Numerical control machining self-adaptive control system based on Internet of Things

Abstract

The invention relates to the technical field of numerical control machining monitoring, and particularly discloses a numerical control machining self-adaptive control system based on the internet of things, which comprises a data acquisition module, an abnormity identification module, a data processing module, a data processing module and a data processing module, and is characterized in that the data acquisition module acquires various parameters of numerical control machining equipment in real time through a high-precision sensor; the influence degree analysis module identifies whether abnormal fluctuation exists in the machining process or not, the influence degree analysis module evaluates the influence degree of the abnormal fluctuation in the machining process on the basis of changes of the cutting speed and the feeding amount and carries out training and prediction in combination with a machine learning model, and the self-adaptive control and feedback adjustment module judges a result on the basis of the abnormal fluctuation. According to the system and the method, the machining precision and the equipment stability can be effectively improved under real-time monitoring, the fault risk can be reduced, the service life of the equipment can be prolonged, the maintenance cost can be reduced, and intelligent optimization and self-adaptive control of the machining process can be realized.

Description

Technical Field

[0001] The present invention relates to the technical field of numerical control machining monitoring, and in particular to a numerical control machining adaptive control system based on the Internet of Things. Background Art

[0002] As the requirements for product precision and production efficiency in modern manufacturing continue to increase, CNC machining technology is increasingly used in the manufacturing process. CNC machining equipment can accurately control various parameters in the cutting process, such as cutting speed, feed rate, tool temperature and vibration frequency, so as to achieve efficient and high-precision machining. However, in complex machining processes, fluctuations in these parameters may lead to reduced machining quality, excessive wear of equipment, and even failure, seriously affecting production efficiency and machining quality. Therefore, real-time monitoring and intelligent adjustment of CNC machining equipment to ensure that various parameters in the machining process are within a reasonable range is the key to improving production efficiency, reducing equipment failure rate, and extending equipment life.

[0003] The prior art has the following steps:

[0004] The control of CNC machining equipment mainly relies on preset parameter values ​​and simple real-time monitoring methods, and lacks the ability to conduct in-depth analysis and adaptive adjustment of dynamic fluctuations during the machining process. CNC systems in existing technologies mostly rely on empirical parameters and cannot make precise adjustments in real time according to changes in key parameters during the machining process, resulting in the inability to intervene in time when abnormal fluctuations occur. This control method based on fixed rules is not only unable to cope with complex and dynamic machining environments, but may also cause excessive wear, equipment failures and quality problems during the machining process, further increasing the cost of repair and maintenance. Summary of the invention

[0005] The purpose of the present invention is to provide an adaptive control system for CNC machining based on the Internet of Things to solve the problems in the above background.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] The adaptive control system for CNC machining based on the Internet of Things includes:

[0008] A data acquisition module, wherein the data acquisition module is used to collect tool temperature, vibration frequency, cutting speed and feed rate data of CNC machining equipment in real time;

[0009] An abnormality identification module, which analyzes the tool temperature and vibration frequency of the CNC machining equipment and identifies whether there is abnormal fluctuation in the machining process according to the fluctuation degree of the tool temperature and the stability of the vibration frequency of the CNC machining equipment;

[0010] An influence degree analysis module, which calculates the influence degree of the changes in cutting speed and feed rate on abnormal fluctuations in the machining process by analyzing the cutting speed and feed rate of the CNC machining equipment;

[0011] An impact degree classification module, wherein the impact degree classification module classifies the impact degree into severe impact and slight impact according to the impact degree analysis result;

[0012] An adaptive control and feedback adjustment module dynamically adjusts the cutting speed and feed rate based on the serious impact judgment result.

[0013] As a further solution of the present invention: the identification of whether there is abnormal fluctuation during the processing specifically includes:

[0014] Analyze the tool temperature data of CNC machining equipment, calculate the tool temperature anomaly coefficient according to the fluctuation degree of tool temperature, and use it to evaluate the degree of tool temperature anomaly;

[0015] Analyze the vibration frequency data of CNC machining equipment, calculate the tool vibration frequency abnormality coefficient according to the real-time fluctuation amplitude of the vibration frequency, and use it to evaluate the degree of tool vibration abnormality;

[0016] The tool temperature abnormality coefficient and the tool vibration frequency abnormality coefficient are normalized and calculated to obtain a comprehensive abnormality coefficient, and it is determined whether the comprehensive abnormality coefficient is greater than or equal to a preset threshold. If so, there is an abnormal fluctuation, and if not, there is no abnormal fluctuation.

