A glass plate product production line evaluation method and system based on multivariate information

Abstract

The present invention relates to the technical field of production line management, and in particular to a glass plate product production line evaluation method and system based on multivariate information, which detect glass plate products on the production line in turn to generate multivariate detection information, and classify the glass plate products into qualified glass plate products and defective glass plate products according to the multivariate detection information of each glass plate product, and subdivide the defect types of the defective glass plate products, summarize and generate a production quality data collection, and then set multiple evaluation criteria for progressive analysis to generate a production quality analysis data collection; at the same time, continuously receive multivariate feedback information of various mechanical components of the production line, set multiple evaluation criteria for progressive analysis to generate an operation quality analysis data collection; finally, summarize the production quality data collection, the production quality analysis data collection and the operation quality analysis data collection to perform a comprehensive evaluation of the production line; and solve the problem that the existing evaluation method is too simple and cannot provide meaningful reference data for production line improvement.

Description

Technical Field

[0001] The present invention relates to the technical field of production line management, and in particular to a glass plate product production line evaluation method and system based on multivariate information. Background Art

[0002] After the glass raw materials are processed into glass plate products through a special process, due to the excellent properties of glass plate products such as high transparency, high corrosion resistance and high hardness, their application range is quite wide. For example, they can be used in mobile phone screens, tablet computer screens and digital camera lenses in electronic products, or in windows and glass curtain walls in the field of building materials. Glass plate products are needed in many fields.

[0003] Especially for mobile phone screens, the glass cover (i.e. glass product) covering the top of the mobile phone display as a protective layer plays a vital role in mobile phone design. It not only provides physical protection for internal precision components, but also affects the overall appearance, feel and durability of the mobile phone. Therefore, with the rapid development of the smartphone market, the quality requirements for the window area of ​​the mobile phone glass cover are getting higher and higher. In the production process of mobile phone glass cover, defect detection in its window area is an indispensable link.

[0004] However, currently, when defect detection is performed on the window area of ​​the glass cover of a mobile phone (i.e., defect detection of glass plate products), it is generally performed through manual detection or simple intelligent detection, which can only select defective glass cover plates, but it is difficult to achieve automated continuous and accurate detection, and analyze and summarize the types of defects that exist on the production line at a macro level. In addition, the production performance of the production line is simply evaluated by the yield rate of the product, and it is impossible to comprehensively evaluate the production performance of the production line based on multiple aspects such as the defect detection classification results and the operating status of various components of the production line. This is not conducive to optimizing the production line, reducing production costs, and improving the yield rate. A rigorous and scientific glass plate product production line evaluation method or system has not been formed to evaluate the production performance of the glass plate product production line and provide meaningful reference data for users to optimize the production line. Summary of the invention

[0005] In view of the above-mentioned defects, the purpose of the present invention is to propose a glass plate product production line evaluation method and system based on multivariate information, which solves the problem that the existing evaluation method is too simple and cannot provide meaningful reference data for production line improvement.

[0006] To achieve this object, the present invention adopts the following technical solutions:

[0007] A glass plate product production line evaluation method based on multivariate information comprises the following steps:

[0008] A1: inspect the glass plate products on the production line one by one and generate multi-dimensional inspection information one by one;

[0009] A2: According to the multivariate detection information of each glass plate product, the glass plate products are divided into qualified glass plate products and defective glass plate products, and the defective glass plate products are subdivided into defect types. Finally, the data of all qualified glass plate products and defective glass plate products and the multivariate detection information are summarized to generate a production quality data collection;

[0010] A3: Set multiple evaluation criteria to conduct progressive analysis on the production quality data collection to generate a production quality analysis data collection;

[0011] A4: Continuously receive multiple feedback information from various mechanical components of the production line, set multiple evaluation criteria to conduct progressive analysis on the multiple feedback information, and generate a collection of operation quality analysis data;

[0012] A5: Summarize the production quality data collection, production quality analysis data collection and operation quality analysis data collection to conduct a comprehensive evaluation of the production line.

[0013] Furthermore, the step A3 includes the following sub-steps:

[0014] A31: Based on the production quality data collection, the defective glass plate products are divided into multiple inferior quality grades according to the severity of the defects from large to small, and the proportion of the defective glass plate products of each inferior quality grade to the total number of glass plate products is obtained;

[0015] A32: Configure a poor quality weight coefficient for each poor quality grade. The proportion of each poor quality grade is multiplied by the corresponding poor quality weight coefficient and then summed to obtain the production quality assessment value.

[0016] When the production quality assessment value is greater than the preset production quality threshold, a production quality warning is generated and sub-step A35 is executed; otherwise, sub-step A33 is executed;

[0017] A33: Based on the production quality data collection, the ratio of defective glass plate products of each defect type to the total number of glass plate products is compared with the preset hidden danger threshold of the corresponding defect type;

[0018] When the proportion of the defect type is greater than the preset hidden danger threshold, the defect type is marked as a hidden danger defect, and a production quality warning is generated, and sub-step A35 is executed; otherwise, sub-step A34 is executed;

[0019] A34: Configure a defect weight coefficient for each defect type. The proportion of each defect type is multiplied by the corresponding defect weight coefficient and then summed to obtain a comprehensive defect potential risk value. The production quality assessment value and the comprehensive defect potential risk value are calculated through the production quality comprehensive evaluation formula to obtain a production quality comprehensive evaluation value.

