On-line dimensional inspection and compensation control system for manufacturing process of automobile generator shaft

Abstract

The application discloses an online size detection and compensation control system for a manufacturing process of a vehicle generator shaft and relates to the technical field of generator shaft manufacturing. The system comprises an information acquisition module: obtaining past size detection information to obtain past detection information. The size detection method is set to detect the size of a target image based on a reference catalog to obtain a detection result. The size of the size detection result is a fuzzy size, and the image similarity matching can be performed simultaneously to obtain the fuzzy size under each part, so as to quickly detect whether the size of the workpiece to be detected is located in the range of the required size, improve the efficiency of the overall online detection and subsequent compensation control, and predict the compensation method. According to the process information of production and processing, it is judged whether the process of real-time production and processing will cause the subsequent size to deviate, and the equipment of the workpiece with the predicted deviation is compensated, so as to reduce the increase of the product rejection rate caused by the structural deviation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of generator shaft manufacturing, in particular to an online size detection and compensation control system for the manufacturing process of a vehicle generator shaft. BACKGROUND

[0002] A vehicle generator shaft generally refers to a part of the shafting in a vehicle engine that serves as a driving component of the generator. It is a key shaft component that connects the inside and outside of the generator, transmits torque, and supports rotating parts. It usually needs to have high rigidity, good roundness, and dimensional stability to ensure that the generator can stably generate electricity during operation and maintain good coordination with the vehicle system.

[0003] The patent with publication number CN117620775A discloses an online detection and automatic compensation method for piston ring size, which includes the following steps: pre-processing the total piston ring; measuring the axial thickness of a group of piston ring assemblies; transmitting the axial thickness of the group of piston ring assemblies to a numerical control system, which calculates the tool's lower knife point coordinates and axial feed amount based on the axial thickness and stores the data; the numerical control system controls the rotation of the machine tool spindle to drive the next group of piston ring assemblies to rotate, and drives the camera assembly to move axially so that the split position of the next group of piston ring assemblies is opposite to the camera assembly; repeat the above steps until the axial thickness measurement of the last group of piston ring assemblies and the corresponding data storage of the tool's lower knife point coordinates and axial feed amount are completed; the numerical control system compensates for the thickness of each group of piston ring assemblies in turn based on each lower knife point coordinate of the tool and the axial feed amount. The application has high measurement accuracy, high detection efficiency, and low cost.

[0004] The existing system often uses physical measurement and other methods for size detection during size measurement. In the measurement process, too much emphasis is placed on the accuracy of size detection, which leads to a complex data processing process and thus a disadvantage in size detection speed. The process from data collection by the sensor to the output of the measurement result is complex and requires processing a large amount of data. The high complexity of the processing algorithm makes it difficult to achieve millisecond-level closed-loop control, which is not convenient for improving compensation efficiency and thus reduces the usability of online detection for subsequent compensation actions. Moreover, due to the hysteresis of online detection, compensation control needs to adjust based on the lagging error information, which easily causes structural deviation and increases the rate of defective products. Therefore, the present application is proposed. SUMMARY

[0005] The present application aims to provide an online size detection and compensation control system for the manufacturing process of a vehicle generator shaft to solve the problems raised in the background.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solution: an online size detection and compensation control system for the manufacturing process of a vehicle generator shaft, which includes:

[0007] The information acquisition module acquires the past size detection information to obtain past detection information, the past detection information including past detection workpieces, past reference sizes, and past compensation information, and acquires a required measurement position to obtain a required position.

[0008] The information processing module acquires an image of the workpiece to be detected based on the required position to obtain a target image, and establishes a reference catalog for reference to complete size detection based on the past reference sizes by an information processing method.

[0009] The size detection module judges the size of the workpiece to be detected based on the reference catalog in combination with the target image by a size detection method to obtain a detection result.

[0010] The compensation determination module obtains a compensation catalog based on the past compensation information by a compensation exploration method, obtains real-time compensation information based on the detection result in combination with the compensation catalog by a compensation determination method, and outputs the real-time compensation information.

[0011] The compensation optimization module obtains an image of the workpiece to be detected after compensation based on the required position to obtain a result image, and optimizes the real-time compensation information based on the result image by an optimization compensation method.

[0012] The prediction compensation module obtains production process information of the workpiece to be detected and production process information of the past detection workpiece to obtain real-time process information and past process information, integrates the real-time process information and the past process information to obtain process information, and performs prediction compensation on the production workpiece process based on the process information by a prediction compensation method.

[0013] Further, the information processing method comprises: presetting a picture specification and a basic size, obtaining a sub-image of a past detection workpiece under the required position based on the picture specification, obtaining a basic image of the basic size under the required position based on the picture specification, the past reference size comprising a plurality of sub-sizes, the sub-image and the basic image being auxiliary images, the basic size and the sub-size being auxiliary sizes, establishing an association relationship between the auxiliary images and the auxiliary sizes, establishing a sub-catalog for storing the auxiliary images, the auxiliary sizes, and the association relationship with the past detection workpiece as a name, and integrating all the sub-catalogs to obtain the reference catalog.

