tool load through a machine learning algorithm, and then determines whether the tool is failed based on the processing load data and the reference data, thereby realizing real-time monitoring of the tool wear and breakage during the adaptive control of the feed speed of the CNC machine tool, improving the accuracy of tool monitoring and processing efficiency, and reducing processing interruptions and workpiece scrapping caused by tool problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a flow chart of the CNC machine tool tool monitoring method provided by the present invention.

[0017] Figure 2 The present invention provides a flowchart for determining the wear state of a tool during the adaptive control of the feed speed of a numerically controlled machine tool.

[0018] Figure 3 The present invention provides a flow chart for determining the fracture state of a tool during the adaptive control of the feed speed of a numerically controlled machine tool.

[0019] Figure 4 It is a system architecture diagram provided by the present invention.

[0020] Figure 5 It is an adaptive control strategy diagram during the adaptive control of the feed speed of a numerically controlled machine tool provided by the present invention.

[0021] Figure 6 It is a determination strategy diagram for monitoring tool wear during the adaptive control of the feed speed of a numerically controlled machine tool provided by the present invention.

[0022] Figure 7 It is a determination strategy diagram for monitoring tool breakage during the adaptive control of the feed speed of a numerically controlled machine tool provided by the present invention.

[0023] Figure 8 The present invention is a flow chart of a method for monitoring tool wear and breakage during adaptive control of feed speed of a numerically controlled machine tool.

[0024] Fig. 9 The figure is a schematic diagram of the structure of the tool monitoring device for a numerically controlled machine tool provided by the present invention.

[0025] Fig.10 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings 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.

[0027] In the process of CNC machine tool processing, tool wear and fracture are the main failure forms of tools, which are related to manufacturing accuracy and product surface quality, and are key factors affecting processing quality and efficiency. Tool wear and fracture are complex processes that are affected by multiple factors. Cutting force is the main force that the tool bears during the cutting process. When the cutting force is too large, the tool will cause greater deformation and stress concentration, and the heat generated during the cutting process will increase the tool temperature. On the one hand, high temperature will reduce the hardness of the tool material, thereby affecting the cutting performance and wear resistance of the tool. On the other hand, the rapid change of temperature will generate thermal stress inside the tool. When the thermal stress exceeds the strength limit of the tool material, it will cause cracks and fractures in the tool. Different workpiece materials have different mechanical properties and processing difficulties. For example, materials such as alloy steel and titanium alloy with higher hardness have more obvious wear effects on the tool during processing. The high hardness and high strength of these materials will make the tool edge more easily worn, and also increase the cutting force and cutting temperature, further aggravating the loss of the tool. Some materials with high viscosity, such as aluminum alloy, are prone to built-up edge during the cutting process. The built-up edge will continuously impact the tool edge, causing increased tool wear and possibly affecting the machining accuracy.

[0028] In the related art, there are many methods for detecting tool wear and fracture, but they have certain limitations. The direct measurement method usually involves regular shutdown during the machining process, using a measuring tool to directly measure the dimensional change of the tool to determine the wear of the tool. The experience-based prediction method is to establish a tool wear model based on past machining experience and tool usage data to predict the tool life. For example, the tool wear is predicted based on factors such as the tool machining time and the number of workpieces processed. However, the accuracy of this method depends on a large amount of accurate historical data and stable machining conditions. In actual machining processes, machining conditions are often complex and changeable. For example, the material properties of workpieces in different batches may differ, and cutting parameters may also be adjusted according to machining requirements. These factors will lead to a decrease in the accuracy of experience-based prediction methods.

[0029] In related technologies, the tool wear and fracture state monitoring method realizes the identification of tool wear state by establishing a mapping relationship between signal features and tool wear state. Among them, the most commonly used identification methods in the field of tool wear state monitoring are mainly divided into two categories: traditional feature learning and deep learning methods. Traditional feature learning methods include hidden Markov models, support vector machines, random forest algorithms, K-means algorithms, etc. Tool wear state monitoring based on deep learning uses a neural network to express the mapping relationship between features and wear states with a networked topological structure, including convolutional neural networks, long short-term memory neural networks, transfer learning, sparse autoencoding, etc.