[0020] When the comprehensive evaluation value of production quality is greater than the preset comprehensive evaluation threshold of production quality, a production quality warning is generated; otherwise, a production quality qualified prompt is generated;

[0021] A35: Summarize all data from sub-steps A31 to A34 to form a collection of production quality analysis data.

[0022] Furthermore, the step A4 includes the following sub-steps:

[0023] A41: Based on the transmission vibration amplitude value and the transmission speed value in the multivariate feedback information, the deviation value of the transmission speed value from the preset transmission speed value is set as the transmission speed deviation value; the transmission vibration amplitude value and the transmission speed deviation value are used to obtain the operation abnormality judgment value through the operation abnormality judgment formula;

[0024] Marking an operation abnormality judgment value greater than a preset abnormality threshold as an abnormal situation;

[0025] The ratio of the operation abnormality judgment values ​​marked as abnormal situations in a unit time to the total number of operation abnormality judgment values ​​is set as the transmission abnormality value;

[0026] The average of the operation abnormality judgment values ​​marked as abnormal situations within a unit time is set as the transmission abnormality table value;

[0027] The transmission abnormal condition value and the transmission abnormal table value are compared with the preset transmission abnormal condition threshold value and the preset transmission abnormal table threshold value respectively;

[0028] When the transmission abnormality value is greater than the preset transmission abnormality threshold or the transmission abnormal table value is greater than the preset transmission abnormal table threshold, a running quality warning is generated and sub-step A43 is executed in a loop; otherwise, sub-step A42 is executed;

[0029] A42: Based on the conveyor belt start-up time in the multivariate feedback information, the average of all conveyor belt start-up time in unit time is set as the start-up detection value; the conveyor belt start-up time longer than the preset start-up time is marked as an abnormal time, and the proportion of the conveyor belt start-up time marked as an abnormal time in the unit time to the total conveyor belt start-up time is set as the start-up abnormal detection value;

[0030] Based on the glass plate product stopping inspection position in the multivariate feedback information, the deviation value of the glass plate product stopping inspection position from the preset stopping position is set as the transportation deviation value, and the average of the transportation deviation values ​​per unit time is set as the transportation deviation value; the proportion of the transportation deviation values ​​greater than the preset transportation deviation value to the total number of transportation deviation values ​​is set as the transportation abnormal inspection value;

[0031] The transmission abnormal condition value, transmission abnormal table value, start detection value, start abnormal inspection value, transport position deviation value, and transport position abnormal inspection value are used to obtain the comprehensive evaluation value of operation quality through the comprehensive evaluation formula of operation quality;

[0032] When the comprehensive evaluation value of the operation quality is greater than the preset comprehensive evaluation threshold of the operation quality, an operation quality warning is generated; otherwise, an operation quality qualified prompt is generated;

[0033] A43: Summarize all data from sub-steps A41 to A42 to form a collection of run quality analysis data.

[0034] Furthermore, the sub-step A31 includes the following sub-steps:

[0035] A311: Based on the multivariate detection information, the surface of the glass plate product is divided into several regions, each of which has an equal area; the region with defective characteristics is marked as a poor quality region, and the proportion of the poor quality region to the total number of regions is set as the proportion of the poor quality region;

[0036] A312: Mark the area with continuous poor-quality areas as poor-quality areas, and set the total number of poor-quality areas in the largest poor-quality area as the poor-quality aggregation peak;

[0037] A313: The number of poor-quality areas in the poor-quality area that exceeds the preset aggregation threshold is set as the poor-quality hidden danger value;

[0038] A314: The proportion of poor quality areas, the peak value of poor quality concentration and the value of poor quality hidden dangers are calculated through the poor quality evaluation formula to obtain the poor quality value;

[0039] A315: Defective glass plate products are divided into multiple inferior quality grades according to the inferior quality degree value, and the proportion of defective glass plate products of each inferior quality grade to the total number of glass plate products is obtained.

[0040] Furthermore, in the sub-step A34, the comprehensive evaluation formula for production quality is:

[0041]

[0042] Among them, WP is the comprehensive evaluation value of production quality, WS is the production quality assessment value, WL is the comprehensive defect hidden danger value, and sy1 and sy2 are both preset proportional coefficients with values ​​greater than zero.

[0043] Furthermore, in the sub-step A41, the abnormal operation judgment formula is:

[0044] PM=(e1*PZ+e2*PS) / 2;

[0045] Among them, PM is the abnormal operation judgment value, PZ is the transmission vibration amplitude value, PS is the transmission speed deviation value, and e1 and e2 are both preset proportional coefficients with values ​​greater than zero;

[0046] In the sub-step A42, the comprehensive evaluation formula for operation quality is:

[0047]

[0048] Among them, YF is the comprehensive evaluation value of operation quality, QY is the transmission abnormality value, QL is the transmission abnormality table value, FX is the start-up detection value, FW is the start-up abnormality detection value, GP is the transport position deviation value, GS is the transport position abnormality detection value, q1, q2, q3, q4, q5 and q6 are all preset proportional coefficients with values ​​greater than zero.

