Numerical control grinding machine intelligent control method and device combined with real-time feedback

Abstract

The invention discloses a numerical control grinding machine intelligent control method and device combined with real-time feedback, and relates to the technical field of grinding machine control. The method comprises the steps that historical machining records of a plurality of numerical control grinding machines in a target factory are collected and divided, a plurality of grinding machine domains of the same kind are determined, machining deviation analysis is conducted, and a plurality of static compensation coefficient sets are obtained; primary compensation is conducted on the initial machining parameters, and the multiple numerical control grinding machines are controlled to conduct grinding machining; performing fixed-point compensation on the primary compensation parameter set to obtain a secondary compensation parameter set; and the multiple numerical control grinding machines are controlled to conduct grinding till the equipment stops continuous operation or grinding wheel accessories are replaced. The technical problem that in the prior art, in the same workshop, due to the fact that the equipment states are inconsistent, and the machining consistency of batch products is poor along with the change of the machining time and the grinding wheel state is solved, and the technical effects that the machining process of the numerical control grinding machine is accurately controlled in an omnibearing mode, and the machining consistency of the batch products is poor are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of grinder control, and in particular to a method and device for intelligently controlling a numerically controlled grinder combined with real-time feedback. Background Art

[0002] In modern manufacturing, CNC grinding machines are key equipment for precision machining and are widely used in many fields such as mechanical machining, aerospace, and automobile manufacturing. With the rapid development of industrial automation and intelligence, higher and higher requirements are placed on the precision, efficiency, and product consistency of CNC grinding machine processing. In the same workshop, there are often multiple CNC grinding machines equipped for simultaneous processing and production. However, due to factors such as different degrees of equipment aging, subtle differences in manufacturing processes, and different losses during use, even if workpieces with the same product characteristics are processed under the same operating parameters, there are still significant differences in the consistency of products processed by each CNC grinder. At the same time, during the processing of a single CNC grinder, as the processing time increases, the grinding wheel will inevitably wear and its grinding performance will gradually decrease. In addition, after long-term use, the cooling and lubrication properties of the coolant will also decrease. If fixed control parameters are always used during the processing process, it is impossible to adapt to the changes in the performance of the grinding wheel and coolant in time, which will inevitably further affect the processing quality of the product and lead to poor product consistency.

[0003] The existing technology has a technical problem that the processing consistency of batch products is poor due to inconsistent equipment status in the same workshop and changes in the grinding wheel status over time. Summary of the invention

[0004] The present application provides an intelligent control method and device for a CNC grinder combined with real-time feedback, which is used to solve the technical problem in the prior art that the processing consistency of batch products is poor due to inconsistent equipment status in the same workshop and changes in grinding wheel status over processing time.

[0005] In view of the above problems, the present application provides a method and device for intelligent control of a CNC grinder combined with real-time feedback.

[0006] In a first aspect of the present application, a method for intelligent control of a CNC grinding machine combined with real-time feedback is provided, the method comprising: Collect historical processing records of several CNC grinders in the target factory, divide the several CNC grinders according to the historical processing records, determine multiple similar grinder domains, and perform processing deviation analysis on the multiple similar grinder domains respectively to obtain multiple static compensation coefficient sets; determine initial processing parameters according to product feature matching, perform a primary compensation on the initial processing parameters based on the multiple static compensation coefficient sets, and control several CNC grinders to perform grinding processing according to the primary compensation parameter set; at a predetermined adjustment time node, perform fixed-point compensation on the primary compensation parameter set according to a secondary compensation scheme to obtain a secondary compensation parameter set, wherein the secondary compensation is used to compensate for the effects of grinding wheel wear and reduced cooling performance; according to the secondary compensation parameter set, control several CNC grinders to perform grinding processing until the equipment stops continuous operation or the grinding wheel accessories are replaced.

[0007] The second aspect of the present application provides an intelligent control device for a CNC grinding machine combined with real-time feedback, the device comprising: A static compensation coefficient set acquisition module, the static compensation coefficient set acquisition module is used to collect historical processing records of several CNC grinders in the target factory, divide the several CNC grinders according to the historical processing records, determine multiple similar grinder domains, and perform processing deviation analysis on the multiple similar grinder domains respectively to obtain multiple static compensation coefficient sets; a grinding process control module, the grinding process control module is used to determine initial processing parameters according to product feature matching, perform a primary compensation on the initial processing parameters based on the multiple static compensation coefficient sets, and control several CNC grinders to perform grinding according to the primary compensation parameter set; a secondary compensation parameter set acquisition module, the secondary compensation parameter set acquisition module is used to perform fixed-point compensation on the primary compensation parameter set according to the secondary compensation scheme at a predetermined adjustment time node to obtain a secondary compensation parameter set, wherein the secondary compensation is used to compensate for the effects of grinding wheel wear and reduced cooling performance; a secondary compensation parameter set control module, the secondary compensation parameter set control module is used to control several CNC grinders to perform grinding according to the secondary compensation parameter set until the equipment stops continuous operation or the grinding wheel accessories are replaced.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: Collect the historical processing records of several CNC grinders in the target factory, divide several CNC grinders according to the historical processing records, determine multiple similar grinder domains, and perform processing deviation analysis on the multiple similar grinder domains respectively to obtain multiple static compensation coefficient sets; determine the initial processing parameters, perform a primary compensation on the initial processing parameters based on the multiple static compensation coefficient sets, and control several CNC grinders to perform grinding processing; at the predetermined adjustment time node, perform fixed-point compensation on the primary compensation parameter set according to the secondary compensation scheme to obtain a secondary compensation parameter set; according to the secondary compensation parameter set, control several CNC grinders to perform grinding processing until the equipment stops continuous operation or the grinding wheel accessories are replaced. Achieve all-round and precise control of the CNC grinder processing process, and solve the technical effect of poor consistency in batch product processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A schematic flow chart of an intelligent control method for a CNC grinder combined with real-time feedback provided in an embodiment of the present application; Figure 2 A schematic diagram of the structure of an intelligent control device for a CNC grinder combined with real-time feedback provided in an embodiment of the present application.

[0011] Description of the reference numerals: static compensation coefficient set acquisition module 10 , grinding process control module 20 , secondary compensation parameter set acquisition module 30 , secondary compensation parameter set control module 40 . DETAILED DESCRIPTION

[0012] The present application provides an intelligent control method and device for a CNC grinder combined with real-time feedback, which is used to solve the technical problem in the prior art that the processing consistency of batch products is poor due to inconsistent equipment status in the same workshop and changes in grinding wheel status over processing time.

