Industrial manufacturing enterprise data resource assessment management method and system

Abstract

The invention relates to the technical field of enterprise data resource assessment management, in particular to an industrial manufacturing enterprise data resource assessment management method and system.The method comprises the steps that first data, second data and third data of a time window on a mineral water production line are obtained and processed to obtain comprehensive detection data; a mineral water quality detection model is input, image features, weight features and water quality features are extracted through a feature extraction layer through the feature extraction layer, a feature analysis layer and an output layer in sequence, the features are analyzed according to the feature analysis layer, and a first quality score value, a second quality score value and a third quality score value are calculated. And the output layer compares the comprehensive quality score value with a preset threshold value, and judges the quality level of the bottled mineral water. And based on the order quantity, the order returning quantity and the qualified rate of the previous time window and the order quantity of the next time window, determining the production plan quantity of the next time window so as to optimize production management.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise data resource evaluation management, and in particular to a method and system for industrial manufacturing enterprise data resource evaluation management. Background Art

[0002] In modern industrial manufacturing enterprises, data resources have become the core assets of enterprise operations. Enterprise data resources cover data generated in various links, from production processes, supply chain management, quality control, market demand analysis to customer service.

[0003] By effectively managing and utilizing data resources, enterprises can achieve more scientific decision-making, more efficient operations and more forward-looking risk management, laying a solid foundation for the long-term development of enterprises. If enterprise data resources are not used reasonably, it is easy to lead to overproduction of enterprises, thus affecting the production efficiency of enterprises.

[0004] Therefore, how to reasonably utilize enterprise data resources is of great significance to effectively evaluate enterprise product quality and improve enterprise production and management efficiency. However, in the production process of traditional industrial manufacturing enterprises, traditional data resource processing methods are difficult to effectively integrate multi-source data, resulting in insufficient enterprise product quality evaluation and lack of scientificity in enterprise production plans, which easily leads to waste of enterprise resources or low production efficiency.

[0005] Therefore, a data resource evaluation management method and system for industrial manufacturing enterprises are proposed. Summary of the invention

[0006] The purpose of the present invention is to provide a data resource evaluation management method and system for industrial manufacturing enterprises. The present invention relates to the technical field of enterprise data resource evaluation management, and specifically to a data resource evaluation management method and system for industrial manufacturing enterprises. The present invention obtains the first data, the second data and the third data of a time window on the mineral water production line, processes the three data, obtains comprehensive detection data, inputs the mineral water quality detection model, sequentially passes through the feature extraction layer, the feature analysis layer and the output layer, extracts image features, weight features and water quality features through the feature extraction layer, analyzes these features according to the feature analysis layer, calculates the first quality score value, the second quality score value and the third quality score value, and obtains the comprehensive quality score value, and the output layer compares the comprehensive quality score value with the preset threshold value to judge the quality level of the bottled mineral water. Based on the order quantity, return quantity, qualified rate of the previous time window and the order quantity of the next time window, the production plan quantity of the next time window is determined to optimize production management.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A data resource evaluation and management method for an industrial manufacturing enterprise, comprising:

[0009] The first data, the second data and the third data of a time window on the production line of a bottled mineral water enterprise are obtained; the first data includes the bottle body image, the label image and the bottle cap image of the bottled mineral water taken by the camera; the second data includes the bottle body weight, the water weight and the total weight of the bottled mineral water obtained by the weighing sensor; the third data includes the pH value, the mineral content and the conductivity of the bottled mineral water obtained by the water quality monitoring equipment; the first detection data, the second detection data and the third detection data are obtained by preprocessing the three kinds of data, and the three kinds of detection data are integrated; the comprehensive detection data is obtained and input into the mineral water quality detection model;

[0010] The model passes through the feature extraction layer, feature analysis layer and output layer in sequence; the feature extraction layer extracts features from the comprehensive detection data to obtain image features, weight features and water quality features; the feature analysis layer analyzes the three features in combination with the attention mechanism to obtain a first quality score value, a second quality score value and a third quality score value, and obtains a comprehensive quality score value based on the three quality scores; the output layer is used to compare the comprehensive quality score value with a preset threshold value to determine the quality level of the bottled mineral water; the quality levels include qualified and unqualified;

[0011] The production volume of bottled mineral water in the next time window is determined based on the company's order volume, order return volume, and pass rate for bottled mineral water in the previous time window and the order volume in the next time window.