[0049] Furthermore, in the sub-step A314, the poor quality evaluation formula is:

[0050]

[0051] Among them, XM is the proportion of poor quality areas, XK is the poor quality concentration peak, XS is the poor quality hidden danger value, XF is the poor quality degree value, and uk1, uk2 and uk3 are all preset proportional coefficients with values ​​greater than zero.

[0052] Furthermore, the step A1 includes: collecting high-resolution surface images and spectral data of the glass plate product, and aggregating them to form multivariate detection information.

[0053] Furthermore, step A2 includes: using a machine learning algorithm based on multivariate detection information to classify glass plate products with defective features as defective glass plate products, and the rest as qualified glass plate products, and then subdividing the defective glass plate products into defective glass plate products of different defect types according to specific defect features in the multivariate detection information, and aggregating the classified glass plate product data and the multivariate detection information to form a production quality data collection.

[0054] A glass plate product production line evaluation system based on multivariate information, comprising a multivariate information acquisition module, an identification and classification module, a production line production quality analysis module, a production line control module, a production line operation quality analysis module and a management terminal;

[0055] The multivariate information collection module is used to detect glass plate products and generate multivariate detection information;

[0056] The identification and classification module is used to receive the multivariate detection information, classify the glass plate products into qualified glass plate products or defective glass plate products according to the multivariate detection information, and subdivide the defect types of the defective glass plate products; after the multivariate information collection module has detected all the glass plate products, the data of all the qualified glass plate products and the defective glass plate products and the multivariate detection information are summarized to generate a production quality data collection;

[0057] The production line production quality analysis module is used to receive the production quality data collection, set multiple evaluation criteria to perform progressive analysis on the production quality data collection, and generate a production quality analysis data collection;

[0058] The production line operation quality analysis module is used to receive multiple feedback information from the management terminal, set multiple evaluation criteria to perform progressive analysis on the multiple feedback information, and generate an operation quality analysis data collection;

[0059] The management terminal is used to control the production line control module to drive the various mechanical components of the production line, and to receive the multi-dimensional feedback information from the various mechanical components of the production line from the production line control module; it is also used to receive and summarize the production quality data collection, the production quality analysis data collection and the operation quality analysis data collection.

[0060] The technical solution provided by the present invention may include the following beneficial effects: to achieve a comprehensive analysis of the production quality and operation quality of the production line, the production line needs to collect multivariate detection information (i.e., data including image data, spectral data, and other data reflecting the quality of the glass plate products) when detecting glass plate products, and classify each glass product produced on the production line in detail based on the multivariate detection information, so as to set multiple evaluation criteria for progressive analysis, thereby obtaining data for comprehensive and rigorous evaluation of production quality (i.e., a collection of production quality data); and when the production line is running, the multivariate feedback information (such as transmission speed, conveyor belt start and stop time, and other data reflecting the operation status of each mechanical component of the production line) fed back by each mechanical component of the production line (through various sensors, limit switches, etc.) is used to set multiple evaluation criteria for progressive analysis, thereby obtaining data for comprehensive and rigorous evaluation of operation quality (i.e., a collection of operation quality data); and finally, all the data of the above evaluation process are summarized, so that the user can learn the specific problems of the production quality and operation quality of the production line through these data, provide meaningful reference data for the user to optimize the production line, and make targeted improvement measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 The present invention is a flowchart of a glass plate product production line evaluation method based on multivariate information according to one embodiment of the present invention.

[0062] Figure 2 The present invention is a schematic diagram of a glass plate product production line evaluation system based on multivariate information according to one embodiment of the present invention.

[0063] Among them: multi-information collection module 1, identification and classification module 2, production line production quality analysis module 3, production line control module 4, production line operation quality analysis module 5, management terminal 6. DETAILED DESCRIPTION

[0064] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0065] In the description of the embodiments of the present invention, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0066] In the description of the embodiments of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the embodiments of the present invention can be understood according to specific circumstances.

[0067] Combine the following Figure 1 to Figure 2 , describing a glass plate product production line evaluation method and system based on multivariate information according to an embodiment of the present invention.

[0068] A glass plate product production line evaluation method based on multivariate information comprises the following steps:

[0069] A1: Inspect the glass plate products on the production line one by one and generate multi-dimensional inspection information one by one;

[0070] A2: According to the multivariate detection information of each glass plate product, the glass plate products are divided into qualified glass plate products and defective glass plate products, and the defective glass plate products are subdivided into defect types. Finally, the data of all qualified glass plate products and defective glass plate products and the multivariate detection information are summarized to generate a production quality data collection;

[0071] A3: Set multiple evaluation criteria to conduct progressive analysis on the production quality data collection to generate a production quality analysis data collection;

[0072] A4: Continuously receive multiple feedback information from various mechanical components of the production line, set multiple evaluation criteria to conduct progressive analysis on the multiple feedback information, and generate a collection of operation quality analysis data;

[0073] A5: Summarize the production quality data collection, production quality analysis data collection and operation quality analysis data collection to conduct a comprehensive evaluation of the production line.