[0013] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0014] Embodiment 1, as Figure 1As shown, the present application provides an intelligent control method for a CNC grinder combined with real-time feedback, the method comprising: Step S100: Collect historical processing records of several CNC grinders in the target factory, divide the several CNC grinders according to the historical processing records, determine multiple similar grinder domains, and perform processing deviation analysis on the multiple similar grinder domains respectively to obtain multiple static compensation coefficient sets.

[0015] Specifically, collecting the historical processing records of several CNC grinders in the target factory is the basic work. These historical processing records contain the processing process data for different products at different time points in the past, such as the size information of the processed products, surface roughness data, and the processing parameters used at the time. According to these rich historical processing records, several CNC grinders are divided. First, the mean value of the size characteristics and the mean value of the surface roughness of the products processed by each CNC grinder in the historical time window are obtained from the historical processing records. Then, with the same processed products as constraints, the CNC grinders are clustered multiple times according to these means to obtain multiple clustering results. Then, the co-occurrence frequency of any two CNC grinders in the clustering results is statistically analyzed, and the two CNC grinders with a co-occurrence frequency greater than the predetermined frequency threshold are identified as having a high similarity, and are added to the same grinder domain, and finally multiple similar grinder domains are determined. After determining the same grinder domain, the processing deviation analysis is performed on each domain separately. A similar grinding machine domain (such as the first similar grinding machine domain) is randomly selected, and the mean dimensional deviation and the mean surface roughness deviation within the historical time window are obtained from the historical processing records. These deviation means are input into the parameter static compensation model, and the model outputs a static compensation coefficient set after calculation. This process is repeated until multiple static compensation coefficient sets corresponding to multiple similar grinding machine domains are obtained.

[0016] Step S200: determining initial processing parameters according to product feature matching, performing a primary compensation on the initial processing parameters based on the multiple static compensation coefficient sets, and controlling a plurality of CNC grinders to perform grinding processing according to the primary compensation parameter set.

[0017] Specifically, the initial processing parameters are determined based on the matching of product features. Product features include the material, shape, size requirements, and surface quality standards of the product. By matching these features with the preset processing parameter matching rules, the initial processing parameters suitable for the product processing are determined. Then, the initial processing parameters are compensated once based on the multiple static compensation coefficient sets obtained in the previous steps. Each static compensation coefficient set corrects the common deviations of equipment in the same grinder domain, and these coefficients are applied to the initial processing parameters to obtain a primary compensation parameter set. Finally, several CNC grinders are controlled to perform grinding processing according to the primary compensation parameter set, so that during the initial processing of the equipment, the processing deviations caused by differences between equipment can be compensated to a certain extent, thereby improving the consistency of the product.

[0018] Step S300: at a predetermined adjustment time node, performing fixed-point compensation on the primary compensation parameter set according to a secondary compensation scheme to obtain a secondary compensation parameter set, wherein the secondary compensation is used to compensate for the effects of grinding wheel wear and reduced cooling performance.

[0019] Specifically, the secondary compensation operation is performed at the predetermined adjustment time node. The predetermined adjustment time node is pre-set based on experience or research on the law of grinding wheel wear and coolant performance degradation, such as adjustment after a certain processing time or after processing a certain number of products. When constructing a secondary compensation scheme, first use product characteristics as constraints, query historical processing records, obtain the mean value of grinding wheel wear characteristics and the mean value of cooling performance degradation ratio at multiple continuous operation nodes when the CNC grinder is continuously operating, and combine these means according to the operation nodes to obtain multiple groups of input data, each group of input data corresponds to an operation node. Then input multiple groups of input data into the secondary compensation analysis plug-in for analysis, and the plug-in outputs multiple predetermined secondary compensation coefficient sets after complex internal calculations. Finally, a mapping relationship between the operation node and the predetermined secondary compensation coefficient set is established, and a secondary compensation scheme is constructed based on this mapping relationship. According to the secondary compensation scheme, the primary compensation parameter set is compensated at a fixed point, thereby obtaining a secondary compensation parameter set, which effectively compensates for the impact of grinding wheel wear and cooling performance degradation on processing consistency.

[0020] Step S400: According to the secondary compensation parameter set, a plurality of CNC grinding machines are controlled to perform grinding processing until the equipment stops continuous operation or the grinding wheel accessories are replaced.

[0021] Specifically, several CNC grinders are controlled to perform grinding processing based on the secondary compensation parameter set. During the entire processing process, the secondary compensation parameter set is continuously used until the equipment stops continuous operation (such as completing the production task, equipment failure, etc.) or the grinding wheel accessories are replaced (when the grinding wheel wear reaches a certain level and affects the processing quality). This step ensures that during the operation of the equipment, through reasonable parameter compensation, a high level of processing consistency is always maintained, the quality stability of batch products is guaranteed, production efficiency and product qualification rate are improved, the scrap rate and rework rate caused by changes in equipment performance are reduced, and the economic benefits and product competitiveness of the entire production process are improved.

[0022] In a possible implementation, step S300 further includes: Step S310: At a predetermined monitoring time node, parameter fluctuation ranges and processed product deviations of several CNC grinders are collected, and parameter optimization analysis is performed based on the parameter fluctuation ranges and processed product deviations to determine several sets of cubic compensation coefficients.

[0023] Step S320: dynamically fine-tune the quadratic compensation parameter set according to the plurality of cubic compensation coefficients to obtain a plurality of optimized processing parameters, and control a plurality of CNC grinders to perform grinding operations.

[0024] Specifically, the setting of predetermined monitoring time nodes is the key to ensure timely capture of changes in the processing process. These time nodes can be determined based on experience, equipment operation rules or production process requirements, such as every certain processing time interval, after each batch of product processing is completed, etc. At these specific moments, comprehensive data collection work is carried out on several CNC grinders. The collected parameter fluctuation range covers the fluctuation of key processing parameters such as grinding speed, workpiece speed, feed speed, cutting depth and tool speed. The maximum, minimum and average values ​​of these parameters in a certain period of time are accurately recorded through sensors and data acquisition systems to reflect the dynamic changes in the processing process. At the same time, the deviation of the processed product, including the dimensional deviation and surface roughness deviation of the product, is accurately measured and compared with the design requirements. Then, parameter optimization analysis is performed based on the collected parameter fluctuation range and processing product deviation. To this end, it is necessary to query the historical processing records, from which a sample parameter fluctuation range set, a sample processing product deviation set and multiple sample parameter compensation coefficient sets are collected. Each processing parameter here corresponds to a parameter compensation coefficient, which is used to quantify the impact of parameter adjustment on the processing results. The sample parameter fluctuation range and sample processing product deviation are used as input data, and the sample parameter compensation coefficient is used as supervision information. These rich sample data are used to train the machine learning model. The machine learning model learns from a large amount of historical data and mines the intrinsic relationship between parameter fluctuation and product deviation, and then obtains the parameter optimization analyzer. This analyzer can accurately perform parameter optimization analysis based on the actual parameter fluctuation range and processing product deviation currently collected, and determine several sets of cubic compensation coefficients. Each coefficient set corresponds to a CNC grinder to correct the deviation in its processing process in a targeted manner.