[0014] Further, the size detection method comprises: selecting a subdirectory corresponding to the workpiece to be detected in a reference directory to obtain a target directory, obtaining a required size range, presetting an added range, combining the required size range and the added range to obtain a result range, selecting images in the target directory located in the result range to obtain an image set, the image set comprising a plurality of matching images, extracting sizes of the matching images based on the result range and the target directory to obtain matching sizes, establishing a corresponding relationship between the matching images and the matching sizes, performing similarity matching based on the target image and the matching images in the image set to obtain a matching result, selecting a matching image with the highest similarity to the target image in the matching result to obtain a selected image, extracting a size corresponding to the selected image based on the corresponding relationship to obtain a selected size, judging whether the selected size is located in the required size range to obtain a judgment result, and integrating the judgment result and the selected size to obtain a detection result.

[0015] Further, the compensation exploration method comprises: splitting the previous compensation information to obtain compensation parameters and correction information, extracting compensation parameters for correcting a single part to obtain single parameters, extracting correction information corresponding to the single parameters to obtain single information, integrating the single parameters and the single information to obtain a first sub-compensation directory, extracting compensation parameters for correcting multiple parts to obtain multiple parameters, extracting correction information corresponding to the multiple parameters to obtain multiple information, integrating the multiple parameters and the multiple information to obtain a second sub-compensation directory, and integrating the first sub-compensation directory and the second sub-compensation directory to obtain a compensation directory.

[0016] Further, the compensation determination method comprises: splitting the detection result to obtain a judgment result and a selected size, extracting a target part for which the judgment result is that the selected size is not located in the required size range, judging the number of target parts, when the number of target parts is a single, selecting a first sub-compensation directory in the compensation directory to obtain a first selected directory, extracting single parameters and single information corresponding to the target part in the first selected directory to obtain first selected parameters and first selected information, obtaining a correction difference value of the first selected information to obtain a first correction difference value, obtaining a difference value between the required size range and the selected size to obtain a first target difference value, judging a multiple relationship between the first target difference value and the first correction difference value to obtain a first multiple relationship, adjusting the first selected parameters based on the first multiple relationship to obtain real-time compensation information, and when the number of target parts is multiple, obtaining the real-time compensation information through a parameter adjustment method.

[0017] Further, the parameter adjustment method comprises: selecting a second sub-compensation directory from the compensation directory to obtain a second selected directory, extracting a plurality of parameters and a plurality of information corresponding to the target part in the second selected directory to obtain a second selected parameter and a second selected information, obtaining a correction difference value of the same part in the second selected information to obtain a second correction difference value, obtaining a difference value between the required size range and the selected size to obtain a second target difference value, extracting a minimum multiple relationship of the corresponding second correction difference value and the second target difference value to obtain a second multiple relationship, and selecting the smallest second multiple relationship to adjust the second selected parameter to obtain real-time compensation information.

[0018] Further, the optimization supplement method comprises: importing the result image into the size detection method to obtain a re-detection result, judging whether the re-detection result feeds back that the re-selected size is located in the required size range, judging the relationship between the re-selected size and the required size range when the re-detection result feeds back that the re-selected size is not located in the required size range, presetting a fixed adjustment value, reducing the items for parameter adjustment in the real-time compensation information based on the fixed adjustment value to obtain re-real-time compensation information when the selected size is smaller than the required size range and the re-selected size exceeds the required size range, and when the selected size is larger than the required size range and the re-selected size does not exceed the required size range, and the selected size is smaller than the required size range and the re-selected size does not exceed the required size range, increasing the items for parameter adjustment in the real-time compensation information based on the fixed adjustment value to obtain re-real-time compensation information, obtaining a re-result image output after the re-real-time compensation information, and optimizing the re-result image until the re-detection result feeds back that the re-selected size is located in the required size range.

[0019] Further, the prediction compensation method comprises: obtaining a previous parameter information of a workpiece machining parameter that has been compensated and controlled to obtain a previous parameter information, establishing a correlation between the previous parameter information and previous compensation information, monitoring a real-time workpiece machining parameter to obtain a real-time parameter information, performing similarity matching between the real-time parameter information and the previous parameter information to obtain a similarity result, presetting a similarity threshold, when the similarity result exceeds the similarity threshold, extracting the previous compensation information of the previous parameter information based on the correlation to obtain a result compensation information, determining a target machining equipment based on the result compensation information, obtaining a standard condition of a normal operation state of the target machining equipment, monitoring a working condition of the target machining equipment to obtain a real-time condition, comparing the standard condition and the real-time condition to obtain a comparison result, and when the comparison result feeds back that the working condition of the target machining equipment is abnormal, outputting the previous compensation information to the target machining equipment to complete the prediction compensation on the production workpiece process.