[0030] Current research is mostly applied to situations where machining parameters are constant, the tool moves repeatedly at a constant speed along a straight path, and the machining conditions are designed and predicted in advance. However, in complex time-varying machining conditions, as the tool feeds along the curved path and the workpiece material is continuously removed, the machining parameters will change dramatically, and this change is currently difficult to predict in advance before machining, and it is also impossible to perceive in real time during machining. This strongly variable condition characteristic makes the existing simplified condition monitoring model inapplicable, and accurate monitoring of tool wear becomes very difficult. Due to the real-time changes in machining parameters, these methods have the problem of inaccurate and untimely monitoring during the adaptive control machining process.

[0031] Based on the above problems, the present invention proposes a tool monitoring method for CNC machine tools, specifically, an intelligent monitoring method for tool wear and fracture during adaptive machining of CNC machine tools.

[0032] Combine the following Figure 1-Figure 10 The invention describes a CNC machine tool tool monitoring method, device, equipment, medium and program product.

[0033] Figure 1 It is a flow chart of the CNC machine tool tool monitoring method provided by the present invention, such as Figure 1 As shown, the method includes the following: Step 101: Obtain processing load data of a CNC machine tool.

[0034] A variety of sensors, measuring instruments and data acquisition systems can be used to monitor the processing load data of CNC machine tools. For example, current sensor modules can be installed in key parts such as the spindle motor of a CNC machine tool to collect the current of the spindle motor.

[0035] The machining load data of the CNC machine tool may include spindle load data, such as spindle current, spindle power, spindle torque, load percentage, etc.

[0036] Step 102: input the machining load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model.

[0037] The tool monitoring model is trained based on the processing load sample data. In one embodiment, the tool monitoring model is trained based on the following methods: extracting features from the processing load sample data, constructing a feature index set based on the extracted features; smoothing the feature index set, and using multiple evaluation indicators to evaluate the performance of the features in the smoothed feature index set; based on the performance evaluation results, selecting features related to the prediction results to construct a sensitive feature set; and training the model based on the sensitive feature set to obtain the tool monitoring model.

[0038] For example, when screening high-quality machining load data, by extracting the time domain and frequency domain feature indicators of CNC machine tool spindle current data, spindle load and other data, an alternative multi-frequency domain feature indicator set is constructed to fully reflect the degradation information of tool wear performance. Since the extracted feature indicators may contain some indicators with poor trends or irrelevant to the tool wear degradation process, these indicators have relatively low contribution to the prediction results, so it is necessary to evaluate and screen the feature indicators to obtain a sensitive feature set.

[0039] Specifically, the features of the processing load sample data are extracted, and a feature index set is constructed based on the extracted features. The feature index set is recorded as , the time series is ; Then, the characteristic index set is smoothed to remove short-term fluctuations in the data and highlight the long-term trend or law of the data. For example, the moving average method, exponential smoothing method, and wavelet transform can be used to smooth the characteristic index set. The smoothed characteristic index set can be decomposed into two parts: ,in, and They are the stationary trend term and the random residual term. The stationary trend term reflects the long-term trend or regularity in the data, while the random residual term represents the random fluctuation or noise in the data. Through this decomposition, the model can focus on the stationary trend information and ignore some irrelevant noise, thereby improving the accuracy of the prediction. Furthermore, using the correlation , monotonicity and robustness The three evaluation criteria are linearly combined to obtain the comprehensive evaluation criteria. Finally, the features with higher comprehensive evaluation indexes are selected to form the sensitive feature set of tool wear. .

