Production control method and system based on SPC alarm linkage process locking

Abstract

The invention provides a production control method and system based on SPC alarm linkage process locking, and relates to the technical field of quality management and control, and the method comprises the steps: deploying an SPC sensing node used for collecting technological parameters at a cutting edge position where a machining tool of production equipment is in contact with a workpiece, or at a contact point where a probe of online measurement equipment is in contact with the surface of the workpiece; the method comprises the following steps: acquiring process parameter data of a key production process in real time through an SPC sensing node deployed at a cutting edge position or a contact point; inputting the process parameter data as real-time data or small sample expansion data into an SPC judgment engine; and performing statistical process control analysis on the real-time data or the small sample expansion data through an SPC judgment engine so as to model data clusters which are densely distributed and have stable parameters in the data into a virtual reference curved surface representing a normal state domain. According to the invention, the accuracy of abnormity determination can be improved, and timely management and control of quality abnormity and minimization of production loss can be realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of quality control, in particular to a production control method and system based on SPC alarm linkage process locking. BACKGROUND

[0002] In the field of modern manufacturing precision machining and quality control, statistical process control (SPC) is a core quality assurance technology. It collects key process parameter data, statistically analyzes fluctuation trends, and triggers an alarm when the limit is exceeded to remind manual intervention.

[0003] However, the existing technology has the following technical defects: the linkage fault of SPC alarm and process locking, and the lack of abnormality determination accuracy further exacerbates this problem, making it difficult to control the risk of quality abnormality expansion. Specifically, in the process parameter collection link, the SPC sensing node is deployed away from the machining blade edge or the measurement contact point, and the data cannot truly reflect the real-time state, providing inaccurate data sources for abnormality determination; in the abnormality determination link, single-dimensional threshold judgment or simple statistical analysis is used, which is difficult to build a normal state stability domain and capture the abnormal evolution trajectory of continuous parameter deviation, and has a lag and a false positive rate; especially, the alarm signal is only a warning output, and is not forcibly bound with the process operation permission, relying on manual response, which is easy to cause continuous abnormal production.

[0004] In summary, the main limitation is that there is a lack of connection between the SPC early warning function and the process execution control in terms of determination accuracy, linkage enforcement, and control closed-loop technology. The early warning value cannot be converted into quality loss control effect. In the existing improvement scheme, manual locking process has a lag and subjectivity, and algorithm optimization only focuses on alarm accuracy, and still does not touch the core linkage logic of alarm driving process locking. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a production control method based on SPC alarm linkage process locking, which can solve the linkage fault problem of SPC alarm and process locking in the prior art, and improve the abnormality determination accuracy.

[0006] To solve the above technical problems, the technical solution of the present application is as follows: In one aspect, the production control method based on SPC alarm linkage process locking comprises: Deploying SPC sensing nodes for collecting process parameters at the blade edge position where the machining tool of the production equipment contacts the workpiece, or at the contact point between the probe of the online measurement equipment and the surface of the workpiece; acquiring real-time process parameter data of the key production process through the SPC sensing nodes deployed at the blade edge position or the contact point; The process parameter data is input into the SPC judgment engine as real-time data or small sample expansion data; through the SPC judgment engine, statistical process control analysis is performed on the real-time data or small sample expansion data, so as to cluster the data set with dense distribution and stable parameters in the data into a virtual reference surface representing a normal state domain; the data point sequence showing continuous deviation or trend change in the data is modeled into a virtual trend curve representing an abnormal evolution track; The spatial intersection of the virtual reference surface and the virtual trend curve is calculated to determine a quality abnormal event in the production process, and a binary alarm signal corresponding to the quality abnormal event is generated synchronously; The binary alarm signal is forcibly bound to the operation permission of the corresponding production process, and after the forced binding is completed, a locking instruction is automatically sent to the production process based on the binary alarm signal to prohibit the production process from continuing the machining operation; When the production process is locked and the machining operation is stopped, a controlled manual unlocking mechanism is started, that is, a person with corresponding permission is required to input an abnormal reason and an analysis conclusion, and when the input abnormal reason and analysis conclusion are verified, the lock on the production process is automatically released.

[0007] Further, the SPC sensing node for collecting process parameters is deployed at the cutting edge position of the machining tool in contact with the workpiece or at the contact point of the probe of the online measurement device with the surface of the workpiece; through the SPC sensing node deployed at the cutting edge position or the contact point, the process parameter data of the key production process is obtained in real time, including: Identifying the physical cutting edge position of the machining tool in the key production process or identifying the mechanical contact point of the probe of the online measurement device; Deploying a physical SPC sensing node at the physical cutting edge position or the mechanical contact point; Establishing a communication connection between the physical SPC sensing node and the SPC judgment engine, and simultaneously performing communication protocol matching; After the communication connection and protocol matching are completed, the physical SPC sensing node is registered as a controlled sensing node in the SPC judgment engine, and a registered controlled sensing node is generated; Based on the registered controlled sensing node, a real-time acquisition link for process parameters is established; through the real-time acquisition link, process parameter data is obtained in real time.

[0008] Further, the process parameter data is input into the SPC judgment engine as real-time data or small sample expansion data; through the SPC judgment engine, statistical process control analysis is performed on the real-time data or small sample expansion data, so as to cluster the data set with dense distribution and stable parameters in the data into a virtual reference surface representing a normal state domain; the data point sequence showing continuous deviation or trend change in the data is modeled into a virtual trend curve representing an abnormal evolution track, including: The process parameter data is continuously input to the SPC determination engine as a real-time data stream through real-time acquisition of the link; The SPC determination engine temporarily stores the real-time data stream in an internal data buffer queue; The SPC determination engine performs real-time evaluation and determination on the quantity and distribution state of the real-time data in the data buffer queue; When the real-time data meets the preset data sufficiency condition, the real-time data is directly used as the analysis data source; when the real-time data does not meet the preset data sufficiency condition, a data expansion process is triggered, that is, matching and fusion are performed based on the real-time data and the preset historical database to generate small sample expansion data, and the small sample expansion data is used as the analysis data source; The analysis data source is subjected to statistical process control analysis to model data clusters with dense distribution and stable parameters in the analysis data source as a virtual reference surface representing a normal state domain, and to model data point sequences exhibiting continuous deviation or trend changes in the analysis data source as a virtual trend curve representing an abnormal evolution trajectory.

[0009] Further, the analysis data source is subjected to statistical process control analysis to model data clusters with dense distribution and stable parameters in the analysis data source as a virtual reference surface representing a normal state domain, and to model data point sequences exhibiting continuous deviation or trend changes in the analysis data source as a virtual trend curve representing an abnormal evolution trajectory, including: The analysis data source is preprocessed to eliminate outliers and noise data therein, forming a preprocessed analysis data source; Based on the preprocessed analysis data source, data clusters with dense distribution and stable parameters are identified and extracted, and data point sequences exhibiting continuous deviation or trend changes are also identified and extracted; The data clusters with dense distribution and stable parameters are subjected to multi-dimensional statistical fitting to model a virtual reference surface representing a parameter stable operation interval; The data point sequences exhibiting continuous deviation or trend changes are subjected to trend modeling processing to model a virtual trend curve representing a parameter deviation from a control limit.

[0010] Further, the spatial intersection of the virtual reference surface and the virtual trend curve is calculated to determine a quality abnormal event in the production process, and a binary alarm signal corresponding to the quality abnormal event is simultaneously generated, including: The parametric equation of the modeled virtual reference surface and the trajectory equation of the virtual trend curve are obtained; Based on the parametric equation of the virtual reference surface and the trajectory equation of the virtual trend curve, a system of simultaneous equations for solving the spatial intersection is constructed; The simultaneous equations are solved by using numerical iterative calculation, wherein the solving includes setting initial estimation of intersection point coordinates, and performing iterative calculation based on the simultaneous equations; During the iterative calculation, it is continuously judged whether the calculation result meets the preset convergence condition; when it is judged that the calculation result meets the preset convergence condition, the corresponding convergent solution is output as one or more candidate intersection point coordinates; The effectiveness of all candidate intersection point coordinates is verified; the effectiveness verification includes judging whether each candidate intersection point coordinate is located within the effective definition domain of the virtual reference surface, and judging whether the parameter offset moment corresponding to each candidate intersection point coordinate conforms to the evolution timing logic of the virtual trend curve, to obtain the candidate intersection point coordinates that pass the effectiveness verification; The candidate intersection point coordinates that pass the effectiveness verification are judged as quality abnormal event points representing process out-of-control, that is, the determination result of the quality abnormal event points is obtained; According to the determination result of the quality abnormal event points, a corresponding binary alarm signal is generated.

[0011] Further, the binary alarm signal is forcibly bound to the operation permission of the corresponding production process, and after the forcible binding is completed, a locking instruction is automatically sent to the production process based on the binary alarm signal to prohibit the production process from continuing the processing operation, including: receiving the binary alarm signal sent by the SPC determination engine; based on the preset device and process mapping relationship, the binary alarm signal is associated with one or more specific production devices and corresponding processing processes; For the associated one or more specific production devices, the current operation permission state of the production device at the device control level is verified to obtain a verification result; When the verification result confirms that the operation permission of the production device is in a lockable state, a corresponding device locking instruction is generated according to the verification result; The device locking instruction is sent to the preset target production device, and it is synchronously monitored whether the preset target production device returns an execution confirmation feedback; When the execution confirmation feedback of the device locking instruction returned by the preset target production device is received, it is confirmed that the prohibition operation on the processing process is completed.

[0012] Further, when the production process is locked and the processing operation is stopped, a controlled manual unlocking mechanism is started, that is, personnel with corresponding permissions are required to input abnormal reasons and analysis conclusions, and when the input abnormal reasons and analysis conclusions are verified, the lock on the production process is automatically released, including: After confirming that the processing process is prohibited from operating, a controlled manual unlocking interface is triggered and activated; The abnormal reason text and analysis conclusion text input by the personnel with the preset operation permission are received through the manual unlocking interface; The abnormal reason text and analysis conclusion text are checked for compliance and integrity; When the compliance and integrity check passes, a check pass result is obtained; and a corresponding process unlocking instruction is generated based on the check pass result; The process unlocking instruction is sent to the preset target production equipment, and whether the preset target production equipment returns an execution confirmation feedback is monitored synchronously; When the execution confirmation feedback returned by the preset target production equipment and corresponding to the process unlocking instruction is received, it is confirmed that the release of the processing process lock is completed.