[0020] Compared with the prior art, the present application has the following advantages:

[0021] The online size detection and compensation control system of the vehicle generator shaft manufacturing process establishes a reference catalog for providing a basis for online size detection through the set information processing method, and the size detection result is obtained by performing size detection on the target image based on the reference catalog through the set size detection method, the size of the size detection result is a fuzzy size, and the sizes of multiple parts of the workpiece to be detected can be simultaneously subjected to image similarity matching to obtain, so as to quickly detect whether the size of the workpiece to be detected is located in the range of the required size, improve the efficiency of subsequent compensation control after overall online detection, and through the set prediction compensation method, whether the subsequent size will be deviated is judged according to the process information of production and processing, and the equipment of the workpiece with prediction deviation is compensated, so as to reduce the increase of product rejection rate caused by structural deviation.

[0022] Meanwhile, the online detection process is an image comparison process, multiple target images can be simultaneously processed in a multi-thread processing manner to improve the efficiency of online size detection, through the set compensation determination method, real-time compensation information under the detection result is obtained according to the detection result and the compensation catalog, so as to perform compensation control after online size detection, improve the quality of the vehicle generator shaft manufacturing, when the second selected catalog does not have multiple parameters and multiple information corresponding to the target part, the target part can be split, when there is a single target part after splitting, the real-time compensation information is obtained through the compensation determination method, after splitting, the number of real-time compensation information is consistent with the number of splitting, and the generation mode of specific real-time compensation information has high flexibility, through the set optimization supplement method, the real-time compensation information is further adjusted to obtain real-time compensation information by judging the relationship between the size before and after compensation and the required size range, so as to optimize the real-time compensation information, form a compensation control closed loop, and further improve the manufacturing quality of the vehicle generator shaft.

[0023] Meanwhile, by linking the multiple relationship of one key error and another error, the adjustment of the sensitive parameter can be quickly amplified when needed, so that the system converges to the target tolerance interval more quickly, the process of adjusting the first selected parameter based on the first multiple relationship to obtain real-time compensation information can preset the gain or reduction limit to avoid overshoot or instability, and in the specific application process, the compensation determination method can be used when multiple size deviations of the same target part are continuously identified, and the specific number is determined according to the actual use, so as to avoid the instability of the vehicle motor shaft manufacturing process caused by frequent output of real-time compensation parameters. BRIEF DESCRIPTION OF DRAWINGS

[0024] Figure 1 It is the overall flowchart of the present application;

[0025] Figure 2 It is the reference catalog structure diagram of the present application;

[0026] Figure 3 Result range structure diagram of the present application;

[0027] Figure 4 Target site split structure diagram of the present application;

[0028] Figure 5 Before and after prediction compensation of the present application;

[0029] Figure 6 Before and after compensation adjustment of the first multiple relationship of the present application with the first selected parameter. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0031] In the manufacturing of vehicle generator shafts, the necessity of online size detection and compensation control is reflected in the following key aspects: improving the stability of size and coaxiality, ensuring the consistency of tolerance distribution in batch production; reducing the rate of defective products and repair costs, reducing rework, waste and process interruption caused by size deviation; improving assembly reliability and life performance, long-term existence of radial deviation and coaxiality of the shaft will cause vibration, thermal stress and bearing wear, compensation helps to improve the size stability under thermal state and working load; realizing efficient process self-adaptation and cost optimization. In summary, online size detection and compensation control is the key link to ensure high precision, low variation, long life and high production efficiency of vehicle generator shafts.

[0032] As shown in Figures 1-6 The present application provides a technical solution: an online size detection and compensation control system for the manufacturing process of vehicle generator shafts, which comprises:

[0033] An information acquisition module: acquiring past size detection information to obtain past detection information, the past detection information including past detection workpieces, past reference sizes and past compensation information, and acquiring required measurement sites to obtain required sites;

[0034] It should be noted that the past detection workpieces are specific workpiece names, specifically the workpieces required for manufacturing vehicle generator shafts, the past reference sizes are specific size sizes, the past compensation information is the information of adjusting parameters for the corresponding processing device after discovering size deviation, the required sites are the sites that need to be detected, and the past detection information is obtained through the work log of past vehicle generator shaft manufacturing.

