A numerical control grinding machine product quality real-time monitoring optimization method based on internet of things

Abstract

The present application relates to machine tool equipment technical field, especially to a kind of based on internet of things's numerical control grinding machine product quality real-time monitoring optimization method.The method includes the following steps: obtaining the multi-node real-time running data of numerical control grinding machine;According to the multi-node real-time running data statistics grinding machine key component working state parameter;Based on multi-node real-time running data and grinding machine key component working state parameter determines product machining quality fluctuation trend;According to product machining quality fluctuation trend detects the tailstock jacking force cooperative change law of grinding machine;According to the tailstock jacking force cooperative change law of grinding machine determines workpiece geometric precision limited trend.The control parameter adjustment behavior of numerical control grinding machine, system response state and quality detection result are associated and integrated, and the historical comparison system of adjusting behavior and response data is constructed, so that the tracking and effect evaluation of key control behavior are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of machine tool equipment, and in particular to a numerical control grinding machine product quality real-time monitoring optimization method based on the Internet of Things. BACKGROUND

[0002] A numerical control grinding machine is a high-precision and high-efficiency automatic machine tool used for machining various complex shapes and high-precision parts. Its main components include the bed and base, spindle components, worktable, grinding wheel frame, numerical control system, cooling system, and lubrication system. These components work together to realize the automation process from program input to grinding. The working principle of the numerical control grinding machine is to input the machining program through the numerical control system to control the movement of the spindle, worktable, and grinding wheel frame. The spindle rotates at high speed to drive the grinding wheel, the worktable moves along multiple axes to bring the workpiece to the machining position, and the grinding wheel frame moves along the feed direction to achieve grinding. The cooling system and lubrication system are used to remove the heat generated during grinding and reduce the friction between components, respectively, to ensure the stability of the machining process and the long-term operation of the machine tool. By deploying vibration sensors, temperature sensors, displacement sensors, etc. at key positions of the numerical control grinding machine, real-time data such as vibration, temperature, and displacement during equipment operation are collected. These data are transmitted to the monitoring center through industrial gateways and other communication devices. The data processing unit of the monitoring center uses time and frequency domain analysis algorithms to deeply analyze the collected data and identify fault characteristic frequencies and abnormal fluctuations. The system has self-learning ability and can continuously optimize the analysis model based on historical data and actual machining conditions to improve the accuracy and timeliness of fault diagnosis. Existing grinding machine monitoring systems mostly use independent data collection and processing methods, and there is a lack of effective data interaction and collaborative analysis capabilities between monitoring nodes. This scattered data situation makes it difficult to fully exploit and utilize the large amount of valuable information generated during equipment operation, limiting the intelligent level of the quality monitoring system. SUMMARY

[0003] Therefore, it is necessary to provide a numerical control grinding machine product quality real-time monitoring optimization method based on the Internet of Things to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a numerical control grinding machine product quality real-time monitoring optimization method based on the Internet of Things includes the following steps:

[0005] Step S1: Obtain multi-node real-time running data of the numerical control grinding machine; according to the multi-node real-time running data, count the working state parameters of the key components of the grinding machine; based on the multi-node real-time running data and the working state parameters of the key components of the grinding machine, determine the product machining quality fluctuation trend;

[0006] Step S2: detecting the cooperative change law of the grinder tailstock clamping force according to the product processing quality fluctuation trend; determining the workpiece geometric precision limited trend according to the grinder tailstock clamping force cooperative change law; determining the grinder processing system linkage disorder characteristics according to the grinder tailstock clamping force cooperative change law and the workpiece geometric precision limited trend;

[0007] Step S3: determining the product quality stability deviation condition based on the grinder processing system linkage disorder characteristics; determining the grinder control parameter deviation degree according to the product quality stability deviation condition and the grinder processing system linkage disorder characteristics; performing real-time adjustment processing on the grinder control parameter deviation degree based on the Internet of Things control system, and obtaining the grinder parameter adjustment execution situation;

[0008] Step S4: evaluating the response effect of the Internet of Things control system based on the grinder parameter adjustment execution situation, and obtaining the Internet of Things control system response defect data; performing grinder quality monitoring system control optimization processing according to the Internet of Things control system response defect data, and obtaining the optimized grinder product quality real-time monitoring system.

[0009] The application is aimed at the problems existing in the actual operation of the traditional numerical control grinding machine system, such as lagging adjustment response, large parameter execution deviation, weak system feedback mechanism, etc. Through the construction of fine Internet of Things control logic and closed-loop optimization process, the depth collaboration and dynamic adaptation of the grinding machine quality monitoring and control system are realized. First, by extracting the grinding machine operation state log and control parameter adjustment record, a historical comparison table of adjustment behavior and system response is established, which enables the system to truly reproduce the causal relationship between various adjustment behaviors and results, providing a high-credibility data basis for subsequent data analysis and strategy optimization. This mapping mechanism effectively makes up for the problem of relying solely on result analysis while ignoring the details of adjustment behavior in the past, improving the system's understanding of the background and results of adjustment actions. Based on the history of adjustment behavior, the system further constructs a control parameter execution tag library based on parameter type, response effectiveness level and execution time. This tag library not only helps the system to structure the historical behavior, but also can be used for sample input of subsequent machine learning models, with high generalization ability and scalability. Through the analysis of the tag library, the system can accurately identify parameters with frequent abnormal response or high adjustment deviation rate, and classify them as optimization items, so that the triggering of optimization strategies no longer depends on artificial experience judgment, but relies on systematic data insight, which greatly improves the accuracy and efficiency of identification. For the identified optimization parameter items, the system establishes a scientific parameter priority division standard through the two dimensions of the number of execution failures and the average adjustment deviation value, so that limited optimization resources can be prioritized to control elements that have the greatest impact on system quality, thereby improving the overall control effect. In high-priority parameter processing, the adjustment coefficient is set by the average deviation value, which essentially realizes the dynamic adjustment of the adjustment sensitivity. This adaptive adjustment method avoids the problems of overcorrection or reaction delay, making the adjustment result more stable and accurate. For example, for parameters with large average deviation values, the system significantly narrows the adjustment range to prevent system overshoot or system oscillation, improving the robustness and safety of the system. In addition, the medium-priority parameter items optimize the response time without adjusting the control value, avoiding interference with relatively stable parameters. Shortening the response time can improve the timeliness of the system's control of medium-important parameters and reduce the cumulative impact of response lag, further optimizing the overall dynamic response capability of the grinding machine. Through the hierarchical optimization strategy, the system realizes the differentiated processing of different types of parameter items, balancing adjustment accuracy and response speed, and showing higher adaptive control capability. The optimization results are integrated into a unified control parameter configuration file and uploaded to the system parameter storage module through the Internet of Things communication interface, ensuring that all optimization configurations can take effect quickly after system restart, realizing the automation and efficiency of parameter update. Through the soft restart mechanism, the system quickly loads new parameters without affecting the hardware structure, ensuring the running continuity and stability of the grinding machine quality monitoring system.Meanwhile, the system automatically verifies after the optimization parameter configuration is online, starts a new round of real-time monitoring process, reanalyzes and evaluates the adjustment results, and builds a complete closed-loop feedback mechanism. This mechanism can continuously track the optimization effect and incorporate it into the subsequent historical comparison table and tag library, forming an evolutionary path of optimization-execution-verification-reoptimization. Overall, this method realizes the transformation of grinding machine control parameters from static setting to dynamic optimization by integrating a series of steps such as historical behavior backtracking, tag induction, priority classification processing, adaptive parameter adjustment, response time refinement control, and automatic configuration replacement. The introduction of Internet of Things technology greatly improves the automation level of data acquisition, communication, and execution configuration, reduces the uncertainty caused by human intervention, and significantly improves the adaptability, stability, and adjustment efficiency of CNC grinding machines in complex production environments. Especially in manufacturing scenarios with frequent switching of multiple product types and high-precision machining tasks, this method can automatically adjust key control parameters based on the grinding machine's own running data, effectively ensuring the consistency of product processing quality and equipment operation efficiency, and providing strong technical support for intelligent operation and maintenance of CNC equipment in intelligent manufacturing scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0010] Other features, objects, and advantages of the present application will become more apparent from the following detailed description of non-limiting embodiments made with reference to the drawings:

[0011] Figure 1 Step flowchart of the present application based on the Internet of Things CNC grinding machine product quality real-time monitoring optimization method;

[0012] Figure 2 For Figure 1 Detailed step flowchart of step S1 in the present application;

[0013] Figure 3 For Figure 1 Detailed step flowchart of step S3 in the present application. DETAILED DESCRIPTION

[0014] The technical method of the present application will be described in detail below with reference to the drawings. Obviously, the described embodiments are part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0015] Furthermore, the accompanying drawings are included to provide a further understanding of the present application, and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments of the present application and, together with the description, serve to explain the principles of the present application. In the drawings:

[0016] It is to be understood that, although terms such as "first", "second", and so on can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, without departing from the scope of the exemplary embodiments, a first element can be referred to as a second element, and similarly a second element can be referred to as a first element. The term "and / or" as used herein includes any and all combinations of one or more of the associated associated items.

[0017] To achieve the above object, there is provided Figures 1 to 3 The present application provides a numerical control grinding machine product quality real-time monitoring optimization method based on Internet of Things, the method comprises the following steps:

[0018] Step S1: acquiring multi-node real-time running data of the numerical control grinding machine; according to the multi-node real-time running data, the working state parameters of the key components of the grinding machine are counted; based on the multi-node real-time running data and the working state parameters of the key components of the grinding machine, the product processing quality fluctuation trend is determined;

[0019] In the embodiment of the present application, a numerical control cylindrical grinding machine equipped with an industrial Internet of Things collection module is selected to collect real-time running data of multiple key nodes. The nodes include spindle unit, bed guide rail, tail seat clamping mechanism, grinding wheel feeding mechanism and cooling system, etc. Through the configuration of an industrial Ethernet gateway, various sensors (such as vibration sensors, temperature sensors, position encoders, pressure sensors, current / voltage collection modules) are connected to the data collection system, and the running state data is acquired every 0.5 seconds. Subsequently, based on the time window statistical method, the axial clamping pressure of the tail seat clamping device, the stability of the grinding wheel spindle speed, and the fluctuation amplitude of the cooling liquid flow rate are extracted as state parameters, and normalized processing is performed. By introducing a time sequence correlation analysis model (such as the weighted trend mean and standard deviation comparison method in the sliding window), the fluctuation of each component parameter in the process of machining the same type of parts is analyzed to determine whether the geometric precision (such as roundness, cylindricity) and surface roughness of the batch of products exist periodic changes, so as to extract the fluctuation trend characteristics of the product processing quality, and provide data support for subsequent analysis of system linkage abnormalities.

