Dynamic key area process inspection method and system based on real-time data driving

Abstract

The invention provides a dynamic key area process inspection method and system based on real-time data driving, and relates to the field of process inspection management of equipment production, and the method comprises the steps: dividing to-be-inspected parts into different quality batches according to a parameter value range to which process behavior data belongs; in the same quality batch, a mapping relation between the process behavior data and a machining area of the target part is constructed according to part machining positions corresponding to a plurality of collection timestamps of the process behavior data of the target part produced in the same time period; constructing a target geometric region set based on the mapping relation and fluctuation characteristics of the process behavior data; and determining a defect risk score corresponding to each target geometric region in the target geometric region set according to the risk prediction model, and further determining a target key region of process inspection. By implementing the method, it can be ensured that the inspection resources are more accurately input into a high-risk area, and the use efficiency of the resources in process inspection is improved.

Description

Technical Field

[0001] This application relates to the field of process inspection management in equipment production, and particularly to a dynamic key area process inspection method and system driven by real-time data. Background Art

[0002] With the continuous deepening of the intelligent transformation of the manufacturing industry, quality inspection in the production and manufacturing process is gradually changing from the traditional static manual inspection mode to an intelligent and automated dynamic inspection mode. In modern manufacturing, product quality inspection has an important impact on the production efficiency and product qualification rate of manufacturing enterprises. Especially in the field of high-precision parts manufacturing, timely detection and prevention of quality defects have become the key link to ensure product quality.

[0003] In the related art, the common practice in industrial production is to determine the key areas and non-key areas of inspection based on the product structure characteristics and functional requirements in the part design stage. Designers will divide the product into different inspection areas and formulate corresponding inspection procedures according to static factors such as the geometric characteristics and stress distribution of the parts. In the actual inspection process, inspectors invest more inspection resources in the key areas according to the area division plan established in the design stage, while adopting a lower-frequency sampling inspection method for the non-key areas.

[0004] However, this way of fixing the inspection area in the design stage has limitations in practical applications. Since the requirements in the inspection procedures are pre-established, the investment of inspection resources often adopts a relatively fixed method. In the actual production process, parts produced in different batches and at different time periods may have different quality conditions. The unified inspection procedure is difficult to flexibly adjust the inspection resources according to the actual situation, which easily causes the problem of low utilization efficiency of inspection resources. Summary of the Invention

[0005] This application provides a dynamic key area process inspection method and system driven by real-time data to address the problem of how to improve the utilization efficiency of inspection resources in process inspection.

[0006] In a first aspect, this application provides a dynamic key area process inspection method driven by real-time data, which is applied to an equipment inspection control system. The method includes: Dividing the parts to be inspected into different quality batches according to the parameter value range to which the process behavior data corresponding to the parts to be inspected belongs. The process behavior data is collected during the production process of the parts to be inspected by multiple types of sensors deployed on the production equipment. Different quality batches correspond to different processing area division rules; In the same quality batch, a mapping relationship between the process behavior data and the processing area of the target part is established according to the process behavior data of the target part produced in the same time period and the part processing positions corresponding to the multiple acquisition time stamps of the process behavior data, wherein the processing area includes multiple processing positions; Based on the mapping relationship, a target processing area whose fluctuation characteristics of process behavior data corresponding to the acquisition time stamp exceed a preset threshold is selected from multiple processing areas corresponding to the target part to construct a target geometric area set, wherein the fluctuation characteristics include fluctuation frequency, fluctuation amplitude and fluctuation trend; Inputting the process behavior data corresponding to each target geometric area in the target geometric area set into a preset risk prediction model, and outputting the defect risk score corresponding to each target geometric area, wherein the risk prediction model is trained based on historical process behavior data and corresponding historical defect risk scores; The target geometric area whose defect risk score exceeds a preset risk threshold is determined as a target critical area for process inspection for resource allocation.

[0007] Through the above embodiments, the system dynamically divides quality batches and key inspection areas through real-time collected process behavior data, breaking through the limitations of fixed inspection areas in traditional inspection methods. The system can intelligently identify the areas that really need to be inspected based on the actual data fluctuations in the production process, and dynamically allocate inspection resources accordingly. This data-driven dynamic inspection method ensures that inspection resources are more accurately invested in high-risk areas, and to a certain extent avoids excessive inspection of low-risk areas, thereby improving the efficiency of the use of inspection resources.

[0008] In some embodiments, before the step of dividing the parts to be inspected into different quality batches according to the parameter value range to which the process behavior data corresponding to the parts to be inspected belongs, the step further includes: Calculate the parameter fluctuation value corresponding to the process behavior data according to the difference between the process behavior data corresponding to all the parts to be inspected and the preset ideal process behavior data; The parameter fluctuation value is divided into a plurality of parameter fluctuation ranges based on a preset ratio to obtain a plurality of corresponding parameter value ranges, and the processing area division granularities corresponding to different parameter value ranges are not all the same.

[0009] Through the above embodiments, the system calculates the degree of deviation between the process behavior data and the ideal data, and establishes a mechanism for dividing different processing areas into different granularities corresponding to different parameter value ranges. This multi-level division method based on the actual processing parameter fluctuations makes the division of the inspection area more targeted, and can select the appropriate inspection granularity according to the actual processing status of different batches of products, avoiding the waste of resources caused by over-coarse or over-fine division.