[0074] The present invention proposes a preferred embodiment of a glass plate product production line evaluation method based on multivariate information, such as Figure 1 As shown in the figure, the quality of the finished glass plate products and the comprehensive operation of the mechanical components of the production line are comprehensively analyzed to evaluate the quality of the glass plate products production line. Users can also learn about the specific problems in the production quality and operation quality of the production line through the analysis process data, which provides meaningful reference data for users to optimize the production line and make targeted improvement measures.

[0075] To this end, in order to achieve a comprehensive analysis of the production quality and operation quality of the production line, it is necessary to collect multivariate detection information (i.e., image data, spectral data, and other data reflecting the quality of glass plate products) when the production line detects glass plate products, and to classify each glass product produced on the production line in detail based on the multivariate detection information, so as to set multiple evaluation criteria for progressive analysis, thereby obtaining data for comprehensive and rigorous evaluation of production quality (i.e., a collection of production quality data); and when the production line is running, the multivariate feedback information (such as transmission speed, conveyor belt start and stop time, and other data reflecting the operation status of various mechanical components of the production line) fed back by various mechanical components of the production line (through various sensors, limit switches, etc.) is used to set multiple evaluation criteria for progressive analysis, thereby obtaining data for comprehensive and rigorous evaluation of operation quality (i.e., a collection of operation quality data); finally, all the data of the above evaluation process are summarized, so that users can learn about the specific problems of the production quality and operation quality of the production line through these data, provide meaningful reference data for users to optimize the production line, and make targeted improvement measures.

[0076] It should be noted that the multivariate information includes multivariate detection information and multivariate feedback information.

[0077] Furthermore, step A3 includes the following sub-steps:

[0078] A31: Based on the production quality data collection, the defective glass plate products are divided into multiple inferior quality grades according to the severity of the defects from large to small, and the proportion of the defective glass plate products of each inferior quality grade to the total number of glass plate products is obtained;

[0079] A32: Configure a poor quality weight coefficient for each poor quality grade. The proportion of each poor quality grade is multiplied by the corresponding poor quality weight coefficient and then summed to obtain the production quality assessment value.

[0080] When the production quality assessment value is greater than the preset production quality threshold, a production quality warning is generated and sub-step A35 is executed; otherwise, sub-step A33 is executed;

[0081] A33: Based on the production quality data collection, the ratio of defective glass plate products of each defect type to the total number of glass plate products is compared with the preset hidden danger threshold of the corresponding defect type;

[0082] When the proportion of the defect type is greater than the preset hidden danger threshold, the defect type is marked as a hidden danger defect, and a production quality warning is generated, and sub-step A35 is executed; otherwise, sub-step A34 is executed;

[0083] A34: Configure a defect weight coefficient for each defect type. The proportion of each defect type is multiplied by the corresponding defect weight coefficient and then summed to obtain a comprehensive defect potential risk value. The production quality assessment value and the comprehensive defect potential risk value are calculated through the production quality comprehensive evaluation formula to obtain a production quality comprehensive evaluation value.

[0084] When the comprehensive evaluation value of production quality is greater than the preset comprehensive evaluation threshold of production quality, a production quality warning is generated; otherwise, a production quality qualified prompt is generated;

[0085] A35: Summarize all data from sub-steps A31 to A34 to form a collection of production quality analysis data.

[0086] In this embodiment, considering that different defects have different degrees of influence on glass plate products, it is possible that minor defects do not affect the quality of glass plate products. Therefore, when comprehensively judging the production quality, corresponding inferior weight coefficients should be configured for defective glass plate products with different degrees of severity, thus providing sub-steps A31 and A32. When the production quality evaluation value is greater than the preset production quality threshold, it indicates that the production quality is relatively serious in general, and it can be directly determined that there is a problem with the production quality (generating a production quality warning).

[0087] If the production quality assessment value is less than or equal to the preset production quality threshold, it does not mean that there is no problem with the production quality. Sub-step A33 is used to judge the defective glass plate products of all defect types to see whether there are one or more hidden defects. If so, it means that although the judgment of sub-step A32 did not find any production quality problems, there are actually hidden dangers in the product, and there is still a problem with the production quality.

[0088] If the judgment of sub-step A33 is still passed, it means that there is no prominent defect type. Then it is necessary to configure defect weight coefficients for all defect types (based on the degree of defect impact) through sub-step A34 to comprehensively judge whether there are hidden dangers in product quality.

[0089] Furthermore, step A4 includes the following sub-steps:

[0090] A41: Based on the transmission vibration amplitude value and the transmission speed value in the multivariate feedback information, the deviation value of the transmission speed value from the preset transmission speed value is set as the transmission speed deviation value; the transmission vibration amplitude value and the transmission speed deviation value are used to obtain the operation abnormality judgment value through the operation abnormality judgment formula;

[0091] Marking an operation abnormality judgment value greater than a preset abnormality threshold as an abnormal situation;

[0092] The ratio of the operation abnormality judgment values ​​marked as abnormal situations in a unit time to the total number of operation abnormality judgment values ​​is set as the transmission abnormality value;

[0093] The average of the operation abnormality judgment values ​​marked as abnormal situations within a unit time is set as the transmission abnormality table value;

[0094] The transmission abnormal condition value and the transmission abnormal table value are compared with the preset transmission abnormal condition threshold value and the preset transmission abnormal table threshold value respectively;