[0020] Step S2: detecting the cooperative change law of the grinder tailstock clamping force according to the product processing quality fluctuation trend; determining the workpiece geometric precision limitation trend according to the grinder tailstock clamping force cooperative change law; determining the grinder processing system linkage disorder feature according to the grinder tailstock clamping force cooperative change law and the workpiece geometric precision limitation trend;

[0021] The embodiment of the application is based on the processing quality fluctuation trend data, and the grinder tailstock clamping force cooperative change law is analyzed by modeling in this step. By analyzing the continuous output data sequence of the tailstock clamping force sensor, the weak disturbance mode of the workpiece clamping state at different time periods is identified when the same workpiece is processed, and the spindle current change and the grinding wheel displacement compensation trend in the grinding process are combined. Using the dynamic time warping (DTW) and principal component cooperative analysis method, the coupling relationship between the clamping force and other component parameters is identified, and the synchronization feature of the change is extracted. For example, when the clamping force periodically decreases after processing 4 products, and at the same time, the workpiece tail end roundness precision deteriorates, it can be judged that the workpiece geometric precision limitation trend appears. On this basis, the fuzzy clustering method is further used to analyze the linkage decoupling of the multi-node parameters, to identify whether the grinder processing system has linkage disorder performance, such as the unsynchronized response of the “tailstock clamping-main shaft stiffness-grinding wheel feeding” three, the unsynchronized disturbance frequency and the like, and to establish the corresponding linkage disorder feature label, such as “cooperative index decrease by more than 20%” as the determination standard, to provide the basis for the subsequent stability control deviation identification.

[0022] Step S3: determining the product quality stability deviation condition based on the grinder processing system linkage disorder feature; determining the grinder control parameter deviation degree according to the product quality stability deviation condition and the grinder processing system linkage disorder feature; performing real-time adjustment processing on the grinder control parameter deviation degree based on the Internet of Things control system, to obtain the grinder parameter adjustment execution situation;

[0023] The embodiment of the application further combines the surface quality data and geometric error trend of the machining batch product at different time periods to construct a product quality stability deviation identification model by using the identified linkage imbalance features. The empirical mode decomposition method (EMD) is used to perform feature decomposition on the workpiece cylindricity error sequence to judge the correspondence between the deviation trend and the linkage imbalance features. For example, it is found that the cylindricity error is generally higher in the tail section than in the head section and has a synchronous relationship with the tail seat clamping fluctuation, which can preliminarily judge that the stability deviation is caused by the tail seat imbalance. Further, the control parameter deviation degree is calculated by combining the deviation between the control system set parameters and the real-time acquisition values. For example, the tail seat pneumatic clamping pressure set value is 0.6 MPa, and the actual fluctuation range is 0.55 MPa to 0.58 MPa, and the deviation rate is more than 3%, which is recorded as "moderate deviation". The PID closed-loop adjustment of the deviation parameters is performed in real time by deploying the industrial internet of things control system, and the adjustment results include the key indicators such as adjustment response time, adjustment stability time, error convergence amplitude, etc., and are uploaded to the cloud database. The system records a complete parameter adjustment execution situation, such as "clamping force compensation 0.02 MPa, response time 3.2 seconds, and adjustment stability is good", which lays a foundation for system response performance evaluation.

[0024] Step S4: perform internet of things control system response effect evaluation based on the grinding machine parameter adjustment execution situation to obtain internet of things control system response defect data; perform grinding machine quality monitoring system control optimization processing according to the internet of things control system response defect data to obtain an optimized grinding machine product quality real-time monitoring system.

[0025] The embodiment of the application constructs an internet of things control system response effect evaluation model based on the grinding machine parameter adjustment execution situation. By setting response defect evaluation indexes, including whether the response lag time exceeds the threshold value, whether the machining quality improvement amplitude after adjustment meets the standard, whether the control system intervention frequency is too high, etc., the fuzzy comprehensive evaluation method is used to perform weighted calculation on each index to extract system response defect data. For example, if the system adjustment response time of a batch is more than 5 seconds on average, and the machining precision improvement is less than 10%, it is recorded as "response efficiency insufficient type defect". Then, the response defect data is analyzed by cause and effect diagram and the control strategy is reconstructed, such as using an adaptive parameter preloading model, introducing a disturbance feedforward suppression strategy, etc., to perform online optimization on the PID parameters or feedback gain matrix in the control system to obtain an optimized grinding machine product quality real-time monitoring system. The system supports dynamic self-learning function, which can quickly adjust the control strategy according to historical data when machining new type products, and real-time mark the key response node state. The specific application scenario is the grinding scene of high-precision aero-engine shaft parts, and under the support of the optimization system, the yield rate is improved from 94.6% to 98.1%, the number of grinding process quality complaints is reduced by more than 70%, and the manufacturing quality stability and intelligent regulation and control capability are significantly improved.

[0026] Preferably, step S1 comprises the following steps:

[0027] Step S11: synchronously collect data through the sensor network installed on the tailstock, spindle box, transverse feed mechanism, grinding wheel frame and workbench to obtain multi-node real-time operation data, wherein the multi-node real-time operation data includes tailstock clamping force data, spindle speed data, feed position data, grinding wheel vibration data and workbench temperature data;

[0028] In order to realize synchronous collection of multi-node real-time operation data of the numerical control grinding machine, an industrial sensor network is arranged in the structure of the grinding machine, including: a micro pneumatic pressure sensor is installed on the tailstock to measure the real-time clamping force of the clamping cylinder with an accuracy of 0.01 MPa, a high-speed rotary encoder and an axle current sensor are embedded in the spindle box to obtain the spindle speed and current load change, the sampling accuracy of the speed reaches ±5 revolutions per minute, a grating ruler and a servo drive feedback sampling device are installed in the transverse feed mechanism to collect the X-axis feed displacement and feed instruction deviation, a three-axis micro vibration sensor is arranged on the grinding wheel frame to record the vibration intensity of the grinding wheel during grinding, the frequency response range is 10 Hz to 10 kHz, and a thermocouple temperature sensor array is pasted on the bottom of the workbench to monitor the temperature rise change and thermal expansion trend. All sensors are connected with the field industrial gateway through the Modbus-TCP protocol, the collection cycle is set to 5 times per second, the collected data is automatically uploaded to the edge processing platform, and multi-node real-time operation data including tailstock clamping force data, spindle speed data, feed position data, grinding wheel vibration data and workbench temperature data is generated, which is used for subsequent processing process feature recognition and quality trend analysis. The collection system has been deployed on the grinding process of a certain aerospace high-precision shaft part, and the collected data is used to detect the failure of thermal deformation compensation.

[0029] Step S12: according to the multi-node real-time operation data, the working state parameters of the key components of the grinding machine are counted;

[0030] In step S11, the multi-node real-time running data is obtained, and in this step, the running state parameters of the key components of the grinding machine are counted and classified, and a parameter database is established by using state feature extraction and grade division method. In the specific implementation process, first, the mean value, range and standard deviation of each type of sensor data in the sliding time window are statistically processed, for example, the tailstock clamping force data is taken as a statistical window of 10 seconds, the average clamping pressure, maximum fluctuation amplitude and the like in the window are calculated, and compared with the historical stable value. When the parameters of a certain component deviate from its stable running interval for a long time, for example, the average value of the grinding wheel vibration intensity is continuously higher than 2.0g (unit of gravitational acceleration), the state of the component is marked as “abnormal vibration”. All key components such as tailstock, spindle, grinding wheel frame, feed mechanism and the like generate “state grade labels” in turn, for example, “tailstock clamping state: normal”, “spindle speed state: slight fluctuation”, “grinding wheel vibration state: severe abnormality”, thereby forming a “key component state parameter set” in the current time period, which is used as the basis data for subsequent judgment of collaborative deviation and system fluctuation analysis. This method has been applied in the maintenance data system of multiple models to realize the digital modeling and running classification of the grinding machine state.

[0031] Step S13: calculating the data fluctuation intensity based on the time sequence change of each node data in the multi-node real-time running data; and calculating the component collaborative deviation based on the change trend of the state grade of each component in the working state parameters of the key components of the grinding machine;

[0032] In the embodiment of the application, the time sequence of the original sensor data collected in step S11 and the key component state parameters extracted in step S12 are used to analyze the dynamic response characteristics in the machining process. First, the time sequence fluctuation intensity analysis model is used to calculate the fluctuation intensity of each sensor data in a certain period, that is, the ratio of the difference between the maximum value and the minimum value to the average value in a fixed time window is taken as the fluctuation intensity index of the node. For example, the fluctuation intensity of the grinding wheel vibration data is 0.85, which indicates that the response of this part is intense. Secondly, the running consistency between multiple components is analyzed in combination with the historical change trend of the state grade. The “component collaborative deviation” index is introduced to measure whether the state grade of the key components changes synchronously, for example, in a certain batch of machining, the state of the spindle is marked as “normal” for 5 consecutive periods, while the tailstock clamping state frequently switches between “mild deviation” and “normal”, which indicates that there is a deviation in the collaborative control of the two. By statistically processing the difference coefficient of the state grade between all key components, the degree of collaborative deviation is quantified, and the greater the difference, the worse the collaboration. This index provides a basis for subsequent product quality trend analysis. Application examples show that this method can effectively find the problem of size consistency decline caused by the slight loosening of the tailstock in the grinding of precision parts such as hydraulic valve core.

[0033] Step S14: correlation analysis is performed based on the data fluctuation intensity and the component synergy deviation, so as to obtain the fluctuation trend of the product processing quality, wherein the correlation analysis is specifically that when the data fluctuation intensity exceeds the set threshold value and the component synergy deviation presents an increasing trend, it is determined that the product processing quality has an unstable fluctuation trend, otherwise it is determined that the product processing quality maintains a stable trend.

[0034] After the fluctuation intensity and the synergy deviation are analyzed, the correlation between the two is determined in this step, so as to determine whether the product processing quality has a fluctuation trend. The specific method is to establish a "data fluctuation-synergy-quality trend" three-factor correlation analysis model, and set multiple empirical threshold parameters: when the node fluctuation intensity index exceeds 0.7, and the synergy deviation value of at least two key components is higher than 0.5, the system automatically determines that the current processing state is "quality fluctuation unstable"; otherwise, it is determined as "quality stable". Taking the machining of a certain aero-engine main shaft part as an example, in the continuous machining process of 30 products, it is recorded that the grinding wheel vibration intensity value rises to 1.1, and the synergy deviation of the tailstock and the main shaft increases. The analysis model combines the historical label comparison of the machining data to determine that the cylindricity deviation of the product at this time has approached the upper limit of the process tolerance, triggering the early warning mechanism. Finally, the analysis result is fed back to the Internet of Things platform control end, so as to guide whether to carry out real-time adjustment or planned shutdown maintenance, and to improve the product consistency and system response sensitivity. This method is especially suitable for online processing quality control of precision workpieces that need to maintain high dimensional consistency, such as hydraulic, aeronautical and precision instrument manufacturing fields.

[0035] Preferably, step S12 comprises the following steps:

[0036] Step S121: calculating the tailstock working load mean value and load fluctuation coefficient according to the tailstock clamping force data, so as to determine the tailstock component working state level;

[0037] The tailstock top pressing force data is collected in real time, and the data is further processed to determine the working state level of the tailstock. First, a fixed time analysis window (for example, every 60 seconds) is set, and the average value of all top pressing force data in the window is calculated to obtain the tailstock working load average, which is used to measure whether the actual clamping capacity of the tailstock clamping system in the machining process is stable; at the same time, the ratio of the standard deviation of the top pressing force data in the window to the average value is calculated as the "load fluctuation coefficient", which is used to judge whether there is instability caused by compression gas source fluctuation or mechanical gap in the tailstock clamping process. Taking a certain numerical control grinding machine as an example, the average value of the top pressing force in the normal working state should be 0.45 MPa, and the fluctuation coefficient should be less than 5%; if the average value is detected to be lower than 0.4 MPa or the fluctuation coefficient exceeds 10%, the system automatically marks the tailstock state level as "abnormal" or "critical". Finally, the working state level of the tailstock component is divided into "normal", "mild deviation" and "serious abnormality" according to the working condition standard, and the result will be used for subsequent state parameter combination modeling.