[0013] In a second aspect, a production control system based on SPC alarm linkage process locking is provided, which executes the method, including: A deployment module is configured to deploy SPC sensing nodes for collecting process parameters at the cutting edge position where the processing tool of the production equipment contacts the workpiece, or at the contact point between the probe of the online measurement equipment and the surface of the workpiece; and the process parameter data of the key production process is obtained in real time through the SPC sensing nodes deployed at the cutting edge position or the contact point; A modeling module is configured to input the process parameter data as real-time data or small-sample expansion data into an SPC judgment engine; and the SPC judgment engine is configured to perform statistical process control analysis on the real-time data or the small-sample expansion data, to model the data cluster with dense distribution and stable parameters as a virtual reference surface representing a normal state domain, and to model the data point sequence showing continuous deviation or trend change as a virtual trend curve representing an abnormal evolution track; A judgment module is configured to determine a quality abnormal event in the production process by calculating the spatial intersection of the virtual reference surface and the virtual trend curve, and to generate a binary alarm signal corresponding to the quality abnormal event synchronously; A locking module is configured to forcibly bind the binary alarm signal and the operation permission of the corresponding production process, and to automatically send a locking instruction to the production process based on the binary alarm signal after the forced binding is completed, to prohibit the production process from continuing the processing operation; A release module is configured to start a controlled manual unlocking mechanism when the production process is locked and the processing operation is stopped, that is, to require the personnel with the corresponding permission to input the abnormal reason and analysis conclusion, and to automatically release the lock on the production process when the input abnormal reason and analysis conclusion are verified.

[0014] In a third aspect, a computing device includes a memory and a processor. One or more computer programs are stored in the memory, and the one or more computer programs include instructions; when the instructions are executed by the processor, the computing device executes the method according to the first aspect.

[0015] In a fourth aspect, a computer-readable storage medium is configured to store a computer program, and the computer program is configured to execute the method according to the first aspect.

[0016] The above scheme of the present application at least has the following beneficial effects: Because the SPC sensing node is arranged at the cutting tool edge or the contact point of the online measurement probe to obtain real-time accurate process parameters, the data is modeled into a virtual reference surface and a virtual trend curve by the SPC determination engine, and the spatial intersection point is solved to realize accurate abnormality determination, the binary alarm signal is forcibly bound with the process operation permission to automatically trigger the process locking, and the controlled manual unlocking mechanism is set to complete the technical means of closed-loop control, so that the core technical problems of the linkage fault of SPC alarm and process locking, insufficient accuracy of abnormality determination, and quality abnormality expansion caused by dependence on manual response in the prior art are effectively overcome, and the technical effects of improving the accuracy and timeliness of production process abnormality determination, preventing abnormal production from continuing from the source, reducing the generation of unqualified products and resource waste, realizing the automation and closed-loop of quality control, and improving the production efficiency and quality stability are achieved. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 It is a production control method flow diagram based on SPC alarm linkage process locking; Figure 2 It is a production control system diagram based on SPC alarm linkage process locking; Figure 3 It is a schematic diagram of a computing device. DETAILED DESCRIPTION

[0018] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0019] The embodiments of the present application propose a production control method based on SPC alarm linkage process locking, which comprises: Step 1, deploying SPC sensing nodes for collecting process parameters at the edge position of the machining tool of the production equipment in contact with the workpiece, or at the contact point of the probe of the online measurement equipment with the workpiece surface; through the SPC sensing nodes deployed at the edge position or the contact point, real-time process parameter data of the key production process is obtained; Step 2, inputting the process parameter data as real-time data or small sample expansion data into the SPC judgment engine; through the SPC judgment engine, statistical process control analysis is performed on the real-time data or small sample expansion data, so as to model the data cluster with dense distribution and stable parameters in the data as a virtual reference surface representing the normal state domain; the data point sequence showing continuous deviation or trend change in the data is modeled as a virtual trend curve representing the abnormal evolution track; Step 3, determining the quality abnormal event in the production process by calculating the spatial intersection of the virtual reference surface and the virtual trend curve, and synchronously generating a binary alarm signal corresponding to the quality abnormal event; Step 4, forcibly binding the binary alarm signal with the operation permission of the corresponding production process, and after the forced binding is completed, automatically sending a locking instruction to the production process based on the binary alarm signal to prohibit the production process from continuing the machining operation; Step 5, when the production process is locked and the machining operation is stopped, starting a controlled manual unlocking mechanism, that is, requiring personnel with corresponding permissions to input abnormal reasons and analysis conclusions, and when the input abnormal reasons and analysis conclusions are verified, the locking of the production process is automatically released.

[0020] In the embodiment of the application, by deploying SPC sensing nodes at the machining edge or the measurement contact point, accurate process parameter data can be obtained in real time; by constructing a virtual reference surface and a virtual trend curve with the help of an SPC judgment engine, the accuracy and timeliness of quality abnormality determination are effectively improved by calculating the spatial intersection; the binary alarm signal is forcibly bound with the process operation permission, which can automatically trigger the process locking and avoid the continuous abnormal production caused by the lag of manual response; in combination with the controlled manual unlocking mechanism, the abnormal problems can be effectively disposed, a complete quality control closed loop is formed, the production of unqualified products is reduced, the resource waste is reduced, and the stability and automation control level of the production process are improved.

[0021] In a preferred embodiment of the application, the above step 1 can include: Step 1.1, identifying the physical edge position of the processing tool in the key production process, or identifying the mechanical contact point of the on-line measuring device probe, specifically including: first, combing the process files of the key production process, clarifying the processing requirements and quality control points corresponding to the process, combining with the technical parameter manual of the production equipment or on-line measuring equipment, extracting the design reference coordinates and installation positioning requirements of the processing tool or probe; for the processing tool, an industrial high-resolution CCD camera is used to continuously scan along the length direction and thickness direction of the edge, ensuring that the scanning range completely covers all working areas of the edge; through gradient analysis of image gray value, the gray mutation feature points of the edge edge and the tool body are identified, and after eliminating the false feature points caused by surface reflection and stains, the effective feature points under multiple perspectives are cross-verified and fitted, specifically: the feature point coordinates collected under different perspectives are mapped to the device coordinate system, the abnormal points deviating from the overall distribution are removed through data consistency verification, and the remaining effective points are curve fitted to obtain the actual three-dimensional coordinates of the physical edge; for the on-line measuring equipment, a standard test piece with known size precision and suitable for the measuring range of the equipment is selected, the approach path is preset according to the device operation specification and the placement position of the standard test piece, the path is a straight line or a segmented moving path that avoids the edge burrs and irregular areas of the test piece, the probe is controlled to approach the test piece surface at a low speed, and the speed setting needs to avoid the influence of contact impact on the contact point determination accuracy; when the built-in force feedback sensor of the probe detects that the contact pressure reaches the preset threshold, the preset threshold is determined according to the measurement accuracy grade of the probe and the material characteristics of the test piece, the mechanical coordinates fed back by the device control end at this time are immediately recorded, so as to accurately determine the mechanical contact point position of the probe.

[0022] Step 1.2, deploying physical SPC sensing nodes at the physical edge position or mechanical contact point, specifically including: first, according to the space size and surrounding environment of the physical edge or mechanical contact point, selecting a micro physical SPC sensing node with a volume not exceeding 10 cubic centimeters and a protection level reaching IP67, and the measurement range needs to cover the normal fluctuation interval of the corresponding process parameters; then fix the sensing node at a position 5-10 mm away from the contact point, the mounting surface of the bracket needs to be ground to ensure that the error of the fit degree with the device surface is not more than 0.01 mm; for working conditions such as cutting fluid splashing and high temperature environment, a high-temperature resistant ceramic protective sleeve is installed for the physical SPC sensing node, a flexible sealing pad is filled between the sleeve and the node, and a high-temperature resistant sealant is used to seal the interface to prevent external medium from invading and affecting the sensing performance.

[0023] Step 1.3, the physical SPC sensor node is connected with the SPC judgment engine, and the communication protocol is matched, which specifically includes: the communication interface specification of the SPC judgment engine is determined, the stable communication connection with the physical SPC sensor node is established to ensure the reliability of the data transmission link; the communication protocol specification of the physical SPC sensor node is called to confirm the supported protocol type, and the compatible protocol list is called in the communication configuration interface of the SPC judgment engine, and the protocol type consistent with both sides is selected for matching; the core communication parameters such as transmission rate, data bit, stop bit and check bit are configured, wherein the transmission rate needs to consider the data real-time and transmission stability, and the check bit adopts the check mode which can effectively identify the transmission error; after the parameter configuration is completed, the test data packet is sent to the physical SPC sensor node, the response data fed back by the node is received and analyzed, the verification is carried out by checking the integrity and accuracy of data transmission, and the protocol matching is successful if there is no abnormality in continuous multiple tests.

[0024] Step 1.4, after the communication connection and protocol matching are completed, the physical SPC sensor node is registered as a controlled sensor node in the SPC judgment engine to generate the registered controlled sensor node, which specifically includes: the physical SPC sensor node automatically sends a registration request to the SPC judgment engine after starting, and the request contains device unique identifier, deployment location coordinates, measurement parameter type, measurement accuracy level, firmware version number and other key information; after the SPC judgment engine receives the request, it first queries the node management database, which is a core data storage unit for storing the basic information, permission configuration and running state of the accessed sensor node, and supports real-time query and update; through the node management database, it is confirmed that the device unique identifier is not registered repeatedly and meets the access permission requirements, and the matching of the deployment location coordinates and the key process is verified, and whether the measurement parameter type is within the support range of the judgment engine; after the verification is passed, the node is allocated a unique controlled identifier code which contains key distinguishing information such as process number and node type, and a node-specific communication encryption key is generated to ensure data transmission security; the registration information, configuration parameters, controlled identifier code and encryption key of the node and other data are entered into the node management database to establish a dynamically updated node state file, and the transformation from the physical SPC sensor node to the controlled sensor node is completed.