[0095] When the transmission abnormality value is greater than the preset transmission abnormality threshold or the transmission abnormal table value is greater than the preset transmission abnormal table threshold, a running quality warning is generated and sub-step A43 is executed in a loop; otherwise, sub-step A42 is executed;

[0096] A42: Based on the conveyor belt start-up time in the multivariate feedback information, the average of all conveyor belt start-up time in unit time is set as the start-up detection value; the conveyor belt start-up time longer than the preset start-up time is marked as an abnormal time, and the proportion of the conveyor belt start-up time marked as an abnormal time in the unit time to the total conveyor belt start-up time is set as the start-up abnormal detection value;

[0097] Based on the glass plate product stopping inspection position in the multivariate feedback information, the deviation value of the glass plate product stopping inspection position from the preset stopping position is set as the transportation deviation value, and the average of the transportation deviation values ​​per unit time is set as the transportation deviation value; the proportion of the transportation deviation values ​​greater than the preset transportation deviation value to the total number of transportation deviation values ​​is set as the transportation abnormal inspection value;

[0098] The transmission abnormal condition value, transmission abnormal table value, start detection value, start abnormal inspection value, transport position deviation value, and transport position abnormal inspection value are used to obtain the comprehensive evaluation value of operation quality through the comprehensive evaluation formula of operation quality;

[0099] When the comprehensive evaluation value of the operation quality is greater than the preset comprehensive evaluation threshold of the operation quality, an operation quality warning is generated; otherwise, an operation quality qualified prompt is generated;

[0100] A43: Summarize all data from sub-steps A41 to A42 to form a collection of run quality analysis data.

[0101] In this embodiment, the glass plate products in the production line are usually transported by a conveyor belt. The transmission vibration amplitude value and the transmission speed value of the glass plate product conveyor belt are collected, the vibration degree of the conveyor belt and the degree of deviation of the transmission speed at an instant are determined, and a comprehensive judgment is made whether it is abnormal at the instant (the abnormal judgment formula of the most recent real situation is obtained by collecting a large amount of data and performing software simulation to weigh the impact of the vibration degree and the degree of deviation of the transmission speed on the operation, and the operation abnormality judgment value is obtained), and then the severity of the abnormal situation at the instant in unit time is calculated, that is, it is judged by the transmission abnormality value and the transmission abnormality table value. If any one of them exceeds the preset value, it is considered that the production line operation quality is poor (generating an operation quality warning).

[0102] If the transmission abnormality value and the transmission abnormality table value do not exceed the preset value, it only means that there is no abnormality in terms of the vibration of the conveyor belt or the speed deviation. It cannot completely represent that there is no abnormality in the operation of the production line. It is necessary to combine the startup time from stop to restart during the inspection of the glass plate products, and the degree of deviation of the glass plate products until they stop to make a comprehensive judgment. By collecting a large amount of data and performing software simulation, the comprehensive evaluation formula for the operation quality of the most recent actual situation is obtained, and the degree of influence of each parameter is obtained to obtain the comprehensive evaluation value of the operation quality for the comprehensive evaluation standard of the operation quality.

[0103] Further, sub-step A31 includes the following sub-steps:

[0104] A311: Based on the multivariate detection information, the surface of the glass plate product is divided into several regions, each of which has an equal area; the region with defective characteristics is marked as a poor quality region, and the proportion of the poor quality region to the total number of regions is set as the proportion of the poor quality region;

[0105] A312: Mark the area with continuous poor-quality areas as poor-quality areas, and set the total number of poor-quality areas in the largest poor-quality area as the poor-quality aggregation peak;

[0106] A313: The number of poor-quality areas in the poor-quality area that exceeds the preset aggregation threshold is set as the poor-quality hidden danger value;

[0107] A314: The proportion of poor quality areas, the peak value of poor quality concentration and the value of poor quality hidden dangers are calculated through the poor quality evaluation formula to obtain the poor quality value;

[0108] A315: Defective glass plate products are divided into multiple inferior quality grades according to the inferior quality degree value, and the proportion of defective glass plate products of each inferior quality grade to the total number of glass plate products is obtained.

[0109] In the present embodiment, in sub-step A31, the method for dividing defective glass plate products into multiple inferior grades according to the severity of defects from large to small is preferably: the inferior grade value is obtained by the inferior area proportion value (the overall defect degree), the inferior concentration peak value (the maximum defect concentration degree) and the inferior potential risk value (the number of serious defects) through the inferior grade evaluation formula, and the inferior grade is divided according to the inferior grade value. This method of dividing the inferior grades takes into account more comprehensive and rigorous considerations.

[0110] It should be noted that the inferiority evaluation formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent actual situation. The preset parameters in the formula are set by technical personnel in this field according to actual conditions. The larger the calculated inferiority value is, the inferiorer the corresponding glass plate product is (one of the implementation methods of sub-step A315). An inferiority threshold value can be optionally set, and the inferiority threshold value is used as the dividing line. When the inferiority value exceeds the inferiority threshold value, it is a high-inferior defective glass plate product, otherwise it is a low-inferior defective glass plate product. The proportion values ​​are respectively obtained, and sub-step A34 can be performed.