[0038] Step S122: calculating spindle running stability and speed consistency index based on spindle speed data to determine the working state level of the spindle component;

[0039] The embodiment of the application extracts "running stability" and "speed consistency index" by analyzing the spindle speed time series data based on the spindle speed data. Running stability is used to reflect whether the spindle exists jitter or load fluctuation, which can be represented by the ratio of the difference between the maximum and minimum values of the speed in the sliding time window to the target set speed; the speed consistency index is used to measure whether the speed control of the spindle is uniform during the machining of batch parts, and the calculation method is the ratio of the standard deviation of the average speed in different workpiece machining periods to the target value. In actual application, if the spindle set speed is 3000 rpm, the stability fluctuation is 1.5% during machining in a certain section, and the consistency index is controlled within 0.8%, then the spindle running state can be evaluated as "stable", if both of them exceed 3%, then it is "abnormal" state. The above state level is written into the edge computing module through model rules, which can determine the working state level of the spindle in real time and provide support for the overall grinding machine quality control.

[0040] Step S123: determining the working state level of the feed system component according to the feed position data feed precision deviation value and the feed speed uniformity index;

[0041] In this embodiment of the invention, the status judgment of the feed system relies on the feed position data collected by the transverse feed mechanism, combined with command and actual feedback for deviation analysis. This step extracts the difference between the "target position" and the "actual reached position" in each feed path, and calculates the average error over a certain time window as the "feed accuracy deviation value." Furthermore, the rate of change of speed at each sampling point during continuous feed is used to calculate the "feed speed uniformity index," with a lower index indicating more uniform feed. For a certain model of CNC grinding machine, under a feed accuracy requirement of 0.001mm, the acceptable operating condition is a feed deviation not exceeding ±0.002mm and a speed uniformity index controlled within 0.05. If data analysis of a batch reveals a feed error of 0.005mm and frequent speed fluctuations, the feed system status level is set to "slight deviation" or "severe deviation" to guide equipment maintenance and servo calibration, thereby preventing dimensional deviations caused by abnormal feed.

[0042] Step S124: Calculate the working balance and vibration intensity index of the grinding wheel based on the grinding wheel vibration data, thereby obtaining the working status level of the grinding wheel system components;

[0043] This invention analyzes grinding wheel vibration data to evaluate its "working balance" and "vibration intensity index." First, the triaxial vibration data is synthesized into a total vibration vector, and its effective value (RMS) is calculated as the "vibration intensity index." Then, the balance of vibration amplitudes in the X, Y, and Z directions is compared to determine whether the grinding wheel is unbalanced due to eccentricity, wear, or improper clamping, i.e., "working balance." If the three-axis vibration amplitudes differ significantly (e.g., the Z-axis amplitude is more than twice that of the X-axis amplitude), it indicates a directional deviation and poor working balance. Taking a heavy-duty grinding wheel as an example, an RMS value below 0.3g is considered normal, while a value exceeding 0.8g indicates severe vibration; an imbalance coefficient greater than 0.4 indicates eccentric clamping. Based on this, the system classifies the grinding wheel's condition and issues maintenance warnings such as "needs dressing" or "needs dynamic balancing" to prevent excessive surface roughness during machining.

[0044] Step S125: Use the workbench temperature data to assess the risk of thermal deformation and temperature stability indicators, thereby determining the operating status level of the thermal control system components;

[0045] The embodiment of the present application utilizes the temperature sensor data arranged by the workbench to perform state evaluation of the thermal control system. The specific method comprises: counting the temperature rising rate of each temperature collection point in the processing start-to-end cycle, and calculating the maximum temperature difference as a "temperature stability index"; at the same time, the thermal deformation risk is evaluated through an empirical model, for example, when the temperature rise of the table surface exceeds 15℃ and the temperature difference of different points exceeds 8℃, it is considered that there is a "high thermal deformation risk". In addition, whether the temperature change of the heat-sensitive area (such as the contact surface of the main guide rail and the workpiece) tends to be stable within 5 minutes is judged to determine whether the system has the "rapid thermal equilibrium" capability. If the temperature rise trend is slow and the temperature difference is less than 5℃ in a certain processing, the state level is "good thermal control", otherwise, it is prompted that "thermal control is unstable", and cooling or shutdown needs to be started. The state level is well applied in the fine grinding process of the aero-engine parts.

[0046] Step S126: combining the tailstock component working state level, the spindle component working state level, the feeding system component working state level, the grinding wheel system component working state level, and the thermal control system component working state level to form the grinding machine key component working state parameter.

[0047] After the aforementioned five component state level determinations are completed, the embodiment of the present application combines these state levels to form a "grinding machine key component working state parameter". The system passes through a state fusion engine, and "tailstock state level", "spindle state level", "feeding system state level", "grinding wheel system state level", and "thermal control system state level" correspond to one item in the five-tuple respectively, and records the time stamp, processing task number, workpiece number and other information, to generate a labeled state parameter set. For example, the state parameter combination at a certain moment is [normal, normal, slight deviation, serious abnormality, good thermal control], which represents that the grinding wheel vibration problem is the main quality hidden danger in this cycle. The combination data will be used as the core input variable for calculating the component coordination deviation and establishing the fluctuation prediction model in the subsequent steps, and is used for chart display of the data visualization system, so that the equipment maintenance personnel can quickly locate the key failure node and take process intervention measures in advance.

[0048] Preferably, the detection of the grinding machine tailstock clamping force coordination change rule according to the product processing quality fluctuation trend in step S2 comprises:

[0049] According to the product processing quality fluctuation trend, the numerical sequence of the tailstock clamping force data in the time window is extracted, the clamping force numerical difference between the continuous time points is calculated, and the clamping force change gradient is obtained;

[0050] The mean and standard deviation of the clamping force change gradient are calculated, so as to determine the tailstock clamping force reference change state;

[0051] The time stamp corresponding relationship between the tailstock chucking force and the spindle speed is extracted based on the tailstock chucking force reference change state, and the numerical value ratio change trend of the two groups of data at the same time point is calculated;

[0052] The synchronization proportion is obtained by analyzing the increase and decrease synchronization of the tailstock chucking force and the feeding position in the tailstock chucking force reference change state through a sliding time window, and counting the proportion of the same direction change times and the reverse direction change times.

[0053] The response delay time is determined according to the numerical value ratio change trend and the synchronization proportion, so as to detect the tailstock chucking force cooperative change rule of the grinding machine.

[0054] According to the quality fluctuation trend in the grinding machine product processing process, such as size tolerance change or surface roughness abnormality, a corresponding time window range is selected, for example, when surface quality fluctuation is found to be larger during the processing of a certain batch of parts, the original force value data collected by the tailstock clamping force sensor in the time period is extracted to form a clamping force value sequence arranged in time sequence. Then, the clamping force difference between two consecutive time points in the sequence is calculated to obtain the clamping force change gradient in each time period, so as to reflect the change degree and fluctuation characteristics of the force value of the tailstock during continuous clamping. The "clamping force" referred to here is the force applied by the tailstock to clamp the workpiece to prevent axial movement, and the unit is usually Newton (N), and the change gradient represents the variation amplitude of the force value in the time dimension. After obtaining the clamping force change gradient sequence, statistical analysis is performed on the gradient sequence in the entire time window, specifically including calculating the mean and standard deviation, wherein the mean reflects the average level of the tailstock clamping force fluctuation in the time period, and the standard deviation represents the dispersion degree or instability degree of the fluctuation, and the two statistical values together constitute the "reference change state" of the tailstock clamping force, which is used as a reference standard for subsequent judgment of the cooperative response relationship. If the mean is stable but the standard deviation is large, it indicates that there is frequent fluctuation problem in the tailstock clamping control, which may affect the clamping stability of the workpiece and further cause the processing quality fluctuation. After obtaining the reference change state, the spindle speed data with the same time stamp as the tailstock clamping force data is extracted, that is, at each time point when the tailstock clamping force is recorded, the current speed value recorded by the spindle speed sensor is obtained. By pairing and calculating the ratio of the values of each pair of clamping force and speed data at the same time point, and constructing the ratio change curve with time evolution, whether the tailstock clamping action remains consistent or there is lag under different spindle speeds is evaluated. For example, when the spindle speed increases, whether the clamping force increases in time to resist the increasing centrifugal force and cutting reaction force, the change trend of the ratio can be used to judge whether the cooperation is normal. On the basis of the above, in order to further analyze the response delay problem of the tailstock clamping force to the processing action, the sliding time window technology is introduced, that is, in a time sliding window of a certain length, the relative relationship between the tailstock clamping force change trend and the feed position change trend in the feed system is observed, and the synchronization of the two is judged. In each sliding window, the number of times that the clamping force and the feed position increase (or decrease) simultaneously or change in opposite directions is counted, the number of times of the same direction and opposite direction changes in the entire time period is accumulated, and the proportion is calculated to obtain the synchronization proportion value. The higher the synchronization proportion is, the closer the cooperation between the tailstock and the feed system is, which is beneficial to improving the processing stability. Finally, the numerical ratio change trend and the synchronization proportion are comprehensively analyzed, and the trend change point observed in the sliding window is combined to quantify the average delay time of the tailstock in response to the spindle and feed action change, that is, the "response delay time". The shorter the response delay time is, the more timely the system reacts, and the more high-precision cooperative processing can be realized.Through such collaborative law detection on the processing process of multiple batches of products, an optimization model can be further established to realize adaptive adjustment of the grinding machine tailstock control parameters based on the Internet of Things, and improve the real-time and intelligent level of overall product quality control.

[0055] Preferably, the determination of the workpiece geometric precision limitation trend in step S2 according to the grinding machine tailstock clamping force collaborative change law comprises:

[0056] According to the grinding machine tailstock clamping force collaborative change law, a continuous clamping force numerical sequence is extracted, and the local maximum value in the numerical sequence is identified as a peak value and the local minimum value is identified as a valley value;

[0057] Based on the identified peak value and valley value, the numerical difference between the peak value and the valley value is calculated, and the maximum numerical difference is recorded as the clamping force fluctuation amplitude range;

[0058] When the workpiece diameter is 10mm-200mm, a proportionality coefficient is generated based on the clamping force fluctuation amplitude range and the workpiece diameter;

[0059] According to the proportionality coefficient, a radial runout prediction value is calculated, when the proportionality coefficient is between 0.001-0.01, then the radial runout prediction value = proportionality coefficient x workpiece diameter x 0.5; when the proportionality coefficient is greater than 0.01, then the radial runout prediction value = proportionality coefficient x workpiece diameter x 0.8;

[0060] Based on the response delay time in the grinding machine tailstock clamping force collaborative change law and the current feed speed, a numerical multiplication operation is performed to obtain a workpiece axial displacement theoretical deviation;

[0061] According to the workpiece axial displacement theoretical deviation, a maximum deviation of the workpiece axial displacement is calculated;

[0062] The radial runout prediction value is compared with the preset radial runout allowable value, when the prediction value is greater than 80% of the allowable value, the radial precision risk level is marked as pre-warning;

[0063] The maximum deviation of the axial displacement is compared with the preset axial displacement allowable value, when the deviation is greater than 80% of the allowable value, the axial precision risk level is marked as pre-warning;

[0064] According to the combination state of the radial precision risk level and the axial precision risk level, the workpiece geometric precision limitation trend is determined, when the radial precision risk level and the axial precision risk level are both pre-warning, the geometric precision limitation trend is determined as serious limitation; when only one level is pre-warning, it is determined as slight limitation, and when both levels are normal, it is determined as stable precision.