[0025] Step 1.5, based on the registered controlled sensing node, establish a real-time acquisition link for the process parameters; through the real-time acquisition link, the process parameter data is acquired in real time, specifically including: based on the communication parameters of the controlled sensing node and the controlled identification code, an independent data transmission channel is created in the SPC judgment engine, the channel uses a dedicated port number and an encryption key to ensure the security and independence of data transmission; according to the processing rhythm of the key production process, the acquisition frequency of the process parameters is set, the acquisition frequency is set to 100Hz for high-speed machining process and 50Hz for ordinary machining process, and the acquisition frequency can be dynamically adjusted through the configuration interface of the judgment engine; the controlled sensing node acquires temperature, pressure, vibration and other process parameter data in real time according to the set frequency, encapsulates the data into a standard data frame after internal AD conversion, and transmits the data to the SPC judgment engine through the established communication link; after the judgment engine receives the data, it first verifies the data integrity through the CRC32 check code, and after the verification is passed, the data is stored in the specified real-time database partition, and the data transmission state in the node state file is updated, so that each set of acquisition data can be accurately and in real time received and stored by the judgment engine.

[0026] In the embodiment of the application, by accurately identifying the physical cutting tool edge position or the online measurement equipment probe mechanical contact point, a physical SPC sensing node is directly deployed at the key contact position, and the communication connection establishment, protocol matching and controlled node registration process are matched, which effectively guarantees the stable linkage of the sensing node and the SPC judgment engine, and further constructs a reliable process parameter real-time acquisition link, which can acquire accurate process parameter data reflecting the real production state in real time, and provides strong data source support for the abnormal analysis and modeling of the SPC judgment engine, and improves the reliability and accuracy of the entire production control method from the acquisition source.

[0027] In a preferred embodiment of the application, the above step 2 can include: Step 2.1, process parameter data is continuously input to the SPC judgment engine as a real-time data stream through a real-time acquisition link, specifically including: the real-time acquisition link presets a fixed transmission frequency according to the processing rhythm of the key production process to ensure that data transmission is synchronized with the production process; the temperature, pressure, vibration, and other process parameter data collected by the physical SPC sensing node are packaged according to a unified standard; each frame of data contains a timestamp accurate to milliseconds, the unique identifier of the controlled sensing node, the type code of the specific process parameter, and the calibrated measured value, forming a standardized data frame with a unified structure; these data frames are transmitted to the dedicated data receiving port of the SPC judgment engine through the previously established stable communication link; during transmission, the real-time monitoring data receiving state is monitored to determine whether there is packet loss or data damage by checking the integrity identifier of the data frame; if an anomaly is detected, a retransmission request is immediately sent to the controlled sensing node, with a maximum of three retransmissions to ensure that each set of process parameter data is input to the SPC judgment engine in order and completeness according to the acquisition timing.

[0028] Step 2.2, the SPC judgment engine temporarily stores the real-time data stream in the internal data buffer queue, specifically including: after the SPC judgment engine receives the standardized data frame, it strictly follows the first-in-first-out rule to store the data frame in the internal data buffer queue; the initial capacity of the data buffer queue is set according to the historical data generation rate of similar production processes, and can be dynamically adjusted according to the actual data transmission volume in production; when the data input rate is consistently higher than the processing rate, the queue automatically expands to avoid data overflow; when the data volume is low for a long time, the queue shrinks to save system resources; the queue has a built-in data survival time control mechanism, with a preset survival time matching the typical processing cycle of the process, usually set to 5 minutes, and historical data exceeding this time will be automatically cleaned to ensure that the queue always stores the latest data to be processed; at the same time, the running state of the data buffer queue is monitored in real time, and when the remaining capacity of the queue is less than 20% of the total capacity, a capacity warning signal is triggered to prioritize processing of early-in-queue data, shorten the data waiting time in the queue, and ensure stable operation and data processing efficiency of the queue.

[0029] Step 2.3, real-time evaluation and judgment of the number and distribution state of real-time data in the data buffer queue by the SPC judgment engine, specifically including: first, real-time double evaluation of the data in the data buffer queue; statistics the total amount of data in the current queue, and compares it with the preset minimum analysis data amount threshold, which is determined based on the effective data amount statistical results of multiple historical analyses of the same type of production process, and can meet the basic statistical analysis requirements; then, numerical distribution analysis of the data in the queue is carried out, and the standard deviation and range of the data are calculated to determine whether the data is distributed within the reasonable numerical interval required by the process; at the same time, the data is divided into several equal segments according to the numerical range, and the data proportion in each segment is calculated, if the difference between the data proportions of each segment is not more than 10%, it is determined that the data is evenly distributed; combining the results of the two dimensions of whether the total amount of data meets the standard and whether the distribution is uniform, the conditions for effective statistical analysis of real-time data are comprehensively evaluated; the evaluation process is synchronized with the data entering the queue, and a new evaluation is performed every time a certain amount of data is added, to ensure that clear evaluation and judgment results are generated in a timely manner.

[0030] Step 2.4, when the real-time data meets the preset data sufficiency condition, the real-time data is directly used as the analysis data source; when the real-time data does not meet the preset data sufficiency condition, the data expansion process is triggered, that is, the real-time data is matched and fused with the preset historical database to generate small sample expansion data, which is used as the analysis data source, specifically including: the preset data sufficiency condition is that the total amount of data reaches the minimum analysis data amount threshold, and the data distribution is uniform, and the difference between the proportions of each segment is not more than 10%; when the real-time data in the data buffer queue meets both conditions, all real-time data is extracted from the queue and sorted by timestamp to be used as the data source for subsequent statistical process control analysis; when the real-time data does not meet any condition, the data expansion process is triggered immediately, and the preset historical database is called. The preset historical database is stored according to production process type, process parameter category, and working condition level, and records historical process parameter data of the same type of production process under different normal working conditions. All historical data has undergone effectiveness verification, denoising processing and standardization packaging in the early stage to ensure data quality; according to the process parameter type, production working condition environment parameter and processing stage characteristics of the current real-time data, similarity matching is carried out in the historical database; extract historical data with a matching degree higher than 85%, and supplement it to the real-time data in chronological order to form a continuous and complete data set; through this fusion method, data gaps are filled and data distribution is optimized, and finally small sample expansion data with sufficient data amount and reasonable distribution is generated, which is used as the analysis data source.

[0031] Step 2.5, performing statistical process control analysis on the analysis data source to model data clusters with intensive distribution and stable parameters in the analysis data source as a virtual reference surface representing a normal state domain, and model data point sequences presenting continuous deviation or trend change in the analysis data source as a virtual trend curve representing an abnormal evolution track.

[0032] In the embodiment of the present application, the process parameter data is continuously input into the SPC judgment engine and temporarily stored in the data buffer queue through real-time acquisition of the link, which can guarantee the continuity and stability of the data flow and provide a basis for subsequent analysis. Through real-time evaluation of the number and distribution state of the data, real-time data or small sample expansion data generated by matching and fusion of historical database are flexibly selected as the analysis data source, effectively solving the analysis deviation problem caused by insufficient data in the small sample scenario, and ensuring the sufficiency and reliability of the data source. On this basis, through statistical process control analysis, stable data clusters are modeled as virtual reference surfaces, and trend data sequences are modeled as virtual trend curves, which can accurately represent the normal state domain of the production process and effectively capture the abnormal evolution track of the parameters, improving the accuracy and foresight of the abnormal judgment and laying a foundation for accurate identification of quality abnormal events.

[0033] In a preferred embodiment of the present application, the above step 2.5 can include: Step 2.51, preprocessing the analysis data source to eliminate outliers and noise data therein to form a preprocessed analysis data source, specifically including: preprocessing to eliminate data interference and ensure data authenticity as the core target, first combining the process specifications of similar production processes with historical data statistics to determine the reasonable value range of each process parameter; based on the statistical distribution principle, calculating the mean and standard deviation of all data points in the analysis data source, and marking data points with values outside the range of mean plus or minus three times the standard deviation as outliers, which are usually caused by sensor instantaneous failure or external sudden interference and do not reflect the true production state; checking the marked outliers one by one to confirm that there is no reasonable cause and then directly eliminating them; for random noise in the data, a fixed length time window matching the production rhythm is set based on the sliding window smoothing principle, and the window size is determined according to the data acquisition frequency to ensure that each window contains enough data samples; the arithmetic mean of all data points in each window is calculated, and the arithmetic mean is used to replace the original data points in the window, and the local average processing weakens the influence of random fluctuations on the data; after the continuous operation of outlier elimination and noise smoothing, the preprocessed analysis data source with data fluctuations conforming to the actual production and without invalid interference components is obtained.

[0034] Step 2.52, based on the pre-processed analysis data source, identify and extract the data cluster with dense distribution and stable parameters, and identify and extract the data point sequence showing continuous offset or trend change, specifically including: based on the pre-processed analysis data source, using the principle of density clustering to identify stable data clusters, first set the data point space distance threshold and the adjacent point number threshold, the data point space distance threshold and the adjacent point number threshold are determined according to the normal fluctuation range of the process parameters; calculate the spatial distance of each data point with all other data points, count the number of adjacent data points of each data point within the set distance threshold, when the number of adjacent points of a certain data point reaches the set threshold, it is included in the candidate dense area; calculate the numerical fluctuation amplitude of all data points in the candidate dense area, quantify the fluctuation degree by the coefficient of variation, the coefficient of variation is the ratio of the standard deviation to the mean, when the coefficient of variation is lower than the pre-set critical value, it means that the data parameters in the dense area are stable, and it is determined as a data cluster with dense distribution and stable parameters and is extracted completely; at the same time, using the time series trend identification principle to extract the abnormal data point sequence, sorting the pre-processed data according to the time stamp, calculating the numerical difference between the continuous adjacent data points, setting the continuous difference direction consistency threshold and the cumulative offset threshold; when the difference values of continuous multiple data points always maintain the same positive or negative direction, and the cumulative sum of all difference values reaches the cumulative offset threshold, it is determined that the group of continuous data points form a trend change sequence, which is regarded as a data point sequence showing continuous offset or trend change.

