Numerical control grinding machine product quality real-time monitoring and optimizing method fused with Internet of Things

Abstract

The invention provides a numerical control grinding machine product quality real-time monitoring and optimizing method fused with the Internet of Things, and relates to the technical field of numerical control grinding machine control. Processing parameter retrieval is carried out; performing frequent item mining on the historical processing data; performing fluctuation centralized identification on the historical vibration parameter set according to the vibration parameter tolerance interval; performing deviation analysis on the historical process quality parameters by using the process quality parameter tolerance interval; performing process monitoring step length identification based on the vibration parameter fluctuation coefficient and the process quality deviation coefficient; and the monitoring branch is used for extracting and analyzing real-time monitoring data and optimally controlling the numerical control grinding machine. By means of the method and device, the technical problem that in the prior art, due to the fact that the adaptability of monitoring step length adjustment is insufficient, key parameters are difficult to adjust in time, and the product quality is affected can be solved, through deep analysis of historical data, refined management and optimization of the machining process of the numerical control grinding machine are achieved, and the product quality is improved.

Description

Technical Field

[0001] The present application relates to the field of CNC grinder control technology, and in particular to a method for real-time monitoring and optimization of CNC grinder product quality integrating the Internet of Things. Background Art

[0002] As the core equipment of modern precision manufacturing industry, CNC grinding machine is widely used in the processing of high-precision parts such as automotive parts and aerospace components due to its high degree of automation and flexibility. In the precision manufacturing industry, product quality is the core of enterprise competitiveness. Any slight error may lead to performance deviation or even scrap of the finished product. Therefore, it is crucial to establish an effective real-time monitoring mechanism to ensure the product quality during the CNC grinding process. However, the response speed, data analysis capability and adaptability of the existing monitoring system of CNC grinding machine may not be sufficient to quickly adjust the monitoring strategy to adapt to the rapid changes in parameters. Due to the inadaptability of the monitoring step, the changes in key parameters may not be captured in time, resulting in the inability to take immediate measures to adjust the parameters when they exceed the allowable range. The key parameters cannot be adjusted in time, affecting the quality of the product and increasing the defective rate. Secondly, the inadaptability of the monitoring step limits the ability of process optimization. In the process link that requires fine adjustment, if the monitoring step is too large, it may not be able to identify small parameter changes, thus affecting the optimization of the process and the improvement of product quality.

[0003] In summary, the prior art has a technical problem that the monitoring step adjustment is not sufficiently adaptable to specific process requirements, and changes in key parameters cannot be captured in a timely manner, resulting in difficulty in adjusting parameters in a timely manner, thereby affecting product quality. Summary of the invention

[0004] The purpose of this application is to provide a real-time monitoring and optimization method for CNC grinding machine product quality integrating the Internet of Things, so as to solve the technical problems in the prior art that the monitoring step adjustment is not adaptable enough to specific process requirements, the changes of key parameters cannot be captured in time, resulting in difficulty in adjusting parameters in time, thus affecting product quality.

[0005] In view of the above problems, the present application provides a real-time monitoring and optimization method for the product quality of CNC grinders integrated with the Internet of Things, wherein the real-time monitoring and optimization method for the product quality of CNC grinders integrated with the Internet of Things includes: obtaining a basic data set of a target grinder product, wherein the basic data set includes a product quality parameter set and product specification and model information; performing processing parameter retrieval of the CNC grinder based on the product specification and model information and the product quality parameter set, and determining K processing procedures, K process quality parameter tolerance intervals, and K processing procedure vibration parameter tolerance intervals; using the product specification and model information and the K processing procedures as indexes, performing frequent item mining on the historical processing data of the CNC grinder, and obtaining K historical processing procedure vibration parameter sets and K historical processing procedure quality parameter sets; and performing frequent item mining on the K historical processing data according to the K processing procedure vibration parameter tolerance intervals. The process vibration parameter set is subjected to concentrated identification of fluctuations to determine the vibration parameter fluctuation coefficients of K processing processes; the K historical process quality parameters are subjected to deviation analysis using the K process quality parameter tolerance intervals to determine the K process quality deviation coefficients; the monitoring step lengths of the K processing processes are identified based on the K processing process vibration parameter fluctuation coefficients and the K process quality deviation coefficients, and K process monitoring branches are constructed according to the identification results; the Internet of Things real-time monitoring data of the K processing processes are extracted and analyzed using the K process monitoring branches to determine K quality monitoring results, wherein the K quality monitoring results include K real-time processing process vibration abnormality monitoring results and K real-time process quality abnormality monitoring results; the CNC grinding machine is optimized and controlled based on the K real-time processing process vibration abnormality monitoring and the K real-time process quality abnormality monitoring results.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: By acquiring a basic data set of a target grinding machine product, wherein the basic data set includes a product quality parameter set and product specification and model information; performing a CNC grinding machine processing parameter search based on the product specification and model information and the product quality parameter set, and determining K processing procedures, K process quality parameter tolerance intervals, and K processing procedure vibration parameter tolerance intervals; using the product specification and model information and the K processing procedures as indexes, performing frequent item mining on the historical processing data of the CNC grinding machine, and obtaining K historical processing procedure vibration parameter sets and K historical processing procedure quality parameter sets; performing fluctuation concentration identification on the K historical processing procedure vibration parameter sets according to the K processing procedure vibration parameter tolerance intervals, and determining the vibration parameter fluctuation coefficients of the K processing procedures ; Use the K process quality parameter tolerance intervals to perform deviation analysis on the K historical process quality parameters to determine K process quality deviation coefficients; identify the monitoring step lengths of the K processing processes based on the K processing process vibration parameter fluctuation coefficients and the K process quality deviation coefficients, and construct K process monitoring branches according to the identification results; use the K process monitoring branches to extract and analyze the IoT real-time monitoring data of the K processing processes to determine K quality monitoring results, wherein the K quality monitoring results include K real-time processing process vibration abnormality monitoring results and K real-time process quality abnormality monitoring results; optimize the control of the CNC grinding machine based on the K real-time processing process vibration abnormality monitoring and K real-time process quality abnormality monitoring results. In other words, by conducting in-depth analysis of historical data, identifying the vibration parameter fluctuation coefficient and quality deviation coefficient, and dynamically determining the monitoring step, the accuracy and flexibility of monitoring can be improved; by building a process monitoring branch, extracting and analyzing real-time monitoring data, and determining the abnormal vibration monitoring results of the processing process and the abnormal quality monitoring results of the process, the CNC grinder can be optimized and controlled, realizing the refined management and optimization of the CNC grinder processing process and improving product quality.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented according to the contents of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0009] Figure 1 A flow chart of a method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things for this application; Figure 2 This is a flow chart of constructing a process monitoring branch in the real-time monitoring and optimization method of CNC grinder product quality integrating the Internet of Things in this application. DETAILED DESCRIPTION