[0017] As a further solution of the present invention: the process of obtaining the tool temperature anomaly coefficient is:

[0018] Perform empirical mode decomposition on the collected tool temperature time series data and decompose it into several intrinsic mode functions and residual components;

[0019] Calculate the fluctuation energy of each intrinsic mode function; calculate the ratio of the fluctuation energy of all high-frequency intrinsic mode functions to the fluctuation energy of all intrinsic mode functions to obtain the abnormal fluctuation energy ratio;

[0020] The tool temperature anomaly coefficient is obtained by calculating the ratio of the integral of the absolute fluctuation amplitude of the residual component to the integral of the absolute fluctuation amplitude of the temperature data and summing them with the abnormal fluctuation energy ratio.

[0021] As a further solution of the present invention: the process of obtaining the abnormal coefficient of the tool vibration frequency is as follows:

[0022] Collect tool vibration frequency signal data of CNC machining equipment, perform short-time Fourier transform on the vibration signal data, use time window function to segment the signal, obtain the distribution of the signal in different time and frequency ranges, and obtain the instantaneous frequency characteristics;

[0023] By calculating the energy of the part of the time-frequency feature that exceeds the preset high-frequency threshold, the energy distribution characteristics of the vibration signal in the high-frequency range are identified, and the high-frequency energy value is calculated; by calculating the overall time-frequency feature of the vibration signal, the total energy value of the vibration signal is obtained;

[0024] The tool vibration frequency anomaly coefficient is calculated based on the ratio of the high-frequency energy value to the total energy value.

[0025] As a further solution of the present invention: the calculation of the influence of the changes in cutting speed and feed rate on the abnormal fluctuations in the machining process specifically includes:

[0026] The cutting speed and feed rate of CNC machining equipment are collected in real time. The cutting speed abnormality characteristic index is calculated according to the real-time fluctuation amplitude of the cutting speed of the CNC machining equipment. The feed rate abnormality characteristic index is calculated according to the change of the feed rate of the CNC machining equipment. A comprehensive feature vector is constructed by the cutting speed abnormality characteristic index, feed rate abnormality characteristic index and comprehensive abnormality coefficient as the input of the machine learning model. The influence coefficient is used as the output of the model. The model is trained. According to the output of the model, the degree of influence of abnormal fluctuations in the machining process is judged.

[0027] As a further solution of the present invention: the process of obtaining the abnormal characteristic index of cutting speed is:

[0028] The cutting speed data of CNC machining equipment is collected in real time; the cutting speed signal is decomposed into multiple modal components with different frequency characteristics by the variational modal decomposition method, the energy of each modal signal is calculated, and the energy of each mode is calculated by ratio calculation with the total energy of all modes to obtain the standardized energy coefficient of each mode. According to the standardized energy coefficient and center frequency of the mode, the cutting speed abnormal characteristic index is calculated. The calculation expression is: ;

[0029] In the formula, Indicates The normalized energy coefficient of each mode is Indicates The center frequency of the mode, is the maximum frequency in the signal, Indicates the abnormal characteristic index of cutting speed, Indicates the total number of modes.

[0030] As a further solution of the present invention: the process of obtaining the feed rate abnormal characteristic index is as follows:

[0031] Collect the feed data of CNC machining equipment in real time, and obtain the error value according to the square of the difference between the real-time feed data point and the ideal feed data point;

[0032] Construct a cumulative error matrix, in which the value of each element is equal to the error value of the current point plus the minimum cumulative error value of its three neighboring points;

[0033] Based on the dynamic programming method, we start from the end of the cumulative error matrix and trace back to the starting point to find the path with the minimum total error. The path length is recorded as ,The total error on the path is calculated by accumulating point by point, which is called the path distance;

[0034] The feed rate abnormal characteristic index is the ratio of path distance to path length.

[0035] As a further solution of the present invention: the construction process of the machine learning model is:

[0036] The cutting speed abnormal characteristic index, feed rate abnormal characteristic index and comprehensive abnormal coefficient are used to construct a comprehensive feature vector as the input of the machine learning model. The machine learning model is trained based on the historical abnormal characteristic index, feed rate abnormal characteristic index and comprehensive abnormal coefficient as training data, and the real-time comprehensive feature vector is input into the trained gradient boosting decision tree model. The model outputs the influence coefficient of abnormal fluctuation according to the input feature vector.