[0035] Step 2.53, by multi-dimensional statistical fitting on the data cluster with dense distribution and stable parameters, model the virtual reference surface representing the parameter stable running interval, specifically including: the modeling of the virtual reference surface takes the normal state as the core of multi-dimensional parameter coordination, first, select data points from the stable data cluster, only keep the data points with a deviation of less than 1 standard deviation from the cluster mean value, and eliminate the edge points caused by accidental fluctuations to ensure data purity; divide the filtered data into training set and validation set in the ratio of 7:3, the training set is used to build the model, and the validation set is used to verify the accuracy throughout the process; set each key process parameter as an independent spatial coordinate axis, for example, temperature corresponds to X axis, pressure corresponds to Y axis, and vibration corresponds to Z axis, each data point in the training set is accurately mapped to a three-dimensional coordinate point in space according to the parameter values of each dimension; calculate the covariance matrix between each parameter dimension to determine the linear or nonlinear correlation degree between parameters, such as the positive correlation degree of temperature and pressure, and the weak correlation between vibration and temperature, to determine the parameter coordination constraint rule.

[0036] In combination with the correlation characteristics between parameters, an initial multi-dimensional space surface is constructed by using a multi-dimensional data point fitting method; by integrating the distribution characteristics and correlation rules of each dimension data, the basic form of the initial surface is determined, including the overall curvature trend, the initial boundary range of each dimension, and other core parameters; start the iteration training, and in each iteration, first substitute the coordinate points in the validation set into the current surface model one by one to determine whether each data point falls inside the surface; for normal data points falling outside the surface, record the direction and distance of the deviation from the surface: if deviating along the X-axis direction, expand the boundary of the surface in the X-axis direction by 1.2 times the deviation distance; if deviating along the Y-axis or Z-axis, adjust the boundary range of the corresponding axis according to the same logic; if the data points are excluded due to excessive local surface curvature, reduce the surface curvature in that area to make the surface fit the data distribution better; for edge noise points contained in the surface, shrink the local boundary by 0.8 times the distribution range of the noise point cluster to improve the accuracy of the model.

[0037] The iteration termination condition is set as that the pass rate of normal data in the validation set is stable at more than 98% for 3 consecutive iterations, and the adjustment amplitude of the surface parameters in each iteration is less than 0.001; during the iteration process, the cooperative constraint relationship of each parameter dimension needs to be recalculated after each parameter adjustment to ensure that the surface form always meets the inherent correlation rules between parameters; after the iteration is completed, the final surface parameters such as overall curvature and boundary range of each dimension are solidified to form a virtual reference surface, and the space region surrounded by the virtual reference surface is the stable operating interval of each process parameter combination.

[0038] Step 2.54, for the data point sequence showing continuous deviation or trend change, trend modeling is performed to model a virtual trend curve representing the deviation of the parameter from the control limit, which specifically includes: the modeling of the virtual trend curve focuses on accurately depicting the parameter abnormal evolution track, the trend data point sequence is strictly sorted according to the time stamp order to ensure that the time sequence logic is consistent with the actual production, and then the sequence is divided into a training set and a validation set according to an 8:2 ratio, the training set is used to fit the curve parameters, and the validation set is used to test the fitting effect and predict the reliability; the training set is subjected to trend type determination: the numerical change rate of each two consecutive data points is calculated, the standard deviation of all change rates is calculated, if the standard deviation is less than a preset threshold, such as 0.02, and the cumulative deviation amount is approximately proportional to the time, it is determined as a linear trend; if the standard deviation is greater than the threshold, or the cumulative deviation amount is not linearly related to the time, it is determined as a nonlinear trend.

[0039] For linear trend fitting training: build a linear relationship model with data point values as the dependent variable and timestamps as the independent variable; calculate the sum of squared deviations of each data point in the training set from the initial fitted straight line, and iteratively adjust the slope and intercept of the straight line using the gradient descent idea, setting the step size of each parameter adjustment to 0.01; during adjustment, if the sum of squared deviations is large, increase the adjustment amplitude of the slope or intercept in the direction of reducing the deviation; if the sum of squared deviations is close to the target value, reduce the adjustment amplitude until the sum of squared deviations decreases to a pre-set minimum value, such as 0.005, and the deviation change amount of continuous 2 iterations is less than 0.001; for non-linear trends, use the sliding window method to identify trend inflection points: set the window size to 5 data points, calculate the average of the change rate in each window, and when the average difference of the change rate of adjacent two windows exceeds a pre-set threshold such as 0.03, determine that the window boundary is a trend inflection point; divide the training set into multiple approximately linear segments according to the inflection points, and perform linear fitting on each segment to obtain the temporary fitting curve of each segment and the corresponding slope and intercept parameters.

[0040] Introduce a smooth transition coefficient to iteratively optimize the connection of each segment temporary curve: calculate the change rate, i.e. derivative, of the segment curves before and after the connection point; adjust the transition coefficient to keep the change rates before and after the connection point consistent, avoiding sudden inflection points; after each adjustment, substitute the validation set data into the complete curve, calculate the fitting deviation of the curve from the validation set data points, if the deviation is greater than a pre-set threshold such as 0.04, re-adjust the segment boundaries or transition coefficient; if the deviation is less than the threshold and there is no significant decrease in continuous 2 iterations, stop training; after training is completed, substitute the last 30% future data points of the validation set into the curve extension part to test the prediction accuracy, to ensure that the curve extension direction is consistent with the actual data trend; solidify the final curve parameters, including the slope and intercept of the linear trend, the segment parameters and transition coefficient of the non-linear trend, and finally generate a virtual trend curve, which can completely reproduce the parameter change trajectory that has occurred and accurately predict the future evolution direction.

[0041] In the embodiment of the present application, by preprocessing the outliers and noise data of the analysis data source, the data quality is effectively purified, providing a reliable data basis for modeling analysis; based on the preprocessed data, the data clusters with dense distribution and stable parameters are accurately identified and extracted, and the data point sequence showing continuous offset or trend change is realized, realizing the clear separation of normal state data and abnormal trend data; through multidimensional statistical fitting, the stable data cluster is modeled as a virtual reference surface, which clearly defines the normal parameter running interval of the production process, and at the same time, the abnormal data sequence is converted into a virtual trend curve through trend modeling, which intuitively presents the abnormal evolution trajectory of the parameter, and the two modeling results together improve the accurate description ability of the production state, effectively reducing the lag and misjudgment rate of abnormal judgment, providing core technical support for accurate identification of quality abnormal events, and helping to realize the forward-looking quality control of the production process.

[0042] In a preferred embodiment of the present application, step 3 can include: Step 3.1, obtaining the parametric equation of the virtual reference surface which has been modeled, and the trajectory equation of the virtual trend curve, specifically including: the parametric equation of the virtual reference surface is constructed based on the cooperative constraint relationship of multidimensional process parameters, the core of which is to map the normal distribution range of three-dimensional process parameters through two-dimensional independent parameters, and the specific equation form is X=p1u+p2v+p3uv+p 4， Y=q1u+q2v+q3uv+q4, Z=r1u+r2v+r3uv+r4; wherein X represents the temperature parameter, Y represents the pressure parameter, and Z represents the vibration parameter, which are the core process dimensions directly affecting the processing quality; u and v are two-dimensional independent control parameters of the surface, the value range of which is 0 to 1, u corresponds to the normal fluctuation interval mapping value of the temperature dimension, v corresponds to the normal fluctuation interval mapping value of the pressure dimension, and the combination of the two can cover the entire space range of the surface; p1 to p4, q1 to q4, r1 to r4 are the core coefficients of the surface, which are obtained by the surface iterative training process, p1 reflects the linear change weight of temperature with u, p2 reflects the linear change weight of temperature with v, p3 reflects the change weight of temperature with the cooperative action of u and v, and p4 is the reference offset of temperature, q series and r series coefficients correspond to the same change weight and reference offset of pressure and vibration parameters respectively, which together reflect the distribution rule and correlation characteristics of each process parameter in the normal domain.

[0043] The trajectory equation of the virtual trend curve is constructed based on the parameter time sequence evolution law, and the core is to describe the offset change of the process parameter with time, and the specific equation form is X(t)=a1t+a2t 2 +a3, Y(t)=b1t+b2t 2 +b3, Z(t)=c1t+c2t 2+ c3; wherein t is a time parameter, in seconds, ranging from the initial time when the trend curve starts to record to the preset future prediction time; X(t), Y(t), Z(t) represent the predicted values of temperature, pressure, vibration parameters at different times, respectively; a1 to a3, b1 to b3, c1 to c3 are core coefficients of the curve, which are obtained by curve iterative training process, a1 reflects the linear offset rate of temperature, a2 reflects the nonlinear offset acceleration of temperature, and a3 is the reference value of temperature at the initial time of the trend; the coefficients of the b series and the c series correspond to the same offset rate, acceleration and initial reference value of the pressure and vibration parameters, respectively, which completely characterize the evolution characteristics of the parameters over time.

[0044] During the acquisition process, the SPC judgment engine accesses the built-in model storage module, which is specially used to store various model parameters after training and verification. First, the virtual reference surface identifier and the virtual trend curve identifier corresponding to the current process are retrieved, and the corresponding surface coefficients p1 to p4, q1 to q4, r1 to r4 and curve coefficients a1 to a3, b1 to b3, c1 to c3 are called out according to the identifiers. The integrity and validity of the coefficients are checked one by one, and after confirming that there is no missing or error, the coefficients are substituted into the above preset equation forms, respectively, to finally generate the virtual reference surface parameterization equation and the virtual trend curve trajectory equation which can be directly used for calculation.