[0065] On the basis of the established tailstock cap force collaborative variation law of the grinding machine, the embodiment of the application further extracts the continuous cap force numerical sequence in the grinding process, which is collected by the Internet of Things pressure sensor in real time, and the sampling frequency is 100 times per second to ensure the details of capturing the force value changes. In the numerical sequence, the peak value and the valley value in each local time period are identified by using the sliding window extreme value detection algorithm, that is, when the cap force value at a certain time is greater than the values of several time points before and after it, it is recorded as a peak value, and vice versa. This operation can effectively identify the periodic fluctuation law of the tailstock cap force in the grinding process, in which the peak value often corresponds to the tailstock just completing a clamping process, and the valley value may represent the lowest clamping state after the tailstock force is released or the feed disturbance. After completing the peak and valley value extraction, the numerical difference between each pair of adjacent peak and valley values is calculated to obtain the amplitude of single force fluctuation. By comparing all the fluctuation difference values, the maximum difference value is selected as the "cap force fluctuation amplitude range" in this time period. This parameter reflects the most severe clamping instability behavior that the tailstock may produce in the grinding process, and is an important basis for subsequent precision prediction. For example, if the maximum cap force fluctuation is identified as 35N during processing, it means that the minimum and maximum difference of the clamping force on the workpiece reaches this value, which may cause significant positioning deviation or deformation. When processing workpieces of different diameters, the stability of the tailstock clamping force has different degrees of influence on the geometric precision of the product. Therefore, when the system identifies that the current processed workpiece diameter is within the range of 10mm to 200mm, the "proportion coefficient" is generated by ratio calculation of the aforementioned cap force fluctuation amplitude range and the workpiece diameter, which is used to normalize the influence intensity of the tailstock force fluctuation. The proportion coefficient is a dimensionless value, and its physical meaning is the instability degree of the clamping force on the unit diameter workpiece. For example, if the current fluctuation amplitude is 30N and the processing diameter is 60mm, the proportion coefficient is 0.5N / mm. The proportion coefficient can be used to further calculate the radial runout prediction value of the workpiece. Radial runout refers to the circumferential eccentric runout of the workpiece during rotation due to improper installation, uneven clamping or tailstock clamping change. In the implementation process, when the proportion coefficient is between 0.001 and 0.01, the estimation is performed by multiplying the workpiece diameter by 0.5; when the proportion coefficient is greater than 0.01, the runout prediction weight is amplified by 0.8. Taking an actual scenario as an example, if the processing diameter is 100mm and the proportion coefficient is 0.012, the runout prediction value is 0.96mm, which means that there may be nearly 1mm of eccentric runout during rotation, which will significantly affect the consistency of the outer circle size. At the same time, to further evaluate the influence of tailstock clamping stability on the axial position control of the workpiece, the response delay time calculated in the aforementioned tailstock cap force collaborative variation law of the grinding machine (for example, 0.15 seconds) is multiplied by the feed speed set in the current numerical control system (for example, set to 50mm / s) to obtain the "theoretical deviation amount of workpiece axial displacement".The deviation represents the maximum displacement deviation of the workpiece in the feed direction that may be caused by the tailstock response not being timely, for example, 7.5 mm, which is the maximum position deviation theoretically caused by the tailstock action lag. Based on the theoretical deviation, the maximum deviation of the axial displacement is further identified according to the current machining process and the tailstock adjustment logic, considering factors such as the actual tailstock feedback speed, the grinding feed mode (such as segmented feed or continuous feed), etc., and fine-tuning is performed, for example, a tailstock compression feedback response correction factor is added, so that the final output of the "maximum deviation of the axial displacement" is more close to the actual deviation effect. Taking the above case as an example, the corrected maximum deviation may be 6.2 mm. The system will then compare the calculated radial run-out prediction value with the "radial run-out allowed value" preset in the current product specification file (for example, the allowed value is 1.2 mm), and when the prediction value exceeds 80% of the allowed value (i.e. 0.96 mm), the system will mark the "radial precision risk level" as a warning state. This logic can prompt the operator in time that the tailstock clamping state has hidden dangers that may cause product precision problems. Similarly, the maximum deviation of the axial displacement is compared with the set "axial displacement allowed value" (for example, 7.0 mm), and when the deviation exceeds 80% of the value (i.e. 5.6 mm), it is also marked as an axial precision risk warning. The system interface will prompt with a red icon, and automatically link the tailstock control module to start the clamping compensation algorithm or pause the next workpiece machining task. Finally, the system combines the two risk level states to form a comprehensive evaluation of the "workpiece geometric precision limitation trend". When both the radial precision and the axial precision are in the warning state, it is determined to be "severely limited", and the tailstock clamping parameters need to be corrected or the tailstock pressing mechanism needs to be replaced; when only one of them is in the warning state, it is determined to be "slightly limited", and the system allows the machining to continue but requires to strengthen the quality sampling frequency; when both of them are normal, it is determined that the current tailstock state has a stable effect on the geometric precision, which is "precision stable". This method realizes the real-time prediction and grading warning of the geometric precision risk of the grinding machine, and is one of the key quality control mechanisms based on data fusion of the Internet of Things in the present application.

[0066] Preferably, the step S2 of determining the grinding machine processing system linkage disorder feature according to the tailstock pressing force cooperative change rule and the workpiece geometric precision limitation trend comprises:

[0067] The response delay time value is extracted from the tailstock pressing force cooperative change rule, and the response delay time is divided by the current machining cycle time to obtain the delay time proportion;

[0068] The radial run-out prediction value and the maximum deviation of the axial displacement extracted from the workpiece geometric precision limitation trend are subjected to weighted summation operation, wherein the radial run-out prediction value weight is 0.6 and the maximum deviation of the axial displacement weight is 0.4, to obtain a comprehensive deviation index;

[0069] The delay time proportion is multiplied by the comprehensive deviation index to obtain a preliminary misadjustment coefficient;

[0070] The difference between the synchronism proportion and the standard synchronism proportion 85% is calculated based on the grinding machine tailstock clamping force cooperative change law to obtain a synchronism deviation degree;

[0071] The radial accuracy risk level and the axial accuracy risk level in the workpiece geometric accuracy limited trend are valued, and when both levels are early warning, the value is 2, when only one level is early warning, the value is 1, and when both levels are normal, the value is 0, to obtain an accuracy risk coefficient;

[0072] The linkage misadjustment reference value is multiplied by the accuracy risk coefficient to obtain an adjusted misadjustment coefficient;

[0073] Based on the grinding machine tailstock clamping force cooperative change law, the clamping force fluctuation range value is divided by the workpiece diameter to obtain a unit diameter fluctuation rate;

[0074] The adjusted misadjustment coefficient and the unit diameter fluctuation rate are weighted and averaged, wherein the weight of the adjusted misadjustment coefficient is 0.7, and the weight of the unit diameter fluctuation rate is 0.3, to obtain a grinding machine processing system linkage misadjustment feature.

[0075] The latest delay time value is extracted from the response delay time data obtained based on the analysis of the top pressing force coordination change law of the grinder tailstock, which represents the time difference between the tailstock top pressing force adjustment action and the actual workpiece response, usually in seconds. Then, by dividing the response delay time by the single machining cycle time defined by the numerical control system in the current machining process (for example, the overall time required to machine a workpiece, usually in seconds), a dimensionless "delay time ratio" is calculated to describe the proportion of the response delay in the overall machining cycle. For example, if the delay time is 0.2 seconds and the cycle time is 5 seconds, the delay time ratio is 0.04, i.e. the delay time accounts for 4% of the overall machining cycle, which is used to measure the real-time response of the tailstock action. Based on the radial run-out prediction value and the maximum axial displacement deviation obtained in the previous calculation of the workpiece geometric precision limitation trend, the weights of 0.6 and 0.4 are assigned respectively, and the weighted sum is calculated to obtain the "comprehensive deviation index". This operation uses weight coefficients to reflect the different effects of the two errors on the overall workpiece accuracy. Radial run-out has a more direct and obvious effect on workpiece roundness, so it is given a higher weight. In specific implementation, the system retrieves the radial run-out prediction value and the maximum axial displacement deviation data of the current machining from the database, multiplies them by the corresponding weights, and then adds them. For example, if the radial run-out prediction value is 0.8 mm and the maximum axial displacement deviation is 5.0 mm, the comprehensive deviation index is 0.6 multiplied by 0.8 plus 0.4 multiplied by 5.0, i.e. 0.48 plus 2.0, totaling 2.48, which reflects the overall level of machining error. Multiply the delay time ratio and the comprehensive deviation index obtained by the foregoing calculation to obtain the "preliminary misadjustment coefficient". This coefficient represents the degree of system misadjustment caused by response delay and error comprehensive effect in the machining process, and is an important value for evaluating the linkage performance of the grinder tailstock. In operation, the delay time ratio and the comprehensive deviation index are passed as input parameters to a specially designed monitoring software module to calculate the misadjustment coefficient in real time. For example, if the delay time ratio is 0.04 and the comprehensive deviation index is 2.48, the preliminary misadjustment coefficient is 0.0992. Based on the top pressing force coordination change law of the grinder tailstock, the real-time "synchronization ratio" is extracted by monitoring the synchronization ratio of the top pressing force and the spindle and feeding system, which represents the same direction frequency of the top pressing force fluctuation and the mechanical action increase and decrease change, usually expressed in percentage. According to industry experience, the standard synchronization ratio is set to 85%, which is used as a reference value for the ideal synchronization state. The absolute value of the difference between the current synchronization ratio and 85% is calculated to obtain the "synchronization deviation degree", which quantifies the degree of deviation of the current system synchronization from the ideal state. For example, if the real-time synchronization ratio is 78%, the synchronization deviation degree is 7%, indicating that there is a small deviation in synchronization, which may lead to machining error.According to the radial accuracy risk level and the axial accuracy risk level in the workpiece geometry precision limited trend analysis result, the risk level is mapped to a numerical "accuracy risk coefficient". The assignment rule is: when both radial and axial are in the early warning state, assign a value of 2, indicating a serious risk; when only one is in the early warning state, assign a value of 1, indicating a moderate risk; when both are normal, assign a value of 0, indicating a low risk. The numerical risk coefficient facilitates subsequent algorithm for risk quantification and decision support. For example, if the current detection is radial accuracy early warning and axial normal, the accuracy risk coefficient is 1. Multiply the pre-set "linkage misadjustment reference value" (this value is set according to historical processing data and experience, for example, 0.05) with the above-mentioned accuracy risk coefficient to obtain the "adjusted misadjustment coefficient", which reflects the weighted influence of the system misadjustment degree after considering the risk level. For example, if the linkage misadjustment reference value is 0.05 and the accuracy risk coefficient is 2, the adjusted misadjustment coefficient is 0.1, which can be used as a reference threshold for the control system to adjust the linkage strategy. Based on the cooperative variation law of the grinder tailstock clamping force, from the previously obtained top tight force fluctuation range data, divide by the current workpiece diameter to calculate the "unit diameter fluctuation rate", which is used to describe the clamping force fluctuation intensity on the unit workpiece size. The calculation process of this value includes extracting the currently measured clamping force fluctuation amplitude (such as 30 Newton) and the workpiece diameter (such as 100 millimeters) from the Internet of Things data platform, and obtaining the fluctuation rate (0.3 N / mm) through numerical division, which provides a standardized reference for subsequent linkage evaluation. Finally, the adjusted misadjustment coefficient and the unit diameter fluctuation rate are weighted and averaged according to the weights 0.7 and 0.3 to obtain the final "grinder processing system linkage misadjustment feature". This feature comprehensively reflects the system response delay, error risk level and force fluctuation intensity, and is used as a key diagnostic index of the grinder intelligent monitoring system to guide processing parameter adjustment and maintenance decision. For example, if the adjusted misadjustment coefficient is 0.1 and the unit diameter fluctuation rate is 0.3, the final feature value is 0.7 times 0.1 plus 0.3 times 0.3, which is approximately equal to 0.16, representing that the current system linkage has a certain misadjustment, and the operator is suggested to pay attention to the clamping device state.