[0111] Furthermore, in sub-step A34, the comprehensive evaluation formula for production quality is:

[0112]

[0113] Among them, WP is the comprehensive evaluation value of production quality, WS is the production quality assessment value, WL is the comprehensive defect hidden danger value, and sy1 and sy2 are both preset proportional coefficients with values ​​greater than zero.

[0114] In this embodiment, the above-mentioned comprehensive production quality evaluation formula is obtained by collecting a large amount of data and performing software simulation, which is a formula of the most recent actual situation.

[0115] Furthermore, in sub-step A41, the operation abnormality judgment formula is:

[0116] PM=(e1*PZ+e2*PS) / 2;

[0117] Among them, PM is the abnormal operation judgment value, PZ is the transmission vibration amplitude value, PS is the transmission speed deviation value, and e1 and e2 are both preset proportional coefficients with values ​​greater than zero;

[0118] In sub-step A42, the comprehensive evaluation formula for operation quality is:

[0119]

[0120] Among them, YF is the comprehensive evaluation value of operation quality, QY is the transmission abnormality value, QL is the transmission abnormality table value, FX is the start-up detection value, FW is the start-up abnormality detection value, GP is the transport position deviation value, GS is the transport position abnormality detection value, q1, q2, q3, q4, q5 and q6 are all preset proportional coefficients with values ​​greater than zero.

[0121] In this embodiment, the above-mentioned abnormal operation judgment formula and operation quality comprehensive evaluation formula are obtained by collecting a large amount of data and performing software simulation, which is a formula of the most recent actual situation, and the combination of the two is more in line with the evaluation system of step A4.

[0122] Furthermore, in sub-step A314, the poor quality evaluation formula is:

[0123]

[0124] Among them, XM is the proportion of poor quality areas, XK is the poor quality concentration peak, XS is the poor quality hidden danger value, XF is the poor quality degree value, and uk1, uk2 and uk3 are all preset proportional coefficients with values ​​greater than zero.

[0125] In this embodiment, the above-mentioned inferior quality evaluation formula is obtained by collecting a large amount of data and performing software simulation, which is a formula of the most recent actual situation.

[0126] Furthermore, step A1 includes: collecting high-resolution surface images and spectral data of the glass plate product, and aggregating them to form multivariate detection information.

[0127] In this embodiment, the multivariate detection information preferably includes high-resolution surface images and spectral data of the glass plate products, which can comprehensively evaluate the quality of the glass plate products from the surface and the inside, and has higher defect detection accuracy; for this purpose, the production line should be equipped with high-definition cameras and spectrometers, and the spectrometer should be able to cover visible light and part of the ultraviolet light bands, so as to more deeply reflect the internal conditions of the glass plate products.

[0128] Furthermore, step A2 includes: using a machine learning algorithm to classify glass plate products with defective features as defective glass plate products based on multivariate detection information, and the rest as qualified glass plate products, and then subdividing the defective glass plate products into defective glass plate products of different defect types according to specific defect features in the multivariate detection information, and aggregating the classified glass plate product data and the multivariate detection information to form a production quality data collection.

[0129] In this embodiment, a machine learning algorithm (such as a convolutional neural network CNN, a support vector machine SVM, etc.) is used to determine whether the glass plate product has defects based on multivariate detection information, and the defects of the glass plate product are automatically identified and classified (such as scratches, cracks, bubbles, foreign matter, and other defect types). That is, the defects are divided into different categories according to their shape, size, color, and other characteristics. The refined classified data and multivariate detection information are aggregated to form a production quality data collection, which can have sufficient data volume for production quality analysis.

[0130] A glass plate product production line evaluation system based on multivariate information, comprising a multivariate information acquisition module 1, an identification and classification module 2, a production line production quality analysis module 3, a production line control module 4, a production line operation quality analysis module 5 and a management terminal 6;

[0131] The multi-information collection module 1 is used to detect glass plate products and generate multi-information detection information;

[0132] The identification and classification module 2 is used to receive multivariate detection information, classify the glass plate products into qualified glass plate products or defective glass plate products according to the multivariate detection information, and subdivide the defect types of defective glass plate products; after the multivariate information collection module 1 has detected all glass plate products, the data of all qualified glass plate products and defective glass plate products and the multivariate detection information are summarized to generate a collection of production quality data;

[0133] The production line production quality analysis module 3 is used to receive the production quality data collection, set multiple evaluation criteria to perform progressive analysis on the production quality data collection, and generate the production quality analysis data collection;

[0134] The production line operation quality analysis module 5 is used to receive multiple feedback information from the management terminal 6, set multiple evaluation criteria to perform progressive analysis on the multiple feedback information, and generate a collection of operation quality analysis data;

[0135] The management terminal 6 is used to control the production line control module 4 to drive the various mechanical components of the production line, and to receive multiple feedback information from the various mechanical components of the production line from the production line control module 4; it is also used to receive a collection of summarized production quality data, a collection of production quality analysis data, and a collection of operation quality analysis data.

[0136] In this embodiment, a preferred embodiment of a glass plate product production line evaluation system based on multivariate information is also proposed, such as Figure 2As shown, the function of the multi-information acquisition module 1 corresponds to step A1 in the above-mentioned evaluation method, the function of the identification and classification module 2 corresponds to step A2 in the above-mentioned evaluation method, the function of the production line production quality analysis module 3 corresponds to step A3 in the above-mentioned evaluation method, the functions of the production line control module 4 and the production line operation quality analysis module 5 correspond to step A4 in the above-mentioned evaluation method, and the function of the management terminal 6 corresponds to step A5 in the above-mentioned evaluation method, thereby building a feasible evaluation system for implementing the above-mentioned evaluation method.