[0010] In some embodiments, after the step of dividing the parameter fluctuation values into multiple parameter fluctuation ranges based on a preset ratio to obtain corresponding multiple parameter value ranges, the method further includes: Based on the machining feature dimensions and machining precision requirements of the part, construct multi-level fine-grained division levels for the machining area, and each of the fine-grained division levels corresponds to a different area division size; Set preset parameter ranges corresponding to each of the fine-grained division levels based on the distribution characteristics of the parameter fluctuation values in the historical machining data; Calculate the overlapping ratio of the parameter fluctuation values within each of the parameter value ranges in the preset parameter ranges; Determine the fine-grained division level corresponding to the preset parameter range with the largest overlapping ratio as the fine-grained division level corresponding to the parameter value range.

[0011] Through the above embodiments, the system constructs multi-level fine-grained division levels based on the machining feature dimensions and precision requirements of the part, and determines the optimal division level through historical data analysis. This adaptive area division method not only considers the technical requirements of the product itself but also combines the distribution characteristics of the actual production data, making the area division more scientific and reasonable and capable of better guiding the allocation of inspection resources.

[0012] In some embodiments, before the step of constructing the mapping relationship between the process behavior data and the machining area of the target part according to the process behavior data of the target part produced in the same time period and the machining positions of the parts corresponding to multiple acquisition timestamps of the process behavior data in the same quality batch, the method further includes: Obtain the production trajectory data of the target part, where the production trajectory data includes the movement trajectory and machining speed of the machining tool of the production equipment relative to the target part; Calibrate the machining positions of the target part corresponding to different acquisition timestamps based on the production trajectory data to obtain the calibrated machining positions of the part.

[0013] Through the above embodiments, the system realizes the precise calibration and calibration of the machining positions of the parts by collecting and analyzing the movement trajectory data of the production equipment. This dynamic position tracking mechanism ensures the precise correspondence between the process behavior data and the actual machining area, providing a reliable data basis for subsequent area division and risk assessment.

[0014] In some embodiments, the step of constructing a target geometric area set by selecting, based on the mapping relationship, a target machining area whose fluctuation characteristics of the process behavior data corresponding to the acquisition timestamp exceed a preset threshold from multiple machining areas corresponding to the target part specifically includes: Based on the mapping relationship, for each of the one or more consecutive acquisition timestamps corresponding to each processing area, construct a process behavior data sequence corresponding to each of the consecutive acquisition timestamps; When it is detected that the fluctuation characteristic of any process behavior data sequence corresponding to the processing area exceeds a preset threshold, mark the processing area as a target processing area; Construct a target geometric area set based on all the target processing areas.

[0015] Through the above embodiments, the system establishes a method for analyzing the fluctuation characteristics of process data based on time series. By monitoring the data fluctuation conditions in different areas, potential high-risk areas are accurately identified. This continuous data monitoring mechanism can timely capture abnormal fluctuations in the processing process and effectively prevent the occurrence of quality problems.

[0016] In some embodiments, after the step of determining the target key area of the process inspection for resource allocation by taking the target geometric area whose defect risk score exceeds the preset risk threshold, the following steps are further included: Obtain the candidate key areas and non-key areas determined in the design stage of the part to be inspected; Adjust the first spatial overlapping part of the target key area and the non-key area to be the candidate key area.

[0017] Through the above embodiments, the system integrates and optimizes the dynamically identified target key area with the area determined in the design stage, and establishes an inspection area determination mechanism that takes into account both design requirements and actual production conditions. This method not only ensures that necessary design requirements are met, but also can flexibly adjust the inspection strategy according to actual production situations.

[0018] In some embodiments, after the step of adjusting the first spatial overlapping part of the target key area and the non-key area to be the candidate key area, the following steps are further included: Divide the area priorities corresponding to different target geometric areas in the candidate key area according to the risk assessment results of the part to be inspected in the design stage, and the area priority corresponding to the first spatial overlapping part is the lowest; Adjust the area priority according to the second spatial overlapping part of the target key area and the candidate key area, so that the area priority corresponding to the second spatial overlapping part is the highest.

[0019] Through the above embodiments, the system introduces a dynamic adjustment mechanism for area priorities. By analyzing the spatial overlapping relationships of different types of key areas, intelligent allocation of inspection priorities is achieved. This method of priority division based on multi-dimensional analysis ensures that inspection resources can cover high-risk areas to the greatest extent, further improving the inspection efficiency.

[0020] In a second aspect, the present application provides an equipment inspection control system, the equipment inspection control system comprising: one or more processors and a memory; The memory is coupled to the one or more processors, and the memory is used to store computer program code, and the computer program code includes computer instructions. The one or more processors call the computer instructions so that the equipment inspection control system can implement a dynamic critical area process inspection method based on real-time data driving provided in the above embodiment, which will not be repeated here.

[0021] In a third aspect, the present application provides a computer-readable storage medium comprising instructions. When the instructions are executed on an equipment inspection control system, the equipment inspection control system can implement a dynamic critical area process inspection method based on real-time data drive provided in the above-mentioned embodiment, which will not be repeated here.

[0022] In a fourth aspect, the present application provides a computer program product. When the computer program product runs on an equipment inspection control system, the equipment inspection control system can implement a dynamic critical area process inspection method based on real-time data drive provided in the above embodiment, which will not be repeated here.