[0045] Step 3.2, based on the parameterization equation of the virtual reference surface and the trajectory equation of the virtual trend curve, a system of simultaneous equations for solving the spatial intersection point is constructed, which specifically includes: the core principle of constructing the system of simultaneous equations is that the spatial intersection point needs to satisfy two constraint conditions at the same time, that is, the intersection point is on the virtual reference surface and on the virtual trend curve, so the two equation systems need to be fused; first, it is clear that the trajectory equation X(t), Y(t), Z(t) of the virtual trend curve describes the dynamic trajectory of the parameter changing with time, and the parameterization equation of the virtual reference surface describes the static spatial range of the parameter under normal operation. The coordinates (X, Y, Z) corresponding to the intersection point are the critical state that the parameter is about to deviate from the normal range, and t is the specific time when the critical state occurs; based on this logic, X(t), Y(t), Z(t) in the trajectory equation of the virtual trend curve are substituted into X, Y, Z variables in the parameterization equation of the virtual reference surface, respectively, so that the parameter expressions in three dimensions are established at the same time, thereby constructing a system of simultaneous equations about the three unknown variables, and the specific form is: u 、 v 、 t a 1 t + a 2 t 2 +a 3 =p ​1 u + p 2 v + p 3 uv + p 4 ； b 1 t + b 2 t 2 +b 3 =q 1 u + q 2 v + q 3 uv + q 4 ； c 1 t + c 2 t 2 +c 3 =r 1 u+r 2 v+r 3 uv + r 4 ； All coefficients in the equation set are known quantities, wherein p 1 to p 4 , q 1 to q 4 , r 1 to r 4 is the solidification coefficient of the virtual reference surface, a 1 to a 3 , b 1 to b 3 , c 1 to c 3 is the solidification coefficient of the virtual trend curve, which has been verified for accuracy through prior training; the unknown variable is u , v ,t , u and v is the control parameter of the virtual reference surface, t is the time parameter, and the three together determine the spatial coordinates of the intersection point and the occurrence time; the essence of solving the equation set is to find one or more sets of u , v , t values, so that the calculation results on both sides of the three equations are completely equal, and the core of the solving process is to gradually approach the true solution of the unknown variable through numerical methods, and finally obtain the solution u 0 ，v 0 ，t 0 ) is substituted into the original equation, the process parameter coordinates of the intersection point X 0 ，Y 0 ，Z 0 ) can be directly converted, that is , , , and the coordinates and the corresponding t 0 together constitute the critical condition for the occurrence of quality abnormalities, providing quantitative basis for abnormality determination.

[0046] Step 3.3, the simultaneous equations are solved by numerical iterative calculation, wherein the solution includes setting the initial estimate of the intersection point coordinates, and simultaneously based on the simultaneous equations, iterative calculation is carried out, which specifically includes: using Newton-Raphson numerical iterative algorithm to solve the simultaneous equations, the core principle of Newton-Raphson numerical iterative algorithm is to approximate the Taylor expansion of the function, by continuously correcting the estimate of the unknown variable, gradually reducing the deviation between the calculation result and the true solution of the equation, and finally approaching the accurate value; first, set the initial estimate of the unknown variables u, v and t, the estimate setting needs to be based on the actual production scene and model characteristics to ensure reasonableness, the initial value of u is 0.5, corresponding to the middle position of the normal fluctuation range in the temperature dimension, the initial value of v is 0.5, corresponding to the middle position of the normal fluctuation range in the pressure dimension, and the initial value of t is half of the difference between the current production time and the starting time of the trend curve, while ensuring that the initial value of t is not less than the starting time of the trend curve and not greater than the preset maximum prediction time, to avoid the divergence of iteration caused by the deviation of the initial estimate from the reasonable search range.

[0047] The iterative calculation process is advanced step by step according to fixed logic. In the first step, the initial estimates (u0, v0, t0) are substituted into the three equations of the simultaneous equation set to calculate the difference between the left and right values of each equation. The difference is the residual of the equation, and the size of the residual reflects the deviation of the current estimate from the true solution. If the residual is zero, it means that the current estimate is the true solution. In the second step, the Jacobian matrix of the equation set is constructed. The Jacobian matrix is composed of the partial derivatives of each equation with respect to each unknown variable. Specifically, the derivative expressions of each equation with respect to u, v, and t are obtained by differentiation, and the initial estimates are substituted into the derivative expressions to obtain the specific matrix values. The Jacobian matrix reflects the influence of small changes in each unknown variable on the residual of each equation, providing a directional basis for variable correction. In the third step, the correction amount of the variable is calculated based on the Jacobian matrix and the residual. By solving the linear equation set whose product of the Jacobian matrix and the correction amount equals the negative residual, the correction amounts of u, v, and t are obtained. The sign of the correction amount determines the adjustment direction of the variable estimate, and the size of the correction amount determines the adjustment amplitude, ensuring that the adjusted residual can be reduced. In the fourth step, the calculated correction amounts are added to the current u, v, and t estimates to obtain new variable estimates (u1, v1, t1). The above process of substituting and calculating the residual, constructing the Jacobian matrix, solving the correction amount, and updating the estimate is repeated. Each iteration is based on the estimate of the previous round to optimize the variable estimate, making it gradually approach the true solution until the preset convergence condition is met.

[0048] Step 3.4, in the iterative calculation process, continuously judge whether the calculation result meets the preset convergence condition; when it is determined that the calculation result meets the preset convergence condition, the corresponding convergent solution is output as one or more candidate intersection point coordinates, specifically including: the setting of the convergence condition aims to ensure the accuracy and efficiency of the solution, specifically including two core judgment standards, one is the difference between the variable estimates obtained by the adjacent two iterations, and the other is the absolute value of the residual of the equation set after iteration; the threshold of the variable estimate difference is set to , that is, the difference between the previous and subsequent estimates of the three variables u, v, and t is less than the threshold, and the residual threshold is set to , that is, the absolute value of the residual of the three equations is less than the threshold; in the iterative calculation process, the variable difference and the residual are calculated synchronously after updating the variable estimate each time, and compared with the corresponding threshold; if only one of the two standards is met, iteration continues; if both standards are met, it is determined that the calculation result reaches the convergence state, the iteration stops, and the current convergent variable estimate (u0, v0, t0) is output as a group of solutions; according to the nonlinear characteristics of the equation set, there may be multiple convergent solutions, and all solutions that meet the convergence condition will be output in sequence according to the search path during the iteration process. Each solution corresponds to a group of candidate intersection point coordinates, that is, (X0, Y0, Z0) is obtained by substituting (u0, v0) into the surface equation, and the complete candidate intersection point coordinates (X0, Y0, Z0, t0) are formed by combining the corresponding t0.

[0049] Step 3.5, validity check is performed on all candidate intersection point coordinates; the validity check includes judging whether each candidate intersection point coordinate is located within the effective definition domain of the virtual reference curve, and judging whether the parameter offset moment corresponding to each candidate intersection point coordinate conforms to the evolution timing logic of the virtual trend curve, to obtain candidate intersection point coordinates that pass the validity check, specifically including: the core of the validity check is to screen out candidate intersection points that conform to the physical meaning and timing logic, to avoid invalid solutions affecting the determination result; first, it is judged whether the candidate intersection point coordinate is located within the effective definition domain of the virtual reference curve; the effective definition domain of the curve is jointly defined by the value range of parameters u and v and the reasonable interval of process parameters; the effective range of u and v is 0 to 1, and at the same time, the process parameters X, Y and Z need to be within the reasonable value interval set in step 2.51; if the u or v of the candidate intersection point exceeds the range of 0 to 1, or X, Y and Z exceed the reasonable interval, the intersection point is determined to be invalid; secondly, it is judged whether the parameter offset moment t0 corresponding to the candidate intersection point conforms to the evolution timing logic of the virtual trend curve; the timing logic requires that t0 needs to be greater than the start moment of the trend curve, and less than the preset maximum prediction time, and at the same time, t0 needs to be later than the current production moment, to ensure that the intersection point corresponds to a possible future abnormality, rather than a historical already-occurred state; if t0 does not conform to the above timing requirements, the intersection point is determined to be invalid; after the above two checks, all candidate intersection point coordinates that conform to the requirements are retained, and invalid intersection points are excluded, to obtain candidate intersection point coordinates that pass the validity check, to ensure the reliability of the subsequent determination result.

[0050] Step 3.6, the candidate intersection point coordinates that pass the validity check are determined as quality abnormal event points representing process out-of-control, i.e. the determination result of the quality abnormal event points is obtained, specifically including: the candidate intersection point coordinates that pass the validity check have the physical meaning that the intersection point of the parameter evolution trajectory represented by the virtual trend curve and the normal parameter running interval represented by the virtual reference curve; the intersection point means that when the production process runs to t0 moment, the process parameters will reach the state of (X0, Y0, Z0), which is exactly the boundary point of the normal running interval; after exceeding the boundary point, the parameters will deviate from the normal domain, and the production process enters the out-of-control state; based on this core logic, the candidate intersection point coordinates that pass the check are directly determined as quality abnormal event points representing process out-of-control; if there are multiple effective intersection points, they are sorted in the order of t0 from small to large, wherein the intersection point with the smallest t0 is the earliest possible quality abnormal event point, as the core determination result; the remaining intersection points are auxiliary warning information, which together constitute the complete quality abnormal event point determination result, and clearly indicate the specific moment and parameter state of the abnormality.

[0051] Step 3.7, generating a corresponding binary alarm signal according to the determination result of the quality abnormal event point, specifically including: the generation of the binary alarm signal is based on the determination result of the quality abnormal event point, adopts a standardized digital signal format, ensures the stability and compatibility of signal transmission, and sets the high level 1 of the binary signal to represent the existence of quality abnormal early warning, and the low level 0 represents the normal production state; if there is an effective quality abnormal event point in the determination result, and the earliest abnormal time t0 is within the preset early warning time window, the SPC determination engine immediately generates a high-level binary alarm signal, which contains the encoding of the core abnormal information, including the abnormal time t0 and the corresponding process parameter coordinates (X0, Y0, Z0), and the encoding adopts a fixed-length binary bit representation to ensure that the receiving end can accurately analyze; if there is no effective intersection in the determination result, or the earliest abnormal time t0 exceeds the early warning time window, a low-level binary alarm signal is generated, indicating that the current parameter evolution trend will not cause quality abnormality in the short term; the generated binary alarm signal is transmitted in real time to the subsequent process control flow through the established communication link, and a clear trigger signal is provided for process locking control.