[0040] After obtaining the sensitive feature set, the convolutional neural network model is used to train the tool wear state monitoring model. Specifically, the collected sensitive feature set of tool wear is used as the input data set, and the upper limit value, upper limit time, lower limit value and lower limit time of the tool load are obtained as the output data set through experimental data collection and calculation. For example, a sample set of 16 features of 5 groups of experiments is constructed, totaling 7452, 80% of which are randomly selected as training sets, and the remaining 20% ​​are test sets. The features of the cutting depth matrix are continuously extracted from the input data through the convolution process, and the matrix dimension is changed through average pooling. Finally, the matrix is ​​expanded into a one-dimensional array through the fully connected layer and the regression output layer to obtain a reference data array of size 4×1.

[0041] After the tool monitoring model is trained, the processing load data is input into the tool monitoring model to obtain reference data of the tool load output by the tool monitoring model, wherein the reference data may include the upper limit value, upper limit time, lower limit value and lower limit time of the tool load.

[0042] Optionally, the tool's machining load data can help predict the tool's remaining service life. For example, by observing the trend of load changes, the model can identify in advance whether the tool is close to failure, so that maintenance or replacement can be performed in advance. At the same time, by analyzing the tool load data, the cutting parameters (such as cutting speed, feed rate, cutting depth, etc.) can be optimized to make the machining process more efficient and the tool life longer. In addition, tool load monitoring can help maintain a stable load state during the cutting process to avoid excessive load that leads to reduced machining accuracy or machining failure.

[0043] Step 103: Determine tool failure information of the CNC machine tool based on a comparison result of the machining load data and the reference data.

[0044] The processing load data and the reference data are compared to obtain a comparison result. In one embodiment, the processing load data and the reference data are converted from the time series domain to the processing length domain; after the conversion to the processing length domain, the processing load data and the reference data are aligned; and the aligned processing load data and the reference data are compared to obtain a comparison result.

[0045] It is understandable that in the original time series domain, the processing load data and reference data are usually collected based on time (or based on the control cycle of the machine tool). The processing length domain refers to a coordinate system based on the distance the cutting tool moves during the processing, rather than time. Since the tool moves along the workpiece surface during the processing, its load and other status indicators are related to factors such as the processing progress, cutting depth, and feed rate.

[0046] For example, the tool movement trajectory, feed rate or actual cutting length can be obtained through the control signal or sensor information of the CNC machine tool. Then, according to the relationship between time and cutting process, the time series data at each moment is associated with the corresponding processing length value, so that the processing length value corresponding to each time point in the time series can be estimated by interpolation methods (such as linear interpolation, spline interpolation, etc.), thereby realizing the conversion of processing load data and reference data from the time series domain to the processing length domain.

[0047] The purpose of alignment is to ensure that the machining load data and reference data are compared at the same machining length position. When aligning data, it is necessary to ensure that the machining load data and reference data match in the machining length domain. For example, if the sampling points of the machining load data and reference data are not exactly the same, the sampling points of one set of data can be interpolated to the sampling points of the other set of data through interpolation methods to achieve data alignment. If the sampling rates of the two sets of data are different, the two sets of data can be resampled so that they have consistent values ​​at the same machining length points to achieve data alignment.

[0048] The embodiment of the present invention converts the machining load data and reference data from the time series domain to the machining length domain and performs alignment comparison, so that the data comparison is more accurate, the nonlinear relationship between the time domain and the machining process is eliminated, and the data is ensured to be compared at the same machining position to avoid errors caused by time inconsistency. At the same time, the accuracy of tool performance analysis is improved, which helps to better optimize machining parameters and improve machining quality and stability.

[0049] Tool failure information can include tool wear and tool breakage. Tool wear refers to the phenomenon that the tool edge gradually loses its sharpness and performance declines due to friction with the workpiece, cutting force and material influence during long-term use. Tool breakage refers to the phenomenon that the tool material suddenly breaks or cracks due to excessive cutting force, impact force, thermal stress and other effects during the processing.