[0023] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By collecting real-time process behavior data, a dynamic quality batch division and key area identification mechanism has been established. The system can intelligently identify areas that need to be inspected and dynamically allocate inspection resources in real time based on actual data fluctuations in the production process, ensuring that inspection resources are accurately invested in high-risk areas, avoiding excessive inspection of low-risk areas, and not missing new potential risk areas, fundamentally improving the efficiency of inspection resource utilization.

[0024] 2. By establishing the corresponding relationship between the process parameter fluctuation value and the processing area division granularity, combined with the part processing feature size, precision requirements and historical data distribution characteristics, a multi-level fine-grained adaptive area division system is constructed. This mechanism achieves accurate position calibration through production trajectory data, and continuously monitors data fluctuation characteristics based on time series analysis, so that the division of the inspection area not only meets the technical requirements, but also can be flexibly adjusted according to the actual production status, providing a reliable spatial positioning basis for accurate inspection.

[0025] 3. The dynamically identified target key areas are intelligently integrated with the preset areas in the design stage, and a dynamic adjustment mechanism for inspection priorities based on multi-dimensional analysis is established by analyzing the spatial overlap relationships of different types of key areas. This integration and optimization strategy not only ensures the satisfaction of design requirements but also can flexibly adjust the inspection strategy according to the actual production situation, maximizing the accuracy of inspection coverage and the efficiency of resource utilization. Brief Description of the Drawings

[0026] Figure 1 is a flowchart of a method for dynamically inspecting key areas in the process based on real-time data in an embodiment of the present application; Figure 2 is another flowchart of a method for dynamically inspecting key areas in the process based on real-time data in an embodiment of the present application; Figure 3 is a schematic structural diagram of an entity device of an equipment inspection control system in an embodiment of the present application. Detailed Embodiments

[0027] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular forms "a", "an", "the", "above-mentioned", "the foregoing", "this" are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations including one or more of the listed items.

[0028] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and should not be construed as implying or indicating relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0029] For ease of understanding, the method provided in this embodiment is described in terms of a process below. Please refer to Figure 1 , which is a flowchart of a method for dynamically inspecting key areas in the process based on real-time data in an embodiment of the present application.

[0030] S101. Divide the parts to be inspected into different quality batches according to the parameter value ranges of the process behavior data corresponding to the parts to be inspected.

[0031] Among them, process behavior data refers to the physical parameters collected by various types of sensors deployed on production equipment during the production process of the parts to be inspected. These data reflect the process status during the part production process, including but not limited to temperature, vibration, pressure, etc.

[0032] This step is executed before the system conducts in-process inspection of the parts and is applied to the quality inspection scenario of mass-produced parts on the production line. By dividing the quality batches, more targeted inspection plans can be formulated according to the characteristics of different batches.

[0033] Specifically, the equipment inspection control system first receives the process behavior data of the parts to be inspected collected by various types of sensors deployed on the production equipment, and then analyzes these data to determine the parameter value range to which each data belongs. According to different parameter value ranges, the parts to be inspected are divided into different quality batches. For example, when the temperature parameter value range during the production of a certain type of part is A, the produced parts are divided into quality batch 1; when the temperature parameter value range is B, the produced parts are divided into quality batch 2. Different quality batches correspond to different processing area division rules.

[0034] In some embodiments, the quality batches can be divided according to the process behavior data in various ways: Optionally, the system pre-sets multiple parameter value range intervals, and compares the collected process behavior data with these intervals. If the data falls within a certain interval, the corresponding parts are divided into the quality batch corresponding to that interval. For example, it is set that the temperature value range [80 - 90]°C corresponds to quality batch A, and [91 - 100]°C corresponds to quality batch B. When the temperature data of a certain part during production is detected to be 85°C, this part is divided into quality batch A.

[0035] Optionally, use a clustering algorithm to analyze the process behavior data. First, import the process behavior data of all parts to be inspected into the clustering algorithm model. The algorithm can automatically divide the data into different clusters according to the similarity between the data, and each cluster corresponds to a quality batch. For example, through the K-means clustering algorithm, according to the comprehensive similarity of multiple process behavior data such as temperature and pressure, the parts are divided into 3 quality batches, and each batch corresponds to different process characteristics.

[0036] S102. In the same quality batch, construct a mapping relationship between the process behavior data and the processing area of the target part according to the process behavior data of the target part produced in the same time period and the part processing positions corresponding to multiple acquisition timestamps of the process behavior data.

[0037] Among them, the acquisition timestamp is used to mark the specific time point of the acquisition of process behavior data. Through the timestamp, the machining progress of the part at the time of data acquisition can be determined; the part machining position represents the specific machining position of the target part on the production equipment at different acquisition time points, and is used to determine the spatial position where the process behavior occurs; the machining area is a set composed of multiple part machining positions, representing different machining parts in the part production process.

[0038] Specifically, the equipment inspection control system collects the process behavior data of the target part in the same quality batch within the same time period, as well as the corresponding acquisition timestamps and part machining position information of these data. Then, the process behavior data under each timestamp is associated with the corresponding part machining position, and the data with the same or similar machining positions are grouped into one machining area, thereby establishing a mapping relationship between the process behavior data and the machining area of the target part. For example, in a certain quality batch, the temperature data collected for the target part at 10:00 is 85°C, and the part machining position at this time is point A. At 10:05, the temperature data collected is 86°C, and the machining position is point B. If points A and B are close, these two data can be associated with the corresponding positions and built into the mapping relationship of the same machining area.