[0010] This application provides a real-time monitoring and optimization method for CNC grinder product quality that integrates the Internet of Things, solving the technical problem in the prior art that the monitoring step adjustment is not adaptable enough to specific process requirements, and the changes in key parameters cannot be captured in time, resulting in difficulty in adjusting parameters in time, thus affecting product quality. Through in-depth analysis of historical data, identification of vibration parameter fluctuation coefficients and quality deviation coefficients, dynamic determination of monitoring steps, and improved monitoring accuracy and flexibility; constructing process monitoring branches, extracting and analyzing real-time monitoring data, and determining abnormal vibration monitoring results and process quality monitoring results of the processing process, the CNC grinder is optimized and controlled, achieving refined management and optimization of the CNC grinder processing process, and improving product quality.

[0011] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0012] For examples, please see the attached Figure 1 The present application provides a method for real-time monitoring and optimization of CNC grinding machine product quality by integrating the Internet of Things, wherein the method for real-time monitoring and optimization of CNC grinding machine product quality by integrating the Internet of Things specifically comprises the following steps: Step 1: Obtain a basic data set of a target grinding machine product, wherein the basic data set includes a product quality parameter set and product specification model information.

[0013] Specifically, the basic data of the target grinder products are collected, including product quality parameters and product specification model information. The basic data set contains all the necessary information of the target grinder products, and the product quality parameter set contains parameters directly related to the quality of the target grinder products, such as processing accuracy, surface roughness, grinding efficiency, grinding force, thermal deformation, etc. The product specification model information refers to the specific model and specifications of the grinder processing products, including the model, size, technical specifications, configuration options, etc. of the grinder. By obtaining basic data, it is helpful to identify fluctuations and potential problems in product quality, thereby providing a basis for real-time monitoring and optimization.

[0014] Step 2: Based on the product specification model information and the product quality parameter set, the processing parameters of the CNC grinder are retrieved to determine K processing steps, K process quality parameter tolerance intervals and K processing step vibration parameter tolerance intervals.

[0015] Specifically, multiple processing records in the historical processing memory of the CNC grinder are obtained, the pairwise similarities of these processing records are calculated, and the number of similarities greater than or equal to the similarity threshold is calculated. If the number is too large, greater than the preset number, it means that the types of products processed by the grinder are too few, and additional historical processing records with rich samples are needed; conversely, if the number is less than the preset number, it means that the existing historical processing memory is already diverse and rich enough, and the number of combinations is insufficient to indicate that the types of products processed by the grinder are too few, and there is no need to obtain cloud-based processing record collection instructions. According to product specifications and processing requirements, the key processing procedures and appropriate processing parameters that need to be monitored are retrieved from the finalized historical processing memory, including but not limited to grinding speed, grinding depth, feed speed, grinding force, etc.

[0016] K processing procedures refer to multiple specific processing steps determined according to product specifications and quality requirements. Each process has its own specific processing methods and parameters. The process quality parameter tolerance range refers to the allowable product quality parameter variation range set for each processing procedure to ensure quality stability and consistency, and define the acceptable range of product quality parameters. Exceeding this range is considered as quality abnormality. For example, for the rough grinding process, the dimensional tolerance tolerance range may be set to ±0.05 mm. The process vibration parameter tolerance range corresponds to each process and is the allowable variation range of vibration parameters (such as amplitude and frequency) set for each process to control processing stability. Vibration parameters are important indicators for measuring the operating status of equipment, and their abnormalities may lead to product quality problems. For example, for the fine grinding process, the vibration level tolerance range may be set to no more than 0.2 mm. By setting the quality parameter tolerance range, processing defects and machine tool wear caused by improper parameters can be reduced, product quality can be controlled, and the production of defective products can be reduced.

[0017] Step 3: Using the product specification model information and the K processing procedures as indexes, perform frequent item mining on the historical processing data of the CNC grinder to obtain K historical processing procedure vibration parameter sets and K historical process quality parameter sets.

[0018] Specifically, product specification model information and K processing procedures are used as indexes to filter and classify historical processing data based on this information. The historical data is classified according to products and procedures, and the vibration parameters and quality parameters that frequently appear in the K processing procedures are identified. Frequent item mining is a data mining technology used to find frequently occurring patterns, associations or sequences from a large amount of data. In CNC grinding machines, it means finding the patterns of vibration parameters and quality parameters that frequently appear in historical processing data. Historical processing data refers to the data recorded in the past CNC grinding process, including vibration parameters, quality parameters, etc. The K historical processing process vibration parameter sets and the K historical process quality parameter sets represent the statistical results of vibration and quality parameters in the past processing process, respectively. By performing frequent item mining on historical processing data, the laws and trends in the processing process are discovered, providing a basis for optimizing the current processing parameters, which helps to improve processing efficiency, reduce processing defects, and extend the service life of machine tools.

[0019] Step 4: performing concentrated fluctuation identification on the K historical processing process vibration parameter sets according to the K processing process vibration parameter tolerance intervals, and determining the vibration parameter fluctuation coefficients of the K processing processes.