[0025] The secondary compensation parameter set is dynamically fine-tuned based on several determined third-order compensation coefficients. Since the secondary compensation parameter set is determined based on factors such as grinding wheel wear and coolant performance degradation, there will still be other dynamic changes that affect processing consistency during the actual processing process. The third-order compensation coefficient further corrects these real-time changes. The third-order compensation coefficient is calculated with the corresponding parameters of the secondary compensation parameter set, for example, according to a certain weighted summation or other reasonable calculation method, to obtain the optimized processing parameters. These optimized processing parameters fully consider many factors such as the current actual operating status of the equipment, the degree of grinding wheel wear, the coolant performance, and the dynamic changes during the processing. Finally, according to these optimized processing parameters, several CNC grinders are controlled to perform grinding operations to ensure that during the entire processing process, the CNC grinder can continuously adjust the processing parameters according to real-time feedback, always maintain a high-precision processing state, effectively improve the dimensional consistency and surface roughness consistency of batch products, improve product quality and production efficiency, and reduce production costs.

[0026] In a possible implementation, step S310 further includes: Step S311: query historical processing records, collect sample parameter fluctuation range sets, sample processing product deviation sets and multiple sample parameter compensation coefficient sets, where the processing parameters include grinding speed, workpiece speed, feed speed, cutting depth and tool speed, and each processing parameter corresponds to a parameter compensation coefficient.

[0027] Step S312: Taking the sample parameter fluctuation range and the sample processing product deviation as input and the sample parameter compensation coefficient as supervision, the machine learning model is trained using the sample parameter fluctuation range set, the sample processing product deviation set and multiple sample parameter compensation coefficient sets to obtain a parameter optimization analyzer.

[0028] Step S313: using the parameter optimization analyzer, performing parameter optimization analysis according to the parameter fluctuation range and the processed product deviation.

[0029] Specifically, we must first deeply explore the valuable data resource of historical processing records. Historical processing records cover the processing details of various products by CNC grinders during long-term operation in the past. For the processing parameter of grinding speed, by carefully analyzing the speed change data under different processing tasks, the maximum, minimum and average values ​​in each processing cycle are determined, so as to construct a sample parameter fluctuation interval set of grinding speed. For example, when processing a specific part, the initial setting of the grinding speed is 2000 rpm. During the processing, due to various factors, the speed fluctuates between 1950 rpm and 2050 rpm. This fluctuation range data is a sample parameter fluctuation interval of grinding speed. In the same way, the fluctuation intervals of processing parameters such as workpiece speed, feed speed, cutting depth and tool speed are collected, and finally summarized to form a complete sample parameter fluctuation interval set. At the same time, a comprehensive quality inspection is carried out on the historical processing products, and the difference between the actual size and the design size of each product, as well as the degree of deviation between the surface roughness and the standard roughness are accurately measured. These size deviation and surface roughness deviation data are sorted out to obtain a sample processing product deviation set. For example, the design size of a batch of processed products is 100 mm, and the actual measured size fluctuates between 99.8 mm and 100.2 mm. These deviation data become part of the sample processed product deviation set. In addition, in order to provide effective supervision information for subsequent machine learning model training, it is necessary to obtain multiple sample parameter compensation coefficient sets from historical processing records. For each processing parameter, the compensation coefficients used in the past to adjust the parameter to improve the quality of the processed product are reviewed. These compensation coefficients are closely related to the fluctuations in processing parameters and product deviations. For example, when it is found that the fluctuation of the grinding speed leads to product size deviation, the compensation coefficient used to correct the grinding speed in the past is recorded. For different processing conditions and deviation types, multiple sample parameter compensation coefficient sets are formed. The coefficients in each set correspond to the corresponding processing parameters, which lays a solid data foundation for subsequent model learning and parameter optimization analysis.

[0030] The data of grinding speed fluctuation range, workpiece speed fluctuation range, feed speed fluctuation range, cutting depth fluctuation range and tool speed fluctuation range in the sample parameter fluctuation range set, as well as the dimensional deviation and surface roughness deviation data in the sample processing product deviation set are used as the input feature vector of the model. These input data comprehensively reflect the dynamic changes of the CNC grinder in the past processing process and the resulting product quality deviation. Then, a neural network model is constructed, which includes an input layer, a hidden layer and an output layer. The number of neurons in the input layer corresponds to the dimension of the input feature vector, which is used to receive the parameter fluctuation range and processing product deviation data. The hidden layer sets a certain number of neurons (which can be adjusted according to the data complexity and model performance requirements) and uses the connection weights and activation functions (such as ReLU functions) between neurons to learn the complex nonlinear relationship in the data. The output layer outputs the same number of predicted compensation coefficients as the number of processing parameters, and each coefficient corresponds to a processing parameter. During the training process, the sample parameter compensation coefficient set is used as the supervisory information, where the compensation coefficient corresponding to each processing parameter clearly indicates the adjustment direction and amplitude to be made for the corresponding parameter fluctuation and product deviation. The model calculates the predicted compensation coefficient based on the input data and compares it with the actual sample parameter compensation coefficient. By calculating the error between the two (such as the mean square error), the back propagation algorithm is used to back propagate the error from the output layer to the hidden layer and the input layer, and the connection weights between neurons are adjusted, so that the compensation coefficient predicted by the model gradually approaches the actual sample parameter compensation coefficient. As the training continues, the model's learning of the sample data continues to deepen, and gradually masters the complex nonlinear relationship between the parameter fluctuation range, the deviation of the processed product and the parameter compensation coefficient. When the performance indicators of the model on the training set and the validation set (such as the mean square error, accuracy, etc.) reach the expected standards, it indicates that the model has the ability to accurately predict the compensation coefficient based on the input parameter fluctuation range and the deviation of the processed product. At this time, the training is completed, and the obtained parameter optimization analyzer can receive the real-time collected parameter fluctuation range and the deviation data of the processed product, and quickly output the corresponding optimized compensation coefficient, which provides a key basis for the dynamic fine-tuning of the subsequent CNC grinder processing parameters, thereby realizing the precise control of the processing process and effectively improving the processing consistency and quality stability of batch products. In addition to the neural network algorithm, support vector machine, decision tree regression and other algorithms can also be applied to the model training process according to actual conditions to meet the needs of different scenarios.