[0039] In some embodiments, the mapping relationship can be established in multiple ways: Optionally, first establish an empty mapping table, and the columns of the table are process behavior data, acquisition timestamp, and part machining position respectively. Then traverse all the process behavior data records of the target part in the same quality batch, and fill the data in each record into the mapping table according to the column definition. Finally, group the data in the mapping table according to the proximity of the part machining positions, and each group corresponds to one machining area, thus completing the construction of the mapping relationship. For example, set the distance between adjacent positions less than 1 cm as being close, and group the mapping table data according to this rule.

[0040] S103. Based on the mapping relationship, select the target machining areas whose fluctuation characteristics of the process behavior data corresponding to the acquisition timestamps exceed the preset threshold from the multiple machining areas corresponding to the target part to construct the target geometric area set.

[0041] After successfully establishing the mapping relationship between the process behavior data and the machining area (i.e., step S102 is completed), when it is necessary to identify the machining areas that may have quality risks, this step is executed.

[0042] Specifically, based on the mapping relationship established in S102, for each processing area of the target part, the equipment inspection control system obtains the corresponding process behavior data and its acquisition timestamp. Then, it analyzes the fluctuation characteristics of these data (including fluctuation frequency, fluctuation amplitude, and fluctuation trend), and compares the fluctuation frequency, fluctuation amplitude, and fluctuation trend with the corresponding preset thresholds. When any one of the fluctuation characteristics of the process behavior data corresponding to a certain processing area exceeds the preset threshold, that processing area is marked as the target processing area. Finally, all the marked target processing areas are integrated to construct a set of target geometric areas. For example, if the preset fluctuation amplitude threshold is 5°C and the fluctuation amplitude of the temperature data in a certain processing area reaches 6°C, then that processing area is determined as the target processing area, and all such areas form the set of target geometric areas.

[0043] It should be noted that the process behavior data includes multiple physical parameters. When the fluctuation characteristics corresponding to any physical parameter exceed the corresponding preset threshold, the processing area corresponding to that physical parameter is determined as the target processing area.

[0044] S104: Input the process behavior data corresponding to each target geometric area in the set of target geometric areas into a preset risk prediction model, and output the defect risk score corresponding to each target geometric area.

[0045] Specifically, the equipment inspection control system obtains the set of target geometric areas obtained in S103, and for each target geometric area in the set, extracts the corresponding process behavior data. Then, these process behavior data are sequentially input into the preset risk prediction model. The risk prediction model analyzes and processes the input data according to the relationship between the process behavior data and the defect risk learned through pre-training. Finally, it outputs the defect risk score corresponding to each target geometric area. For example, if the process behavior data corresponding to the target geometric area A shows large temperature fluctuations and unstable pressure, the risk prediction model combines the association between similar process behaviors and the occurrence of defects in historical data and gives the defect risk score of this area as 80 points (out of 100), indicating a relatively high defect risk in this area.

[0046] It should be noted that when constructing the risk prediction model, the system collects a large amount of historical process behavior data collected by various types of sensors during the production process, including parameters such as temperature and pressure, and records the corresponding historical defect risk scores. Then, the data is preprocessed to handle missing values and normalize or standardize it to eliminate data scale differences. Select an appropriate model according to the data characteristics, such as a multi-layer perceptron neural network for handling complex non-linear relationships. After dividing the data into a training set and a test set, start training. Calculate the predicted values through forward propagation, calculate the error using a loss function, and then update the weights using backpropagation and gradient descent methods to make the model fit the data. Then, evaluate the model using the test set and calculate metrics such as accuracy and mean square error to measure the performance. Optimize according to the evaluation results, such as increasing the data volume in case of overfitting and increasing the model complexity in case of underfitting until the model performance meets the expectations.

[0047] In some embodiments, the output of the defect risk score using the risk prediction model can be achieved in various ways: Optionally, a neural network model is used as the risk prediction model. First, preprocess the process behavior data of the target geometric region to standardize or normalize the data to meet the input requirements of the neural network. Then, input the processed data into the trained neural network model, and the neurons in the model perform layer-by-layer calculations and transmissions on the data according to the pre-trained weights and thresholds. Finally, obtain the corresponding defect risk score through the output layer of the model. For example, using a multi-layer perceptron (MLP) neural network, the input layer receives the process behavior data, the hidden layer performs feature extraction and complex non-linear transformations, and the output layer gives the defect risk score.

[0048] S105. Determine the target key area for process inspection of the target geometric region whose defect risk score exceeds the preset risk threshold, and allocate resources.

[0049] Specifically, the equipment inspection control system obtains the defect risk scores of each target geometric region obtained in step S104, and at the same time reads the preset risk threshold set in advance. Compare the defect risk score of each target geometric region with the preset risk threshold one by one. When the defect risk score of a certain target geometric region is greater than the preset risk threshold, determine that target geometric region as the target key area for process inspection. After determining the target key area, the system allocates inspection resources to these areas according to the pre-established resource allocation strategy. For example, the preset risk threshold is 60 points. If the defect risk score of target geometric area B is 75 points, then area B is determined as the target key area, and the system can arrange more inspection personnel, more advanced detection equipment, and a longer inspection time to inspect this area.