[0137] It should be noted that the multi-information acquisition module 1 is composed of a high-definition camera, a spectrometer and various sensors, and the identification and classification module 2, the production line production quality analysis module 3, the production line control module 4 (that is, the production line general control module, responsible for receiving or sending signals to various motors, sensors, switches and other components of the production line, so that multi-dimensional feedback information can be obtained) and the production line operation quality analysis module 5 are integrated in the lower computer, and the management terminal 6 is the upper computer.

[0138] A glass plate product production line evaluation method based on multivariate information according to an embodiment of the present invention and other components and operations of the system thereof are well known to those skilled in the art and will not be described in detail herein.

[0139] In the description of this specification, the description with reference to the terms "embodiment", "example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0140] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A glass plate product production line evaluation method based on multivariate information, characterized in that: The following steps are involved: A1: Inspect the glass plate products on the production line one by one and generate multi-dimensional inspection information one by one; A2: According to the multivariate detection information of each glass plate product, the glass plate products are divided into qualified glass plate products and defective glass plate products, and the defective glass plate products are subdivided into defect types. Finally, the data of all qualified glass plate products and defective glass plate products and the multivariate detection information are summarized to generate a production quality data collection; A3: Set multiple evaluation criteria to conduct progressive analysis on the production quality data collection to generate a production quality analysis data collection; A4: Continuously receive multiple feedback information from various mechanical components of the production line, set multiple evaluation criteria to conduct progressive analysis on the multiple feedback information, and generate a collection of operation quality analysis data; A5: Summarize the production quality data collection, production quality analysis data collection and operation quality analysis data collection to conduct comprehensive evaluation of the production line; The step A1 comprises: collecting high-resolution surface images and spectral data of the glass plate product, and aggregating them to form multivariate detection information; The step A3 comprises the following sub-steps: A31: Based on the production quality data collection, the defective glass plate products are divided into multiple inferior quality grades according to the severity of the defects from large to small, and the proportion of the defective glass plate products of each inferior quality grade to the total number of glass plate products is obtained; A32: Configure a poor quality weight coefficient for each poor quality grade. The proportion of each poor quality grade is multiplied by the corresponding poor quality weight coefficient and then summed to obtain the production quality assessment value. When the production quality assessment value is greater than the preset production quality threshold, a production quality warning is generated and sub-step A35 is executed; otherwise, sub-step A33 is executed; A33: Based on the production quality data collection, the ratio of defective glass plate products of each defect type to the total number of glass plate products is compared with the preset hidden danger threshold of the corresponding defect type; When the proportion of the defect type is greater than the preset hidden danger threshold, the defect type is marked as a hidden danger defect, and a production quality warning is generated, and sub-step A35 is executed; otherwise, sub-step A34 is executed; A34: Configure a defect weight coefficient for each defect type. The proportion of each defect type is multiplied by the corresponding defect weight coefficient and then summed to obtain a comprehensive defect potential risk value. The production quality assessment value and the comprehensive defect potential risk value are calculated through the production quality comprehensive evaluation formula to obtain a production quality comprehensive evaluation value. When the comprehensive evaluation value of production quality is greater than the preset comprehensive evaluation threshold of production quality, a production quality warning is generated; otherwise, a production quality qualified prompt is generated; A35: Summarize all data from sub-steps A31 to A34 to form a collection of production quality analysis data; The step A4 comprises the following sub-steps: A41: Based on the transmission vibration amplitude value and the transmission speed value in the multivariate feedback information, the deviation value of the transmission speed value from the preset transmission speed value is set as the transmission speed deviation value; the transmission vibration amplitude value and the transmission speed deviation value are used to obtain the operation abnormality judgment value through the operation abnormality judgment formula; Marking an operation abnormality judgment value greater than a preset abnormality threshold as an abnormal situation; The ratio of the operation abnormality judgment values ​​marked as abnormal situations in a unit time to the total number of operation abnormality judgment values ​​is set as the transmission abnormality value; The average of the operation abnormality judgment values ​​marked as abnormal situations within a unit time is set as the transmission abnormality table value; The transmission abnormal condition value and the transmission abnormal table value are compared with the preset transmission abnormal condition threshold value and the preset transmission abnormal table threshold value respectively; When the transmission abnormality value is greater than the preset transmission abnormality threshold or the transmission abnormal table value is greater than the preset transmission abnormal table threshold, a running quality warning is generated and sub-step A43 is executed in a loop; otherwise, sub-step A42 is executed; A42: Based on the conveyor belt start-up time in the multivariate feedback information, the average of all conveyor belt start-up time in unit time is set as the start-up detection value; the conveyor belt start-up time longer than the preset start-up time is marked as an abnormal time, and the proportion of the conveyor belt start-up time marked as an abnormal time in the unit time to the total conveyor belt start-up time is set as the start-up abnormal detection value; Based on the glass plate product stopping inspection position in the multivariate feedback information, the deviation value of the glass plate product stopping inspection position from the preset stopping position is set as the transportation deviation value, and the average of the transportation deviation values ​​per unit time is set as the transportation deviation value; the proportion of the transportation deviation values ​​greater than the preset transportation deviation value to the total number of transportation deviation values ​​is set as the transportation abnormal inspection value; The transmission abnormal condition value, transmission abnormal table value, start detection value, start abnormal inspection value, transport position deviation value, and transport position abnormal inspection value are used to obtain the comprehensive evaluation value of operation quality through the comprehensive evaluation formula of operation quality; When the comprehensive evaluation value of the operation quality is greater than the preset comprehensive evaluation threshold of the operation quality, an operation quality warning is generated; otherwise, an operation quality qualified prompt is generated; A43: Summarize all data from sub-steps A41 to A42 to form a collection of run quality analysis data.