[0050] In one embodiment, if the comparison result is that the machining load of the machining load data is greater than the upper limit value, and the duration of the machining load being greater than the upper limit value is greater than the upper limit time, then it is determined that the tool is worn. It is understandable that tool wear usually leads to an increase in cutting force. When the tool is worn, the cutting edge is no longer sharp and generates a larger cutting force, which may cause the machining load to be greater than the normal range. If the machining load continues at a higher level, exceeds the upper limit threshold, and this high load state continues for a long period of time, it means that the tool wear has reached a certain extent. Worn tools will lead to a decrease in cutting efficiency, requiring greater force to complete the same cutting task, which will cause an increase in load. A high load state that lasts for a long time indicates that the tool performance has significantly decreased.

[0051] For example, refer to Figure 2, the CNC machine tool performs adaptive control processing, and the judgment module compares the processing load data with the upper limit value of the reference data calculated by machine learning. If the processing load does not exceed the upper limit value, the monitoring system continuously monitors the processing load to determine whether the processing load exceeds the upper limit value; if the processing load exceeds the upper limit value, the judgment module determines whether the duration exceeds the upper limit time. If it exceeds the upper limit time, it determines that the tool is worn; if it does not exceed the upper limit time, continue to monitor whether it exceeds the upper limit value and the upper limit time.

[0052] In one embodiment, if the comparison result is that the machining load of the machining load data is less than the lower limit value, and the duration of the machining load being less than the lower limit value is greater than the lower limit time, it is determined that the tool is broken. It is understandable that when the tool is broken, the tool will instantly lose its cutting ability during the cutting process, causing the load to drop sharply, or even completely no longer bear the stress during the cutting process. At this time, the load will be far below the normal range. If the load is lower than the lower limit value for a long time, and this low load state continues for a period of time, it may be a signal that the tool is broken or seriously damaged. Tool breakage usually causes a sharp decrease in force during the cutting process, especially after the break occurs, the contact between the workpiece and the tool almost disappears, resulting in a significant reduction in load. The continuous low load state indicates that the tool cannot continue to work normally and may have been broken or seriously damaged.

[0053] For example, refer to Figure 3 , the CNC machine tool performs adaptive control processing, and the judgment module compares the processing load data with the lower limit value of the reference data calculated by machine learning. If the processing load is not lower than the lower limit value, the monitoring system continuously monitors the processing load to determine whether the processing load is lower than the lower limit value; if the processing load is lower than the lower limit value, the judgment module determines whether the duration exceeds the lower limit time. If it exceeds the lower limit time, it is determined that the tool is broken; if it does not exceed the lower limit time, it continues to monitor whether it is lower than the lower limit value and the lower limit time.

[0054] By monitoring the machining load data and comparing it with the upper limit value, upper limit time, lower limit value and lower limit time of the tool load, the wear and breakage state of the tool can be effectively judged, production downtime can be reduced and machining quality can be improved.

[0055] The CNC machine tool tool monitoring method provided by the embodiment of the present invention obtains the processing load data of the CNC machine tool; inputs the processing load data into the tool monitoring model to obtain the reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on the processing load sample data; based on the comparison result of the processing load data and the reference data, the tool failure information of the CNC machine tool is determined. The present invention determines the reference data of the tool load through a machine learning algorithm, and then determines whether the tool is failed based on the processing load data and the reference data, thereby realizing real-time monitoring of the tool wear and breakage during the adaptive control of the feed speed of the CNC machine tool, improving the accuracy of tool monitoring and processing efficiency, and reducing processing interruptions and workpiece scrapping caused by tool problems.

[0056] Based on the above embodiment, the method further includes: Step 110, when the machining load of the CNC machine tool is greater than the upper limit of the tool load, reducing the feed speed of the tool; Step 111 , when the machining load of the CNC machine tool is less than the lower limit of the tool load, increasing the feed speed of the tool.