[0031] Using the trained parameter optimization analyzer, the key parameter optimization analysis work is carried out according to the real-time collected parameter fluctuation range and processing product deviation. First, the fluctuation range data of processing parameters such as grinding speed, workpiece speed, feed speed, cutting depth and tool speed obtained from the CNC grinder sensor at the current moment are input into the parameter optimization analyzer. These data reflect the real-time dynamic changes of the equipment operation during the processing process. At the same time, the dimensional deviation and surface roughness deviation data obtained by the precise measurement of the product being processed are also input. The parameter optimization analyzer quickly processes the input real-time data based on the complex mapping relationship between the parameter fluctuation range, processing product deviation and parameter compensation coefficient learned in the training stage. It comprehensively considers the influence of each processing parameter fluctuation on product deviation and their interaction through the internal complex calculation logic. For example, the correlation between grinding speed fluctuation and product dimensional deviation, and the influence of workpiece speed fluctuation on surface roughness deviation are analyzed. Based on these analyses, the parameter optimization analyzer calculates the most suitable compensation coefficient for each processing parameter, forming a set of three compensation coefficients for the current processing state. This coefficient set can accurately guide the subsequent dynamic fine-tuning of the secondary compensation parameter set, so that the CNC grinder can adapt to various changes in the processing process in a timely manner, ensuring that high-precision processing is always maintained throughout the entire processing process, effectively improving the dimensional consistency and surface roughness consistency of batch products, improving product quality and production efficiency, and reducing scrap rate and production costs caused by processing deviations.

[0032] In a possible implementation, step S100 further includes: Step S110: According to the historical processing records, a plurality of dimensional feature averages and a plurality of surface roughness averages of a plurality of processed products of a plurality of CNC grinding machines within a historical time window are obtained.

[0033] Step S120: taking the same processed products as a constraint, clustering a plurality of CNC grinding machines multiple times according to the multiple size feature means and the multiple surface roughness means to obtain multiple clustering results.

[0034] Step S130: Counting the co-occurrence frequencies of any two CNC grinding machines in the plurality of clustering results, and adding any two CNC grinding machines with a co-occurrence frequency greater than a predetermined frequency threshold to a similar grinding machine domain, to obtain the plurality of similar grinding machine domains.

[0035] Specifically, we will dig deep into the key information source of historical processing records. The historical processing records record in detail the rich data of several CNC grinders processing various products in the past period of time. For the multiple processed products processed by each CNC grinder, we will accurately extract their dimensional feature information, such as the diameter, length, aperture and other key dimensional data of the product, and calculate the average value of these dimensional data within the historical time window, so as to obtain multiple dimensional feature means. At the same time, we use high-precision measuring equipment to obtain the surface roughness data of each processed product, and calculate its mean value to obtain multiple surface roughness mean values. These mean data can reflect the average quality level and characteristics of the products processed by each CNC grinder in a specific historical period from a macro perspective.

[0036] Taking the same processed product as the core constraint, we deeply explore and utilize multiple dimensional feature means and multiple surface roughness means, carry out multiple clustering operations on several CNC grinders, and then obtain multiple clustering results. First, for each processed product, the dimensional feature means and surface roughness means corresponding to the CNC grinder processing data related to it are accurately extracted. For example, when processing a certain precision gear, we focus on the dimensional feature means such as the tooth thickness size mean, the tooth top circle diameter mean, and the tooth surface surface roughness mean obtained by each CNC grinder processing the gear. These mean data for specific products are used as the clustering basis, and advanced clustering algorithms (such as hierarchical clustering algorithm, DBSCAN algorithm, etc.) are used for clustering operations. In the clustering process, the algorithm divides them into different categories based on the similarity of the dimensional features and surface roughness features presented by each CNC grinder when processing the product. The similarity measurement standard comprehensively considers the difference between the dimensional feature means and the closeness of the surface roughness means. The smaller the difference and the higher the closeness, the stronger the similarity is considered, and the more likely it is to be clustered into one category. Since there are many different processed products, the above clustering process is repeated for each product, so multiple clustering results will be obtained. Each clustering result clearly reflects the similarity grouping between CNC grinders in terms of processing quality and performance when processing specific products, providing key classification information for subsequent accurate identification of equipment groups with common characteristics and then implementing targeted control strategies, which helps to improve the accuracy and stability of the overall CNC grinder processing in the workshop and achieve high quality consistency in mass production.

[0037] For the multiple clustering results obtained previously, the combination of any two CNC grinders is comprehensively sorted out. For each pair of CNC grinders, the number of times they appear in the same group in different clustering results is carefully counted, which is used as the basis for calculating the co-occurrence frequency. For example, in a scenario with five clustering results, if CNC grinder A and CNC grinder B are classified in the same group in three of the clustering results, then their co-occurrence frequency is 3. Next, the key judgment criterion of the predetermined frequency threshold is introduced. This threshold is set based on an in-depth understanding of the overall performance distribution and similarity requirements of the workshop CNC grinders. The co-occurrence frequency of each pair of CNC grinders is compared with the predetermined frequency threshold one by one. When the co-occurrence frequency of a pair of CNC grinders is greater than the predetermined frequency threshold, it means that the two devices show high similarity and correlation in the processing clusters of multiple products, and they have relatively consistent performance characteristics and quality performance when processing different products. At this time, the two CNC grinders are added to the domain of similar grinders. By counting and screening the co-occurrence frequencies of all CNC grinder combinations, multiple domains of similar grinders are gradually constructed. The CNC grinders in each similar grinding machine domain have stronger commonalities, which provides a solid foundation for subsequent intelligent control measures such as unified compensation coefficient determination and processing parameter optimization for the equipment in the domain, helps to improve the overall CNC grinder processing efficiency and product quality consistency in the workshop, and effectively reduce processing errors and production costs caused by equipment differences.

[0038] In a possible implementation, step S100 further includes: Step S140: randomly selecting a first similar grinding machine domain, and obtaining a mean value of size deviation and a mean value of surface roughness deviation of the first similar grinding machine domain within a historical time window according to the historical processing records.

[0039] Step S150: inputting the mean value of the size deviation and the mean value of the surface roughness deviation into a parameter static compensation model, outputting a first static compensation coefficient set, and adding the first static compensation coefficient set to the plurality of static compensation coefficient sets.

[0040] Specifically, the first similar grinder domain is randomly selected from multiple similar grinder domains that have been constructed. This random selection ensures the representativeness and universality of subsequent analysis. Subsequently, the historical processing records are deeply mined, focusing on the processing data of the CNC grinder in the first similar grinder domain within the historical time window. For the dimensional information of the processed product, the deviation between it and the design standard dimension is accurately calculated, and the mean dimensional deviation is obtained by statistically averaging a large number of dimensional deviation data of processed products. At the same time, for the surface roughness of the product, its deviation from the ideal surface roughness is also measured, and the mean surface roughness deviation is calculated. These deviation mean data comprehensively reflect the average deviation level and quality stability of this similar grinder domain in the historical processing process.