[0020] Specifically, the fluctuation of the vibration parameter set of K historical processing procedures is identified according to the tolerance intervals of the vibration parameters of K processing procedures. The fluctuation of the vibration parameter set of K historical processing procedures is identified, the fluctuation of the vibration parameter is identified, and the distance between the vibration parameter and the upper and lower limits of the tolerance interval at each time point is calculated. If the vibration parameter value fluctuates within the tolerance interval, it is considered that the processing process is stable and the fluctuation value is zero; if it exceeds this range, it is considered that there is fluctuation. For each processing procedure, the degree to which its vibration parameter exceeds the tolerance interval, that is, the fluctuation value, is calculated. For the vibration parameter set of K historical processing procedures, a set of vibration parameter fluctuation values ​​of K historical processing procedures is obtained. The fluctuation value is calculated centrally, including mean calculation and variance calculation, to obtain K fluctuation means, and K centralized identification step lengths are determined. The K fluctuation means are used as the starting point, and the vibration parameter fluctuation value set of K historical processing procedures is iterated according to the K centralized identification step lengths to find the fluctuation value. The iteration is stopped when the preset number of iterations is reached or / and the iteration stop requirement is met, and the vibration parameter fluctuation values ​​of the K iterative historical processing procedures obtained in the last iteration are used as the vibration parameter fluctuation values ​​of the K target historical processing procedures. Calculate the sum of the vibration parameter fluctuation values ​​of the K target historical processing procedures, that is, add up the vibration parameter fluctuation values ​​of the K target historical processing procedures. Divide the vibration parameter fluctuation values ​​of the K target historical processing procedures by the sum, respectively, and obtain the ratio of each fluctuation value in the sum, and use the ratio as the vibration parameter fluctuation coefficient of the K processing procedures. Through the concentrated identification of fluctuations and the determination of the fluctuation coefficient, the fluctuation of the vibration parameters of each processing procedure is quantified, and its influence on product quality is evaluated, which helps to accurately identify and optimize the key factors affecting product quality, thereby improving the stability of the processing process and the consistency of product quality.

[0021] Step 5: Use the K process quality parameter tolerance intervals to perform deviation analysis on the K historical process quality parameters to determine the K process quality deviation coefficients.

[0022] Specifically, for each process, the degree of deviation between the actual quality parameter and the center value (or reference value) of the tolerance interval of the corresponding process quality parameter is calculated. Deviation analysis refers to analyzing the degree of deviation of each parameter value relative to its tolerance interval in the historical process quality parameters. By calculating the distance between the parameter value and the upper and lower limits of the tolerance interval, the fluctuations and potential problems of product quality can be identified. By performing deviation analysis on K processes respectively, K process quality deviation coefficients are obtained. The process quality deviation coefficient is a quantitative indicator used to indicate the degree of deviation of the actual quality parameter relative to the tolerance interval. The calculation of the deviation coefficient usually involves the standard deviation, coefficient of variation or other statistics. By identifying the deviation coefficient, the processing process can be controlled more accurately and the parameters can be adjusted in time to reduce the deviation.

[0023] Step six: Identify the monitoring steps of the K processing steps based on the vibration parameter fluctuation coefficients of the K processing steps and the quality deviation coefficients of the K processing steps, and construct K process monitoring branches according to the identification results.

[0024] Specifically, a monitoring step identifier is constructed, and the vibration parameter fluctuation coefficients of K processing procedures and the quality deviation coefficients of K procedures are taken as inputs. For each processing procedure, the model will output two monitoring step values, namely the first monitoring step and the second monitoring step, where the first monitoring step is smaller than the second monitoring step. For each procedure, a first monitoring channel is set according to the identified first monitoring step, and the first monitoring channel will be responsible for monitoring the processing process more frequently to capture short-term changes. For each procedure, a second monitoring channel is set according to the identified second monitoring step, and the second monitoring channel will be responsible for monitoring the processing process more sparsely to capture long-term trends. The first monitoring channel and the second monitoring channel of each procedure are integrated into the corresponding monitoring branch to ensure that the monitoring branch can process data from two different monitoring channels at the same time. The output of the monitoring channel is connected to the data processing module of the monitoring branch, and a monitoring result analysis fully connected layer is set at the end of the monitoring branch to comprehensively analyze the data from the two monitoring channels. Finally, the monitoring branch corresponding to each procedure, namely K process monitoring branches, is obtained. Each process monitoring branch is an independent monitoring channel responsible for monitoring a specific processing procedure, including a corresponding monitoring channel and a monitoring result analysis fully connected layer. By identifying the monitoring step length based on the vibration parameter fluctuation coefficient and the quality deviation coefficient and constructing the process monitoring branch, the independent monitoring and analysis of the vibration parameter changes of each process can be realized, which helps to improve the accuracy and efficiency of monitoring, thereby more effectively monitoring and optimizing the processing process, improving product quality and production efficiency. Since each process monitoring branch is constructed based on the vibration parameter fluctuation coefficient and process quality deviation coefficient of the process, it can more accurately capture problems in the process, help improve the monitoring efficiency and quality of the entire processing process, and promptly discover and solve problems, thereby improving product quality and production efficiency.

[0025] Step seven: Use the K process monitoring branches to extract and analyze the IoT real-time monitoring data of the K processing processes to determine K quality monitoring results, wherein the K quality monitoring results include K real-time processing process vibration abnormality monitoring results and K real-time process quality abnormality monitoring results.

[0026] Specifically, real-time monitoring data, including vibration data, temperature, humidity, processing speed, etc., are collected from IoT devices (such as sensors, actuators, CNC systems, etc.). These data can reflect the real-time status of the processing process, including the vibration of the processing process and the quality of the process. Using K process monitoring branches, extract data related to the process from the real-time monitoring data of the IoT, and process the data through data filtering, data format conversion and other operations. According to the extracted data, according to the monitoring logic and algorithm in the process monitoring branch, analyze the real-time fluctuation of the vibration parameters of the processing process, determine whether there is vibration abnormality, and record the type, degree, occurrence time and other information of the abnormality. Similarly, analyze the deviation of quality parameters (such as dimensional accuracy, surface roughness, etc.) to determine whether there is quality abnormality, and also record the detailed information of the abnormality. Integrate the vibration abnormality monitoring results and quality abnormality monitoring results into K quality monitoring results, output the monitoring results to the monitoring interface or send an alarm to relevant personnel so that corrective measures can be taken in time. K quality monitoring results (including K real-time abnormal vibration monitoring results of processing procedures and K real-time abnormal quality monitoring results of processing procedures) can reflect the status of K processing procedures in real time, so that production management personnel can timely understand the abnormal situation in the production process, so as to take prompt measures to adjust or repair it, avoid further deterioration of the problem, and thus improve product quality and production efficiency.

[0027] Step eight: optimizing and controlling the CNC grinding machine based on the K real-time abnormal vibration monitoring results of the machining processes and the K real-time abnormal quality monitoring results of the machining processes.