2. The glass plate product production line evaluation method based on multivariate information according to claim 1, characterized in that: The sub-step A31 includes the following sub-steps: A311: Based on the multivariate detection information, the surface of the glass plate product is divided into several regions, each of which has an equal area; the region with defective characteristics is marked as a poor quality region, and the proportion of the poor quality region to the total number of regions is set as the proportion of the poor quality region; A312: Mark the area with continuous poor-quality areas as poor-quality areas, and set the total number of poor-quality areas in the largest poor-quality area as the poor-quality aggregation peak; A313: The number of poor-quality areas in the poor-quality area that exceeds the preset aggregation threshold is set as the poor-quality hidden danger value; A314: The proportion of poor quality areas, the peak value of poor quality concentration and the value of poor quality hidden dangers are calculated through the poor quality evaluation formula to obtain the poor quality value; A315: Defective glass plate products are divided into multiple inferior quality grades according to the inferior quality degree value, and the proportion of defective glass plate products of each inferior quality grade to the total number of glass plate products is obtained.

3. The glass plate product production line evaluation method based on multivariate information according to claim 1, characterized in that: In the sub-step A34, the comprehensive evaluation formula for production quality is: ； Among them, WP is the comprehensive evaluation value of production quality, WS is the production quality assessment value, WL is the comprehensive defect hidden danger value, and sy1 and sy2 are both preset proportional coefficients with values ​​greater than zero.

4. The glass plate product production line evaluation method based on multivariate information according to claim 1, characterized in that: In the sub-step A41, the abnormal operation judgment formula is: PM=(e1*PZ+e2*PS) / 2; Among them, PM is the abnormal operation judgment value, PZ is the transmission vibration amplitude value, PS is the transmission speed deviation value, and e1 and e2 are both preset proportional coefficients with values ​​greater than zero; In the sub-step A42, the comprehensive evaluation formula for operation quality is: ； Among them, YF is the comprehensive evaluation value of operation quality, QY is the transmission abnormality value, QL is the transmission abnormality table value, FX is the start-up detection value, FW is the start-up abnormality detection value, GP is the transport position deviation value, GS is the transport position abnormality detection value, q1, q2, q3, q4, q5 and q6 are all preset proportional coefficients with values ​​greater than zero.

5. The glass plate product production line evaluation method based on multivariate information according to claim 2, characterized in that: In the sub-step A314, the poor quality evaluation formula is: ； Among them, XM is the proportion of poor quality areas, XK is the poor quality concentration peak, XS is the poor quality hidden danger value, XF is the poor quality degree value, and uk1, uk2 and uk3 are all preset proportional coefficients with values ​​greater than zero.

6. The glass plate product production line evaluation method based on multivariate information according to claim 1, characterized in that: The step A2 comprises: using a machine learning algorithm to classify glass plate products with defective features as defective glass plate products and the rest as qualified glass plate products based on multivariate detection information, and then subdividing the defective glass plate products into defective glass plate products of different defect types according to the specific defect features in the multivariate detection information, and aggregating the classified glass plate product data and the multivariate detection information to form a production quality data collection.

7. A glass plate product production line evaluation system based on multivariate information, characterized in that: Applying a glass plate product production line evaluation method based on multivariate information according to any one of claims 1 to 6; It includes multi-information collection module, identification and classification module, production line production quality analysis module, production line control module, production line operation quality analysis module and management terminal; The multivariate information collection module is used to detect glass plate products and generate multivariate detection information; The identification and classification module is used to receive the multivariate detection information, classify the glass plate products into qualified glass plate products or defective glass plate products according to the multivariate detection information, and subdivide the defect types of the defective glass plate products; after the multivariate information collection module has detected all the glass plate products, the data of all the qualified glass plate products and the defective glass plate products and the multivariate detection information are summarized to generate a production quality data collection; The production line production quality analysis module is used to receive the production quality data collection, set multiple evaluation criteria to perform progressive analysis on the production quality data collection, and generate a production quality analysis data collection; The production line operation quality analysis module is used to receive multiple feedback information from the management terminal, set multiple evaluation criteria to perform progressive analysis on the multiple feedback information, and generate an operation quality analysis data collection; The management terminal is used to control the production line control module to drive the various mechanical components of the production line, and to receive the multi-dimensional feedback information from the various mechanical components of the production line from the production line control module; it is also used to receive and summarize the production quality data collection, the production quality analysis data collection and the operation quality analysis data collection.

Images