[0076] Preferably, step S3 comprises the following steps:

[0077] Step S31: According to the grinder processing system linkage misadjustment feature, extract the misadjustment degree quantitative value, and obtain the workpiece processing batch quantity in the corresponding time period to obtain the single batch misadjustment intensity;

[0078] The embodiment of the application is based on the previously calculated grinding machine processing system linkage disorder characteristics, first uses the Internet of Things platform to collect the real-time values of the disorder characteristics, combines the processing log record of the processing batch quantity of the workpiece in the corresponding time period, and performs data correlation processing. The specific operation is to accumulate and sum all the linkage disorder characteristic data of all batches in the same time period, and then divide by the batch quantity of the time period, thereby calculating the "single batch disorder intensity". This index reflects the average intensity of system disorder in each batch processing process. For example, the cumulative disorder characteristics detected in one hour is 10, and the number of workpiece batches completed during the period is 50, so the single batch disorder intensity is 0.2. This step combines real-time data acquisition and batch statistics to quantify the impact of disorder on production batch size, facilitating subsequent quality analysis.

[0079] Step S32: based on the single batch disorder intensity and the preset quality stability reference value, when the single batch disorder intensity is greater than the reference value, record the quality stability deviation frequency, and accumulate and count the occurrence frequency of the deviation frequency in the last 10 batches, thereby obtaining the quality stability deviation data;

[0080] The embodiment of the application compares the calculated single batch disorder intensity with the preset quality stability reference value to monitor whether there is a quality stability deviation. The reference value is usually determined by historical data analysis or process standards, such as 0.15 as the threshold value. When the disorder intensity of a batch exceeds the reference value, the system automatically records a "quality stability deviation frequency". In order to ensure the real-time and continuity of the deviation statistics, the system maintains a sliding window to count the occurrence frequency of the deviation frequency in the last 10 batches, calculates the "quality stability deviation data", i.e. the percentage of the deviation frequency in the total batch number. For example, if there are 5 deviations in 10 batches, the deviation data is 50%. This statistics helps to identify the continuity and regularity of quality deviation.

[0081] Step S33: according to the quality stability deviation data and the grinding machine processing system linkage disorder characteristics, the product quality stability deviation condition is obtained by association matching, when the occurrence frequency is greater than 40% and it is a systematic disorder, the product quality stability deviation condition is determined as a continuous deviation;

[0082] The embodiment of the present application associates the quality stability deviation data obtained in the previous step with the current grinding machine processing system linkage maladjustment features for data correlation analysis. This correlation matching is achieved by constructing a multi-dimensional time series model, and the change trend of both is compared using statistical methods or machine learning algorithms to determine whether the quality deviation is closely related to the system linkage maladjustment. If the occurrence frequency of the deviation data exceeds 40%, and the corresponding maladjustment feature shows a systematic and continuous abnormality (for example, the linkage maladjustment feature exceeds the normal range for consecutive multiple periods), it is determined that the product quality stability deviation condition is "continuous deviation", indicating that there is a persistent linkage problem in the grinding machine processing process, affecting the product quality stability. This judgment supports the development of subsequent control strategies.

[0083] Step S34: determining the grinding machine control parameter deviation degree according to the product quality stability deviation condition and the grinding machine processing system linkage maladjustment features;

[0084] The embodiment of the present application determines the deviation degree of the current grinding machine control parameter according to the product quality stability deviation condition and the grinding machine processing system linkage maladjustment features. The specific method is to input the continuous deviation condition and the numerical value and change trend of the maladjustment feature into the parameter adjustment model, and output the control parameter deviation level, which is usually divided into mild, moderate and severe deviation. For example, when the continuous deviation and the maladjustment feature value are high, it is determined as severe deviation, prompting quick adjustment. This process relies on the real-time detection of the Internet of Things monitoring system on the control parameter and the comparison with the historical model to accurately locate the parameter abnormal range, providing a basis for subsequent adjustment.

[0085] Step S35: real-time adjustment processing of the grinding machine control parameter deviation degree based on the Internet of Things control system, to obtain the grinding machine parameter adjustment execution situation.

[0086] The embodiment of the present application performs real-time automatic adjustment operation for the grinding machine control parameter deviation degree determined in step S34 through the integrated Internet of Things control system. The system receives the deviation level information in real time, calls the preset parameter optimization algorithm and control instructions, and automatically adjusts the key parameters of the grinding machine, such as the tailstock clamping force, the feed speed, the grinding depth, etc., to ensure that the parameters return to the normal range. During the adjustment process, the system continuously monitors the feedback data to form a closed-loop control, ensuring the effectiveness and stability of the adjustment. For example, when severe deviation is detected, the system quickly reduces the clamping force and adjusts the feed speed, while recording the adjustment execution situation and effect. This step realizes intelligent control based on real-time monitoring, improving the grinding machine processing quality and equipment operation efficiency.

[0087] Especially important is that step S34 includes the following steps:

[0088] Step S341: Extract the out-of-adjustment degree quantization value sequence during the persistent deviation period based on the product quality stability deviation condition, calculate the standard deviation of the values in the sequence, and obtain the deviation fluctuation intensity;

[0089] The embodiment of the present application has identified that the product has a "persistent deviation" quality problem in the previous stage. The key of this step is to extract the "out-of-adjustment degree quantization value" in the grinding machine processing system linkage out-of-adjustment feature during the persistent deviation period from the Internet of Things platform to form a time sequence. The quantization value is the out-of-adjustment feature index obtained by weighting the adjusted out-of-adjustment coefficient and the unit diameter fluctuation rate. The system sorts these indexes using time labels to form a continuous out-of-adjustment value sequence such as every minute or every batch. Then, the statistical analysis module is called to calculate the standard deviation of the sequence, and the standard deviation is the "deviation fluctuation intensity" used to describe the stability of the out-of-adjustment degree during the persistent deviation period. For example, if the out-of-adjustment degree quantization value sequence in a certain period of time is [2.1, 2.3, 2.2, 2.5, 3.0], the standard deviation is large, indicating that the fluctuation is violent, reflecting that the stability of the grinding machine control system is poor.

[0090] Step S342: Multiply the deviation fluctuation intensity and the linkage out-of-adjustment frequency in the grinding machine processing system linkage out-of-adjustment feature to calculate the control parameter influence factor;

[0091] After obtaining the deviation fluctuation intensity, the embodiment of the present application further combines the "linkage out-of-adjustment frequency" in the grinding machine processing system linkage out-of-adjustment feature to analyze the sensitivity of the control parameter. The linkage out-of-adjustment frequency refers to the number of times of detecting linkage out-of-adjustment in unit time, usually represented by "number of abnormal detection times per hour". In this step, the system multiplies the deviation fluctuation intensity and the linkage out-of-adjustment frequency to calculate the "control parameter influence factor", which is used to measure the influence degree of the control parameter on the system stability in the processing process. For example, if the deviation fluctuation intensity is 0.35 and the linkage out-of-adjustment frequency is 24 times / hour, the influence factor is 8.4, indicating that the current control parameter change has a strong triggering effect on the system out-of-adjustment, which is the core basis for subsequent parameter analysis and adjustment.

[0092] Step S343: Extract the effective influence segment with a numerical range of 1.2-8.5 based on the control parameter influence factor, and count the duration of the numerical rising trend in the effective influence segment to obtain the numerical rising trend duration;

[0093] The embodiment of the present application extracts the key data change stage from the control parameter influence factor calculated above. The system first filters out the paragraph with the numerical range between 1.2 and 8.5 as the "effective influence paragraph", which represents the control influence stage with actual intervention value in the system. Then, the system detects the "numerical rising trend duration" in the effective influence paragraph through a trend analysis algorithm, that is, how long the control parameter influence factor remains in the rising state. This process realizes trend judgment by combining methods such as sliding window and fitting regression line, and counts the continuous rising time, for example, if the influence factor rises from 2.5 to 7.8 in 20 minutes, the duration is 20 minutes. This information provides time quantification basis for subsequent analysis of parameter deviation accumulation effect.

[0094] Step S344: According to the numerical rising trend duration and the current grinding machine running period, when the numerical rising trend duration exceeds 60% of the running period, the parameter deviation accumulation amount is calculated, wherein the parameter deviation accumulation amount is equal to the product of the control parameter influence factor and the numerical value of the duration;

[0095] The embodiment of the present application compares the numerical rising trend duration calculated in the above step with the current grinding machine running period. The grinding machine running period can be determined according to the daily tool changing period or the average processing time per batch, for example, the current period is 30 minutes. When the trend duration exceeds 60% of the period (i.e. 18 minutes), it is determined that the parameter fluctuation has strong persistence. At this time, the system calculates the "parameter deviation accumulation amount" according to the product of the "control parameter influence factor" and the "rising trend duration", which represents the overall cumulative effect of the control parameter deviation on the system. For example, if the influence factor is 6.5 and the trend duration is 20 minutes, the deviation accumulation amount is 130, indicating that the current control system has a significant and persistent parameter imbalance phenomenon, which requires intervention at the control strategy level.

[0096] Step S345: Determine the grinding machine control parameter deviation degree based on the parameter deviation accumulation amount, when the parameter deviation accumulation amount is greater than 15, it is marked as severe deviation, when the parameter deviation accumulation amount is between 5 and 15, it is marked as moderate deviation, and when the parameter deviation accumulation amount is less than 5, it is marked as slight deviation.