[0052] In the embodiment of the application, by obtaining the parametric equation of the virtual reference surface and the trajectory equation of the virtual trend curve, a system of simultaneous equations is constructed to realize the quantitative association of the normal parameter stable domain and the abnormal evolution trajectory, which provides a mathematical modeling basis for accurate determination of quality abnormalities; by solving the system of simultaneous equations and setting the initial estimate and convergence condition, the stability and accuracy of the solving process are ensured, and the hysteresis defect of the traditional threshold judgment is effectively avoided; the candidate intersection points that meet the effective definition domain of the reference surface and the timing logic of the trend curve are screened out through effectiveness verification, which significantly reduces the risk of misjudgment and ensures the reliability of the determination result; the intersection points that pass the verification are determined as quality abnormal event points and a binary alarm signal is generated, realizing the closed-loop connection from abnormal trend prediction to accurate alarm, not only accurately positioning the specific time and state of the parameter exceeding the normal running interval, but also providing standardized trigger basis for the subsequent alarm and process locking linkage, effectively solving the core problems of insufficient accuracy of abnormal determination, disconnection of early warning and control in the prior art, and effectively improving the forward-looking management and control capability and response efficiency of quality abnormal events.

[0053] In a preferred embodiment of the application, the above-mentioned step 4 can include: Step 4.1, receiving the binary alarm signal issued by the SPC judgment engine, specifically including: continuously monitoring the signal output through the dedicated communication port, which is directly associated with the stable communication link established in advance to ensure the real-time and reliability of signal transmission; receiving the binary alarm signal transmitted in the preset standardized format, the signal contains a state identification field and an abnormal information coding field, after receiving, first perform signal integrity check, through detecting the start bit, stop bit and check bit of the signal, confirm that the signal has not occurred transmission packet loss or damage; then analyze the state identification field in the signal, distinguish the high level monitoring 1 corresponding to the abnormal early warning state and the low level monitoring 0 corresponding to the normal state, if it is an abnormal early warning state, further extract the key data in the abnormal information coding, provide the basis for the association operation, the whole receiving process strictly follows the preset communication protocol, ensures the accuracy of signal analysis.

[0054] Step 4.2, based on the pre-set device and process mapping relationship, associate the binary alarm signal to one or more specific production equipment and corresponding processing procedures, specifically including: built-in device and process mapping relationship database, the mapping relationship is based on production process planning, the corresponding association rules between production equipment and processing procedures established with process layout, clearly record the unique identification of each device corresponding to one or more processing procedure number, procedure name, at the same time mark the process association attribute and priority of the device, ensure the corresponding relationship between the device and the process is clear and traceable, and support dynamic update according to production adjustment; the device and process mapping relationship database pre-stores all the unique identification of the device, corresponding processing procedure number, procedure name, process association attribute and priority information; the core of the association process is to extract the key identification related to production, such as process number or device association code, from the abnormal information coding in the alarm signal; according to the extracted key identification, accurately query in the device and process mapping relationship database, match the corresponding device unique identification and processing procedure information; if the query result corresponds to multiple devices or processes, then according to the preset association priority in the device and process mapping relationship database, preferentially lock the core device and main processing procedure directly involved in the abnormal parameter, ensure the targeting of the association result, avoid irrelevant devices or processes being misassociated.

[0055] Step 4.3, for one or more specific production equipment associated, by checking the current operation permission state of the production equipment at the device control level, to obtain a check result, specifically including: for each target obtained, send a permission state query instruction to its control unit, the instruction contains a unique identifier for precise positioning query target; the control unit receiving the instruction calls the current operation permission state data stored by itself, the operation permission state is set based on the management rules of the control level, mainly divided into two categories of lockable state and non-lockable state, wherein the non-lockable state includes the case of being in emergency maintenance, critical task execution or being controlled by other high-priority instructions; the control unit feeds back the current permission state to the execution end, and the execution end checks the feedback information, and judges whether there is an emergency procedure that must be completed or a processing link that cannot be interrupted, in combination with the production task progress corresponding to the target; comprehensive permission state data and actual production task situation, finally form a clear check result, clearly mark whether each target has the condition of being locked.

[0056] Step 4.4, when the check result confirms that the operation permission of the production equipment is in a lockable state, then generate a corresponding device lock instruction according to the check result, specifically including: when the check result shows that the target operation permission is in a lockable state, start the lock instruction generation process; the generation logic is based on the preset locking rule and object characteristics, wherein the preset locking rule is an instruction generation criterion formulated according to production safety specifications, abnormal influence range and process requirements, which clearly indicates the locking mode and effectiveness requirement corresponding to different abnormal levels; the object characteristics refer to the inherent properties of the target device, such as processing type, running mode, core function and process correlation degree, which directly determine the scope of application and execution logic of the lock instruction; the lock instruction contains core information, including target unique identifier, lock type, emergency lock or regular lock, lock effective time, lock range, overall lock or specific process operation lock and unlock condition prompt; wherein the lock type is determined according to the severity of the abnormality, the emergency lock is suitable for high-risk situations where the parameters are about to exceed the control limit and need to be terminated immediately, and the regular lock is suitable for non-emergency abnormalities and allows the processing step being executed to be stopped after completion; the device lock instruction is generated and packaged in a standardized format after generation, ensuring that the control unit can accurately parse, and at the same time records the instruction generation log, including generation time, instruction content and corresponding alarm signal information.

[0057] Step 4.5, the device locking instruction is sent to the preset target production device, and it is synchronized whether the preset target production device returns an execution confirmation feedback, specifically including: through the established communication link, the encapsulated locking instruction is sent to the control unit of the preset target, the sending process adopts an acknowledgment transmission mechanism, and it is ensured that the instruction is successfully received; at the same time, a feedback monitoring process is started, a fixed timeout time is set, the time is determined according to the response speed and production scene demand, and usually does not exceed 3 monitoring seconds; within the timeout time, the feedback signal of the target is continuously monitored, if the returned signal is received within the specified time, it is preliminarily determined that the instruction has been received, if the feedback is not received within the timeout time, a retry mechanism is automatically started, the locking instruction is resent, at most 2 monitoring times are retried, if the feedback is still not received, communication exception information is recorded and a prompt is sent; in the monitoring process, the instruction sending state is updated in real time, and the whole transmission process is ensured to be controlled.

[0058] Step 4.6, when the execution confirmation feedback returned by the preset target production device and corresponding to the device locking instruction is received, it is confirmed that the inhibition operation of the machining process is completed, specifically including: receiving the execution confirmation feedback signal generated after the target control unit executes the locking instruction, the signal contains key information such as unique identification, locking execution time, current state, and locked; after receiving the feedback signal, the key information in the feedback signal is analyzed, it is verified whether the feedback corresponding identification and the target in the locking instruction are consistent, and it is confirmed whether the locking execution time is within a reasonable range; then the current state and the process state are inquired, it is verified that the target has stopped the related machining operation and cannot start a new machining task, and it is ensured that the locking instruction has actually taken effect; after confirming that there is no error, the state archives and process execution records stored by itself are updated, it is marked that the machining process corresponding to the target has been inhibited, and the locking completion information is synchronized to the production management terminal, the related staff is notified to handle the exception in time, and the whole process forms a closed loop control, so that the effective implementation of the process inhibition operation is ensured.

[0059] In the embodiment of the application, by receiving the standardized binary alarm signal issued by the monitoring SPC monitoring determination engine, and combining the preset device and process mapping relationship, the accurate association of the alarm signal with the specific production device and processing process is realized, the disconnection of the alarm information and the execution link is avoided, and the targeting of the locking operation is ensured; by checking the current operation permission state of the production device, the risks of false locking, invalid locking, etc. are effectively avoided, and the reliability and rationality of the process locking are improved; based on the effective verification result, the device locking instruction is generated and sent to the target device, and the execution confirmation feedback is monitored and synchronized, a complete closed-loop control logic of alarm, association, verification, locking monitoring and confirmation is constructed, the traditional passive early warning is converted into active forced control; this process effectively solves the core defects of the monitoring SPC monitoring alarm and process locking linkage fault in the prior art, and the dependence on manual response leads to the expansion of abnormality, realizes the forced connection of quality abnormality early warning and process execution control, can quickly terminate the risk process at the initial stage of abnormality, and maximizes the reduction of quality loss.

[0060] In a preferred embodiment of the application, step 5 can include: Step 5.1, after confirming that the processing process is prohibited from operation, triggering and activating the controlled manual unlocking interface, specifically including: after confirming that the processing process has been completely prohibited from operation by executing the confirmation feedback, automatically triggering the activation process of the manual unlocking interface, the triggering action is linked with the state signal of the process locking completion, ensuring that the unlocking process is started only after the locking state takes effect; the manual unlocking interface is in a dormant and unavailable state by default, and will be switched to an interactive mode after activation, and the interface will clearly display key information such as the unique identifier of the locked production device, the corresponding processing process number and the locking time, helping authorized personnel quickly locate the operation object; the controlled characteristics of the interface are guaranteed by the permission verification mechanism, and only personnel in the preset permission list have access to the access portal, and unauthorized personnel will be prompted that they have no operation permission when clicking the interface, and the activation process will record the triggering time, the associated locking event number and the interface state in real time, and all information is stored in the set event traceability database.

[0061] Step 5.2, receiving the abnormal reason text and analysis conclusion text input by personnel with preset operation permission through the manual unlocking interface, specifically including: after the manual unlocking interface is activated, a special interactive interface is automatically popped up, the interface is designed by function partition, and independent input boxes for abnormal reason text and analysis conclusion text are set, and clear filling instructions are marked; personnel need to complete permission verification first, and the verification method is double verification of identity and operation permission level, identity can be completed by work number input combined with biological identification, operation permission level needs to match the abnormal handling permission, only the checkers can enter the text input link; abnormal reason text needs to fill in the specific production link of quality abnormality occurrence, the name of the process parameter involved, the parameter offset value and the preliminary judgment of the inducing factor, analysis conclusion text needs to clarify the specific measures of abnormal handling, the responsible department of problem rectification, the time limit for rectification completion and the follow-up prevention scheme; the interface supports real-time text saving function, and the data is automatically backed up to cache every time a certain number of characters are input, to avoid data loss due to operation interruption, and the submit button needs to be clicked after the input is completed to trigger the verification process.