[0057] When the machining load is greater than the upper limit of the tool load, it means that the force the tool bears during the cutting process has exceeded the designed load bearing range. At this time, the tool load is too high, and factors such as pressure and temperature during the cutting process may increase, resulting in increased wear of the tool and even possible tool breakage. By reducing the feed speed of the tool and reducing the amount of material cut per unit time, the burden on the tool is reduced, allowing the tool to withstand lower cutting forces and avoid overload. This helps the tool maintain a longer service life while reducing overheating and wear. Among them, the purpose of reducing the feed speed is to reduce the load at each point in a single cutting process and ensure that the tool works stably within a reasonable load range.

[0058] When the machining load is less than the lower limit of the tool load, it means that the tool is under a lighter load during the cutting process, and the cutting force is too small. The cutting effect and efficiency of the tool may be affected, resulting in low machining efficiency. By increasing the feed speed of the tool, the amount of cutting per unit time can be increased, so that the load of the tool during the cutting process increases, thereby ensuring that the tool works fully and improving cutting efficiency and machining quality. Among them, the purpose of increasing the feed speed is to make the tool work under a higher load and ensure that the cutting force is within a reasonable range, thereby optimizing machining efficiency and tool life.

[0059] The embodiment of the present invention optimizes the machining load by dynamically adjusting the feed speed, ensuring that the tool operates under the best working condition, thereby improving the machining efficiency and extending the tool life.

[0060] In order to further analyze and illustrate the CNC machine tool tool monitoring method provided by the present invention, refer to the following embodiments.

[0061] The embodiment of the present invention specifically provides a CNC machine tool tool wear and fracture monitoring system, referring to Figure 4 The system may include the following functional modules: Monitoring module, using monitoring equipment to obtain CNC machine tool processing load data; A training module calculates reference data through machine learning, such as calculating reference data based on the high-quality processing part of the processing load data obtained by the monitoring equipment; A storage module, used to store reference data such as an upper limit value, an upper limit time, a lower limit value and a lower limit time calculated by a training module; The determination module compares the reference data with the processing load data monitored during the processing based on the adaptive control, and determines that the tool is worn or broken when the processing load data monitored during the processing based on the adaptive control continuously exceeds or falls below the reference data. Specifically, if the processing load data exceeds the upper limit value and lasts for more than the upper limit time, the tool is determined to be worn; if the processing load data is lower than the lower limit value and lasts for more than the lower limit time, the tool is determined to be broken.

[0062] Optionally, the monitoring module may include the following submodules: A sensor module is used to monitor the processing load data of the CNC machine tool, which may include information such as the current size and vibration amplitude of the spindle motor. The module may include a current sensor and an acceleration sensor, etc.; The acquisition card module converts the analog signal of the sensor into a digital signal and transmits the digital signal to the computer; Computer module, displays real-time load data, deploys training module, judgment module and storage module.

[0063] Optionally, the training module may include the following submodules: A data conversion module converts the reference data and the real-time monitored processing load data from the time series domain to the processing length domain for more accurate comparison; The data processing module calculates reference data through a machine learning algorithm based on the high-quality machining load data monitored by the sensor module. The reference data may include the upper limit value, upper limit time, lower limit value and lower limit time of tool wear and fracture judgment. In the specific calculation, the machining load data of the same number of tool uses can be averaged, and the regression formula obtained by data mining based on machine learning can be applied.

[0064] The embodiment of the present invention also provides a method for intelligently monitoring tool wear and fracture of a CNC machine tool. The method may include the following steps: Monitoring processing load data: Monitoring the processing load data of CNC machine tools through sensor modules; Calculate reference data: Based on the monitored high-quality processing load data, the reference data is calculated through machine learning algorithms, which can include upper limit value, upper limit time, lower limit value and lower limit time.