[0037] As a further solution of the present invention: the impact degree is divided into severe impact and slight impact according to the impact degree analysis result, specifically including:

[0038] Determine whether the impact coefficient of the model output is greater than or equal to the preset threshold. If so, it is recorded as a serious impact; if not, it is recorded as a slight impact.

[0039] As a further solution of the present invention: the dynamic adjustment of the cutting speed and feed rate specifically includes:

[0040] According to the results of the impact degree analysis, a dynamic intervention logic is constructed. When performing adjustments, the cutting speed or feed rate value is dynamically modified through a step-by-step iterative method to monitor the changing trend of abnormal features. If the fluctuation amplitude of the abnormal features decreases, the current parameter combination is recorded as the optimized solution. If the fluctuation amplitude does not decrease, the abnormal feature trajectory is re-analyzed, the association rules are updated and iterative adjustments are made until the abnormal fluctuation is stable.

[0041] Beneficial effects of the present invention:

[0042] (1) The present invention uses multi-dimensional data fusion technology to conduct in-depth analysis of dynamic fluctuations in the machining process by accurately and in real time monitoring key process parameters such as tool temperature, vibration frequency, cutting speed and feed rate of CNC machining equipment. When the system detects abnormal fluctuations, it evaluates the potential impact of each parameter fluctuation on the machining quality in real time through efficient abnormality identification and impact degree analysis modules. Based on this analysis, the system can intelligently adjust key machining parameters such as cutting speed and feed rate, automatically implement adaptive adjustment strategies, and optimize various key control variables in the machining process. This innovative control method can effectively prevent machining errors caused by excessive temperature, excessive vibration or unstable cutting speed and feed rate, thereby reducing workpiece deformation, improving surface quality and dimensional accuracy, and ensuring the stability and consistency of the machining process. In addition, through the prediction model based on machine learning and the real-time feedback mechanism, the system can continuously optimize the adjustment strategy to achieve more efficient and accurate CNC machining, thereby effectively improving production efficiency and equipment utilization, and reducing scrap rate and defective products.

[0043] (2) The present invention can accurately identify potential failure risks and issue early warnings for fluctuations during the processing process by real-time monitoring and in-depth analysis of multiple key parameters of CNC machining equipment. The system can timely capture small deviations in equipment operation and evaluate their potential threats to equipment performance by detecting abnormalities in data such as tool temperature, vibration frequency, cutting speed and feed rate. Based on the abnormality identification results, the system can automatically take dynamic adjustment measures, such as timely reducing the cutting speed or feed rate, thereby effectively reducing the impact of excessive wear or abnormal load on the equipment and preventing the equipment from entering a high load or fatigue state. This intelligent control mechanism not only reduces the loss of equipment under harsh working conditions and extends the service life of the equipment, but also effectively reduces the occurrence of sudden failures and reduces the need for frequent maintenance and maintenance costs. At the same time, the system's adaptive adjustment capability enables the equipment to always maintain the optimal operating state, improves processing efficiency and quality, further saves maintenance and operating costs for the enterprise, and improves the reliability and economic benefits of the overall production line. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The present invention will be further described below in conjunction with the accompanying drawings.

[0045] Figure 1 It is a flow chart of the adaptive control system for numerical control machining based on the Internet of Things of the present invention. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0047] See also Figure 1 As shown, the present invention is a CNC machining adaptive control system based on the Internet of Things, comprising:

[0048] A data acquisition module, wherein the data acquisition module is used to collect tool temperature, vibration frequency, cutting speed and feed rate data of CNC machining equipment in real time;

[0049] An abnormality identification module, which analyzes the tool temperature and vibration frequency of the CNC machining equipment and identifies whether there is abnormal fluctuation in the machining process according to the fluctuation degree of the tool temperature and the stability of the vibration frequency of the CNC machining equipment;

[0050] An influence degree analysis module, which calculates the influence degree of the changes in cutting speed and feed rate on abnormal fluctuations in the machining process by analyzing the cutting speed and feed rate of the CNC machining equipment;

[0051] An impact degree classification module, wherein the impact degree classification module classifies the impact degree into severe impact and slight impact according to the impact degree analysis result;

[0052] An adaptive control and feedback adjustment module dynamically adjusts the cutting speed and feed rate based on the serious impact judgment result.