[0041] After obtaining the mean dimensional deviation and surface roughness deviation of the first similar grinder domain in the historical time window, they are input into the parameter static compensation model. This parameter static compensation model can be an intelligent model built based on a machine learning algorithm, or a mapping relationship table summarized based on a large amount of experimental data and engineering experience. If it is a machine learning model, it has learned the complex relationship between many different dimensional deviation mean values, surface roughness deviation mean value combinations and corresponding static compensation coefficient sets during the training phase. When the current dimensional deviation mean value and surface roughness deviation mean value are input, the model will quickly calculate and reason based on its internal neural network structure algorithm logic. Through feature extraction and pattern recognition of the input data, the model predicts the most suitable static compensation coefficient set for the current deviation situation, which includes accurate compensation coefficient adjustments for each processing parameter (such as grinding speed, workpiece speed, feed speed, cutting depth and tool speed, etc.) to correct the processing deviation caused by inherent differences in the equipment. If a mapping relationship table is used, a corresponding relationship table between the dimensional deviation mean value, surface roughness deviation mean value and static compensation coefficient set is established in advance based on rich historical data and practical experience. After inputting the specific mean value of dimensional deviation and mean value of surface roughness deviation, the corresponding first static compensation coefficient set is directly obtained by looking up the table. This method is simple and direct, and can quickly provide a compensation plan based on past mature experience. Regardless of which method is used to obtain the first static compensation coefficient set, it is added to multiple static compensation coefficient sets, providing a key basis for subsequent unified compensation and processing parameter optimization for different similar grinder domains, thereby effectively improving the consistency of CNC grinder processing products, ensuring the quality stability of mass production, improving production efficiency and reducing production costs.

[0042] In a possible implementation, step S300 further includes: Step S330: Taking the product characteristics as constraints, query historical processing records to obtain the average values ​​of multiple grinding wheel wear characteristics at multiple continuous operation nodes when several CNC grinders are continuously operating, and the average values ​​of multiple cooling performance reduction ratios at multiple continuous operation nodes.

[0043] Step S340: combining the multiple grinding wheel wear characteristic mean values ​​and the multiple cooling performance degradation ratio mean values ​​according to the operation nodes to obtain multiple groups of input data, wherein each group of input data corresponds to one operation node.

[0044] Step S350: inputting the plurality of groups of input data into a quadratic compensation analysis plug-in for analysis, and outputting a plurality of predetermined quadratic compensation coefficient sets.

[0045] Step S360: establishing a mapping relationship between a plurality of operation nodes and a plurality of predetermined secondary compensation coefficient sets, and constructing the secondary compensation scheme according to the mapping relationship.

[0046] Specifically, with product characteristics as the key constraints, historical processing records are queried in depth. Aiming at the conditions of several CNC grinders during continuous operation, the mean values ​​of grinding wheel wear characteristics at multiple continuous operation nodes are accurately obtained. Grinding wheel wear characteristics can be measured in a variety of ways, such as the reduction in grinding wheel radius, the degree of change in grinding wheel surface morphology, etc. The average values ​​of these characteristics in each operation node are calculated to obtain multiple mean values ​​of grinding wheel wear characteristics. At the same time, the cooling performance of the coolant during continuous operation is measured, such as the change in the heat dissipation efficiency of the coolant, and the mean value of the decrease ratio relative to the initial performance at each continuous operation node is calculated. The acquisition of these data provides a solid foundation for subsequent analysis.

[0047] The obtained multiple grinding wheel wear characteristic mean values ​​and multiple cooling performance reduction ratio mean values ​​are carefully combined according to the corresponding operation nodes. For example, the grinding wheel wear characteristic mean value under a certain operation node is combined with the cooling performance reduction ratio mean value of the node to form a set of input data, and each set of input data uniquely corresponds to a single operation node. This combination can fully present the status of the two key factors of grinding wheel wear and coolant performance reduction at each operation node, providing comprehensive and accurate input information for the secondary compensation analysis plug-in.

[0048] After the carefully combined multiple sets of input data are input into the secondary compensation analysis plug-in, a series of complex and orderly analysis operations begin inside the plug-in. First, the plug-in extracts features from each set of input data, and deeply analyzes the information contained in the mean value of the grinding wheel wear feature and the mean value of the cooling performance reduction ratio. For the mean value of the grinding wheel wear feature, the plug-in will further analyze the degree of grinding wheel wear, wear mode, and potential impact on machining accuracy and surface quality; for the mean value of the cooling performance reduction ratio, the plug-in will evaluate in detail the effect of the reduced heat dissipation capacity of the coolant on thermal deformation, tool wear, etc. during the machining process. Then, based on the pre-set algorithm models, which are optimized through a large amount of historical data training and engineering experience, the plug-in performs a comprehensive analysis based on the extracted features. It uses machine learning algorithms (such as neural networks) to calculate the compensation coefficients that need to be adjusted for different machining parameters (grinding speed, workpiece speed, feed speed, cutting depth, and tool speed, etc.). For example, when severe wear of the grinding wheel is detected and the performance of the coolant is greatly reduced, in order to maintain machining accuracy and surface quality, the plug-in may calculate compensation coefficients for reducing the grinding speed and appropriately adjusting parameters such as feed speed and cutting depth to reduce the impact of increased wear of the grinding wheel on the workpiece, while considering the heat dissipation problem caused by the reduction in coolant performance to avoid thermal deformation affecting machining accuracy. After a complex calculation and optimization process, the plug-in outputs a predetermined set of secondary compensation coefficients for each set of input data. These coefficient sets accurately specify the amount of compensation adjustment that should be made for each machining parameter under the corresponding operation node, providing a key basis for the subsequent construction of a secondary compensation solution, ensuring that the CNC grinder can maintain stable machining performance in the case of grinding wheel wear and reduced coolant performance, improving product quality consistency, reducing scrap rates, and improving production efficiency.

[0049] A one-to-one mapping relationship between multiple operation nodes and multiple predetermined secondary compensation coefficient sets is established. Based on this mapping relationship, a complete secondary compensation scheme is constructed. In the actual processing process, when the CNC grinder runs to a specific operation node, the corresponding predetermined secondary compensation coefficient set is quickly and accurately obtained according to the mapping relationship, and the processing parameters are adjusted in real time. This ensures that during the entire continuous operation process, the CNC grinder can dynamically adjust the processing parameters according to the changes in grinding wheel wear and coolant performance, effectively maintain the consistency of the processed products, improve product quality and production efficiency, extend the service life of the grinding wheel, and reduce production costs.

[0050] In a possible implementation, step S350 further includes: Step S351: Collect a sample grinding wheel wear feature set and a sample cooling performance reduction ratio set, obtain a set of secondary compensation coefficients under different sample grinding wheel wear features and different sample cooling performance reduction ratios, and construct multiple sample secondary compensation coefficient sets.