[0035] The information processing module obtains a target image based on the image of the required part of the workpiece to be detected, and establishes a reference catalog for reference to complete size detection through an information processing method based on the previous reference size;

[0036] It should be noted that the workpiece to be detected is the workpiece that needs to be detected, and the image of the required part of the workpiece to be detected is obtained to obtain a target image. Specifically, the target image can be obtained by installing a high-resolution camera module on the work station for machining and manufacturing the vehicle generator shaft. When the camera module is fixed on the shaft, the different batches or assembly stages in the production process are calibrated first to ensure that the camera module field of view covers the required parts of the shaft (such as end cover, shaft neck, spline, flange, etc.). The shaft is fine-tuned in the horizontal and vertical directions by the mechanical clamp or programmable turntable on the production line according to the preset angle, so that different required parts are shot at the same angle. Then, the orientation of the shaft is adjusted gradually according to the requirements of the production process. For example, in some processes, the shaft axis needs to face upwards to detect the end face perpendicularity, while in other processes, it needs to be turned to the side to check the spline gap or the concentricity of the shaft neck. Finally, the fixed machine position is uniformly photographed to ensure that the orientation changes of different positions are recorded and evaluated completely, and the reference catalog is established through the set information processing method to provide a basis for online size detection.

[0037] The size detection module judges the size of the workpiece to be detected based on the reference catalog and the target image through a size detection method to obtain a detection result;

[0038] It should be noted that the size detection method set based on the reference catalog for size detection of the target image obtains a detection result, and the online detection process is an image comparison process, which can be specifically understood as matching the target image with all images in the reference catalog and judging whether they are consistent, when consistent, the accurate detection result is obtained, when inconsistent, the fuzzy detection result is obtained, and the specific image comparison process is: using a robust feature extractor (such as SIFT / ORB / AFM, etc.) to extract key points and their descriptors from two images respectively, then matching the descriptors (such as brute force matching, FLANN, BRUTE_FORCE_KNN), removing mismatches through screening (such as ratio test, RANSAC estimation of transformation matrix or intrinsic distance threshold), and finally using distance threshold, matching number, or consistency of robust transformation to judge whether the features are“similar / consistent”. In addition, the average distance of matching, the coverage rate of matching, and the consistency of geometric transformation (such as homography matrix or re-projection error of similarity transformation) can be used to quantitatively evaluate. When the key point clusters and descriptors of two images are highly overlapped and the geometric transformation error is within an acceptable range, it is determined to be consistent. In the specific comparison process, multiple target images can be processed simultaneously in a multi-threaded manner to improve the efficiency of online size detection. By merging the detection results obtained from individual target images, the online detection size results of multiple parts of the workpiece can be obtained.

[0039] The compensation determination module: based on the past compensation information, the compensation directory is obtained by the compensation exploration method, and the real-time compensation information is obtained by the compensation determination method based on the detection result and the compensation directory, and the real-time compensation information is output;

[0040] It should be noted that the compensation exploration method set based on the past compensation information establishes a compensation directory for providing a basis for subsequent acquisition of real-time compensation information, and the compensation determination method is used to obtain real-time compensation information under the detection result according to the detection result and the compensation directory. The real-time compensation information is transmitted to the corresponding production and processing equipment, so that the compensation control after online size detection can be carried out to improve the quality of workpiece processing, and thus the quality of the vehicle generator shaft manufacturing is improved.

[0041] The compensation optimization module: based on the demand part, the result image is obtained based on the detected workpiece image after compensation, and the real-time compensation information is optimized based on the result image by the optimization supplement method;

[0042] It should be noted that the process of obtaining the result image based on the demand site and the compensated workpiece image to be detected is that a high-resolution camera module is installed at the corresponding position of the demand site during the manufacturing and processing of the automobile generator shaft, and the result image is obtained. The result image is the workpiece image to be detected after the real-time compensation information is transmitted to the corresponding production and processing equipment for processing and production. Specifically, the workpiece image to be detected can be obtained by erecting a high-definition camera to obtain the image of the workpiece to be detected in real time. Through the setting of the optimization and compensation method, the real-time compensation information can be optimized based on the result image, and the compensation control closed loop is completed to further improve the manufacturing quality of the automobile generator shaft.

[0043] The prediction compensation module obtains the production process information of the workpiece to be detected and the production process information of the workpiece detected in the past to obtain real-time process information and past process information, integrates the real-time process information and the past process information to obtain process information, and predicts and compensates the production workpiece process based on the process information through a prediction compensation method.

[0044] It should be noted that the production process information of the workpiece detected in the past is the production process information that will cause size deviation, and the real-time process information and the past process information are the process information of the same workpiece in the same processing process, which specifically includes the parameters and time of the processing equipment. Through the setting of the prediction compensation method, it is judged whether the real-time production and processing process will cause the subsequent size to deviate, and when it is judged that there will be deviation, the equipment for producing workpieces is predicted and compensated to reduce the probability of size deviation.

[0045] As shown in Figure 2 The information processing method includes: presetting a picture specification and a basic size, obtaining a workpiece image under a demand site based on the picture specification to obtain a sub-image, obtaining a basic size image under the demand site based on the picture specification to obtain a basic image, the past reference size includes a plurality of sub-sizes, the sub-image and the basic image are auxiliary images, the basic size and the sub-size are auxiliary sizes, an association relationship between the auxiliary images and the auxiliary sizes is established, a sub-directory is established for storing the auxiliary images, the auxiliary sizes and the association relationship with the workpiece detected in the past as the name, and all sub-directories are integrated to obtain a reference directory.