[0053] In the data acquisition module, the tool temperature, vibration frequency, cutting speed and feed rate data of CNC machining equipment are collected in real time, including:

[0054] The tool temperature of CNC machining equipment is collected in real time through high-precision temperature sensors integrated on the tool. These sensors are usually deployed in the contact area between the tool and the workpiece to accurately monitor the temperature change of the tool; the vibration frequency is collected in real time through acceleration sensors installed on the spindle and tool of the machining equipment. The sensors are deployed on the support structure where the tool is fixed, and can accurately capture the fluctuation of the equipment vibration frequency; the cutting speed is measured through online measuring instruments, using laser speed sensors installed near the tool or workbench, and the linear speed of the tool is detected in real time through the reflection principle; the feed rate is collected through displacement sensors installed on the feed shaft. These sensors record the feed rate changes in real time by detecting the position changes of the feed shaft. All these data are connected to the local control platform of the CNC system through the data acquisition interface, and uniformly transmitted to the Internet of Things platform for remote monitoring and analysis.

[0055] In the abnormal identification module, by analyzing the tool temperature and vibration frequency of the CNC machining equipment, according to the fluctuation degree of the tool temperature and the stability of the vibration frequency of the CNC machining equipment, it is identified whether there is abnormal fluctuation in the machining process, including:

[0056] Analyze the tool temperature data of CNC machining equipment, calculate the tool temperature anomaly coefficient according to the fluctuation degree of tool temperature, and use it to evaluate the degree of tool temperature anomaly;

[0057] Analyze the vibration frequency data of CNC machining equipment, calculate the tool vibration frequency abnormality coefficient according to the real-time fluctuation amplitude of the vibration frequency, and use it to evaluate the degree of tool vibration abnormality;

[0058] The tool temperature abnormality coefficient and the tool vibration frequency abnormality coefficient are normalized and calculated to obtain a comprehensive abnormality coefficient, and it is determined whether the comprehensive abnormality coefficient is greater than or equal to a preset threshold value. If so, there is an abnormal fluctuation, and if not, there is no abnormal fluctuation;

[0059] The process of obtaining the tool temperature anomaly coefficient is as follows:

[0060] Performing empirical mode decomposition on the collected tool temperature time series data, decomposing it into several intrinsic mode functions and residual components, each intrinsic mode function represents the temperature fluctuation characteristics within a specific frequency range;

[0061] Calculate the wave energy of each intrinsic mode function, and the calculation expression is: ; In the formula, Indicates the collection time point, Indicates the starting point of the collection time point, represents the end point of the collection time point, represents the number of intrinsic mode functions, represents the intrinsic mode function, Indicates The fluctuation energy of the intrinsic mode functions;

[0062] Calculating the ratio of the fluctuation energy of all high-frequency intrinsic modal functions to the fluctuation energy of all intrinsic modal functions to obtain the abnormal fluctuation energy ratio; the high-frequency intrinsic modal function represents the first three intrinsic modal functions with the highest frequency in the decomposition;

[0063] The absolute fluctuation amplitude integral of the residual component is calculated by ratio with the absolute fluctuation amplitude integral of the temperature data, and then summed with the abnormal fluctuation energy ratio to obtain the tool temperature abnormality coefficient. The calculation expression is: ; In the formula, represents the abnormal fluctuation energy proportion, represents the residual component, represents the tool temperature time series data, Indicates the tool temperature anomaly coefficient.

[0064] The process of obtaining the abnormal coefficient of the tool vibration frequency is as follows:

[0065] Collect tool vibration frequency signal data of CNC machining equipment, perform short-time Fourier transform on the vibration signal data, and use the time window function to perform segmented analysis on the signal to obtain the distribution of the signal in different time and frequency ranges and obtain the instantaneous frequency characteristics;

[0066] By calculating the energy of the part of the time-frequency feature that exceeds the preset high-frequency threshold, the energy distribution characteristics of the vibration signal in the high-frequency range are identified, and the high-frequency energy value is calculated, and the high-frequency energy value reflects the significance of the abnormal characteristics of the vibration signal;

[0067] The calculation expression of the high frequency energy value is: ; In the formula, Represents the high-frequency energy value, represents the high frequency threshold, Indicates the maximum frequency of acquisition, Indicates the starting point of the collection time point, represents the end point of the collection time point, Indicated in Collection points The complex amplitude of the frequency component, represents the frequency component, Indicates the collection time point;

[0068] By calculating the overall time-frequency characteristics of the vibration signal, the total energy value of the vibration signal is obtained, where the total energy represents the energy accumulation of the vibration signal in all frequency ranges;

[0069] The calculation expression of the total energy value is: ; In the formula, Represents the total energy value, Indicates the minimum frequency of acquisition;

[0070] The tool vibration frequency anomaly coefficient is calculated based on the ratio of the high-frequency energy value to the total energy value.