[0062] Step 5.3, the compliance and integrity of the abnormal reason text and analysis conclusion text are checked, specifically including: the compliance and integrity check is based on the preset standardized check rules, the compliance check focuses on checking whether the text content meets the requirements of production safety specification and quality control system, confirming that there is no description of violating process logic, no illegal operation description, and ensuring that the text uses standard technical language, and does not contain invalid characters or interference information; the integrity check is checked one by one according to the preset list of mandatory elements, the abnormal reason text needs to include the core elements of abnormal occurrence time, involved parameter name, parameter offset range, affected production batch, and the analysis conclusion text needs to include specific processing measures, responsibility attribution, rectification time limit, and prevention mechanism; the checking process is pushed forward according to the logic of checking field by field, and the missing mandatory elements or irregular expressions are accurately marked, and the specific rectification prompt is fed back on the interface, such as needing to supplement the missing abnormal occurrence time, needing to modify the illegal expression, if all elements meet the requirements and the expression is regular, it is determined that the verification is passed.

[0063] Step 5.4, when the compliance and integrity check passes, a pass result is obtained; based on the pass result, a corresponding process unlocking instruction is generated, specifically including: when the compliance and integrity check passes, an explicit pass result is automatically generated, which contains the check time, the check personnel identifier, the text audit conclusion and the associated lock event number; based on the pass result, a process unlocking instruction generation process is started, and the generation logic strictly follows the preset unlocking rule to ensure that the instruction and the previous locking instruction are for the same production equipment and processing procedure; the process unlocking instruction contains core information, including the unique identifier of the preset target production equipment, the corresponding processing procedure number, the unlocking authorization code, the unlocking effective time and the unlocking permission validity period, wherein the unlocking authorization code is generated by encrypting the device identifier and the check result to ensure the uniqueness and security of the instruction; after the instruction is generated, it is packaged in a standardized format recognizable by the production equipment control unit, and after packaging, it is stored in association with the previous lock event record, the text entry data and the check result in the set event database to form a complete traceability chain.

[0064] Step 5.5, send the process unlocking instruction to the preset target production equipment, and synchronously monitor whether the preset target production equipment returns an execution confirmation feedback, specifically including: through the stable communication link established in the early stage, the packaged process unlocking instruction is sent to the control unit of the preset target production equipment, and the sending process adopts a confirmation transmission mechanism, that is, after sending the instruction, the device returns a receiving confirmation signal immediately to ensure that the instruction is not lost; at the same time of sending the instruction, a feedback monitoring process is started, and the timeout time is set according to the type of production equipment, the response speed of complex processing equipment is slower, and the timeout time is set to 5 seconds for monitoring, and the timeout time is set to 3 seconds for monitoring for simple equipment; the feedback signal returned by the equipment is continuously monitored within the timeout time, if the receiving confirmation signal sent by the equipment is received within the specified time, the instruction receiving state is recorded as successful; if the feedback is not received within the timeout time, a retry mechanism is automatically started, the first retry interval is 1 second for monitoring, the second retry interval is 2 seconds for monitoring, and the maximum retry is 2 times for monitoring; if the feedback is still not received after two retries, the communication abnormal information is recorded, including the abnormal occurrence time, the equipment identifier and the instruction content, which are synchronized to the production management terminal to prompt the staff to troubleshoot the communication failure.

[0065] Step 5.6, when receiving the return of the preset target production equipment and the execution confirmation feedback of the process unlocking instruction, the unlocking of the processing process is confirmed to be completed, specifically including: receiving the execution confirmation feedback signal returned by the control unit of the preset target production equipment, the signal containing device unique identifier, unlocking execution time, current process state code, unlocking authorization code verification result and other key information; when parsing the feedback signal, first verify whether the device identifier is consistent with the target device in the unlocking instruction, confirm that the unlocking authorization code verification result is passed, and the unlocking execution time is within a reasonable range; then send a state query instruction to the device control unit to verify whether the current running mode of the device has been restored to an operable state and whether the corresponding processing process has the condition to start a new production task; after confirming that the unlocking instruction actually takes effect, update the device state archives, process execution records and unlocking event logs stored by itself, supplement the unlocking operator, unlocking time and unlocking basis and other information in the archives, ensure that all data are associated with the previous locking event and text record; finally, synchronize the unlocking completion information to the production management terminal, the terminal pops up an unlocking success prompt pop-up window, displays the device identifier, process number and unlocking state, notifies the relevant staff that normal production can be resumed, and finally completes the entire unlocking closed loop.

[0066] In the embodiment of the present application, the controlled manual unlocking interface is triggered after the processing process is prohibited, the authorized subject of the unlocking operation is limited, abnormal risks caused by random unlocking of irrelevant personnel are avoided, the safety and standardization of process unlocking are ensured, personnel with preset operation permission are required to input abnormal reason text and analysis text, the abnormal traceability and analysis responsibility are implemented, data support is provided for process optimization and prevention of similar problems, the comprehensiveness and standardization of abnormal analysis are ensured through compliance and integrity verification of the text, and the problem is avoided after unlocking. Repeated; based on the verification result, the process unlocking instruction is generated and the execution confirmation feedback is monitored, the complete unlocking closed loop of locking, analysis, verification, unlocking and confirmation is constructed, the orderly connection of abnormal handling and process recovery is realized, and the defects of no control of unlocking operation, non-standard abnormal analysis and no basis for problem tracing in the prior art are solved. Under the premise of monitoring and ensuring production safety and product quality, the flexibility of process recovery and the systematicness of abnormal handling are considered, and the quality control forms a virtuous cycle of discovery, handling, monitoring and optimization.

[0067] The production control system based on SPC alarm linkage process locking comprises: A deployment module is used to deploy SPC sensing nodes for collecting process parameters at the edge position of the processing tool of the production equipment in contact with the workpiece or at the contact point of the probe of the online measurement equipment with the surface of the workpiece; through the SPC sensing nodes deployed at the edge position or the contact point, the process parameter data of the key production process are obtained in real time; The modeling module is configured to input the process parameter data as real-time data or small sample expansion data into the SPC judgment engine; the SPC judgment engine is configured to perform statistical process control analysis on the real-time data or the small sample expansion data, so as to cluster data sets with dense distribution and stable parameters in the data into a virtual reference surface representing a normal state domain; and the SPC judgment engine is configured to model data point sequences presenting continuous offset or trend change in the data into a virtual trend curve representing an abnormal evolution track. The judgment module is configured to determine a quality abnormal event in the production process by calculating a spatial intersection point of the virtual reference surface and the virtual trend curve, and to synchronously generate a binary alarm signal corresponding to the quality abnormal event. The locking module is configured to forcibly bind the binary alarm signal with an operation permission of a corresponding production process, and to automatically send a locking instruction to the production process based on the binary alarm signal after the forcible binding is completed, so as to prohibit the production process from continuing the processing operation. The unlocking module is configured to start a controlled manual unlocking mechanism when the production process is locked and the processing operation is stopped, that is, to require a person with a corresponding permission to input an abnormal reason and an analysis conclusion, and to automatically unlock the production process when the input abnormal reason and analysis conclusion are verified.

[0068] The production control system according to the embodiment of the present application can correspond to the method described in the embodiment of the present application, and the above and other operations and / or functions of each module of the production control system are respectively used to realize Figure 1 the corresponding flow of the method in the illustrated embodiment, which will not be described herein again for the sake of brevity.

[0069] The embodiment of the present application further provides a computing device. The computing device can be an application server.

[0070] As Figure 3 shown, the figure is a schematic diagram of a computing device provided by the embodiment of the present application. The computing device 700 includes a bus 701, a processor 702, a communication interface 703 and a memory 704. The processor 702, the memory 704 and the communication interface 703 communicate through the bus 701.

[0071] The bus 701 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 3 only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0072] The processor 702 can be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP), etc.

[0073] The communication interface 703 is configured to communicate with the outside. The memory 704 can include a volatile memory, such as a random access memory (RAM), and / or a non-volatile memory, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD), or a solid-state drive (SSD). The memory 704 stores executable code, and the processor 702 executes the executable code to perform the above-mentioned production control method based on the SPC alarm linkage process locking.

[0074] Specifically, in the case of implementing the above-mentioned embodiments of the production control system based on the SPC alarm linkage process locking, and in the case of implementing each module or unit of the production control system based on the SPC alarm linkage process locking by software, the software or program code required to execute the functions of each module / unit in the production control system based on the SPC alarm linkage process locking described in the above-mentioned embodiments can be stored partially or entirely in the memory 704. The processor 702 executes the program code corresponding to each unit stored in the memory 704 to perform the above-mentioned production control method based on the SPC alarm linkage process locking.

[0075] The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium can be any available medium or a data center containing one or more available media that a computing device can store. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid-state disk), etc. The computer-readable storage medium includes instructions that instruct the computing device to perform the above-mentioned production control method based on the SPC alarm linkage process locking.

[0076] The embodiments of the present application further provide a computer program product including one or more computer instructions. When the computer instructions are loaded and executed on a computing device, the computer instructions produce, in whole or in part, the processes or functions described in the embodiments of the present application.

[0077] The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website site, computer or data center to another website site, computer or data center through wired (for example, coaxial cable, optical fiber, digital subscriber line) or wireless (for example, infrared, wireless, microwave, etc.).

[0078] The computer program product is executed by a computer, and the computer executes any one of the production control methods based on the SPC alarm linkage process locking procedure described above. The computer program product can be a software installation package, and when any one of the production control methods based on the SPC alarm linkage process locking procedure described above is needed, the computer program product can be downloaded and executed on the computer.