[0046] It should be noted that the picture specification is determined according to the actual use, and all the pictures obtained need to be consistent with the picture specification. The basic size is the preset size of the workpiece part. By presetting the basic size, the data can be enriched. The process of obtaining the sub-image based on the picture specification of the previous detection workpiece image of the required part and obtaining the basic image based on the picture specification of the basic size image of the required part, that is, obtaining the image of the required part of the previously produced workpiece, including the size qualified workpiece and the size unqualified workpiece, establishing the association between the auxiliary image and the auxiliary size, can be established in the image acquisition stage. A subdirectory is established based on the previously detected workpiece, so that the corresponding subdirectory of the workpiece to be detected can be selected according to the corresponding workpiece to be detected in the future, so as to facilitate information searching and facilitate subsequent online size detection. Figure 2 In the subdirectory 1, the subdirectory 2 and the subdirectory n, the auxiliary image, the auxiliary size and the association corresponding to the workpiece to be detected 1, the workpiece to be detected 2 and the workpiece to be detected n are stored respectively.

[0047] As shown in Figure 3 The size detection method includes: selecting a subdirectory corresponding to the workpiece to be detected in the reference directory to obtain a target directory, obtaining a required size range, presetting an added range, combining the required size range and the added range to obtain a result range, selecting an image in the target directory within the result range to obtain an image set, the image set includes a plurality of matching images, extracting the size of the matching image based on the result range and the target directory to obtain a matching size, establishing a corresponding relationship between the matching image and the matching size, performing similarity matching based on the target image and the matching image in the image set to obtain a matching result, selecting a matching image with the highest similarity to the target image in the matching result to obtain a selected image, extracting the size corresponding to the selected image based on the corresponding relationship to obtain a selected size, judging whether the selected size is within the required size range to obtain a judgment result, and integrating the judgment result and the selected size to obtain a detection result.

[0048] It should be noted that the process of obtaining the required size range, that is, obtaining the size range of the workpiece to be detected, the added range is a specific size added value, and the added range is determined according to the actual use. The process of combining the added range and the size range, that is, Figure 3As shown, a result range larger than the size range is obtained to facilitate online detection of the size of the required part of the workpiece to be detected, the process of selecting images in the target directory located in the result range to obtain an image set, that is, selecting images in the target directory with sizes located in the result range to obtain an image set, and the process of similarity matching based on the target image and the matching image in the image set to obtain a matching result, that is, image processing, specifically similarity matching, obtaining the fuzzy size of the target image by extracting the image with the highest similarity, selecting the size as the fuzzy size, and obtaining a judgment result by judging whether the fuzzy size is located in the required size range, the process of obtaining the selected size is only obtained by image similarity matching, and the sizes of multiple parts of the workpiece to be detected can be simultaneously subjected to image similarity matching to obtain the fuzzy size of each part, so as to facilitate rapid online detection of whether the size of the workpiece to be detected is located in the required size range, and improve the efficiency of subsequent compensation control after overall online detection.

[0049] As shown in Figure 1 The compensation determination method includes: splitting the detection result to obtain a judgment result and a selected size, extracting the part whose selected size is not located in the required size range to obtain a target part, judging the number of target parts, when the number of target parts is single, selecting a first sub-compensation directory in the compensation directory to obtain a first selected directory, extracting the single parameter and the single information corresponding to the target part in the first selected directory to obtain a first selected parameter and a first selected information, obtaining a correction difference value of the first selected information to obtain a first correction difference value, obtaining a difference value between the required size range and the selected size to obtain a first target difference value, judging the multiple relationship between the first target difference value and the first correction difference value to obtain a first multiple relationship, adjusting the first selected parameter based on the first multiple relationship to obtain real-time compensation information, and when the number of target parts is multiple, obtaining real-time compensation information by the parameter adjustment method.

[0050] It should be noted that by distinguishing the compensation parameters of single parts and multiple parts, the efficiency of obtaining real-time compensation parameters can be improved subsequently. Specifically, the more the number of past compensation information, the higher the accuracy of obtaining real-time compensation parameters subsequently.

[0051] As shown in Figure 1 The compensation determination method includes: splitting the detection result to obtain a judgment result and a selected size, extracting the part whose selected size is not located in the required size range to obtain a target part, judging the number of target parts, when the number of target parts is single, selecting a first sub-compensation directory in the compensation directory to obtain a first selected directory, extracting the single parameter and the single information corresponding to the target part in the first selected directory to obtain a first selected parameter and a first selected information, obtaining a correction difference value of the first selected information to obtain a first correction difference value, obtaining a difference value between the required size range and the selected size to obtain a first target difference value, judging the multiple relationship between the first target difference value and the first correction difference value to obtain a first multiple relationship, adjusting the first selected parameter based on the first multiple relationship to obtain real-time compensation information, and when the number of target parts is multiple, obtaining real-time compensation information by the parameter adjustment method.