[0071] It should be noted that the window function used in the short-time Fourier transform adopts a Gaussian window, the purpose of which is to smooth the signal segmentation data and improve the accuracy of time-frequency analysis.

[0072] The calculation expression of the comprehensive abnormal coefficient is:

[0073] ;

[0074] In the formula, represents the comprehensive anomaly coefficient, Indicates the tool temperature anomaly coefficient, Indicates the abnormal coefficient of tool vibration frequency, and is the preset scale factor, and and Both are greater than 0.

[0075] In the impact degree analysis module, the impact degree of the changes in cutting speed and feed rate on abnormal fluctuations in the machining process is calculated by analyzing the cutting speed and feed rate of the CNC machining equipment, including:

[0076] The cutting speed and feed rate of CNC machining equipment are collected in real time. The cutting speed abnormality characteristic index is calculated according to the real-time fluctuation amplitude of the cutting speed of the CNC machining equipment. The feed rate abnormality characteristic index is calculated according to the change of the feed rate of the CNC machining equipment. A comprehensive feature vector is constructed by the cutting speed abnormality characteristic index, feed rate abnormality characteristic index and comprehensive abnormality coefficient as the input of the machine learning model. The influence coefficient is used as the output of the model. The model is trained. According to the output of the model, the degree of influence of abnormal fluctuations in the machining process is judged.

[0077] The process of obtaining the abnormal cutting speed characteristic index is as follows:

[0078] Real-time collection of cutting speed data of CNC machining equipment. Cutting speed data indicates the speed change of the equipment tool during the cutting process and is a time series signal that changes with time.

[0079] The cutting speed signal is decomposed into multiple modal components with different frequency characteristics by using the variational modal decomposition method. The target solution of the variational modal decomposition includes: decomposing the cutting speed signal into multiple intrinsic mode functions, each mode corresponds to a frequency component; the frequency component of each mode is represented by its central frequency and amplitude, and there is no frequency overlap between the modes, satisfying the frequency independence; minimizing the signal decomposition error by optimizing the objective function, and extracting different frequency components;

[0080] Calculate the energy of each modal signal, the calculation expression is: , where Indicates the collection time point, represents the number of modal signals, represents the total number of acquisition time points, Indicates A modal signal, Indicates The energy of each mode;

[0081] The energy of each mode is calculated by ratio with the total energy of all modes to obtain the standardized energy coefficient of each mode. According to the standardized energy coefficient and center frequency of the mode, the abnormal characteristic index of cutting speed is calculated. The calculation expression is: ;

[0082] In the formula, Indicates The normalized energy coefficient of each mode is Indicates The center frequency of the mode, is the maximum frequency in the signal, Indicates the abnormal characteristic index of cutting speed, Indicates the total number of modes.

[0083] It should be noted that the cutting speed anomaly detection method based on variational mode decomposition can effectively detect and quantify abnormal fluctuations in the cutting process, providing strong support for the optimization and fault prediction of CNC machining processes.

[0084] The process of obtaining the feed rate abnormality characteristic index is as follows:

[0085] Collect feed data of the CNC machining equipment in real time, use it as a real-time feed data sequence, and calculate the error value between the real-time feed data point and the reference data point point by point, wherein the error value is the square of the difference between the real-time feed data point and the ideal feed data point;

[0086] In the cumulative error matrix, the value of each element is equal to the error value of the current point plus the minimum cumulative error value of its three neighboring points. The cumulative error value of the current matrix position is determined by the sum of the current error value and the cumulative error of the previous time point;

[0087] Based on the dynamic programming method, we start from the end of the cumulative error matrix and trace back to the starting point to find the path with the minimum total error. The path length is recorded as ,The total error on the path is calculated by accumulating point by point, which is called the path distance;

[0088] The feed rate abnormal characteristic index is the ratio of path distance to path length.