[0028] Specifically, according to the monitoring results, the processing parameters of the CNC grinder, such as cutting speed, feed rate, cutting depth, etc., are adjusted to reduce vibration and improve quality. For example, for vibration abnormalities, it may be necessary to adjust the cutting strategy or enhance the stability of the machine tool structure; for quality abnormalities, it may be necessary to recalibrate the equipment or replace worn tools. If the monitoring results show potential systemic problems, preventive measures should be taken, such as equipment maintenance, operation training, etc. The automatic control function of the CNC system is used to automatically adjust the processing parameters according to the monitoring results. For complex abnormal situations, the operator may need to make manual adjustments based on the monitoring results. After implementing the optimization control, continue to monitor the processing process to ensure that the adjustment is effective and feedback the results in real time. Record the implementation details of all optimization control measures, including time, parameter changes, results, etc., evaluate the effects of the optimization control measures, such as the degree of vibration reduction, quality improvement, etc., and continuously adjust and optimize the control strategy according to the effect evaluation results to achieve the best production efficiency and product quality. Through the optimization control based on real-time monitoring results, the dynamic adjustment and optimization of the CNC grinder processing process can be achieved, which helps to improve the consistency and stability of product quality, reduce the scrap rate, and improve production efficiency.

[0029] Further, step 2 of this application includes: Acquire a historical processing memory library of the CNC grinding machine, wherein the historical processing memory library includes multiple historical processing records; enumerate the multiple historical processing records in pairs to obtain multiple enumeration combinations; traverse the multiple enumeration combinations to perform similarity identification and determine the similarities of multiple combinations; count the number of combinations whose similarities are greater than or equal to a preset similarity threshold, and if the number of combinations is greater than or equal to the preset number of combinations, obtain a cloud-based processing record collection instruction; supplement the historical processing memory library in the cloud based on the cloud-based processing record collection instruction to obtain a supplemented historical processing memory library; search the supplemented historical processing memory library based on the product specification model information and the product quality parameter set to determine K processing procedures, K process quality parameter tolerance intervals and K processing procedure parameter tolerance intervals.

[0030] Specifically, the historical processing memory library of the CNC grinder is obtained, which records various data and results of past grinder processing, including but not limited to processing parameters, product specifications, quality inspection records, etc. Multiple historical processing records include processing date, time, parameters used, product quality data, equipment status and other information. The records in the historical processing memory library are paired and compared to find similar record combinations. Pairwise enumeration is a comparison method in which each historical processing record is compared with other records in the library to reveal similarities and differences in processing parameters, product quality, equipment status, etc. The degree of similarity is determined by comparing the enumerated record combinations, and is measured by calculating the similarity of processing parameters and quality parameters to obtain multiple combination similarities. Combination similarity refers to the degree of similarity between two record combinations, usually a value between 0 and 1, with 1 indicating exact similarity.

[0031] A similarity threshold is set in advance to determine the boundary value of the similarity between two processing record combinations. The calculated combination similarity is statistically analyzed to determine the number of combinations whose similarity is greater than or equal to the preset similarity threshold. The preset number of combinations is another pre-set standard used to determine whether the number of combinations whose similarity exceeds the threshold is sufficient. If the number of combinations is greater than or equal to the preset number of combinations, that is, the number is too large, it indicates that the data diversity of the historical processing memory is insufficient, that is, the types of products processed by the grinder are too few, and there is no way to comprehensively consider various production situations to give suitable processing parameters for the target grinder products, and further data supplementation is required. Therefore, a processing record collection instruction is generated and sent to the cloud to collect more diverse supplementary data. For example, assuming that the preset similarity threshold is 0.8, the statistical results show that there are 6 pairs of combinations with a similarity greater than or equal to 0.8. If the preset number of combinations is 5, since the statistical result is 6, it indicates that the diversity is insufficient, and an instruction needs to be issued to collect additional processing records from the cloud.

[0032] According to the cloud processing record collection instructions, additional historical processing records are obtained from the cloud server to supplement and improve the existing historical processing memory library, and a supplementary historical processing memory library with rich samples is determined. The supplementary historical processing memory library refers to a more comprehensive processing data set that is expanded by combining local information after cloud data collection. Use product specification model information and product quality parameter sets to search in the historical processing memory library supplemented by the cloud to identify the key processing procedures that need to be monitored and the appropriate processing parameters, including but not limited to grinding speed, grinding depth, feed speed, grinding force, etc. K processing procedures refer to multiple specific processing steps determined according to product specifications and quality requirements, and each procedure has its specific processing methods and parameters. The process quality parameter tolerance range refers to setting the allowable product quality parameter variation range for each processing procedure to ensure quality stability and consistency, and define the acceptable range of product quality parameters. Exceeding this range is considered as quality abnormality. Through cloud supplementation, the data diversity of the historical processing memory library is increased, which helps to determine the processing parameters more accurately.

[0033] Further, step 4 of this application includes: Based on the tolerance intervals of the vibration parameters of the K processing procedures, fluctuations are identified on the sets of vibration parameters of the K historical processing procedures, and K sets of vibration parameter fluctuation values ​​of the historical processing procedures are determined; the sets of vibration parameter fluctuation values ​​of the K historical processing procedures are traversed to perform mean calculation, and K fluctuation means are determined; the variances of the sets of vibration parameter fluctuation values ​​of the K historical processing procedures are respectively calculated, and K concentrated identification steps are determined according to the calculation results; starting from the K fluctuation means, the sets of vibration parameter fluctuation values ​​of the K historical processing procedures are iterated according to the K concentrated identification steps, and when a preset number of iterations is reached or / and an iteration stop requirement is met, the iteration is stopped to obtain K target historical processing procedure vibration parameter fluctuation values; calculations are performed based on the K target historical processing procedure vibration parameter fluctuation values ​​to determine the vibration parameter fluctuation coefficients of the K processing procedures.

[0034] Furthermore, the present application also includes the following steps: The K target historical processing step vibration parameter fluctuation values ​​are respectively divided by the sum of the K target historical processing step vibration parameter fluctuation values, and the ratio is used as the vibration parameter fluctuation coefficient of the K processing steps.