[0052] It should be noted that the process of determining the number of target sites, i.e. determining the number of target sites under a single workpiece to be detected, when the number of target sites is single, the first sub-compensation directory is selected to obtain the first selected directory, the first correction difference value is obtained by obtaining the correction difference value of the first selected information, i.e. obtaining the difference value of the size before and after compensation control in the first selected information, obtaining the difference value of the required size range and the selected size to obtain the first target difference value, the difference value of the optimal size in the required size range and the selected size can be obtained to obtain the first target difference value, or other values in the required size range, which is determined according to actual use, by linking the multiple relationship of one key error (such as the difference value of the size before and after the change) and another error, the adjustment of sensitive parameters can be quickly amplified when needed, so that the system converges to the target tolerance interval more quickly, the first selected parameter is adjusted based on the first multiple relationship to obtain real-time compensation information, the gain or reduction limit can be preset to avoid overshoot or instability, in the specific application process, the compensation determination method can be used when multiple size deviations with the same target site are continuously identified, the specific number is determined according to actual use, to avoid frequent output of real-time compensation parameters leading to unstable vehicle motor shaft manufacturing process, such as Figure 6 As shown, the range composed of the target upper limit and the target lower limit is the required size range, and the K value is the multiple relationship of the first target difference value and the first correction difference value, i.e. the first multiple relationship, the first selected parameter is adjusted based on the first multiple relationship and applied to quickly return the size to the required size range, the specific adjustment process is subject to actual use, Figure 6 The adjustment process in the middle is to adjust multiple times to make the size information tend to the middle value of the required size range, and the change line segment of the size information has various forms, Figure 6 To reflect the rapid adjustment of the size to the required size range.

[0053] The parameter adjustment method comprises: selecting a second sub-compensation directory in the compensation directory to obtain a second selected directory, extracting multiple parameters and multiple information corresponding to the target site in the second selected directory to obtain second selected parameters and second selected information, obtaining the correction difference value of the same site in the second selected information to obtain the second correction difference value, obtaining the difference value of the required size range and the selected size to obtain the second target difference value, extracting the minimum multiple relationship of the corresponding second correction difference value and the second target difference value to obtain the second multiple relationship, and selecting the smallest second multiple relationship to adjust the second selected parameter to obtain real-time compensation information.

[0054] It should be noted that when the target quantity is multiple, the second selected directory is obtained by selecting the second sub-compensation directory, and the second correction difference value is obtained by acquiring the correction difference value of the second selected information. The process is consistent with the process of obtaining the first correction difference value by acquiring the correction difference value of the first selected information. Specifically, the same part needs to be noted for the same part difference acquisition. In specific use, when the second selected directory does not have multiple parameters and multiple information corresponding to the target part, the target part can be split, for example, when the number of target parts is four, the target parts can be split into two or three to select the corresponding second selected parameter and second selected information. The specific target part splitting is as shown in Figure 4 The positions of the split target part 1, the target part 2, the target part 3, and the target part 4 can be exchanged. The process of acquiring real-time compensation information is unchanged. When a single target part is split, the real-time compensation information is acquired by the compensation determination method. After splitting, the number of real-time compensation information is consistent with the number of splits. The flexibility of the generation mode of the real-time compensation information is high.

[0055] As shown in Figure 1 The optimization supplement method includes: importing the result image into the size detection method to obtain a re-detection result, judging whether the re-detection result feedbacks that the re-selected size is located in the required size range, judging the relationship between the re-selected size and the required size range when the re-detection result feedbacks that the re-selected size is not located in the required size range, presetting a fixed adjustment value, when the selected size is smaller than the required size range and the re-selected size exceeds the required size range, and when the selected size is greater than the required size range and the re-selected size exceeds the required size range, the project in the real-time compensation information for parameter adjustment is reduced based on the fixed adjustment value to obtain re-real-time compensation information, when the selected size is greater than the required size range and the re-selected size does not exceed the required size range, and when the selected size is smaller than the required size range and the re-selected size does not exceed the required size range, the project in the real-time compensation information for parameter adjustment is increased based on the fixed adjustment value to obtain re-real-time compensation information, and the re-result image obtained after the re-real-time compensation information is acquired is output. The re-result image is optimized until the re-detection result feedbacks that the re-selected size is located in the required size range.

[0056] It should be noted that the re-detection result and the re-selected size are the detection results generated by importing the result image into the size detection method. The size of the fixed adjustment value is determined according to the actual use. The real-time compensation information is further adjusted to obtain the re-real-time compensation information by judging the relationship between the size before and after compensation and the required size range, so as to optimize the real-time compensation information and form a compensation control closed loop.