[0097] The embodiment of the application classifies and judges the grinding machine control parameter deviation degree according to the parameter deviation cumulative amount calculated above. The Internet of Things monitoring platform marks different numerical intervals through automatic rules. When the deviation cumulative amount is greater than 15, it is marked as “serious deviation”, which indicates that the current grinding machine has a large risk, and the feed speed or the clamping force curve needs to be adjusted immediately. When the cumulative amount is between 5 and 15, it is marked as “moderate deviation”, at which time a mild adjustment strategy can be adopted, such as fine-tuning the spindle load compensation parameter. When the cumulative amount is less than 5, it is marked as “slight deviation”, which can be monitored regularly or corrected slowly in the next batch. The classification result is directly fed back to the intelligent Internet of Things control system, which realizes closed-loop adaptive optimization with automatic control logic. This step not only improves the control accuracy, but also guarantees the product processing stability and batch consistency.

[0098] Preferably, step S35 comprises the following steps:

[0099] Step S351: According to the grinding machine control parameter deviation degree, the corresponding adjustment instruction is issued through the Internet of Things control system to obtain the Internet of Things control system adjustment instruction. If the grinding machine control parameter deviation degree is serious deviation, the adjustment range is 20% of the current parameter value. If the grinding machine control parameter deviation degree is moderate deviation, the adjustment range is 10% of the current parameter value. If the grinding machine control parameter deviation degree is slight deviation, the adjustment range is 5% of the current parameter value.

[0100] After obtaining the determination result of the grinding machine control parameter deviation degree in the above step, the embodiment of the application generates the corresponding automatic adjustment instruction through the Internet of Things control platform and issues it to the grinding machine execution unit for real-time control. For example, when it is judged that the control parameter deviation degree in a certain grinding batch is “serious deviation”, the system will automatically calculate the adjustment range according to the current set value of the parameter. For example, if the current value of the spindle speed is 1800 rpm, the system sets the adjustment range to be 20% of it, that is, to increase or decrease 360 rpm, forming the adjustment target value of 2160 rpm or 1440 rpm (the specific adjustment direction is determined based on the deviation trend in the previous stage). When the deviation degree is “moderate deviation” or “slight deviation”, the adjustment range is calculated as 10% and 5% of the current value respectively. The adjustment instruction is transmitted to the grinding machine system through the PLC controller or the edge computing gateway, and the execution feedback mechanism is started to ensure the synchronization and traceability of the adjustment response.

[0101] Step S352: Based on the execution time of the Internet of Things control system adjustment instruction and the parameter deviation cumulative amount, the numerical value is divided to calculate the adjustment response efficiency, and the actual parameter change after adjustment is recorded;

[0102] The embodiment of the present application records the "execution time" of the adjustment instruction in real time after the Internet of Things platform successfully issues the adjustment instruction, that is, the time (usually in seconds or milliseconds) from issuing the command to the completion of the actual change of the grinding machine parameters. This data can be obtained by the feedback module or high-frequency data acquisition device of the grinding machine. At the same time, the system uses the "parameter deviation cumulative amount" calculated in the previous stage as the base value, divides it by the execution time, and obtains the "adjustment response efficiency". This efficiency reflects the speed of repairing unit deviation, and is an important indicator for judging the response sensitivity and control link efficiency of the system. In addition, the system also needs to record the "actual parameter change amount" in synchronization, that is, the actual change degree of the grinding machine parameters after changing from the initial set value, such as setting the adjustment target as increasing the spindle speed by 180 rpm in a certain adjustment, but the system finally only increases by 150 rpm, so the actual change amount is 150 rpm. This data will be used to verify the adjustment execution effect in the subsequent steps.

[0103] Step S353: According to the numerical comparison between the actual parameter change amount and the preset expected adjustment amplitude, the adjustment execution deviation rate is calculated, and when the adjustment execution deviation rate is less than 15%, it is determined that the grinding machine parameter adjustment execution is adjustment effective, otherwise it is determined as adjustment abnormal, so as to obtain the grinding machine parameter adjustment execution.

[0104] After obtaining the "actual parameter change amount", the system compares the value with the "expected adjustment amplitude" set before, so as to calculate the "adjustment execution deviation rate". This deviation rate reflects the response accuracy of the system after executing the instruction, that is, the relative difference between the actual adjustment amplitude and the expected adjustment amplitude. For example, if the expected adjustment amplitude is 200 rpm, and the actual adjustment is only 170 rpm, the adjustment execution deviation rate is 15%, and the system compares this value with the preset tolerance threshold 15%. If the deviation rate is less than the threshold, the system determines that the adjustment is "adjustment effective", which means that the control parameter change and the control strategy are basically consistent, and the adjustment action is reliable. If the deviation rate is greater than 15%, it is determined as "adjustment abnormal", which indicates that there may be problems such as actuator failure, instruction delay, network interference or insufficient mechanical rigidity. The judgment result is finally recorded as "grinding machine parameter adjustment execution" in the Internet of Things platform database, and can be used for subsequent adaptive control strategy optimization and device health management model training.

[0105] Preferably, step S4 comprises the following steps:

[0106] Step S41: According to the grinding machine parameter adjustment execution, the adjustment response efficiency and the adjustment execution deviation rate are extracted, and the proportion of the number of times of adjustment effective and the number of times of adjustment abnormal in the last 5 adjustment operations is counted, so as to obtain the system response success rate;

[0107] On the basis of the grinding machine parameter adjustment execution, adjustment response efficiency and adjustment execution deviation rate obtained in the foregoing steps, the system processes the results of the last five continuous adjustment operations by means of the Internet of Things data analysis module. In the specific implementation, the system first extracts the adjustment response efficiency and adjustment execution deviation rate corresponding to each adjustment operation, and counts the five operations based on the determination results of "adjustment effective" and "adjustment abnormal". For example, if three of the five adjustments are effective and two are abnormal, the system response success rate is 60%. The operation can be dynamically updated by setting a sliding time window. Whenever a new adjustment execution is completed, the success rate is updated by re-counting, for the purpose of real-time monitoring of the execution stability and reliability of the Internet of Things control system.

[0108] Step S42: When the system response success rate is lower than the preset standard value, record the number of response defect occurrences, and extract the corresponding adjustment execution deviation rate value sequence, calculate the average value of the deviation rate values, and obtain the average execution deviation;

[0109] When the system response success rate is detected to be lower than the preset standard value (for example, 70%), the system marks this case as a "response defect", and counts the event in the "response defect occurrence number". At the same time, the system extracts the adjustment execution deviation rate values in the five adjustment operations to form a value sequence, for example, the deviation rates in a certain sampling period are 10%, 17%, 13%, 21% and 9% in turn. Then, the system averages the value sequence to obtain the "average execution deviation", which is 14% in this example. This data reflects the consistency deviation degree of the Internet of Things control system in the actual adjustment process, and provides a basis for subsequent defect severity quantification.

[0110] Step S43: Perform numerical multiplication operation according to the response defect occurrence number and the average execution deviation to obtain the response defect severity index;

[0111] The embodiment of the present application performs numerical multiplication operation based on the "response defect occurrence number" and "average execution deviation" obtained in the previous stage to quantize the "response defect severity index". The index is used to describe the combined strength of system response defects, that is, the comprehensive influence of defect frequency and deviation amplitude. For example, in a certain monitoring stage, the system response defect occurs 5 times, and the average execution deviation is 14%, so the response defect severity index is 5 times 14, which is 70. Through the index, it can be accurately judged whether there is a systemic problem in the overall adjustment stability of the control system, which is conducive to the early warning of the risk of equipment control precision decline.

[0112] Step S44: Based on the response defect severity index, count the cumulative occurrence frequency of the response defect in the last 20 adjustment periods to obtain the response defect frequency;

[0113] The embodiment of the present application further performs periodic trend evaluation on the "response defect severity index", and the specific method is to statistically calculate the "cumulative occurrence frequency" of the index in the last 20 adjustment cycles based on the sliding window technology, that is, how many times there are response defects in 20 cycles. For example, if it is found that there are response defect events in 8 cycles in 20 cycles, then the response defect frequency is 40%. The frequency data can be automatically recorded and updated by setting a periodic event log collection module, and is a key reference index for quantifying system performance fluctuations.

[0114] Step S45: weighted sum calculation based on the response defect severity index and the response defect frequency, wherein the weight of the response defect severity index is 0.6 and the weight of the response defect frequency is 0.4, to obtain the Internet of Things control system response defect data;

[0115] The embodiment of the present application takes the "response defect severity index" and "response defect frequency" obtained in the previous step as weighted input variables, and performs weighted sum to construct "Internet of Things control system response defect data", with weights of 0.6 and 0.4, respectively, that is, the system pays more attention to the severity than the frequency alone. In actual implementation, the system performs the weighted calculation process through the embedded control rule model, for example, if the severity index is 70 and the defect frequency is 40%, then the final response defect data is 70 multiplied by 0.6 plus 40 multiplied by 0.4, which is 58. The result is a unified score of system response performance, providing a basis for the next step of control optimization.

[0116] Step S46: performing grinding machine quality monitoring system control optimization processing according to the Internet of Things control system response defect data to obtain an optimized grinding machine product quality real-time monitoring system.

[0117] After obtaining the "Internet of Things control system response defect data", the system compares it with the preset performance threshold, and when it is found that the score exceeds the set risk limit (such as 60 points), the "grinding machine quality monitoring system control optimization processing" is automatically triggered, which mainly includes the following operations: 1) adjusting the adjustment sensitivity of the control parameter to narrow the response bandwidth to reduce the risk of large-scale misadjustment; 2) optimizing the actuator driving frequency and control loop refresh frequency to enhance real-time performance; 3) automatically selecting a redundant parameter adjustment channel according to the adjustment execution deviation characteristics to perform compensation control; 4) re-optimizing the PID parameters involved in the control rule model or calling a deep learning sub-model for nonlinear adjustment prediction. Through the above control optimization operations, the system can continuously improve the monitoring accuracy of the grinding machine product quality, realize closed-loop optimization of the grinding process quality stability, and finally output an optimized grinding machine product quality real-time monitoring system.

[0118] Preferably, step S46 includes the following steps:

[0119] Step S461: According to the Internet of Things control system response defect data, the control parameter adjustment record and the grinder operation state log in the corresponding adjustment period are extracted, and a history comparison table of adjustment behavior and system response is generated;

[0120] Based on the "Internet of Things control system response defect data" obtained in the foregoing steps, the system automatically extracts the adjustment period data corresponding to each response defect event, specifically including the control parameter adjustment record involved in the adjustment process (such as the adjustment value and timestamp of parameters such as spindle speed, grinding wheel feed speed, and cooling liquid flow) and the grinder operation state log (including real-time operation indexes such as vibration amplitude, temperature rise, and load fluctuation of the grinder working condition). These data are collected in real time through the Internet of Things control gateway device embedded in the system, and are associated and matched with the defect event, and finally a "history comparison table of adjustment behavior and system response" is generated. The comparison table takes time axis as the main line, and forms a one-to-one correspondence between the adjustment behavior of each control parameter and the corresponding system response result (whether the adjustment is effective, response time, deviation rate, etc.), which is used for subsequent adjustment effect tracing and model optimization.