[0050] Optionally, the system can utilize intelligent optimization algorithms for resource allocation. For example, using the genetic algorithm, first, the information of the target key areas (such as defect risk scores, area locations, etc.) is encoded to form an initial population. Then, a fitness function is defined to evaluate the quality of each individual (i.e., the resource allocation scheme). The fitness function can comprehensively consider factors such as inspection costs and inspection effects. Next, the population is continuously optimized through genetic operations such as selection, crossover, and mutation, and finally, the optimal resource allocation scheme is obtained. For example, after multiple iterations, the genetic algorithm determines that for a certain high-risk target key area, 3 experienced inspectors and 2 high-precision detection devices are allocated to achieve the best inspection effect.

[0051] In the above embodiment, the system dynamically divides quality batches and key inspection areas through real-time collected process behavior data, breaking through the limitation of fixed inspection areas in traditional inspection methods. The system can intelligently identify the areas that truly need to be focused on for inspection based on the actual data fluctuations during the production process, and accordingly perform dynamic allocation of inspection resources. This data-driven dynamic inspection method ensures that inspection resources are more precisely invested in high-risk areas, and to a certain extent, avoids over-inspection of low-risk areas, thereby improving the utilization efficiency of inspection resources.

[0052] The following provides a further and more specific process description of the method provided in this embodiment. Please refer to Figure 2 , which is another process schematic diagram of a dynamic key area process inspection method based on real-time data drive in the embodiment of the present application.

[0053] S201. Calculate the parameter fluctuation value corresponding to the process behavior data based on the difference between the process behavior data corresponding to all parts to be inspected and the preset ideal process behavior data.

[0054] Among them, the preset ideal process behavior data is preset and represents the data standard that the process behavior should present during the production process of parts under ideal production conditions, and is used to measure the deviation of actual production data.

[0055] This step is executed before dividing the quality batches according to the process behavior data of the parts to be inspected, and is applicable to production scenarios with high requirements for production process stability and product quality consistency. By calculating the parameter fluctuation value, the fluctuation situation of the production process can be initially evaluated, providing a data basis for subsequent quality batch division and inspection area division.

[0056] Specifically, the equipment inspection control system obtains the process behavior data corresponding to all parts to be inspected, and at the same time calls the preset ideal process behavior data. For each piece of process behavior data, it is compared with the corresponding preset ideal process behavior data, and the difference between the two is calculated. For example, if the actual temperature data during the production of a part to be inspected is 90°C and the preset ideal temperature data is 85°C, then the fluctuation value of this temperature parameter is 90 - 85 = 5°C. Calculate the differences corresponding to all process behavior data to obtain the parameter fluctuation values corresponding to the process behavior data, and these fluctuation values reflect the differences between the actual production process and the ideal state.

[0057] S202. Divide the parameter fluctuation values into multiple parameter fluctuation ranges based on a preset ratio to obtain corresponding multiple parameter value ranges.

[0058] The equipment inspection control system divides the parameter fluctuation values obtained in step S201 according to a preset ratio set in advance. For example, the preset ratio is to divide the parameter fluctuation values into three intervals. The fluctuation value in the range of 0 - 2 is one interval, 2 - 5 is one interval, and greater than 5 is one interval. These three intervals are the parameter fluctuation ranges. Then, according to these parameter fluctuation ranges, determine the parameter value ranges corresponding to the process behavior data. For example, for the temperature parameter, when the parameter fluctuation value is in the 0 - 2 interval, the corresponding actual temperature parameter value range may be 83 - 87°C; when the parameter fluctuation value is in the 2 - 5 interval, the corresponding temperature parameter value range may be 80 - 90°C, etc. At the same time, set different machining area division granularities for different parameter value ranges. For example, for the parameter value range with small fluctuations (corresponding to the parameter fluctuation value of 0 - 2), a larger division granularity can be adopted, and the machining area is divided into fewer but larger areas; for the parameter value range with large fluctuations (corresponding to the parameter fluctuation value greater than 5), a smaller division granularity is adopted, and the machining area is divided into more and smaller areas for more detailed inspection.

[0059] Optionally, the system can configure the preset ratio and the corresponding parameter value range rules in the rule engine. For example, it is set that when the parameter fluctuation value is less than or equal to 2, the lower limit of the corresponding parameter value range is the ideal value - 3, and the upper limit is the ideal value + 3; when the parameter fluctuation value is greater than 2 and less than or equal to 5, the lower limit of the corresponding parameter value range is the ideal value - 5, and the upper limit is the ideal value + 5, etc. Then, input the calculated parameter fluctuation value into the rule engine, and the rule engine automatically divides the parameter fluctuation range according to the preset rules and determines the corresponding parameter value range. Finally, store the division results in the database for subsequent use.

[0060] In the above embodiments, the system established a mechanism where different ranges of parameter values correspond to different granularities of machining area division by calculating the deviation degree between the process behavior data and the ideal data. This multi-level division method based on the fluctuations of actual machining parameters makes the division of the inspection area more targeted. It can select an appropriate inspection granularity according to the actual machining status of products in different batches, avoiding resource waste caused by overly coarse or overly fine division.

[0061] S203. Based on the machining feature dimensions and machining accuracy requirements of the part, construct a multi-level fine-grained division hierarchy for the machining area.