[0079] The above describes the preferred embodiments of the present application. It should be noted that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A production control method based on SPC alarm linkage process locking, characterized by, The method comprises: deploying SPC sensing nodes for collecting process parameters at the cutting edge position where the machining tool of the production equipment contacts the workpiece, or at the contact point of the probe of the online measurement equipment and the surface of the workpiece; acquiring process parameter data of a key production process in real time through the SPC sensing nodes deployed at the cutting edge position or the contact point; inputting the process parameter data as real-time data or small-sample expansion data into an SPC judgment engine; performing statistical process control analysis on the real-time data or small-sample expansion data through the SPC judgment engine, so as to model data clusters with dense distribution and stable parameters in the data as virtual reference surfaces representing normal state domains; modeling data point sequences presenting continuous offset or trend changes in the data as virtual trend curves representing abnormal evolution trajectories; determining quality abnormal events in the production process by calculating the spatial intersection of the virtual reference surfaces and the virtual trend curves, and synchronously generating a binary alarm signal corresponding to the quality abnormal events; forcibly binding the binary alarm signal and the operation permission of the corresponding production process, and automatically sending a locking instruction to the production process based on the binary alarm signal after the forced binding is completed, so as to prohibit the production process from continuing machining operation; starting a controlled manual unlocking mechanism when the production process is locked and machining operation is stopped, that is, requiring personnel with corresponding permissions to input abnormal reasons and analysis conclusions, and automatically releasing the lock on the production process when the input abnormal reasons and analysis conclusions are verified.

2. The SPC alarm linkage process locking-based production control method according to claim 1, characterized by, deploying SPC sensing nodes for collecting process parameters at the cutting edge position where the machining tool of the production equipment contacts the workpiece, or at the contact point of the probe of the online measurement equipment and the surface of the workpiece; acquiring process parameter data of a key production process in real time through the SPC sensing nodes deployed at the cutting edge position or the contact point, including: identifying the physical cutting edge position of the machining tool in the key production process, or identifying the mechanical contact point of the probe of the online measurement equipment; deploying physical SPC sensing nodes at the physical cutting edge position or the mechanical contact point; establishing a communication connection between the physical SPC sensing nodes and the SPC judgment engine, and simultaneously performing communication protocol matching; after the communication connection and protocol matching are completed, registering the physical SPC sensing nodes as controlled sensing nodes in the SPC judgment engine, and generating the registered controlled sensing nodes; based on the registered controlled sensing nodes, establishing a real-time acquisition link for process parameters; acquiring process parameter data in real time through the real-time acquisition link.

3. The SPC alarm linkage process locking-based production control method according to claim 2, characterized by, inputting the process parameter data as real-time data or small-sample expansion data into an SPC judgment engine; performing statistical process control analysis on the real-time data or small-sample expansion data through the SPC judgment engine, so as to model data clusters with dense distribution and stable parameters in the data as virtual reference surfaces representing normal state domains; modeling data point sequences presenting continuous offset or trend changes in the data as virtual trend curves representing abnormal evolution trajectories, including: inputting the process parameter data as a real-time data stream through the real-time acquisition link to the SPC judgment engine continuously; the SPC judgment engine temporarily stores the real-time data stream in an internal data buffer queue; The SPC judgment engine is used to evaluate and judge the number and distribution state of real-time data in the data buffer queue in real time. When it is judged that the real-time data meet the preset data sufficiency condition, the real-time data are directly used as the analysis data source; when it is judged that the real-time data do not meet the preset data sufficiency condition, a data expansion process is triggered, that is, matching and fusion are performed based on the real-time data and the preset historical database to generate small sample expansion data, and the small sample expansion data are used as the analysis data source. The analysis data source is subjected to statistical process control analysis, so as to model a data cluster with dense distribution and stable parameters in the analysis data source as a virtual reference surface representing a normal state domain, and model a data point sequence with continuous deviation or trend change in the analysis data source as a virtual trend curve representing an abnormal evolution trajectory.

4. The SPC alarm linkage process locking-based production control method according to claim 3, characterized by, The analysis data source is subjected to statistical process control analysis, so as to model a data cluster with dense distribution and stable parameters in the analysis data source as a virtual reference surface representing a normal state domain, and model a data point sequence with continuous deviation or trend change in the analysis data source as a virtual trend curve representing an abnormal evolution trajectory, including: The analysis data source is preprocessed to remove outliers and noise data, thereby forming a preprocessed analysis data source; Based on the preprocessed analysis data source, a data cluster with dense distribution and stable parameters is identified and extracted, and a data point sequence with continuous deviation or trend change is also identified and extracted; The data cluster with dense distribution and stable parameters is subjected to multidimensional statistical fitting to model a virtual reference surface representing a parameter stable operation interval; The data point sequence with continuous deviation or trend change is subjected to trend modeling processing to model a virtual trend curve representing a parameter deviation from a control limit.

5. The SPC alarm linkage process locking-based production control method according to claim 4, characterized by, A quality abnormal event in a production process is judged by calculating the spatial intersection point of the virtual reference surface and the virtual trend curve, and a binary alarm signal corresponding to the quality abnormal event is generated synchronously, including: The parametric equation of the modeled virtual reference surface and the trajectory equation of the virtual trend curve are obtained; Based on the parametric equation of the virtual reference surface and the trajectory equation of the virtual trend curve, a system of simultaneous equations for solving the spatial intersection point is constructed; The system of simultaneous equations is solved, wherein the solving includes setting an initial estimate of the intersection point coordinates, and performing iterative calculation based on the system of simultaneous equations; During the iterative calculation process, it is continuously judged whether the calculation result meets the preset convergence condition; when it is judged that the calculation result meets the preset convergence condition, the corresponding convergence solution is output as one or more candidate intersection point coordinates; All candidate intersection point coordinates are subjected to effectiveness verification; the effectiveness verification includes judging whether each candidate intersection point coordinate is located within the effective definition domain of the virtual reference surface, and judging whether the parameter deviation time corresponding to each candidate intersection point coordinate meets the evolution timing logic of the virtual trend curve, so as to obtain the candidate intersection point coordinates that pass the effectiveness verification; The candidate intersection point coordinates that pass the effectiveness verification are judged as quality abnormal event points representing process out-of-control, thereby obtaining the judgment result of the quality abnormal event points. According to the determination result of the quality abnormal event point, a corresponding binary alarm signal is generated.

6. The SPC alarm-linked process lockout-based production control method according to claim 5, characterized by, The binary alarm signal is forcibly bound with the operation permission of the corresponding production process. After the forced binding is completed, a lock instruction is automatically sent to the production process based on the binary alarm signal to prohibit the production process from continuing the processing operation, including: Receiving the binary alarm signal sent by the SPC determination engine; Based on the pre-set device and process mapping relationship, the binary alarm signal is associated with one or more specific production devices and the corresponding processing process; For the associated one or more specific production devices, the current operation permission state of the production device at the device control level is checked to obtain a checking result; When the checking result confirms that the operation permission of the production device is in a lockable state, a corresponding device lock instruction is generated according to the checking result; The device lock instruction is sent to the pre-set target production device, and whether the pre-set target production device returns an execution confirmation feedback is monitored synchronously; When the execution confirmation feedback of the device lock instruction returned by the pre-set target production device is received, it is confirmed that the prohibition operation of the processing process is completed.

7. The SPC alarm-linked process lockout-based production control method according to claim 6, characterized by, When the production process is locked and the processing operation is stopped, a controlled manual unlocking mechanism is started, that is, personnel with corresponding permissions are required to input abnormal reason and analysis conclusion. When the input abnormal reason and analysis conclusion are verified, the lock of the production process is automatically released, including: After confirming that the processing process is prohibited from operating, a controlled manual unlocking interface is triggered and activated; Through the manual unlocking interface, the abnormal reason text and analysis conclusion text input by personnel with pre-set operation permission are received; The abnormal reason text and analysis conclusion text are checked for compliance and integrity; When the compliance and integrity check is passed, a check pass result is obtained; a corresponding process unlocking instruction is generated based on the check pass result; The process unlocking instruction is sent to the pre-set target production device, and whether the pre-set target production device returns an execution confirmation feedback is monitored synchronously; When the execution confirmation feedback of the process unlocking instruction returned by the pre-set target production device is received, it is confirmed that the lock of the processing process is released.

8. A production control system based on SPC alarm linkage procedure locking, characterized by, The system executes the method of any one of claims 1 to 7, including: A deployment module is used to deploy SPC sensing nodes for collecting process parameters at the cutting edge position where the processing tool of the production device contacts the workpiece, or at the contact point between the probe of the online measurement device and the surface of the workpiece; through the SPC sensing nodes deployed at the cutting edge position or the contact point, the process parameter data of the key production process is obtained in real time; A modeling module is used to input the process parameter data into the SPC determination engine as real-time data or small sample expansion data; through the SPC determination engine, statistical process control analysis is performed on the real-time data or small sample expansion data to model the data cluster with dense distribution and stable parameters as a virtual reference surface representing the normal state domain; the data point sequence showing continuous deviation or trend change is modeled as a virtual trend curve representing the abnormal evolution track; A determination module is configured to determine a quality abnormal event in a production process by calculating a spatial intersection between a virtual reference surface and a virtual trend curve, and to generate a binary alarm signal corresponding to the quality abnormal event synchronously; A locking module is configured to forcibly bind the binary alarm signal with an operation permission corresponding to a production procedure, and to automatically send a locking instruction to the production procedure based on the binary alarm signal to prohibit the production procedure from continuing processing operation after the binding is completed; A release module is configured to start a controlled manual release mechanism when the production procedure is locked and stops processing operation, i.e., to require a person with a corresponding permission to input an abnormal reason and an analysis conclusion, and to automatically release the lock on the production procedure when the input abnormal reason and analysis conclusion are verified.

9. A computing device, comprising: comprises a memory and a processor; wherein one or more computer programs are stored in the memory, and the one or more computer programs comprise instructions; when the instructions are executed by the processor, the computing device performs the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium is configured to store a computer program, and the computer program is configured to perform the method of any one of claims 1 to 7.

Images