[0065] Tool status determination: The real-time monitored machining load data is compared with the reference data. During the adaptive control machining process, if the machining load data continues to exceed or fall below the reference data, the tool wear or breakage is monitored.

[0066] refer to Figure 5 When the machining load exceeds the upper load limit (part A), the CNC machine tool can reduce the machining load by adaptively reducing the feed speed; when the machining load is lower than the lower load limit (part B), the CNC machine tool can increase the machining load by adaptively increasing the feed speed.

[0067] refer to Figure 6 , if the machining load temporarily exceeds the upper limit but does not exceed the upper limit time, the determination module can determine that there is no tool wear (Part C). That is, if the machining load exceeds the upper limit but does not continue to exceed the upper limit time, the determination module does not determine that there is tool wear. The monitoring module monitors tool wear when the machining load exceeds the upper limit, and determines tool wear when the upper limit time is exceeded.

[0068] refer to Figure 7 The monitoring module determines that the tool is broken when the machining load is lower than the lower limit and the duration exceeds the lower limit. If the machining load is temporarily lower than the lower limit, but the time of maintaining this state does not exceed the lower limit, the monitoring module does not determine that the tool is broken at this time.

[0069] In summary, in the adaptive feed speed control process of CNC machine tools, the feed speed changes according to the current machining load, and the corresponding machining length (or cutting position) also changes. Therefore, it is impossible to monitor tool wear and breakage by simply comparing the upper or lower limits of the time domain machining load data.

[0070] In order to overcome the above problems, an embodiment of the present invention first obtains the processing load data of the CNC machine tool, then calculates the reference data of the tool load based on machine learning, and monitors the tool breakage and tool wear by comparing the reference data with the current processing load data. At the same time, the reference data and the processing load data are converted from the time series domain to the processing length domain to improve the correspondence when comparing the processing load data.

[0071] Take the five-axis adaptive vertical machining center as an example and combine Figure 8 The embodiments of the present invention are described in detail.

[0072] Step 1: Monitor the processing load data. Install current sensor modules at key locations such as the spindle motor of the CNC machine tool to ensure that the processing load data can be accurately collected. The acquisition card module transmits the collected data to the data processing module in real time.

[0073] Step 2: Calculate reference data through machine learning. The reference data calculation unit in the training module receives the processing load data from the sensor module and selects high-quality processing load data. According to the tool life information, these data are processed and the reference data, including the upper limit value, upper limit time, lower limit value and lower limit time, are calculated through the machine learning algorithm and stored in the storage module.

[0074] During adaptive control, since the feed speed changes according to the current machining load, the corresponding machining length also changes, so the reference data cannot be directly compared with the machining load data detected during workpiece machining. Therefore, the reference data can be obtained based on the machining length calculation.

[0075] Feature extraction is performed on the processing load sample data, and a feature index set is constructed based on the extracted features; the feature index set is smoothed, and the performance of the features in the smoothed feature index set is evaluated using multiple evaluation indicators; based on the performance evaluation results, features related to the prediction results are selected to construct a sensitive feature set; model training is performed based on the sensitive feature set to obtain a tool monitoring model.

[0076] Step 3: Tool status determination. During the CNC machine tool adaptive control processing, the determination module receives the processing load data from the sensor module in real time and compares it with the reference data in the storage module. When the processing load data exceeds the upper limit and lasts for more than the upper limit, the determination module determines that the tool is worn and issues a corresponding alarm signal. When the processing load data is lower than the lower limit and lasts for more than the lower limit, the determination module determines that the tool is broken and issues a corresponding alarm signal. At the same time, the CNC machine tool can be controlled to stop processing to avoid further losses.