[0035] Specifically, the vibration parameter sets of K historical processing procedures are subjected to fluctuation identification, the fluctuation of the vibration parameters is identified, and the distance between the vibration parameters and the upper and lower limits of the tolerance interval at each time point is calculated. If the vibration parameter value fluctuates within the tolerance interval, the processing process is considered to be stable and the fluctuation value is zero; if it exceeds this range, it is considered that there is fluctuation. For each processing procedure, the degree to which its vibration parameter exceeds the tolerance interval, that is, the fluctuation value, is calculated. The fluctuation value can be an absolute value or a percentage relative to the width of the tolerance interval. All fluctuation values ​​in the vibration parameter fluctuation value set of each processing procedure are averaged, that is, all fluctuation values ​​of the processing procedure are accumulated and divided by the number of fluctuation values ​​in the process to obtain the mean. The calculation process is repeated for K processing procedures, and K fluctuation means are finally obtained, which represent the average fluctuation level of the vibration parameters of each processing procedure. For each fluctuation value, the square of the difference between it and the mean is calculated, and then these square values ​​are accumulated, and the accumulated result is divided by the number of fluctuation values ​​in the process to obtain the variance. The calculation process is repeated for K processing procedures, and K variances are finally obtained. The variance values ​​of each process are analyzed to identify processes with larger and smaller fluctuations.

[0036] Variance is a statistic that measures the degree of fluctuation of a set of data. The larger the variance, the more drastic the fluctuation of the data and the higher the uncertainty. A large variance means a large degree of fluctuation of the data and high uncertainty, indicating that there are large unstable factors in the processing process, such as machine tool vibration, material inconsistency, etc. According to the size of the variance, the identification step size is determined. For high-variance processes, in order to more accurately identify the target fluctuation value, a smaller step size is used for fine detection. A larger step size may cause the target fluctuation value to be missed during the iteration process, while a smaller step size can help analyze the data and reduce the possibility of missing the target fluctuation value. For low-variance processes, since the data fluctuation is small, the step size can be appropriately increased to save resources.

[0037] Taking K fluctuation means as the starting point, iterate in the K historical processing process vibration parameter fluctuation value sets according to K centralized identification steps to find the fluctuation value. The iteration effect is evaluated by comparing the volume of the iteration neighborhood with the fluctuation mean neighborhood. After each iteration, check whether the preset number of iterations is reached, or whether the neighborhood volume meets the stop requirement. When the conditions are met, the iteration process is stopped, and the K iterative historical processing process vibration parameter fluctuation values ​​obtained in the last iteration are used as the K target historical processing process vibration parameter fluctuation values. Add the K target historical processing process vibration parameter fluctuation values ​​to obtain the sum, and divide the K target historical processing process vibration parameter fluctuation values ​​by the sum to obtain the ratio of each fluctuation value in the sum, and use the ratio as the vibration parameter fluctuation coefficient of the K processing processes. By performing fluctuation identification and iterative calculation on the historical processing process vibration parameter set, the vibration parameter fluctuation of each processing process is determined, which helps to improve the stability of the processing process and the consistency of product quality.

[0038] Furthermore, the present application also includes the following steps: Taking the K fluctuation means as the starting point, iterating in the K historical processing procedure vibration parameter fluctuation value sets according to the K concentrated identification steps, obtaining K iterative historical processing procedure vibration parameter fluctuation values; constructing K fluctuation mean neighborhoods and K iteration neighborhoods, wherein the K fluctuation mean neighborhoods are historical processing procedure vibration parameter fluctuation values ​​in the K historical processing procedure vibration parameter fluctuation value sets whose distance to the K fluctuation means is less than or equal to the fluctuation means of the K concentrated identification steps; judging whether the neighborhood volume of the K iteration neighborhoods is greater than or equal to the neighborhood volume of the K fluctuation mean neighborhoods, if so, the iteration stop requirement is not met, continuing to take the K iterative historical processing procedure vibration parameter fluctuation values ​​as the starting point, iterating in the K historical processing procedure vibration parameter fluctuation value sets according to the K concentrated identification steps, until reaching a preset number of iterations or / and satisfying the iteration stop requirement, stopping the iteration, and taking the K iterative historical processing procedure vibration parameter fluctuation values ​​obtained in the last iteration as the K target historical processing procedure vibration parameter fluctuation values.

[0039] Specifically, K fluctuation means are used as the starting point, and K centralized identification steps are followed to iterate in the K historical processing process vibration parameter fluctuation value sets. For each iteration, the fluctuation value found is recorded. In data processing, iteration refers to the repetition of a series of operations, each operation is based on the result of the previous one. In each iteration, according to the current iteration starting point and step size, the closest fluctuation value is found from the historical processing process vibration parameter fluctuation value set as the starting point for the next iteration. The iterative process continues until the preset number of iterations is reached, or the iteration stop requirements are met, and finally K iterative historical processing process vibration parameter fluctuation values ​​are obtained.

[0040] Construct K fluctuation mean neighborhoods, i.e., the historical processing process vibration parameter fluctuation values ​​whose distances to the K fluctuation means in the K historical processing process vibration parameter fluctuation values ​​set are less than or equal to the fluctuation means of the K centralized identification steps. Specifically, the fluctuation mean of each process is taken as the center point, the distance between each value and each fluctuation mean is calculated, and the fluctuation values ​​whose distances are less than or equal to the fluctuation mean of the centralized identification step are retained to form K fluctuation mean neighborhoods. Similar to the K fluctuation mean neighborhoods, construct K iteration neighborhoods, i.e., the iterative historical processing process vibration parameter fluctuation values ​​whose distances to the K iteration fluctuation means in the K iteration historical processing process vibration parameter fluctuation values ​​set are less than or equal to the fluctuation mean of the K centralized identification steps. Specifically, according to the fluctuation values ​​obtained in the iterative process, a new center point is calculated, and the distance between each fluctuation value and the new center point is calculated. The fluctuation values ​​whose distances are less than or equal to the fluctuation mean of the centralized identification step are retained to form K iteration neighborhoods.