[0062] The equipment inspection control system obtains the machining feature dimensions and machining accuracy requirement information of the part. For example, for a precision shaft part, its machining feature dimensions may include the diameter and length of the shaft, and the machining accuracy requirement may be that the diameter tolerance is within ±0.01 mm. Based on this information, the system constructs a multi-level fine-grained division hierarchy for the machining area. Suppose three levels are constructed. Level 1 corresponds to a larger area division size and is applicable to areas with relatively low machining accuracy requirements and small changes in machining feature dimensions; the division size of Level 2 is the second; Level 3 corresponds to the smallest area division size and is used for areas with extremely high machining accuracy requirements and critical machining feature dimensions. For the shaft part, the non-critical parts of the shaft may be divided into Level 1, and the machining area is divided according to a larger size range; the critical parts such as the areas near the bearing fit are divided into Level 3 and divided according to a smaller size range to ensure more detailed inspection of the machining accuracy of the critical parts.

[0063] S204. Based on the distribution characteristics of the parameter fluctuation values in the historical machining data, set the preset parameter range corresponding to each fine-grained division level.

[0064] Specifically, the equipment inspection control system collects and collates historical processing data and extracts the parameter fluctuation values therein. Then, statistical analysis is performed on the parameter fluctuation values to calculate their distribution characteristics, such as calculating statistical quantities such as the mean value and standard deviation, and observing the distribution pattern. For example, it is found that the parameter fluctuation values of a certain type of part in a certain processing area show a normal distribution, with a mean value of 3 and a standard deviation of 1. According to these distribution characteristics and in combination with the fine-grained division levels constructed in step S203, a preset parameter range is set for each level. For level 1 (coarser division granularity), the preset parameter range can be set as the mean value ± 2 times the standard deviation, that is, 1 - 5; for level 2, it is set as the mean value ± 1.5 times the standard deviation, that is, 1.5 - 4.5; for level 3 (finest division granularity), it is set as the mean value ± 1 times the standard deviation, that is, 2 - 4. In this way, different fine-grained division levels correspond to different preset parameter ranges, enabling a more accurate judgment of the quality status of the processing area based on the comparison between the actual process behavior data and the preset parameter ranges during the subsequent inspection process.

[0065] S205. Calculate the overlapping proportion of the parameter fluctuation values within each parameter value range in the preset parameter ranges; determine the fine-grained division level corresponding to the preset parameter range with the largest overlapping proportion as the fine-grained division level corresponding to the parameter value range.

[0066] This step is executed after completing the construction of the multi-level fine-grained division levels of the processing area (S203), setting the preset parameter ranges corresponding to each fine-grained division level (S204), and determining the parameter value ranges (S202).

[0067] Specifically, for each parameter value range determined in step S202, the equipment inspection control system obtains all the parameter fluctuation values within that range. Then, these parameter fluctuation values are compared with the preset parameter ranges set for each fine-grained division level in step S204. Taking a certain parameter value range as an example, assume there are 100 parameter fluctuation values within this range, among which 60 fall within the preset parameter range of level 1, 30 fall within the preset parameter range of level 2, and 10 fall within the preset parameter range of level 3. Then, the overlapping proportions of this parameter value range with the preset parameter ranges of level 1, level 2, and level 3 are 60%, 30%, and 10% respectively. The system calculates the overlapping proportions of all parameter value ranges with each preset parameter range, and determines the fine-grained division level corresponding to the preset parameter range with the largest overlapping proportion as the fine-grained division level corresponding to this parameter value range. For example, if for a certain parameter value range, the overlapping proportion with the preset parameter range of level 2 is the largest, then level 2 is determined as the fine-grained division level corresponding to this parameter value range, which can make the regional division granularity more in line with the current production actual situation.

[0068] In the above embodiments, the system constructs a multi-level fine-grained division level based on the processing feature dimensions and precision requirements of the parts, and determines the optimal division level through historical data analysis. This adaptive region division method not only considers the technical requirements of the product itself but also combines the distribution characteristics of the actual production data, making the region division more scientific and reasonable and capable of better guiding the allocation of inspection resources.

[0069] S206. Obtain the production trajectory data of the target part; calibrate the machining positions of the target part corresponding to different acquisition timestamps based on the production trajectory data to obtain the calibrated machining positions of the part.

[0070] The equipment inspection control system obtains the production trajectory data of the target part through the communication interface or data storage system of the production equipment. These data may exist in the form of files, database records, or real-time data streams. Then, the system reads the motion trajectory and machining speed information in these production trajectory data. For example, obtain the changes in the X, Y, and Z-axis coordinates of the machining tool relative to the target part and the corresponding speed data within a certain period of time. Then, according to the acquisition timestamp, associate the production trajectory data with the part machining position data. Since factors such as equipment vibration and measurement errors may affect the accuracy of the part machining position data in actual production, the system calibrates the machining positions of the part corresponding to different acquisition timestamps based on the motion trajectory and speed information in the production trajectory data using specific algorithms. For example, correct the measured machining positions of the part according to the change rules of the machining speed and motion trajectory. Finally, obtain the calibrated machining positions of the part, and these position information is more accurate, laying a foundation for constructing an accurate mapping relationship between process behavior data and machining regions.

[0071] Optionally, the system can also utilize sensor fusion technology and algorithm compensation. First, install multiple sensors on the production equipment, such as position sensors, speed sensors, etc., to collect the production trajectory data of the target part in real time. Then, fuse and process the data collected by these sensors to remove noise and outliers. For example, use the Kalman filter algorithm to fuse the data of multiple sensors to improve the accuracy of the data. Then, based on the fused data, combined with the pre-established equipment motion model and error model, calibrate the part machining position data. For example, compensate and adjust the measured machining positions of the part by analyzing factors such as transmission error and positioning error in the equipment motion model to obtain the calibrated machining positions of the part.