[0077] The embodiment of the present invention can monitor the wear and fracture of the tool in real time during the adaptive control processing of the CNC machine tool, thereby improving the processing quality and efficiency. The reference data is calculated by a machine learning algorithm, and the data is converted to the processing length domain for comparison, thereby improving the accuracy of monitoring. The modular design of the system makes it easy to implement and maintain, and has strong scalability. Based on this, the CNC machine tool tool wear and fracture intelligent monitoring system and method proposed in the embodiment of the present invention can monitor the wear and fracture of the tool in real time during the adaptive control of the feed speed of the CNC machine tool, thereby improving the processing quality, reducing the processing interruption and workpiece scrapping caused by tool problems, and is suitable for the processing of various materials and composite materials.

[0078] The tool monitoring device for a numerically controlled machine tool provided by the present invention is described below. The tool monitoring device for a numerically controlled machine tool described below and the tool monitoring method for a numerically controlled machine tool described above can be referred to each other.

[0079] refer to Fig. 9 The CNC machine tool tool monitoring device provided by the present invention includes an acquisition module 901, a prediction module 902 and a determination module 903.

[0080] An acquisition module 901 is used to acquire processing load data of a CNC machine tool; Prediction module 902, used for inputting the processing load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on the processing load sample data; The determination module 903 is used to determine the tool failure information of the numerically controlled machine tool based on the comparison result of the machining load data and the reference data.

[0081] The tool monitoring device for a CNC machine tool provided by an embodiment of the present invention obtains the processing load data of the CNC machine tool; inputs the processing load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on the processing load sample data; and based on the comparison result of the processing load data and the reference data, the tool failure information of the CNC machine tool is determined. The present invention determines the reference data of the tool load through a machine learning algorithm, and then determines whether the tool has failed based on the processing load data and the reference data, thereby realizing real-time monitoring of tool wear and breakage during the adaptive control of the feed speed of the CNC machine tool, improving the accuracy of tool monitoring and processing efficiency, and reducing processing interruptions and workpiece scrapping caused by tool problems.

[0082] In one embodiment, the reference data includes an upper limit value and an upper limit time of the tool load; the tool failure information includes tool wear; the determination module 903 is specifically used to: If the comparison result is that the machining load of the machining load data is greater than the upper limit value, and the duration for which the machining load is greater than the upper limit value is greater than the upper limit time, it is determined that the tool is worn.

[0083] In one embodiment, the reference data further includes a lower limit value and a lower limit time of the tool load; the tool failure information further includes tool breakage; the determination module 903 is specifically used to: If the comparison result is that the machining load of the machining load data is less than the lower limit value, and the duration for which the machining load is less than the lower limit value is greater than the lower limit time, it is determined that the tool is broken.

[0084] In one embodiment, the determination module 903 is further specifically configured to: The processing load data and the reference data are converted from the time series domain to the processing length domain; after the conversion to the processing length domain, the processing load data and the reference data are aligned; and the aligned processing load data and the reference data are compared to obtain the comparison result.

[0085] In one embodiment, the CNC machine tool tool monitoring device further includes a model training module for: The processing load sample data is subjected to feature extraction, and a feature index set is constructed based on the extracted features; the feature index set is smoothed, and the performance of the features in the smoothed feature index set is evaluated using multiple evaluation indicators; based on the performance evaluation results, features related to the prediction results are selected to construct a sensitive feature set; model training is performed based on the sensitive feature set to obtain the tool monitoring model.

[0086] In one embodiment, the CNC machine tool tool monitoring device further includes a feed speed adaptive module, which is used to: When the machining load of the CNC machine tool is greater than the upper limit of the tool load, the feed speed of the tool is reduced; when the machining load of the CNC machine tool is less than the lower limit of the tool load, the feed speed of the tool is increased.

[0087] Fig.10 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.10 As shown, the electronic device may include: a processor 1010, a communication interface 1020, a memory 1030 and a communication bus 1040, wherein the processor 1010, the communication interface 1020 and the memory 1030 communicate with each other through the communication bus 1040. The processor 1010 may call the logic instructions in the memory 1030 to execute the CNC machine tool tool monitoring method, which includes: obtaining the processing load data of the CNC machine tool; inputting the processing load data into the tool monitoring model to obtain the reference data of the tool load output by the tool monitoring model; the tool monitoring model is obtained by training based on the processing load sample data; and determining the tool failure information of the CNC machine tool based on the comparison result of the processing load data and the reference data.