[0041] The neighborhood volume refers to the amount of data contained in the neighborhood. It is determined whether the neighborhood volume of the K iterative neighborhoods is greater than or equal to the neighborhood volume of the K fluctuation mean neighborhoods to determine whether the iteration stop requirement is met. If it is greater than or equal to, it means that the current iteration has not met the requirement of sufficient refinement and does not meet the iteration stop requirement. It is necessary to continue to use the vibration parameter fluctuation values ​​of the K iterative historical processing procedures as the starting point and iterate in the K historical processing procedure vibration parameter fluctuation value sets according to K centralized identification steps. The fluctuation value obtained in the latest iteration is used as the starting point and iterates again according to the K centralized identification steps. Until the number of iterations reaches the preset number of iterations, or the iteration stop requirement is met, such as when the iteration volume of the K iterative neighborhoods is less than the neighborhood volume of the K fluctuation mean neighborhoods, the iteration is stopped. The vibration parameter fluctuation values ​​of the K iterative historical processing procedures obtained in the last iteration are used as the vibration parameter fluctuation values ​​of the K target historical processing procedures. Through multiple iterations of refinement, a more accurate description of the fluctuation characteristics is obtained, and the comparison of the volume of the iterative neighborhood and the fluctuation mean neighborhood is used to assist in identifying potential anomalies. The number of iterations and stop conditions are preset to avoid excessive calculation and improve efficiency.

[0042] Further, as attached Figure 2 As shown, step six of this application includes: A monitoring step identifier is constructed, and the monitoring step identifier is used to analyze the vibration parameter fluctuation coefficients of the K processing processes and the quality deviation coefficients of the K processes to determine K first monitoring steps and K second monitoring steps, wherein each first monitoring step is smaller than the corresponding second monitoring step; K first monitoring channels and K second monitoring channels are constructed using the K first monitoring steps and the K second monitoring steps respectively; the K first monitoring channels and the K second monitoring channels are respectively connected in parallel, and the output layers of the K first monitoring channels and the K second monitoring channels are connected to the input layers of the K monitoring result analysis fully connected layers to construct the K process monitoring branches.

[0043] Specifically, the monitoring step length identifier is a model based on a feedforward neural network, which receives the vibration parameter fluctuation coefficient and the process quality deviation coefficient of the processing process as input and outputs the predicted monitoring step length. The training process of the monitoring step length identifier includes supervised training of the network using sample data, and continuously adjusting the network parameters by comparing the difference between the predicted step length and the actual step length until the output converges to a predetermined accuracy. According to the design principle of the feedforward neural network, a neural network is constructed that can input the vibration parameter fluctuation coefficient and the process quality deviation coefficient and output the monitoring step length. The number of nodes in the input layer should be equal to the total number of vibration parameter fluctuation coefficients and process quality deviation coefficients, and contain one or more hidden layers. Each hidden layer uses an appropriate activation function, and the output layer should have two nodes, representing the first monitoring step length and the second monitoring step length respectively. During the training process, by comparing the difference between the network output and the actual monitoring step length, the network parameters are adjusted using optimization algorithms such as gradient descent. When the network performance reaches the preset standard or the number of iterations reaches the preset value, the training is stopped.

[0044] The vibration parameter fluctuation coefficient and process quality deviation coefficient of K processing procedures are analyzed using the trained monitoring step identifier. For each processing procedure, the identifier outputs two monitoring steps, and the first monitoring step is smaller than the corresponding second monitoring step. The first monitoring step is used to quickly monitor the processing process to quickly respond to possible abnormal situations; while the second monitoring step is used to monitor the processing process in detail to more accurately evaluate product quality. According to the first monitoring step of each processing procedure, K first monitoring channels are constructed. According to the second monitoring step of each processing procedure, K second monitoring channels are constructed. The K first monitoring channels and the K second monitoring channels are connected in parallel, so that each monitoring channel can independently collect data and perform preliminary processing to maintain their respective monitoring characteristics. A fully connected layer is designed, whose input layer node number is the sum of the output layer node number of the K first monitoring channels and the K second monitoring channels, and the output layers of the K first monitoring channels and the K second monitoring channels are connected to the input layer of the monitoring result analysis fully connected layer.

[0045] The fully connected layer can receive output information from different monitoring channels. Through its internal neuron connections and calculations, it can comprehensively judge the state of the process, thereby constructing K process monitoring branches. Each process monitoring branch can conduct a comprehensive monitoring result analysis for a specific process, including a comprehensive evaluation of various monitoring data such as the fluctuation coefficient of vibration parameters of the processing process and the process quality deviation coefficient. The two monitoring channels and a fully connected layer of each processing process constitute a monitoring branch, and K process monitoring branches are obtained. By constructing such a process monitoring branch structure, the efficiency and accuracy of monitoring are improved. The data of different monitoring channels are integrated and analyzed in the fully connected layer, avoiding the limitations of a single monitoring method. At the same time, constructing an independent monitoring branch for each process can monitor specific processes more accurately, making the monitoring results more targeted and practical.

[0046] Furthermore, the present application also includes the following steps: A plurality of sample processing step vibration parameter fluctuation coefficients, a plurality of sample process quality deviation coefficients and a plurality of sample monitoring step sets are obtained as training data, wherein each monitoring step set includes a sample first monitoring step and a sample second monitoring step; the training data is used to perform supervised training on a framework constructed based on a feedforward neural network, and during the training, the network parameters are continuously adjusted according to the output until convergence, so as to obtain the trained monitoring step identifier.

[0047] Specifically, multiple sample processing steps vibration parameter fluctuation coefficients, multiple sample process quality deviation coefficients and multiple sample monitoring step sets are obtained, the sample processing step vibration parameter fluctuation coefficient represents a quantitative indicator of the degree of vibration parameter fluctuation of each processing step, the sample process quality deviation coefficient represents a quantitative indicator of the degree of deviation of each process quality parameter from its ideal value, and the sample monitoring step set is obtained from multiple samples, including a first monitoring step of a sample and a second monitoring step of a sample, which is used to determine a certain rhythm or frequency during the monitoring process. The vibration parameter fluctuation coefficient, process quality deviation coefficient and monitoring step are integrated into a training data set, and each data sample should contain the corresponding vibration parameter fluctuation coefficient, quality deviation coefficient and monitoring step.