[0121] Step S462: The history comparison table is classified and summarized based on the adjustment parameter type, response effectiveness level, and execution time to form a control parameter execution tag library;

[0122] After the comparison table is constructed, the system further uses a classification summary algorithm to analyze the information in the table, mainly according to three dimensions for clustering processing: first, the adjustment parameter type dimension, that is, the parameter type is classified, for example, feed speed type, grinding pressure type, temperature control type, etc.; second, the response effectiveness level dimension, which is divided into "high response", "general response", and "invalid response" and other levels according to whether the adjustment is effective and the deviation rate; and third, the execution time dimension, which divides the behavior into different time bins (such as less than 5 seconds, 5-10 seconds, and more than 10 seconds) according to the time required from the execution of the adjustment instruction to the completion of the response. After the three-dimensional classification, the system can construct a "control parameter execution tag library", and each tag in the tag library represents a typical response mode of a certain type of adjustment behavior, providing a quantifiable reference basis for abnormal behavior identification.

[0123] Step S463: Based on the control parameter execution tag library, the parameter type with abnormal response or high execution deviation rate in the adjustment behavior is identified and marked as a parameter item to be optimized;

[0124] The embodiment of the application performs a tag library based on the above control parameters, and introduces a combined strategy of rule matching and data mining to identify and adjust abnormal behaviors. First, the response effectiveness level and execution deviation rate of each type of adjustment parameter in all historical records are statistically analyzed. When it is found that the abnormal response proportion of a certain type of parameter (such as invalid response exceeding 30%) or the average execution deviation rate is higher than the system average level (such as exceeding 20%), the parameter type is automatically identified as a parameter type with potential control problems. Such parameters are marked as “to-be-optimized parameter items” by the system, and the corresponding behavior sample quantity, fault cycle distribution, and additional information such as involved machine type are recorded. Taking an actual application as an example, if it is found that the “grinding wheel speed control” type parameter has serious response delay and generally high deviation rate in multiple cycles, the system automatically lists it in the to-be-optimized list, indicating that there is a problem with the parameter weight or threshold setting in the control model.

[0125] Step S464: The execution failure number and average adjustment deviation value of the to-be-optimized parameter items in the continuous running cycle are obtained to form a parameter optimization candidate list.

[0126] After identifying the to-be-optimized parameter items, the system further tracks the behaviors of these parameter items in continuous cycles, counts the number of execution failures, that is, the number of occurrences of “adjustment abnormalities” determined by the system in the past several continuous running cycles (such as the last 50 adjustment operations), and calculates the average adjustment deviation value corresponding to these operations. For example, for the “main shaft torque adjustment” parameter, there are 15 adjustment abnormalities in the past 50 adjustments, and the average deviation value is 18%. The system forms a “parameter optimization candidate list” based on this, which contains the abnormal frequency, average deviation, priority, and other information of each to-be-optimized parameter, for subsequent judgment of which parameters most urgently need to be adjusted and guidance of selection of optimization strategies.

[0127] Step S465: Based on the parameter optimization candidate list, an adjustment parameter optimization configuration process is performed to obtain optimized control parameter settings and upload them to the grinding machine quality monitoring system to replace the original parameters, thereby obtaining an optimized grinding machine product quality real-time monitoring system.

[0128] The embodiment of the application enters the "adjustment parameter optimization configuration processing" stage according to the parameter optimization candidate list. In the specific implementation process, the system first calls the intelligent parameter adaptive module to generate optimization suggestions according to the parameter behavior characteristics and fault performance in the list. For example, for the "cooling liquid flow adjustment parameter", which is analyzed and determined to have high abnormal frequency and poor response stability, the system can adjust its initial setting range (such as increasing the default flow upper limit from 2.5 liters / minute to 3.0 liters / minute), and optimize the adjustment step and feedback update period to enhance the system's rapid response capability. The optimized control parameter settings are uploaded to the grinding machine quality monitoring system by the Internet of Things control master station, replacing the original parameter configuration. The update process synchronously triggers version log recording and model retraining marking, ensuring that the system uses the optimal configuration in future adjustment, and finally forms an optimized grinding machine product quality real-time monitoring system, realizing intelligent control closed-loop improvement of equipment.

[0129] Especially important is that step S465 includes the following steps:

[0130] According to the parameter optimization candidate list, the number of execution failures and the average adjustment deviation value of each parameter to be optimized are extracted, and sorted in descending order of the number of failures. The parameter items with more than 8 failures are marked as high-priority optimization items, and the parameter items with 3-8 failures are marked as medium-priority optimization items.

[0131] Based on the average adjustment deviation value of the high-priority optimization items, a parameter adjustment coefficient is calculated. When the average adjustment deviation value is greater than 25%, the parameter adjustment coefficient is 0.7; when the average adjustment deviation value is between 15% and 25%, the parameter adjustment coefficient is 0.8; and when the average adjustment deviation value is less than 15%, the parameter adjustment coefficient is 0.9.

[0132] According to the parameter adjustment coefficient, the original parameter threshold adjustment processing is performed on the high-priority optimization items, and the adjusted parameter threshold is matched and combined with the corresponding response time setting to form an optimized parameter combination.

[0133] Based on the number of execution failures of the medium-priority optimization items and the preset reference failure number, a numerical comparison is performed. When the number of execution failures exceeds the reference value, the response time of the corresponding parameter is shortened by 10%; when the number of execution failures is equal to the reference value, the response time is shortened by 5%, to obtain a time adjustment scheme for the medium-priority parameters.

[0134] The optimized parameter combination and the time adjustment scheme are integrated to generate a complete control parameter configuration file, which is uploaded to the parameter storage module of the grinding machine quality monitoring system through the Internet of Things communication interface. After the parameter replacement operation, the system is restarted for verification, thereby obtaining an optimized grinding machine product quality real-time monitoring system.

[0135] The execution failure number and average adjustment deviation value of each parameter to be optimized in the list are extracted and processed by the system, and the parameter behavior record analysis is performed by using the built-in data traversal and screening module of the system. The execution failure number refers to the total number of times that the system judges that the parameter adjustment is invalid or does not achieve the expected response within a specified monitoring period. The average adjustment deviation value represents the average proportion of the error between the actual response value and the target set value in each invalid adjustment. The system sorts all parameter items in descending order of execution failure number, and marks the priority according to the set priority division rule: when the failure number is greater than 8, it is marked as a "high-priority optimization item"; when the failure number is between 3 and 8, it is marked as a "medium-priority optimization item". For example, for the "grinding wheel shaft cooling liquid flow rate" adjustment parameter, there are 9 failure behaviors in the last 50 runs, which are judged by the system as high-priority items; while the "grinding wheel feed speed" parameter fails 5 times during this period, it is marked as a medium-priority optimization item. For the parameter items marked as high-priority, the system further analyzes the average adjustment deviation value and calculates the corresponding "parameter adjustment coefficient" for each item according to the set classification standard. The parameter adjustment coefficient is a scaling coefficient used by the system to scale the original control parameter threshold range, which reflects the optimization degree of the adjustment sensitivity. The system judges the deviation interval of each high-priority parameter in turn, and if the average adjustment deviation value is greater than 25%, the parameter adjustment coefficient is set to 0.7, indicating that the parameter deviation is large and the adjustment range or amplitude needs to be significantly reduced; if the deviation value is between 15% and 25%, the coefficient is set to 0.8; if it is less than 15%, it is set to 0.9. For example, the average deviation value of the "main shaft torque" parameter is 28%, and the system sets its adjustment coefficient to 0.7 and uses it for the next original threshold revision. The original control parameter threshold of the high-priority optimization item is adjusted according to the parameter adjustment coefficient obtained above. Threshold adjustment refers to resetting the upper and lower limit range of the control parameter to optimize the adjustment behavior and make the parameter response more stable and effective. The system compresses the adjustment range by multiplying the corresponding parameter adjustment coefficient, for example, the original upper limit of "main shaft torque" is 80 Nm and the lower limit is 40 Nm, if the adjustment coefficient is 0.7, the new upper limit is set to the original value multiplied by 0.7, i.e. 56 Nm, and the lower limit is set to 28 Nm in proportion. After adjustment, the system performs parameter combination matching according to the optimal response time (i.e. the time setting with fast response and small deviation) matched by each parameter in the historical adjustment behavior, to form an "optimized parameter combination" including the optimized parameter value and the recommended response time setting, so as to improve the dynamic response accuracy and efficiency of the system. For medium-priority optimization items, the system does not perform parameter threshold adjustment, but focuses on the optimization of "response time". Response time refers to the time taken by the system to complete parameter adjustment after receiving the control instruction, which is used to measure the execution efficiency of the control system.The system sets a preset reference failure number, for example 5 times, and compares the value with the actual failure number of each medium priority parameter item: if the failure number is higher than 5 times, the response time corresponding to the parameter will be shortened by 10%; if the failure number is equal to 5 times, the response time is shortened by 5%; if it is lower than 5 times, no processing is performed. Taking the "grinding wheel feed speed" parameter as an example, the failure number is 6 times, which is higher than the reference value, so the original response time is set to 6 seconds, and the optimized response time is set to 5.4 seconds. The system integrates the response time optimization results of all medium priority parameters to form a "time adjustment scheme". Finally, the high priority optimization parameter combination and the medium priority time adjustment scheme are integrated and processed to generate a unified format "control parameter configuration file". The file is a structured file that can be recognized by the system, usually in JSON or XML format, containing all the optimized parameter values, response times and control tag information. After the configuration file is generated, it is uploaded to the parameter storage module of the grinding machine quality monitoring system through the built-in Internet of Things communication interface module, replacing the original parameter setting. After uploading is completed, the system automatically triggers the control system soft restart process to ensure that the new parameters take effect and enter the real-time monitoring state. The system will run again under the new configuration and automatically monitor the response effect of the first adjustment cycle to form a closed-loop feedback mechanism, thereby realizing the control optimization of the numerical control grinding machine product quality real-time monitoring system.

[0136] Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, the scope of the application being defined by the attached claims and not by the above description, therefore all variations falling within the meaning and scope of the equivalent elements of the application file are intended to be included in the present application.