[0089] The construction process of the machine learning model is as follows:

[0090] The cutting speed abnormality feature index, the feed rate abnormality feature index and the comprehensive abnormality coefficient are used to construct a comprehensive feature vector as an input of a machine learning model; the machine learning model is a gradient boosting decision tree;

[0091] Based on historical processing data and its corresponding processing results, a training data set containing input comprehensive feature vectors and output abnormal fluctuation influence coefficients is constructed. The feature vector data comes from the parameter calculation results of the actual processing process, and the influence coefficient is obtained based on the evaluation of processing quality, marking the specific impact of abnormal fluctuations in processing on the results.

[0092] Based on the historical abnormal characteristic index, feed rate abnormal characteristic index and comprehensive abnormal coefficient as training data, a training data set containing input comprehensive characteristic vector and output abnormal fluctuation influence coefficient is constructed. The characteristic vector data comes from the parameter calculation results of the actual machining process, and the influence coefficient is obtained based on the evaluation of machining quality, marking the specific influence of abnormal fluctuations in machining on the results.

[0093] The comprehensive feature vector constructed after processing the real-time collected data is input into the trained gradient boosting decision tree model, and the model outputs the abnormal fluctuation impact coefficient according to the input feature vector.

[0094] In the impact degree classification module, the impact degree is divided into severe impact and slight impact according to the impact degree analysis results, including:

[0095] Determine whether the impact coefficient of the model output is greater than or equal to the preset threshold. If so, it is recorded as a serious impact; if not, it is recorded as a slight impact.

[0096] In the adaptive control and feedback adjustment module, the cutting speed and feed rate are dynamically adjusted based on the serious impact judgment results, including:

[0097] Based on the cutting speed abnormal characteristic index, feed rate abnormal characteristic index, tool temperature abnormal coefficient, vibration frequency abnormal coefficient and comprehensive abnormal coefficient, a multi-dimensional abnormal characteristic mapping is constructed to express the real-time change trend of each abnormal characteristic as a characteristic trajectory curve, and extract the significant inflection point, slope change and peak distribution of each trajectory to intuitively reflect the influence of abnormal characteristics on the machining process. In the process of characteristic analysis, by analyzing the time series fluctuation amplitude and change rate of the abnormal coefficient, it is judged whether the abnormal impact presents an increasing trend or periodic fluctuation.

[0098] Based on multi-dimensional feature analysis, dynamic association rule mining technology is used to explore the potential relationship between changes in cutting speed and feed rate and abnormal coefficients. By mining the co-occurrence pattern of abnormal features in the processing process, the influence of cutting speed and feed rate on different abnormal features is identified. For example, it is found that a sudden change in cutting speed may cause abnormal vibration or an unstable feed rate may cause abnormal temperature. This process generates a set of association rules that describe the causal relationship between processing parameters and abnormal features.

[0099] Based on the mined association rule set and the actual status of the processing process, a dynamic intervention logic rule set is constructed, including:

[0100] If the vibration frequency increases significantly and is accompanied by an increase in tool temperature, gradually reduce the cutting speed and adjust the feed rate slightly after monitoring the temperature trend to stabilize;

[0101] If the comprehensive abnormal coefficient rises rapidly and the temperature abnormal coefficient remains stable, the feed rate is adjusted first to keep the cutting speed unchanged. The dynamic intervention logic realizes the conditional decision of parameter adjustment in the form of rule triggering, ensuring that the adjustment is targeted and accurate.

[0102] The cutting speed and feed rate adjusted by the dynamic intervention logic are executed, and the changes of abnormal features are monitored in real time. If the fluctuation amplitude of the abnormal features is monitored to continue to decrease, the current adjustment parameter combination is recorded as a feasible solution for the optimization strategy; if the fluctuation amplitude does not decrease significantly or even increases, the multi-dimensional abnormal feature trajectory is re-analyzed, new association rules are mined, and the intervention logic is updated until the abnormal features tend to stabilize.