[0048] A feedforward neural network is a common neural network structure in which information is transmitted in only one direction, from the input layer to the hidden layer and then to the output layer. The SLOWFAST network is a special neural network structure that combines slow and fast feature extraction paths for processing video data and is used to construct a monitoring step identifier. A feedforward neural network with two hidden layers is used as a monitoring step identifier. The input layer receives the vibration parameter fluctuation coefficient of the processing process, the process quality deviation coefficient, and the monitoring step set as input. The hidden layer uses the SLOWFAST network structure to extract features on different time scales. For example, the slow stream is used to process the vibration parameter fluctuation coefficient and the quality deviation coefficient, which change slowly and can use fewer layers and larger time intervals to capture long-term features; the fast stream is used to process the monitoring step, which changes quickly and can use more layers and smaller time intervals to capture short-term features. The output layer outputs two values, corresponding to the first monitoring step and the second monitoring step. By comparing the difference between the predicted step and the actual step, the network continuously adjusts its parameters until the output converges to the predetermined accuracy. When the network converges, it means that the training is completed. The network obtained at this time is a trained monitoring step length identifier, which can accurately output the monitoring step length according to the input processing process vibration parameter fluctuation coefficient and process quality deviation coefficient. Through the trained network monitoring step length identifier, according to the processing process vibration parameter fluctuation coefficient and process quality deviation coefficient, the appropriate monitoring step length can be accurately predicted and identified, which helps to improve the real-time and accuracy of product quality monitoring, so as to timely adjust the processing parameters, optimize the processing process, and improve the consistency and stability of product quality.

[0049] Furthermore, the present application also includes the following steps: Obtaining a feedback monitoring window; identifying the stability of the optimization control in the feedback monitoring window, and obtaining first control warning information when the identification result does not meet the requirements.

[0050] Specifically, a time window is determined, called the feedback monitoring window. The feedback monitoring window is a specific time interval or data range used to monitor the effect of the optimized control of the CNC grinder. The setting of this window needs to take into account the characteristics of the grinding process, the cycle of control adjustment, and the time scale that can effectively reflect the stability. For example, for a CNC grinder, the feedback monitoring window may be set according to the time to process a batch of workpieces, so that the effect of the optimized control can be observed within a complete processing cycle. In the monitoring window, relevant process parameters, equipment status, product quality and other data are collected, including vibration data of the processing process, workpiece size and shape accuracy, equipment operating status parameters, etc. The collected data is analyzed to evaluate the stability of the optimized control, and key parameters such as vibration, temperature, cutting force, etc. of the processing process are analyzed to determine whether the processing process is stable and whether the optimized control strategy is effective. For example, the machining accuracy stability can be evaluated by analyzing the fluctuation of the dimensional accuracy and surface roughness of the workpiece processed within the feedback monitoring window; the control parameter stability can be evaluated by observing whether the adjusted control parameters (such as the proportional, integral, and differential parameters of the PID controller) fluctuate within a reasonable range during the optimization control process; the operating status parameters (such as vibration, temperature, etc.) of the key components of the grinder (such as the grinding wheel motor, workpiece fixture, etc.) are monitored. If the fluctuation of the parameters exceeds the normal range, it may affect the machining quality and control stability. Set the standards for stability identification, which can be thresholds based on historical data, industry standards, or internal regulations of the enterprise to identify whether the data indicates that the control process is stable. If the identification result shows that the process after optimization control is unstable, such as vibration exceeding the threshold, it is considered that the stability does not meet the requirements, and the first control warning information is generated, including the warning level, warning reason, recommended adjustment measures, and timestamp. The warning information can be sent in a variety of ways, such as popping up a warning box on the operation interface of the CNC grinder, sending text messages or emails to relevant operators and maintenance personnel. Once the warning information is received, the operator can take corresponding measures according to the warning content, such as suspending processing, adjusting control parameters, checking equipment components, etc., to ensure that the optimization control of the CNC grinder can meet the requirements of stable operation. By obtaining the feedback monitoring window and identifying the stability of the optimization control, the real-time monitoring of the optimization control effect is realized, which helps to timely discover possible problems in the optimization control and take corresponding measures to adjust, thereby ensuring the stability and quality of the processing process, and avoiding inaccurate optimization decisions caused by unclear monitoring standards.

[0051] In summary, the real-time monitoring and optimization method for CNC grinding machine product quality integrated with the Internet of Things provided in this application has the following technical effects: By acquiring a basic data set of a target grinding machine product, wherein the basic data set includes a product quality parameter set and product specification and model information; performing a CNC grinding machine processing parameter search based on the product specification and model information and the product quality parameter set, and determining K processing procedures, K process quality parameter tolerance intervals, and K processing procedure vibration parameter tolerance intervals; using the product specification and model information and the K processing procedures as indexes, performing frequent item mining on the historical processing data of the CNC grinding machine, and obtaining K historical processing procedure vibration parameter sets and K historical processing procedure quality parameter sets; performing fluctuation concentration identification on the K historical processing procedure vibration parameter sets according to the K processing procedure vibration parameter tolerance intervals, and determining the vibration parameter fluctuation coefficients of the K processing procedures ; Use the K process quality parameter tolerance intervals to perform deviation analysis on the K historical process quality parameters to determine K process quality deviation coefficients; identify the monitoring step lengths of the K processing processes based on the K processing process vibration parameter fluctuation coefficients and the K process quality deviation coefficients, and construct K process monitoring branches according to the identification results; use the K process monitoring branches to extract and analyze the IoT real-time monitoring data of the K processing processes to determine K quality monitoring results, wherein the K quality monitoring results include K real-time processing process vibration abnormality monitoring results and K real-time process quality abnormality monitoring results; optimize the control of the CNC grinding machine based on the K real-time processing process vibration abnormality monitoring and K real-time process quality abnormality monitoring results. In other words, by conducting in-depth analysis of historical data, identifying the vibration parameter fluctuation coefficient and quality deviation coefficient, and dynamically determining the monitoring step, the accuracy and flexibility of monitoring can be improved; by building a process monitoring branch, extracting and analyzing real-time monitoring data, and determining the abnormal vibration monitoring results of the processing process and the abnormal quality monitoring results of the process, the CNC grinder can be optimized and controlled, realizing the refined management and optimization of the CNC grinder processing process and improving product quality.