[0088] In addition, the logic instructions in the above-mentioned memory 1030 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0089] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the CNC machine tool tool monitoring method provided by the above-mentioned methods, and the method includes: obtaining processing load data of the CNC machine tool; inputting the processing load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on processing load sample data; based on the comparison result of the processing load data and the reference data, determining the tool failure information of the CNC machine tool.

[0090] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it is implemented to execute the CNC machine tool tool monitoring method provided by the above-mentioned methods, the method comprising: obtaining processing load data of the CNC machine tool; inputting the processing load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on processing load sample data; and determining the tool failure information of the CNC machine tool based on the comparison result of the processing load data and the reference data.

[0091] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0092] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A CNC machine tool tool monitoring method, characterized in that: include: Obtain processing load data of CNC machine tools; Inputting the processing load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on the processing load sample data; Based on the comparison result between the machining load data and the reference data, tool failure information of the numerically controlled machine tool is determined.

2. The CNC machine tool tool monitoring method according to claim 1, characterized in that: The reference data includes an upper limit value and an upper limit time of the tool load; The tool failure information includes tool wear; the determining of the tool failure information of the CNC machine tool based on the comparison result of the machining load data and the reference data includes: If the comparison result is that the machining load of the machining load data is greater than the upper limit value, and the duration for which the machining load is greater than the upper limit value is greater than the upper limit time, it is determined that the tool is worn.

3. The CNC machine tool tool monitoring method according to claim 1, characterized in that: The reference data also includes a lower limit value and a lower limit time of the tool load; the tool failure information also includes tool breakage; the tool failure information of the numerical control machine tool is determined based on the comparison result of the processing load data and the reference data, including: If the comparison result is that the machining load of the machining load data is less than the lower limit value, and the duration for which the machining load is less than the lower limit value is greater than the lower limit time, it is determined that the tool is broken.

4. The CNC machine tool tool monitoring method according to claim 1, characterized in that: The comparison result between the processing load data and the reference data is determined based on the following method: Converting the processing load data and the reference data from the time series domain to the processing length domain; After conversion into the processing length domain, aligning the processing load data with the reference data; The aligned machining load data and the reference data are compared to obtain the comparison result.

5. The CNC machine tool tool monitoring method according to claim 1, characterized in that: The tool monitoring model is trained based on the following method: Extracting features from the processing load sample data, and constructing a feature index set based on the extracted features; Smoothing the feature index set, and using multiple evaluation indicators to evaluate the performance of features in the smoothed feature index set; Based on the performance evaluation results, features related to the prediction results are selected to construct a sensitive feature set; Model training is performed based on the sensitive feature set to obtain the tool monitoring model.

6. The CNC machine tool tool monitoring method according to claim 1, characterized in that: The method further comprises: When the machining load of the CNC machine tool is greater than the upper limit of the tool load, reducing the feed speed of the tool; When the machining load of the numerically controlled machine tool is less than a lower limit value of the tool load, the feed speed of the tool is increased.

7. A CNC machine tool tool monitoring device, characterized in that: include: An acquisition module, used for acquiring processing load data of a CNC machine tool; A prediction module, used for inputting the processing load data into a tool monitoring model to obtain reference data of the tool load output by the tool monitoring model; the tool monitoring model is trained based on the processing load sample data; A determination module is used to determine tool failure information of the numerically controlled machine tool based on a comparison result of the machining load data and the reference data.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the CNC machine tool tool monitoring method according to any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the CNC machine tool tool monitoring method according to any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the CNC machine tool tool monitoring method according to any one of claims 1 to 6 is implemented.

Images