[0072] S207. After determining the target key area for process inspection, obtain the candidate key areas and non-key areas determined in the design stage of the part to be inspected; adjust the first spatial overlap part between the target key area and the non-key area to the candidate key area.

[0073] Specifically, after determining the target key areas for process inspection, the equipment inspection control system obtains the information on the candidate key areas and non-key areas of the parts to be inspected from relevant documents, databases or systems in the design phase. Then, through spatial analysis algorithms or relevant geometric calculation methods, the first spatial overlapping part between the target key areas and the non-key areas is determined. For example, using the spatial analysis function in 3D modeling software or spatial databases, the overlapping part of the two areas in space is found. For this overlapping area, the system adjusts it to a candidate key area, which means that in the subsequent inspection process, this area will be subject to more stringent inspection and attention, changing its original inspection strategy defined as a non-key area in the design phase and making the division of inspection areas more in line with the quality risk situation in actual production.

[0074] Optionally, the system can combine computer-aided design (CAD) software with a data analysis system. First, import the part model in the design phase (including information on candidate key areas and non-key areas) into the CAD software. Then, also import the data of the target key areas determined by the equipment inspection control system into the software. Using the spatial analysis tools of the CAD software, accurately calculate the first spatial overlapping part between the target key areas and the non-key areas. Finally, through interaction with the data analysis system, update the attribute information of this overlapping area to a candidate key area and store it in the corresponding database for subsequent inspection process calls.

[0075] S208. Divide the area priorities corresponding to different target geometric areas in the candidate key areas according to the risk assessment results of the parts to be inspected in the design phase.

[0076] The equipment inspection control system obtains the risk assessment results of the parts to be inspected in the design phase, which usually exist in the form of documents, tables or database records. Then, for each target geometric area in the candidate key areas, divide its area priority according to the risk assessment results. For example, if the risk assessment results show that a certain target geometric area involves the core function of the part and the probability of quality problems is relatively high, then the priority of this area will be set relatively high. For the target geometric area corresponding to the first spatial overlapping part, regardless of its risk assessment situation in the design phase, its area priority will be set to the lowest. This is because this area was originally designed as a non-key area. Although it is adjusted to a candidate key area due to actual production conditions, compared with other areas recognized as key in the design phase, its risk is relatively low, so the lowest priority is given. In this way, the system determines a clear inspection priority order for each target geometric area in the candidate key areas, providing a basis for the reasonable allocation of subsequent inspection resources.

[0077] S209. Adjust the region priority according to the second spatial overlap between the target key region and the candidate key region, so that the region corresponding to the second spatial overlap has the highest priority.

[0078] The equipment inspection control system first determines the second spatial overlap between the target key region and the candidate key region, which can be achieved through spatial analysis algorithms or related tools. For example, the Boolean operation function in 3D modeling software can be used to find the intersection of the two region sets. For this second spatial overlap region, regardless of the priority assigned to it in step S208, the system adjusts its region priority to the highest. For example, in the previous priority division, a certain target geometric region originally had a medium priority, but it is in the second spatial overlap part, so the system raises its priority to the highest. This is because this part of the region not only meets the definition of the key region in the design stage but is also identified as a high-risk region in actual production. Therefore, the highest inspection priority needs to be given to ensure that this part of the region receives the most sufficient inspection resource investment and key attention, and the quality risk is minimized to the greatest extent.

[0079] In the above embodiment, the system introduces a dynamic adjustment mechanism for region priority. By analyzing the spatial overlap relationship of different types of key regions, the intelligent allocation of inspection priority is realized. This priority division method based on multi-dimensional analysis ensures that the inspection resources can cover high-risk regions to the greatest extent and further improves the inspection efficiency.

[0080] The equipment inspection control system of the embodiment of the present invention is applied to an electronic device. Figure 3 The schematic diagram of the architecture of the electronic device suitable for implementing the embodiment of the present invention is shown.

[0081] It should be noted that Figure 3 The shown electronic device is only an example and should not bring any limitation to the functions and usage scope of the embodiment of the present invention.

[0082] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed through instructions (computer programs), or the relevant hardware can be controlled through instructions (computer programs). The instructions can be stored in a computer-readable storage medium and loaded and executed by a processor. The electronic device of this embodiment includes a storage medium and a processor. Among them, multiple instructions are stored in the storage medium, and these instructions can be loaded by the processor to execute any step of the method provided by the embodiment of the present invention.

[0083] Specifically, the storage medium and the processor are electrically connected directly or indirectly to achieve data transmission or interaction. For example, these components can be electrically connected to each other through one or more signal lines. The storage medium stores computer-executable instructions for implementing the data access control method, including at least one software function module that can be stored in the storage medium in the form of software or firmware. The processor executes various functional applications and data processing by running the software programs and modules stored in the storage medium. The storage medium can be, but is not limited to, a random access memory (RAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc. Among them, the storage medium is used to store programs, and the processor executes the programs after receiving the execution instructions.

[0084] Furthermore, the software programs and modules in the above storage medium may further include an operating system, which may include various software components and / or drivers for managing system tasks (such as memory management, storage device control, power management, etc.), and may communicate with various hardware or software components to provide a running environment for other software components. The processor can be an integrated circuit chip with signal processing capabilities. The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc., which can implement or execute the various methods, steps, and logic flow block diagrams disclosed in this embodiment. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.