[0052] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0053] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. A method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things, characterized in that: include: Acquire a basic data set of a target grinding machine product, wherein the basic data set includes a product quality parameter set and product specification model information; Based on the product specification model information and the product quality parameter set, a processing parameter search of a CNC grinding machine is performed to determine K processing steps, K process quality parameter tolerance intervals, and K processing step vibration parameter tolerance intervals; Using the product specification model information and the K processing procedures as indexes, frequent item mining is performed on the historical processing data of the CNC grinder to obtain K historical processing procedure vibration parameter sets and K historical process quality parameter sets; Performing concentrated fluctuation identification on the K historical processing process vibration parameter sets according to the K processing process vibration parameter tolerance intervals to determine the K processing process vibration parameter fluctuation coefficients; Using the K process quality parameter tolerance intervals, the K historical process quality parameters are subjected to deviation analysis to determine the K process quality deviation coefficients; Identify the monitoring steps of the K processing steps based on the vibration parameter fluctuation coefficients of the K processing steps and the quality deviation coefficients of the K processing steps, and construct K process monitoring branches according to the identification results; Utilizing the K process monitoring branches to extract and analyze the IoT real-time monitoring data of the K processing processes, and determining K quality monitoring results, wherein the K quality monitoring results include K real-time processing process vibration abnormality monitoring results and K real-time process quality abnormality monitoring results; The CNC grinding machine is optimized and controlled based on the K real-time abnormal vibration monitoring results of the machining processes and the K real-time abnormal quality monitoring results of the processes.

2. The method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things as claimed in claim 1 is characterized in that: include: Acquire a historical processing memory library of the CNC grinding machine, wherein the historical processing memory library includes a plurality of historical processing records; Enumerating the plurality of historical processing records in pairs to obtain a plurality of enumeration combinations; Traversing the multiple enumeration combinations to perform similarity identification, and determining the similarities of the multiple combinations; Counting the number of combinations whose similarities are greater than or equal to a preset similarity threshold, and if the number of combinations is greater than or equal to a preset number of combinations, obtaining a cloud processing record collection instruction; Supplementing the historical processing memory library in the cloud based on the cloud processing record collection instruction to obtain a supplemented historical processing memory library; Based on the product specification model information and the product quality parameter set, a search is performed in the supplementary historical processing memory library to determine K processing procedures, K process quality parameter tolerance intervals and K processing procedure parameter tolerance intervals.

3. The method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things as claimed in claim 1 is characterized in that: The vibration parameter sets of the K historical processing steps are subjected to concentrated fluctuation identification according to the tolerance intervals of the vibration parameters of the K processing steps, and the fluctuation coefficients of the vibration parameters of the K processing steps are determined, including: Based on the tolerance intervals of the vibration parameters of the K processing steps, fluctuation identification is performed on the vibration parameter sets of the K historical processing steps to determine the vibration parameter fluctuation value sets of the K historical processing steps; Traversing the K historical processing steps vibration parameter fluctuation value sets to perform mean calculation, and determining K fluctuation means; Calculating the variances of the K historical processing steps' vibration parameter fluctuation value sets respectively, and determining K centralized identification step lengths according to the calculation results; Taking the K fluctuation means as the starting point, iterating in the K historical processing process vibration parameter fluctuation value sets according to the K centralized identification steps, stopping the iteration when a preset number of iterations is reached or / and an iteration stop requirement is met, and obtaining K target historical processing process vibration parameter fluctuation values; Calculation is performed based on the K target historical processing step vibration parameter fluctuation values ​​to determine the K processing step vibration parameter fluctuation coefficients.

4. The method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things as claimed in claim 3 is characterized in that: Taking the K fluctuation means as the starting point, iterating in the K historical processing process vibration parameter fluctuation value sets according to the K centralized identification steps, stopping the iteration when a preset number of iterations is reached or / and an iteration stop requirement is met, and obtaining K target historical processing process vibration parameter fluctuation values, including: Taking the K fluctuation means as the starting point, iterating in the K historical machining process vibration parameter fluctuation value sets according to the K centralized identification steps to obtain K iterative historical machining process vibration parameter fluctuation values; Constructing K fluctuation mean neighborhoods and K iteration neighborhoods, wherein the K fluctuation mean neighborhoods are historical processing process vibration parameter fluctuation values ​​whose distance to the K fluctuation means in the K historical processing process vibration parameter fluctuation value sets is less than or equal to the fluctuation means of K concentrated identification steps; It is determined whether the neighborhood volumes of the K iterative neighborhoods are greater than or equal to the neighborhood volumes of the K fluctuation mean neighborhoods. If so, the iteration stop requirement is not met, and the K iterative historical processing step vibration parameter fluctuation values ​​are continued to be taken as the starting point. The K historical processing step vibration parameter fluctuation value sets are iterated according to the K concentrated identification steps until a preset number of iterations is reached or / and the iteration stop requirement is met. The iteration is stopped, and the K iterative historical processing step vibration parameter fluctuation values ​​obtained in the last iteration are used as the K target historical processing step vibration parameter fluctuation values.

5. The method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things as claimed in claim 3 is characterized in that: The K target historical processing step vibration parameter fluctuation values ​​are respectively divided by the sum of the K target historical processing step vibration parameter fluctuation values, and the ratio is used as the vibration parameter fluctuation coefficient of the K processing steps.

6. The method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things as claimed in claim 1 is characterized in that: include: Constructing a monitoring step length identifier, and using the monitoring step length identifier to analyze the vibration parameter fluctuation coefficients of the K processing steps and the quality deviation coefficients of the K processes, to determine K first monitoring step lengths and K second monitoring step lengths, wherein each first monitoring step length is smaller than the corresponding second monitoring step length; The K first monitoring steps and the K second monitoring steps respectively construct K first monitoring channels and K second monitoring channels; The K first monitoring channels and the K second monitoring channels are respectively connected in parallel, and the output layers of the K first monitoring channels and the K second monitoring channels are connected to the input layers of the K monitoring result analysis fully connected layers to construct the K process monitoring branches.

7. The method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things as claimed in claim 6 is characterized in that: include: Acquire multiple sample processing process vibration parameter fluctuation coefficients, multiple sample process quality deviation coefficients and multiple sample monitoring step sets as training data, wherein each monitoring step set includes a sample first monitoring step and a sample second monitoring step; The framework constructed based on the feedforward neural network is supervised and trained using the training data. During the training, the network parameters are continuously adjusted according to the output until convergence, thereby obtaining the trained monitoring step length identifier.

8. The method for real-time monitoring and optimization of CNC grinding machine product quality integrating the Internet of Things as claimed in claim 1 is characterized in that: include: Get feedback monitoring window; The stability of the optimized control is identified in the feedback monitoring window, and when the identification result does not meet the requirements, the first control early warning information is obtained.

Images