[0103] Working principle of the invention: The invention provides an adaptive control system for CNC machining based on the Internet of Things, which aims to realize abnormal fluctuation identification and adaptive adjustment of the machining process by real-time monitoring and analysis of key parameters of CNC machining equipment (such as tool temperature, vibration frequency, cutting speed and feed rate). The key modules of the system include data acquisition module, abnormal identification module, impact degree analysis module, impact degree division module and adaptive control and feedback adjustment module. The data acquisition module collects tool temperature, vibration frequency, cutting speed and feed rate data in real time by deploying high-precision sensors. The abnormal identification module analyzes the fluctuation of tool temperature and vibration frequency, and determines whether there is abnormal fluctuation in machining by calculating the abnormal coefficient. The impact degree analysis module further calculates the influence of abnormal fluctuation on the machining process through the changes in cutting speed and feed rate, and uses a machine learning model (such as a gradient boosting decision tree) to construct the abnormal characteristics of cutting speed and feed rate and the comprehensive abnormal coefficient into a comprehensive feature vector, and trains and predicts to determine the influence of abnormal fluctuation in machining. The impact degree division module divides the impact degree into severe impact and slight impact based on the model output. Finally, the adaptive control and feedback adjustment module adjusts the cutting speed and feed rate according to the impact results, and uses the dynamic intervention logic rule set to ensure that the various parameters in the processing process are adjusted accurately and timely. Through dynamic analysis and association rule mining, the system can achieve real-time response to abnormal fluctuations, adjust processing parameters, optimize processing quality, and thus improve the stability and processing accuracy of the equipment.

[0104] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. The adaptive control system for CNC machining based on the Internet of Things is characterized by: include: A data acquisition module, wherein the data acquisition module is used to collect tool temperature, vibration frequency, cutting speed and feed rate data of CNC machining equipment in real time; An abnormality identification module, which analyzes the tool temperature and vibration frequency of the CNC machining equipment and identifies whether there is abnormal fluctuation in the machining process according to the fluctuation degree of the tool temperature and the stability of the vibration frequency of the CNC machining equipment; An influence degree analysis module, which calculates the influence degree of the changes in cutting speed and feed rate on abnormal fluctuations in the machining process by analyzing the cutting speed and feed rate of the CNC machining equipment; An impact degree classification module, wherein the impact degree classification module classifies the impact degree into severe impact and slight impact according to the impact degree analysis result; An adaptive control and feedback adjustment module dynamically adjusts the cutting speed and feed rate based on the serious impact judgment result.

2. The adaptive control system for CNC machining based on the Internet of Things according to claim 1 is characterized in that: The identification of whether there is abnormal fluctuation during the processing specifically includes: Analyze the tool temperature data of CNC machining equipment, calculate the tool temperature anomaly coefficient according to the fluctuation degree of tool temperature, and use it to evaluate the degree of tool temperature anomaly; Analyze the vibration frequency data of CNC machining equipment, calculate the tool vibration frequency abnormality coefficient according to the real-time fluctuation amplitude of the vibration frequency, and use it to evaluate the degree of tool vibration abnormality; The tool temperature abnormality coefficient and the tool vibration frequency abnormality coefficient are normalized and calculated to obtain a comprehensive abnormality coefficient, and it is determined whether the comprehensive abnormality coefficient is greater than or equal to a preset threshold. If so, there is an abnormal fluctuation, and if not, there is no abnormal fluctuation.

3. The adaptive control system for numerical control machining based on the Internet of Things according to claim 2 is characterized in that: The process of obtaining the tool temperature anomaly coefficient is as follows: Perform empirical mode decomposition on the collected tool temperature time series data and decompose it into several intrinsic mode functions and residual components; Calculate the fluctuation energy of each intrinsic mode function; calculate the ratio of the fluctuation energy of all high-frequency intrinsic mode functions to the fluctuation energy of all intrinsic mode functions to obtain the abnormal fluctuation energy ratio; The tool temperature anomaly coefficient is obtained by calculating the ratio of the integral of the absolute fluctuation amplitude of the residual component to the integral of the absolute fluctuation amplitude of the temperature data and summing them with the abnormal fluctuation energy ratio.

4. The adaptive control system for numerical control machining based on the Internet of Things according to claim 2 is characterized in that: The process of obtaining the abnormal coefficient of the tool vibration frequency is as follows: Collect tool vibration frequency signal data of CNC machining equipment, perform short-time Fourier transform on the vibration signal data, use time window function to segment the signal, obtain the distribution of the signal in different time and frequency ranges, and obtain the instantaneous frequency characteristics; By calculating the energy of the part of the time-frequency feature that exceeds the preset high-frequency threshold, the energy distribution characteristics of the vibration signal in the high-frequency range are identified, and the high-frequency energy value is calculated; by calculating the overall time-frequency feature of the vibration signal, the total energy value of the vibration signal is obtained; The tool vibration frequency anomaly coefficient is calculated based on the ratio of the high-frequency energy value to the total energy value.