[0085] Since the instructions stored in the storage medium can execute the steps in any of the methods provided in the embodiments of the present invention, the beneficial effects of any of the methods provided in the embodiments of the present invention can be achieved. For details, see the previous embodiments and will not be repeated here.

[0086] As described above, it is only the preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A dynamic key area process inspection method based on real-time data driving, which is applied to an equipment inspection control system, and is characterized in that, The method includes: Dividing the parts to be inspected into different quality batches according to the parameter value ranges to which the process behavior data corresponding to the parts to be inspected belong. The process behavior data is collected during the production process of the parts to be inspected by multiple types of sensors deployed on the production equipment. Different quality batches correspond to different processing area division rules; In the same quality batch, constructing a mapping relationship between the process behavior data and the processing areas of the target parts according to the process behavior data of the target parts produced in the same time period and the part processing positions corresponding to multiple acquisition timestamps of the process behavior data. The processing areas include multiple processing positions; Based on the mapping relationship, selecting target processing areas whose fluctuation characteristics of the process behavior data corresponding to the acquisition timestamps exceed a preset threshold from the multiple processing areas corresponding to the target parts to construct a target geometric area set. The fluctuation characteristics include fluctuation frequency, fluctuation amplitude, and fluctuation trend; Inputting the process behavior data corresponding to each target geometric area in the target geometric area set into a preset risk prediction model, and outputting the defect risk score corresponding to each target geometric area. The risk prediction model is trained based on historical process behavior data and corresponding historical defect risk scores; Determining the target geometric areas with defect risk scores exceeding the preset risk threshold as the target key areas for process inspection for resource allocation.

2. The method according to claim 1, wherein Before the step of dividing the parts to be inspected into different quality batches according to the parameter value ranges to which the process behavior data corresponding to the parts to be inspected belong, it further includes: Calculating the parameter fluctuation value corresponding to the process behavior data based on the difference between the process behavior data corresponding to all parts to be inspected and the preset ideal process behavior data; Dividing the parameter fluctuation value into multiple parameter fluctuation ranges based on a preset ratio to obtain corresponding multiple parameter value ranges. The processing area division granularities corresponding to different parameter value ranges are not all the same.

3. The method according to claim 2, characterized in that, After the step of dividing the parameter fluctuation value into multiple parameter fluctuation ranges based on a preset ratio to obtain corresponding multiple parameter value ranges, it further includes: Based on the machining feature dimensions and machining accuracy requirements of the parts, constructing a multi-level fine-grained division level for the processing areas. Each fine-grained division level corresponds to a different area division size; Setting the preset parameter range corresponding to each fine-grained division level based on the distribution characteristics of the parameter fluctuation values in the historical machining data; Calculating the overlapping ratio of the parameter fluctuation values within each parameter value range in the preset parameter range; Determining the fine-grained division level corresponding to the preset parameter range with the largest overlapping ratio as the fine-grained division level corresponding to the parameter value range.

4. The method according to claim 1, characterized in that Before the step of constructing a mapping relationship between the process behavior data and the processing areas of the target parts according to the process behavior data of the target parts produced in the same time period and the part processing positions corresponding to multiple acquisition timestamps of the process behavior data in the same quality batch, it further includes: Obtain the production trajectory data of the target part, where the production trajectory data includes the movement trajectory and processing speed of the processing tool of the production equipment relative to the target part; Based on the production trajectory data, calibrate the machining positions of the target part corresponding to different acquisition timestamps to obtain the calibrated machining positions of the part.

5. The method according to claim 1, characterized in that, The step of constructing the target geometric region set by selecting the target processing region whose fluctuation characteristics of the process behavior data corresponding to the acquisition timestamp exceed a preset threshold from multiple processing regions corresponding to the target part based on the mapping relationship specifically includes: Based on the mapping relationship, construct a process behavior data sequence corresponding to each continuous acquisition timestamp according to one or more continuous acquisition timestamps corresponding to each processing region; When it is detected that the fluctuation characteristics of any process behavior data sequence corresponding to the processing region exceed the preset threshold, mark the processing region as the target processing region; Construct a target geometric region set according to all the target processing regions.

6. The method according to claim 1, characterized in that, After the step of determining the target key region for process inspection and allocating resources for the target geometric region whose defect risk score exceeds the preset risk threshold, it further includes: Obtain the candidate key regions and non-key regions determined during the design stage of the part to be inspected; Adjust the first spatial overlap part between the target key region and the non-key region to be the candidate key region.

7. The method according to claim 6, wherein After the step of adjusting the first spatial overlap part between the target key region and the non-key region to be the candidate key region, it further includes: Divide the region priorities corresponding to different target geometric regions in the candidate key region according to the risk assessment result of the part to be inspected during the design stage, and the region priority corresponding to the first spatial overlap part is the lowest; Adjust the region priority according to the second spatial overlap part between the target key region and the candidate key region, so that the region priority corresponding to the second spatial overlap part is the highest.

8. An equipment inspection control system, characterized in that, The equipment inspection control system includes: one or more processors and a memory; The memory is coupled to the one or more processors, and the memory is used to store computer program code, and the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the equipment inspection control system to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium, comprising instructions, characterized in that, When the instruction runs on the equipment inspection control system, it enables the equipment inspection control system to execute the method described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product runs on the equipment inspection control system, it enables the equipment inspection control system to execute the method described in any one of claims 1-7.

Images