[0051] Step S352: constructing a sample data set based on the sample grinding wheel wear feature set, the sample cooling performance degradation ratio set and multiple sample secondary compensation coefficient sets, and equally dividing the sample data set into M data sets.

[0052] Step S353: using the M data sets to perform supervised training on the BP network respectively, and when convergence is approaching, cross-validating the BP network by using the M data sets alternately to obtain M converged secondary compensation analysis branches.

[0053] Step S354: constructing the quadratic compensation analysis plug-in based on the M convergent quadratic compensation analysis branches, wherein the output of the quadratic compensation analysis plug-in is the average of the output results of the M convergent quadratic compensation analysis branches.

[0054] Specifically, through a variety of detection means and data collection methods, a wide range of sample grinding wheel wear feature sets and sample cooling performance reduction ratio sets are collected. For the sample grinding wheel wear features, high-precision measuring instruments can be used to monitor key indicators such as surface morphology changes, abrasive grain shedding, and reduction in grinding wheel size at different wear stages, and these data are summarized to form a sample grinding wheel wear feature set. At the same time, professional coolant performance testing equipment is used to measure the cooling efficiency reduction ratio and lubrication performance reduction degree of the coolant under different usage time or working conditions, and construct a sample cooling performance reduction ratio set. Then, for different combinations of sample grinding wheel wear features and sample cooling performance reduction ratios, based on a large amount of experimental data and actual processing experience, the corresponding secondary compensation coefficient sets are obtained. These coefficient sets clarify how to adjust the grinding speed, workpiece speed, feed speed, cutting depth, and tool speed and other processing parameters to ensure processing quality under specific wear and cooling performance conditions, and then construct multiple sample secondary compensation coefficient sets.

[0055] The carefully collected sample grinding wheel wear feature set, sample cooling performance reduction ratio set and the corresponding multiple sample secondary compensation coefficient sets are integrated. The sample grinding wheel wear feature set contains the characteristic data of the grinding wheel at different wear levels obtained by various advanced detection technologies, such as the change of grinding wheel surface texture, abrasive wear state, and reduction of grinding wheel diameter; the sample cooling performance reduction ratio set covers the reduction ratio data of key indicators such as cooling efficiency and lubrication performance of the coolant under long-term use or different working conditions relative to the initial state. These data are combined with the corresponding sample secondary compensation coefficient set (which contains the compensation adjustment amount of the processing parameters under different wear and cooling performance conditions) according to certain rules to construct a sample data set that comprehensively reflects the relationship between grinding wheel wear, cooling performance change and compensation coefficient. In order to ensure the comprehensiveness, accuracy and generalization ability of the subsequent training, this sample data set needs to be equally divided. The sample data set is evenly divided into M data sets, and each data set is as consistent as possible in terms of data size, feature distribution, etc. For example, if there are N sample data points in the sample data set, then each divided data set contains approximately N / M data points. In the division process, the diversity of sample data should be fully considered to ensure that each data set covers different degrees of grinding wheel wear characteristics, cooling performance degradation ratios, and corresponding compensation coefficients, so that each data set can be used independently for model training, while working together to improve the performance of the final model. This division method lays a solid foundation for the subsequent use of M data sets to supervise the BP network training, which helps to train a more accurate and reliable secondary compensation analysis model, thereby effectively responding to the problems of grinding wheel wear and coolant performance degradation during CNC grinding machine processing, and ensuring the quality stability and consistency of processed products.

[0056] For each data set, the sample grinding wheel wear characteristics and the sample cooling performance degradation ratio are used as the input of the BP network, and the corresponding sample quadratic compensation coefficient set is used as the expected output to start the supervised training process of the BP network. When the BP network is trained, the transmission results of the input data between the neurons in each layer are calculated by forward propagation to obtain the predicted quadratic compensation coefficient set. Then, the error between the predicted result and the expected output is calculated, and the error is back-propagated from the output layer to the input layer using the back propagation algorithm. The connection weights and thresholds between neurons are adjusted according to the error, and the parameters of the network are continuously optimized to reduce the prediction error. As the training continues, the error of the BP network model corresponding to each data set on the validation set gradually decreases and tends to converge, which indicates that the model has gradually learned the complex mapping relationship between the input data (grinding wheel wear characteristics and cooling performance degradation ratio) and the output data (quadratic compensation coefficient set). When each BP network model tends to converge, in order to further improve the stability and generalization ability of the model, the M data sets are interactively cross-validated, and each data set is used as a validation set in turn, and the remaining M-1 data sets are used as training sets to train and evaluate the BP network again. In this process, the network continuously adjusts its own parameters to adapt to the characteristics of different data sets, thereby further optimizing the model. After such a cross-validation process, each data set participated in the training and verification of the model, and finally M converged quadratic compensation analysis branches were obtained. Each branch is a fully trained and optimized BP network model, each of which has the ability to accurately predict the quadratic compensation coefficient set based on the wear characteristics of the grinding wheel and the ratio of cooling performance degradation, and can show good adaptability and stability when facing different data distributions, providing a solid foundation for the subsequent integration and construction of a high-precision quadratic compensation analysis plug-in.

[0057] In the intelligent control system of CNC grinders, the key step to achieve accurate compensation is to build a quadratic compensation analysis plug-in based on the integration of M convergent quadratic compensation analysis branches. After obtaining M convergent quadratic compensation analysis branches, the quadratic compensation analysis plug-in is constructed. In the architectural design of the plug-in, these M branches are integrated into an organic whole. When new grinding wheel wear characteristic mean value and cooling performance degradation ratio mean value data are input into the plug-in, the M convergent quadratic compensation analysis branches will process the input data at the same time. Each branch independently calculates and outputs the corresponding quadratic compensation coefficient set according to the complex mapping relationship between input and output learned during the training process. Then, the plug-in summarizes the output results of these M branches. By calculating the mean of these M output results, the final quadratic compensation coefficient set is obtained as the output of the plug-in. This integration method makes full use of the advantages of multiple branches and effectively reduces the errors and uncertainties that may exist in a single model. Since each branch uses a different data subset during the training process, there are certain differences in their understanding and processing of the data. The prediction information of multiple branches can be integrated by averaging, making the final output quadratic compensation coefficient set more accurate and reliable. This helps to more accurately deal with grinding wheel wear and coolant performance degradation during CNC grinding processing, and adjust processing parameters in real time according to actual working conditions, thereby ensuring the consistency and quality stability of processed products, improving production efficiency, reducing scrap rate, and improving the performance and reliability of the entire CNC grinding system.