[0057] As shown in Figure 1As shown, the prediction compensation method comprises: obtaining workpiece machining parameters for which compensation control has been performed in the past to obtain past parameter information, establishing a correlation between the past parameter information and past compensation information, monitoring real-time workpiece machining parameters to obtain real-time parameter information, performing similarity matching between the real-time parameter information and the past parameter information to obtain a similarity result, predefining a similarity threshold, when the similarity result exceeds the similarity threshold, extracting past compensation information of the past parameter information based on the correlation to obtain result compensation information, determining a target machining device based on the result compensation information, obtaining a standard condition of normal operation of the target machining device, monitoring a working condition of the target machining device to obtain a real-time condition, comparing the standard condition and the real-time condition to obtain a comparison result, when the comparison result feeds back that the working condition of the target machining device is abnormal, outputting the past compensation information to the target machining device to complete prediction compensation on the production workpiece process.

[0058] It should be noted that the past parameter information is obtained by collecting workpiece machining parameters (such as rotation speed, feed amount, temperature, tool wear, etc.) used for compensation control in the history, and by setting the prediction compensation method, the workpiece machining parameters for which compensation control has been performed in the past are used as a reference for similarity matching with real-time workpiece machining parameters, and a similarity threshold is pre-defined. The larger the similarity threshold is, the better it is, and the smaller the error rate of subsequent output of the past compensation information is. When the similarity matching result exceeds the similarity threshold, it is determined that the real-time machining parameter may have the same deviation as the machining parameter of the workpiece for which compensation control has been performed in the past. At this time, the corresponding target machining device that causes the size deviation of the past workpiece can be monitored, and whether the machining device will cause the deviation can be determined by monitoring whether the running condition of the target machining device is a standard condition. When it is determined that the deviation exists, the past compensation information is delivered to the target machining device to complete compensation control of the machining device in advance, so as to further reduce the defective rate of the workpiece and improve the efficiency and quality of the generator shaft manufacturing. Figure 5 As shown, after the prediction compensation method (i.e., the compensation method in the figure) is applied, the defective rate shows a decreasing trend. By identifying the device abnormality in advance and performing compensation, the unplanned downtime of the device is significantly reduced, and the structural deviation is reduced, thereby reducing the defective rate of the product.

[0059] Although the embodiments of the present application have been shown and described, it can be understood by those of ordinary skill in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended embodiments and their equivalents.

Claims

1. An online dimensional inspection and compensation control system for the manufacturing process of automotive generator shafts, the system comprising: Information acquisition module: Obtains previous dimensional inspection information, including previous inspected workpieces, previous reference dimensions, and previous compensation information; obtains the required measurement location to obtain the required location. The system is characterized in that it further includes: Information processing module: Obtains the image of the workpiece to be inspected based on the required location to obtain the target image, and establishes a reference catalog for dimensional inspection based on the previous reference dimensions through information processing methods; Size detection module: Based on the reference catalog and target image, the size detection method is used to determine the size of the workpiece to be inspected and obtain the detection result; Compensation determination module: Based on previous compensation information, a compensation catalog is obtained through compensation exploration methods. Based on the detection results and the compensation catalog, real-time compensation information is obtained through compensation determination methods, and the real-time compensation information is output. Compensation and optimization module: Obtains the compensated image of the workpiece to be inspected based on the required location to obtain the result image, and optimizes the real-time compensation information based on the result image through optimization and supplementation methods; Prediction and compensation module: Obtains production process information of the workpiece to be inspected and production process information of previously inspected workpieces to obtain real-time process information and past process information, integrates real-time process information and past process information to obtain process information, and performs prediction and compensation on the production process of the workpiece based on the process information through prediction and compensation methods. The size detection method includes: selecting a subdirectory in a reference directory corresponding to the workpiece to be detected to obtain a target directory; obtaining the required size range; pre-setting an added range; combining the required size range and the added range to obtain a result range; selecting images in the target directory that are within the result range to obtain an image set, the image set including several matching images; extracting the dimensions of the matching images based on the result range and the target directory to obtain matching dimensions; establishing a correspondence between matching images and matching dimensions; performing similarity matching between the target image and the matching images in the image set to obtain a matching result; selecting the matching image with the highest similarity to the target image in the matching result to obtain a selected image; extracting the dimensions corresponding to the selected image based on the correspondence to obtain the selected dimensions; determining whether the selected dimensions are within the required size range to obtain a judgment result; and integrating the judgment result and the selected dimensions to obtain a detection result.