[0137] The above description is merely one specific implementation of the application, which enables those skilled in the art to understand or implement the application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, the application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An Internet of Things-based real-time monitoring and optimization method for product quality of a CNC grinding machine, characterized in that, The method comprises the following steps: Step S1: acquiring multi-node real-time running data of the numerical control grinding machine; According to the multi-node real-time running data, the working state parameters of the key components of the grinding machine are counted; and based on the multi-node real-time running data and the working state parameters of the key components of the grinding machine, the product machining quality fluctuation trend is determined; Step S2: detecting the tailstock clamping force cooperative change law of the grinding machine according to the product machining quality fluctuation trend; determining the workpiece geometric precision limitation trend according to the tailstock clamping force cooperative change law of the grinding machine; and determining the grinding machine machining system linkage disorder feature according to the tailstock clamping force cooperative change law of the grinding machine and the workpiece geometric precision limitation trend; Wherein, the grinding machine machining system linkage disorder feature is determined according to the tailstock clamping force cooperative change law of the grinding machine and the workpiece geometric precision limitation trend, comprising: The response delay time value is extracted from the tailstock clamping force cooperative change law of the grinding machine, and the response delay time is divided by the current machining cycle time to obtain the delay time proportion; The radial runout prediction value and the maximum deviation amount of axial displacement extracted from the workpiece geometric precision limitation trend are subjected to weighted summation operation, wherein the radial runout prediction value weight is 0.6, and the maximum deviation amount of axial displacement weight is 0.4, to obtain a comprehensive deviation index; The delay time proportion and the comprehensive deviation index are subjected to multiplication operation to obtain a preliminary disorder coefficient; The difference absolute value between the synchronism proportion calculated based on the tailstock clamping force cooperative change law of the grinding machine and the standard synchronism proportion 85% is calculated to obtain a synchronism deviation degree; The radial precision risk level and the axial precision risk level in the workpiece geometric precision limitation trend are assigned values, when both levels are early warning, the value is 2, when only one level is early warning, the value is 1, and when both levels are normal, the value is 0, to obtain a precision risk coefficient; The linkage disorder reference value and the precision risk coefficient are subjected to multiplication operation to obtain an adjusted disorder coefficient; Based on the tailstock clamping force cooperative change law of the grinding machine, the clamping force fluctuation amplitude range value is divided by the workpiece diameter to obtain a unit diameter fluctuation rate; The adjusted disorder coefficient and the unit diameter fluctuation rate are subjected to weighted average calculation, wherein the adjusted disorder coefficient weight is 0.7, and the unit diameter fluctuation rate weight is 0.3, to obtain the grinding machine machining system linkage disorder feature; Step S3: determining the product quality stability deviation condition based on the grinding machine machining system linkage disorder feature; determining the grinding machine control parameter deviation degree according to the product quality stability deviation condition and the grinding machine machining system linkage disorder feature; and based on the Internet of Things control system, the grinding machine control parameter deviation degree is subjected to real-time adjustment processing to obtain the grinding machine parameter adjustment execution situation; Step S4: evaluating the response effect of the Internet of Things control system based on the grinding machine parameter adjustment execution situation to obtain Internet of Things control system response defect data; and according to the Internet of Things control system response defect data, the grinding machine quality monitoring system control optimization processing is performed to obtain an optimized grinding machine product quality real-time monitoring system.

2. The method according to claim 1, wherein, Step S1 comprises the following steps: Step S11: Collecting data through a sensor network installed on the tailstock, spindle box, transverse feed mechanism, grinding wheel frame and workbench to obtain multi-node real-time operation data, wherein the multi-node real-time operation data includes tailstock clamping force data, spindle speed data, feed position data, grinding wheel vibration data and workbench temperature data; Step S12: According to the multi-node real-time operation data, the working state parameters of the key components of the grinding machine are counted; Step S13: Based on the time series change of each node data in the multi-node real-time operation data, the data fluctuation intensity is calculated; and according to the change trend of the state level of each component in the working state parameters of the key components of the grinding machine, the component collaboration deviation is calculated; Step S14: Based on the data fluctuation intensity and the component collaboration deviation, correlation analysis is carried out to obtain the product processing quality fluctuation trend, wherein the correlation analysis is specifically that when the data fluctuation intensity exceeds the set threshold value and the component collaboration deviation shows an increasing trend, it is determined that the product processing quality has an unstable fluctuation trend, otherwise it is determined that the product processing quality maintains a stable trend.

3. The method according to claim 2, wherein, Step S12 includes the following steps: Step S121: According to the tailstock clamping force data, the tailstock working load average and the load fluctuation coefficient are calculated to determine the tailstock component working state level; Step S122: Based on the spindle speed data, the spindle operation stability and the speed consistency index are calculated to determine the spindle component working state level; Step S123: According to the feed position data, the feed accuracy deviation value and the feed speed uniformity index are calculated to determine the feed system component working state level; Step S124: According to the grinding wheel vibration data, the grinding wheel working balance degree and the vibration intensity index are calculated to obtain the grinding wheel system component working state level; Step S125: The thermal deformation risk and temperature stability index are evaluated by using the workbench temperature data to determine the thermal control system component working state level; Step S126: The tailstock component working state level, the spindle component working state level, the feed system component working state level, the grinding wheel system component working state level and the thermal control system component working state level are combined to form the working state parameters of the key components of the grinding machine.

4. The method according to claim 3, wherein, The tailstock clamping force collaborative change rule is detected according to the product processing quality fluctuation trend in step S2, including: According to the product processing quality fluctuation trend, the numerical sequence of the tailstock clamping force data in the time window is extracted, the clamping force numerical difference between the continuous time points is calculated, and the clamping force change gradient is obtained; The mean and standard deviation of the clamping force change gradient are calculated to determine the clamping force reference change state; Based on the tailstock clamping force reference change state, the timestamp corresponding relationship between the clamping force and the spindle speed is extracted, and the numerical ratio change trend of the two groups of data at the same time point is calculated; The synchronization proportion is obtained by analyzing the increase and decrease synchronization of the clamping force and the feed position in the tailstock clamping force reference change state, and counting the proportion of the same direction change times and the reverse change times; According to the numerical ratio change trend and the synchronization proportion, the response delay time is determined to detect the tailstock clamping force collaborative change rule.

5. The method according to claim 4, wherein, The step S2 includes the following steps: According to the tailstock clamping force of the grinding machine, the continuous clamping force value sequence is extracted, and the local maximum value in the value sequence is identified as the peak value and the local minimum value as the valley value; Based on the identified peak value and valley value, the numerical difference between the peak value and the valley value is calculated, and the maximum numerical difference is recorded as the clamping force fluctuation amplitude range; When the workpiece diameter is 10mm- 200mm, a proportional coefficient is generated based on the range of the fluctuation amplitude of the pressing force and the workpiece diameter; According to the proportional coefficient, the radial run-out prediction value is calculated, when the proportional coefficient is between 0.001-0.01, then the radial run-out prediction value=proportional coefficient×workpiece diameter×0.5; when the proportional coefficient is greater than 0.01, then the radial run-out prediction value=proportional coefficient×workpiece diameter×0.8; Based on the response delay time in the tailstock clamping force of the grinding machine and the current feed speed, the value multiplication operation is performed to obtain the theoretical deviation of the axial displacement of the workpiece; According to the theoretical deviation of the axial displacement of the workpiece, the maximum deviation of the axial displacement of the workpiece is calculated; The radial run-out prediction value is compared with the preset radial run-out allowable value, when the prediction value is greater than 80% of the allowable value, the radial precision risk level is marked as pre-warning; The maximum deviation of the axial displacement is compared with the preset allowable value of the axial displacement, when the deviation is greater than 80% of the allowable value, the axial precision risk level is marked as pre-warning; According to the combination state of the radial precision risk level and the axial precision risk level, the geometric precision limiting trend of the workpiece is determined, when the radial precision risk level and the axial precision risk level are both pre-warning, the geometric precision limiting trend is determined as serious limiting; when only one level is pre-warning, it is determined as slight limiting; when both levels are normal, it is determined as stable precision.

6. The IoT-based real-time monitoring and optimization method for product quality of a CNC grinding machine according to claim 5, characterized in that, Step S3 includes the following steps: Step S31: According to the grinding machine processing system linkage disorder feature, the disorder degree quantitative value is extracted, and the workpiece processing batch quantity in the corresponding time period is obtained, to obtain the single batch disorder intensity; Step S32: Based on the single batch disorder intensity and the preset quality stability reference value, when the single batch disorder intensity is greater than the reference value, the quality stability deviation frequency is recorded, and the appearance frequency of the deviation frequency in the continuous 10 batches is accumulated to obtain the quality stability deviation data; Step S33: According to the quality stability deviation data and the grinding machine processing system linkage disorder feature, the product quality stability deviation state is obtained, when the appearance frequency is greater than 40% and the system is disordered, the product quality stability deviation state is determined as continuous deviation; Step S34: According to the product quality stability deviation state and the grinding machine processing system linkage disorder feature, the grinding machine control parameter deviation degree is determined; Step S35: Based on the Internet of Things control system, the grinding machine control parameter deviation degree is adjusted in real time to obtain the grinding machine parameter adjustment execution situation.

7. The IoT-based real-time monitoring and optimization method for product quality of a CNC grinding machine according to claim 6, characterized in that, Step S35 includes the following steps: Step S351: According to the degree of deviation of the grinding machine control parameters, the corresponding adjustment instruction is issued through the Internet of Things control system to obtain the Internet of Things control system adjustment instruction. If the degree of deviation of the grinding machine control parameters is serious deviation, the adjustment range is 20% of the current parameter value; if the degree of deviation of the grinding machine control parameters is moderate deviation, the adjustment range is 10% of the current parameter value; if the degree of deviation of the grinding machine control parameters is slight deviation, the adjustment range is 5% of the current parameter value. Step S352: Based on the execution time of the Internet of Things control system adjustment instruction and the parameter deviation cumulative amount, the numerical value is divided to calculate the adjustment response efficiency, and the actual parameter change amount after adjustment is recorded. Step S353: According to the numerical comparison of the actual parameter change amount and the preset expected adjustment range, the adjustment execution deviation rate is calculated. When the adjustment execution deviation rate is less than 15%, it is determined that the grinding machine parameter adjustment execution is effective, otherwise it is determined to be abnormal, thereby obtaining the grinding machine parameter adjustment execution.

8. The method according to claim 7, wherein, Step S4 includes the following steps: Step S41: According to the grinding machine parameter adjustment execution, the adjustment response efficiency and the adjustment execution deviation rate are extracted, and the proportion of the number of times of adjustment effective and the number of times of adjustment abnormal in the last 5 times of adjustment operation is counted to obtain the system response success rate; Step S42: When the system response success rate is lower than the preset standard value, the number of response defect occurrences is recorded, and the corresponding adjustment execution deviation rate numerical sequence is extracted. The average value of the deviation rate numerical value is calculated to obtain the average execution deviation; Step S43: The response defect severity index is obtained by multiplying the number of response defect occurrences and the average execution deviation; Step S44: Based on the response defect severity index, the cumulative occurrence frequency of the response defect in the last 20 adjustment cycles is counted, thereby obtaining the response defect frequency; Step S45: The response defect severity index and the response defect frequency are weighted and summed, wherein the weight of the response defect severity index is 0.6 and the weight of the response defect frequency is 0.4, thereby obtaining the Internet of Things control system response defect data; Step S46: According to the Internet of Things control system response defect data, the grinding machine quality monitoring system control optimization processing is carried out to obtain the optimized grinding machine product quality real-time monitoring system. 9.The method of claim 8, wherein, Step S46 includes the following steps: Step S461: According to the Internet of Things control system response defect data, the control parameter adjustment record and the grinding machine running state log in the corresponding adjustment period are extracted to generate a historical comparison table of adjustment behavior and system response; Step S462: The historical comparison table is classified and summarized based on the adjustment parameter type, response effectiveness level and execution time to form a control parameter execution tag library; Step S463: Based on the control parameter execution tag library, the parameter type with abnormal response or high execution deviation rate in the adjustment behavior is identified and marked as a to-be-optimized parameter item; Step S464: The to-be-optimized parameter item is counted in the number of execution failures and the average adjustment deviation value in the continuous running period to form a parameter optimization candidate list; Step S465: adjusting parameter optimization configuration processing based on the parameter optimization candidate list, obtaining the optimized control parameter setting and uploading to the grinding machine quality monitoring system to replace the original parameters, and obtaining the optimized grinding machine product quality real-time monitoring system.

Images