[0058] Embodiment 2 is based on the same inventive concept as the previous embodiment - the intelligent control method of a CNC grinder combined with real-time feedback, such as Figure 2 As shown, the present application provides an intelligent control device for a CNC grinder combined with real-time feedback, and the device and method embodiments in the present application are based on the same inventive concept. The device includes: A static compensation coefficient set acquisition module 10 is used to collect historical processing records of several CNC grinders in a target factory, divide the several CNC grinders according to the historical processing records, determine multiple similar grinder domains, and perform processing deviation analysis on the multiple similar grinder domains respectively to obtain multiple static compensation coefficient sets.

[0059] The grinding control module 20 is used to determine initial processing parameters according to product feature matching, perform a primary compensation on the initial processing parameters based on the multiple static compensation coefficient sets, and control a plurality of CNC grinders to perform grinding according to the primary compensation parameter set.

[0060] The secondary compensation parameter set acquisition module 30 is used to perform fixed-point compensation on the primary compensation parameter set according to the secondary compensation scheme at a predetermined adjustment time node to obtain the secondary compensation parameter set, wherein the secondary compensation is used to compensate for the effects of grinding wheel wear and reduced cooling performance.

[0061] The secondary compensation parameter set control module 40 is used to control a plurality of CNC grinding machines to perform grinding processing according to the secondary compensation parameter set until the equipment stops continuous operation or the grinding wheel accessories are replaced.

[0062] Furthermore, the secondary compensation parameter set acquisition module 30 further includes: A cubic compensation coefficient set determination unit is used to collect parameter fluctuation ranges and processed product deviations of several CNC grinders at a predetermined monitoring time node, and perform parameter optimization analysis based on the parameter fluctuation ranges and processed product deviations to determine several cubic compensation coefficient sets.

[0063] The optimized processing parameter acquisition unit is used to dynamically fine-tune the secondary compensation parameter set according to the plurality of third-order compensation coefficients, obtain a plurality of optimized processing parameters, and control a plurality of CNC grinders to perform grinding operations.

[0064] Furthermore, the third compensation coefficient set determination unit further includes: A historical processing record query unit, which is used to query historical processing records, collect sample parameter fluctuation range sets, sample processing product deviation sets and multiple sample parameter compensation coefficient sets, wherein the processing parameters include grinding speed, workpiece speed, feed speed, cutting depth and tool speed, and each processing parameter corresponds to a parameter compensation coefficient.

[0065] A parameter optimization analyzer acquisition unit is used to take sample parameter fluctuation range and sample processing product deviation as input, supervise with sample parameter compensation coefficient, use the sample parameter fluctuation range set, sample processing product deviation set and multiple sample parameter compensation coefficient sets to train a machine learning model to obtain a parameter optimization analyzer.

[0066] A parameter optimization analysis unit is used to utilize the parameter optimization analyzer to perform parameter optimization analysis according to the parameter fluctuation range and the processed product deviation.

[0067] Furthermore, the static compensation coefficient set acquisition module 10 further includes: A surface roughness mean value acquisition unit is used to acquire, based on the historical processing records, a plurality of dimensional feature mean values ​​and a plurality of surface roughness mean values ​​of a plurality of processed products of a plurality of CNC grinders within a historical time window.

[0068] A plurality of clustering result acquisition units are used to perform multiple clustering on a plurality of CNC grinding machines based on the plurality of dimensional feature means and the plurality of surface roughness means under the constraint of the same processed product, so as to obtain a plurality of clustering results.

[0069] A plurality of similar grinder domain acquisition units are used to count the co-occurrence frequencies of any two CNC grinders in the plurality of clustering results, and add any two CNC grinders with frequencies greater than a predetermined frequency threshold to the similar grinder domain to obtain the plurality of similar grinder domains.

[0070] Furthermore, the static compensation coefficient set acquisition module 10 further includes: The first similar grinder domain selection unit is used to randomly select a first similar grinder domain, and obtain a mean value of size deviation and a mean value of surface roughness deviation of the first similar grinder domain within a historical time window according to the historical processing records.

[0071] The first static compensation coefficient set output unit is used to input the size deviation mean and the surface roughness deviation mean into a parameter static compensation model, output a first static compensation coefficient set, and add it to the multiple static compensation coefficient sets.

[0072] Furthermore, the secondary compensation parameter set acquisition module 30 further includes: A grinding wheel wear characteristic mean acquisition unit is used to query historical processing records based on the product characteristics as constraints, obtain multiple grinding wheel wear characteristic means at multiple continuous operation nodes when a number of CNC grinders are continuously operating, and multiple cooling performance reduction ratio means at multiple continuous operation nodes.

[0073] A plurality of input data acquisition units are provided, wherein the plurality of grinding wheel wear characteristic means and the plurality of cooling performance degradation ratio means are combined according to operation nodes to obtain a plurality of input data sets, wherein each input data set corresponds to an operation node.

[0074] A predetermined quadratic compensation coefficient set output unit is used to input the multiple groups of input data into the quadratic compensation analysis plug-in for analysis and output multiple predetermined quadratic compensation coefficient sets.

[0075] A secondary compensation scheme construction unit is used to establish a mapping relationship between multiple operating nodes and multiple predetermined secondary compensation coefficient sets, and construct the secondary compensation scheme according to the mapping relationship.

[0076] Furthermore, the predetermined secondary compensation coefficient set output unit further includes: A sample secondary compensation coefficient set construction unit is used to collect a sample grinding wheel wear feature set and a sample cooling performance degradation ratio set, and obtain secondary compensation coefficient sets under different sample grinding wheel wear features and different sample cooling performance degradation ratios, and construct multiple sample secondary compensation coefficient sets.

[0077] A sample data set construction unit is used to construct a sample data set based on the sample grinding wheel wear feature set, the sample cooling performance degradation ratio set and a plurality of sample secondary compensation coefficient sets, and to divide the sample data set into M data sets with equal values.

[0078] The convergent quadratic compensation analysis branch acquisition unit is used to use the M data sets to perform supervised training on the BP network respectively, and when convergence is approaching, the M data sets are used to cross-validate the BP network to obtain M convergent quadratic compensation analysis branches.

[0079] A quadratic compensation analysis plug-in construction unit, wherein the quadratic compensation analysis plug-in construction unit constructs the quadratic compensation analysis plug-in based on the M convergent quadratic compensation analysis branches, wherein the output of the quadratic compensation analysis plug-in is the average of the output results of the M convergent quadratic compensation analysis branches.