2. The online dimensional detection and compensation control system for the manufacturing process of automotive generator shafts according to claim 1, characterized in that: The information processing method includes: presetting image specifications and basic dimensions; obtaining sub-images from previous inspected workpiece images of the required location based on the image specifications; obtaining basic dimension images of the required location based on the image specifications; previous reference dimensions include several sub-dimensions; sub-images and basic images are auxiliary images; basic dimensions and sub-dimensions are auxiliary dimensions; establishing the association between auxiliary images and auxiliary dimensions; creating subdirectories named after previous inspected workpieces to store auxiliary images, auxiliary dimensions, and associations; and integrating all subdirectories to obtain a reference directory.

3. The online dimensional detection and compensation control system for the manufacturing process of automotive generator shafts according to claim 1, characterized in that: The compensation exploration method includes: splitting previous compensation information to obtain compensation parameters and correction information; extracting compensation parameters for correcting a single part to obtain single parameters; extracting correction information corresponding to the single parameters to obtain single information; integrating single parameters and single information to obtain a first sub-compensation catalog; extracting compensation parameters for correcting multiple parts to obtain multiple parameters; extracting correction information corresponding to the multiple parameters to obtain multiple information; integrating multiple parameters and multiple information to obtain a second sub-compensation catalog; and integrating the first sub-compensation catalog and the second sub-compensation catalog to obtain a compensation catalog.

4. The online dimensional detection and compensation control system for the manufacturing process of automotive generator shafts according to claim 1, characterized in that: The compensation determination method includes: splitting the detection results to obtain a judgment result and a selected size; extracting the parts whose selected size is not within the required size range to obtain target parts; determining the number of target parts; when the number of target parts is single, selecting a first sub-compensation directory in the compensation directory to obtain a first selection directory; extracting the single parameter and single information corresponding to the target part in the first selection directory to obtain a first selection parameter and a first selection information; obtaining the correction difference of the first selection information to obtain a first correction difference; obtaining the difference between the required size range and the selected size to obtain a first target difference; determining the multiple relationship between the first target difference and the first correction difference to obtain a first multiple relationship; adjusting the first selection parameter based on the first multiple relationship to obtain real-time compensation information; when the number of targets is multiple, obtaining real-time compensation information through parameter adjustment methods.

5. The online dimensional detection and compensation control system for the manufacturing process of automotive generator shafts according to claim 4, characterized in that: The parameter adjustment method includes: selecting a second sub-compensation directory in the compensation directory to obtain a second selection directory; extracting multiple parameters and multiple information corresponding to the target part in the second selection directory to obtain a second selection parameter and a second selection information; obtaining the correction difference of the same part in the second selection information to obtain a second correction difference; obtaining the difference between the required size range and the selected size to obtain a second target difference; extracting the minimum multiple relationship between the corresponding second correction difference and the second target difference to obtain a second multiple relationship; and selecting the minimum second multiple relationship to adjust the second selection parameter to obtain real-time compensation information.

6. The online dimensional detection and compensation control system for the manufacturing process of automotive generator shafts according to claim 1, characterized in that: The optimization and supplementation method includes: importing the result image into the size detection method to obtain a re-detection result; determining whether the re-detection result indicates that the re-selected size is within the required size range; when the re-detection result indicates that the re-selected size is not within the required size range; determining the relationship between the re-selected size and the required size range; presetting a fixed adjustment value; when the selected size is smaller than the required size range and the re-selected size exceeds the required size range, and when the selected size is larger than the required size range and the re-selected size exceeds the required size range, adjusting the parameters based on the fixed adjustment value in the real-time compensation information to obtain re-real-time compensation information; when the selected size is larger than the required size range and the re-selected size does not exceed the required size range, and when the selected size is smaller than the required size range and the re-selected size does not exceed the required size range, adjusting the parameters based on the fixed adjustment value in the real-time compensation information to obtain re-real-time compensation information; obtaining the re-result image after the re-real-time compensation information is output; and optimizing the re-result image until the re-detection result indicates that the re-selected size is within the required size range.

7. The online dimensional detection and compensation control system for the manufacturing process of automotive generator shafts according to claim 1, characterized in that: The prediction and compensation method includes: obtaining the workpiece processing parameters that have been previously compensated and controlled to obtain previous parameter information; establishing the correlation between the previous parameter information and previous compensation information; monitoring the real-time workpiece processing parameters to obtain real-time parameter information; performing similarity matching between the real-time parameter information and the previous parameter information to obtain a similarity result; setting a similarity threshold; when the similarity result exceeds the similarity threshold, extracting the previous compensation information from the previous parameter information based on the correlation to obtain result compensation information; determining the target processing equipment based on the result compensation information; obtaining the normal operating status of the target processing equipment to obtain a standard status; monitoring the working status of the target processing equipment to obtain a real-time status; comparing the standard status and the real-time status to obtain a comparison result; when the comparison result indicates that the working status of the target processing equipment is abnormal, outputting the previous compensation information to the target processing equipment to complete the prediction and compensation for the production process of the workpiece.

Images