5. The adaptive control system for numerical control machining based on the Internet of Things according to claim 1 is characterized in that: The influence degree of the change of the calculated cutting speed and feed rate on the abnormal fluctuation in the machining process specifically includes: The cutting speed and feed rate of CNC machining equipment are collected in real time. The cutting speed abnormality characteristic index is calculated according to the real-time fluctuation amplitude of the cutting speed of the CNC machining equipment. The feed rate abnormality characteristic index is calculated according to the change of the feed rate of the CNC machining equipment. A comprehensive feature vector is constructed by the cutting speed abnormality characteristic index, feed rate abnormality characteristic index and comprehensive abnormality coefficient as the input of the machine learning model. The influence coefficient is used as the output of the model. The model is trained. According to the output of the model, the degree of influence of abnormal fluctuations in the machining process is judged.

6. The adaptive control system for numerical control machining based on the Internet of Things according to claim 5 is characterized in that: The process of obtaining the abnormal cutting speed characteristic index is as follows: The cutting speed data of CNC machining equipment is collected in real time; the cutting speed signal is decomposed into multiple modal components with different frequency characteristics by the variational modal decomposition method, the energy of each modal signal is calculated, and the energy of each mode is calculated by ratio calculation with the total energy of all modes to obtain the standardized energy coefficient of each mode. According to the standardized energy coefficient and center frequency of the mode, the cutting speed abnormal characteristic index is calculated. The calculation expression is: ; In the formula, Indicates The normalized energy coefficient of each mode is Indicates The center frequency of the mode, is the maximum frequency in the signal, Indicates the abnormal characteristic index of cutting speed, represents the total number of modes, Indicates the number of modal signals.

7. The adaptive control system for numerical control machining based on the Internet of Things according to claim 5 is characterized in that: The process of obtaining the feed rate abnormality characteristic index is as follows: Collect the feed data of CNC machining equipment in real time, and obtain the error value according to the square of the difference between the real-time feed data point and the ideal feed data point; Construct a cumulative error matrix, in which the value of each element is equal to the error value of the current point plus the minimum cumulative error value of its three neighboring points; Based on the dynamic programming method, we start from the end of the cumulative error matrix and trace back to the starting point to find the path with the minimum total error. The path length is recorded as ,The total error on the path is calculated by accumulating point by point, which is called the path distance; The feed rate abnormal characteristic index is the ratio of path distance to path length.

8. The adaptive control system for numerical control machining based on the Internet of Things according to claim 5 is characterized in that: The construction process of the machine learning model is as follows: The cutting speed abnormal characteristic index, feed rate abnormal characteristic index and comprehensive abnormal coefficient are used to construct a comprehensive feature vector as the input of the machine learning model. The machine learning model is trained based on the historical abnormal characteristic index, feed rate abnormal characteristic index and comprehensive abnormal coefficient as training data, and the real-time comprehensive feature vector is input into the trained gradient boosting decision tree model. The model outputs the influence coefficient of abnormal fluctuation according to the input feature vector.

9. The adaptive control system for numerical control machining based on the Internet of Things according to claim 1 is characterized in that: According to the impact analysis results, the impact levels are divided into severe impact and slight impact, including: Determine whether the impact coefficient of the model output is greater than or equal to the preset threshold. If so, it is recorded as a serious impact; if not, it is recorded as a slight impact.

10. The adaptive control system for numerical control machining based on the Internet of Things according to claim 1, characterized in that: The dynamic adjustment of cutting speed and feed rate specifically includes: According to the results of the impact degree analysis, a dynamic intervention logic is constructed. When performing adjustments, the cutting speed or feed rate value is dynamically modified through a step-by-step iterative method to monitor the changing trend of abnormal features. If the fluctuation amplitude of the abnormal features decreases, the current parameter combination is recorded as the optimized solution. If the fluctuation amplitude does not decrease, the abnormal feature trajectory is re-analyzed, the association rules are updated and iterative adjustments are made until the abnormal fluctuation is stable.

Images