[0080] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0081] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0082] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. An intelligent control method for a CNC grinder combined with real-time feedback, characterized in that: Methods include: Collecting historical processing records of several CNC grinders in a target factory, dividing the several CNC grinders according to the historical processing records, determining a plurality of similar grinder domains, and performing processing deviation analysis on the plurality of similar grinder domains respectively to obtain a plurality of static compensation coefficient sets; Determine initial processing parameters according to product feature matching, perform primary compensation on the initial processing parameters based on the multiple static compensation coefficient sets, and control a plurality of CNC grinders to perform grinding processing according to the primary compensation parameter set; At a predetermined adjustment time node, the primary compensation parameter set is subjected to fixed-point compensation according to a secondary compensation scheme to obtain a secondary compensation parameter set, wherein the secondary compensation is used to compensate for the effects of grinding wheel wear and cooling performance degradation; According to the secondary compensation parameter set, a plurality of CNC grinding machines are controlled to perform grinding processing until the equipment stops continuous operation or the grinding wheel accessories are replaced.

2. The intelligent control method for CNC grinding machine combined with real-time feedback according to claim 1 is characterized in that: The method also includes: At a predetermined monitoring time node, parameter fluctuation ranges and processing product deviations of several CNC grinders are collected, and parameter optimization analysis is performed based on the parameter fluctuation ranges and processing product deviations to determine several sets of cubic compensation coefficients; The secondary compensation parameter set is dynamically fine-tuned according to the plurality of third-order compensation coefficients to obtain a plurality of optimized processing parameters, and a plurality of numerically controlled grinders are controlled to perform grinding operations.

3. The intelligent control method for CNC grinding machine combined with real-time feedback according to claim 2 is characterized in that: Perform parameter optimization analysis based on the parameter fluctuation range and processed product deviation, including: Query historical processing records, collect sample parameter fluctuation range sets, sample processing product deviation sets and multiple sample parameter compensation coefficient sets, where the processing parameters include grinding speed, workpiece speed, feed speed, cutting depth and tool speed, and each processing parameter corresponds to a parameter compensation coefficient; Taking the sample parameter fluctuation range and the sample processing product deviation as input and the sample parameter compensation coefficient as supervision, the machine learning model is trained using the sample parameter fluctuation range set, the sample processing product deviation set and multiple sample parameter compensation coefficient sets to obtain a parameter optimization analyzer; The parameter optimization analyzer is used to perform parameter optimization analysis according to the parameter fluctuation range and the processed product deviation.

4. The intelligent control method for CNC grinding machine combined with real-time feedback according to claim 3 is characterized in that: According to the historical processing records, a number of CNC grinding machines are divided to determine a plurality of similar grinding machine domains, including: According to the historical processing records, multiple dimensional feature averages and multiple surface roughness averages of multiple processed products of several CNC grinders within a historical time window are obtained; Taking the same processed product as a constraint, clustering a number of CNC grinding machines multiple times according to the multiple size feature means and the multiple surface roughness means to obtain multiple clustering results; The co-occurrence frequencies of any two CNC grinding machines in the multiple clustering results are counted, and any two CNC grinding machines with frequencies greater than a predetermined threshold are added to a similar grinding machine domain to obtain the multiple similar grinding machine domains.

5. The intelligent control method for CNC grinding machine combined with real-time feedback according to claim 4 is characterized in that: Perform machining deviation analysis on the multiple similar grinding machine domains respectively to obtain multiple static compensation coefficient sets, including: Randomly selecting a first similar grinding machine domain, and obtaining a mean value of a size deviation and a mean value of a surface roughness deviation of the first similar grinding machine domain within a historical time window according to the historical processing records; The dimension deviation mean value and the surface roughness deviation mean value are input into a parameter static compensation model, and a first static compensation coefficient set is output and added to the plurality of static compensation coefficient sets.

6. The intelligent control method for CNC grinding machine combined with real-time feedback according to claim 1 is characterized in that: Construct a secondary compensation plan, including: Taking the product characteristics as constraints, query historical processing records to obtain the average values ​​of multiple grinding wheel wear characteristics at multiple continuous operation nodes when a number of CNC grinders are continuously operating, and the average values ​​of multiple cooling performance reduction ratios at multiple continuous operation nodes; Combining the plurality of grinding wheel wear characteristic means and the plurality of cooling performance degradation ratio means according to the operation nodes to obtain a plurality of groups of input data, wherein each group of input data corresponds to one operation node; Inputting the plurality of groups of input data into a quadratic compensation analysis plug-in for analysis, and outputting a plurality of predetermined quadratic compensation coefficient sets; A mapping relationship between a plurality of operating nodes and a plurality of predetermined secondary compensation coefficient sets is established, and the secondary compensation scheme is constructed according to the mapping relationship.

7. The intelligent control method for CNC grinding machine combined with real-time feedback according to claim 6 is characterized in that: Build a quadratic compensation analysis plug-in, including: Collecting a sample grinding wheel wear feature set and a sample cooling performance reduction ratio set, and obtaining a set of secondary compensation coefficients under different sample grinding wheel wear features and different sample cooling performance reduction ratios, and constructing multiple sample secondary compensation coefficient sets; Constructing a sample data set based on the sample grinding wheel wear feature set, the sample cooling performance degradation ratio set and multiple sample secondary compensation coefficient sets, and dividing the sample data set into M data sets with equal values; Using the M data sets, respectively, supervise the BP network for training, and when convergence is approaching, cross-validate the BP network by using the M data sets alternately to obtain M convergent secondary compensation analysis branches; The quadratic compensation analysis plug-in is integrated and constructed based on the M convergent quadratic compensation analysis branches, wherein the output of the quadratic compensation analysis plug-in is the average of the output results of the M convergent quadratic compensation analysis branches.

8. The intelligent control device for CNC grinding machine combined with real-time feedback is characterized in that: The device is used to execute the intelligent control method for a CNC grinder combined with real-time feedback according to any one of claims 1 to 7, and the device comprises: A static compensation coefficient set acquisition module, which is used to collect historical processing records of several CNC grinders in a target factory, divide the several CNC grinders according to the historical processing records, determine multiple similar grinder domains, and perform processing deviation analysis on the multiple similar grinder domains respectively to obtain multiple static compensation coefficient sets; A grinding control module, the grinding control module is used to determine initial processing parameters according to product feature matching, perform a primary compensation on the initial processing parameters based on the multiple static compensation coefficient sets, and control a plurality of CNC grinding machines to perform grinding according to the primary compensation parameter set; A secondary compensation parameter set acquisition module, the secondary compensation parameter set acquisition module is used to perform fixed-point compensation on the primary compensation parameter set according to the secondary compensation scheme at a predetermined adjustment time node to obtain a secondary compensation parameter set, wherein the secondary compensation is used to compensate for the effects of grinding wheel wear and cooling performance degradation; A secondary compensation parameter set control module is used to control a plurality of CNC grinders to perform grinding processing according to the secondary compensation parameter set until the equipment stops continuous operation or the grinding wheel accessories are replaced.

Images