[0012] Preferably, the preprocessing process is to perform cropping, denoising, grayscale and normalization on the first data; and to perform missing value processing, standardization and normalization on the second data and the third data.

[0013] Preferably, the comprehensive detection data is:

[0014] AF i =[I i ,WE i ,WQ i ];

[0015] I i =[I bottle,i ,I label,i ,I cap,i ];

[0016] WE i =[WE b,i ,WE w,i ,WE wb,i ];

[0017] WQ i =[WQ p,i ,WQ m,i ,WQ d,i ];

[0018] WQ m,i =[Q ca,i ,Q mg,i ,Q k,i ,Q na,i ,Q hco3,i ,Q so4,i ,Q cl,i ];

[0019] Among them, AF i represents the comprehensive test data of the i-th bottled mineral water; I i represents the first detection data of the i-th bottled mineral water; WE i represents the second detection data of the i-th bottled mineral water; WQ i represents the third detection data of the i-th bottled mineral water; bottle,i I represents the bottle image data of the i-th bottle of mineral water; label,i Represents the label image data of the i-th bottled mineral water; cap,i represents the bottle cap image data of the i-th bottled mineral water; WE b,i represents the weight of the bottle of the i-th bottle of mineral water; WE w,i represents the water weight of the i-th bottle of mineral water; WE wb,i represents the total weight of the i-th bottled mineral water; WQ p,i represents the pH value of the i-th bottled mineral water; WQ m,i represents the mineral content data of the i-th bottled mineral water; WQ d,i represents the conductivity data of the i-th bottled mineral water; Q ca,i , Q mg,i , Q k,i , Q na,i , Q hco3,i , Q so4,i , Q cl,i Represents the content of calcium, magnesium, potassium, sodium, bicarbonate, sulfate and chloride in the i-th bottled mineral water.

[0020] Preferably, the first quality score is:

[0021]

[0022] Among them, R 1,i represents the first score value of the i-th bottled mineral water; LeakyReLU represents the activation function; n 1Indicates the number of image features; m j represents the dimension of the jth image feature; f image,jh represents the hth feature value of the jth image feature; Attention(f image,jh ) represents the attention mechanism weight of the hth feature value of the jth image feature of the i-th bottle of mineral water;

[0023] The second quality score is:

[0024]

[0025] Among them, R 2,i represents the second score of the i-th bottled mineral water; n 2 represents the number of weight features; f weight,k represents the k-th eigenvalue; LeakyReLU represents the activation function; Attention(f weight,k ) represents the attention mechanism weight of the k-th weight feature value;

[0026] The third quality score value is:

[0027]

[0028] Among them, R 2,i represents the third score value of the i-th bottled mineral water; n 3 Indicates the quantity of water quality characteristics; f water,m represents the mth water quality feature value; LeakyReLU represents the activation function; Attention(f weight,m ) represents the attention mechanism weight of the m-th water quality feature value.

[0029] Preferably, the comprehensive quality score is:

[0030] AR i =λ 1 *R 1,i +λ 2 *R 2,i +λ 3 *R 3,i +γ*ln(1+R 1,i *R 2,i *R 3,i );

[0031] Among them, AR i represents the comprehensive quality score of the i-th bottled mineral water; 1 represents the weight of the first quality score; 2 represents the second quality score weight; 3 represents the weight of the third quality score; ln is the logarithm with base e.

[0032] Preferably, the production volume of bottled mineral water in the next time window is:

[0033]

[0034] Among them, MF next Indicates the production volume of bottled mineral water in the next time window; OR next Indicates the order volume of the next time window; BQ pre represents the qualified number of bottled mineral water in the previous time window; ξ represents the compensation coefficient for non-qualified products; ROR pre represents the number of returned orders for bottled mineral water in the previous time window; ψ represents the adjustment coefficient for returned orders; θ represents the feedback coefficient; OR pre represents the order volume of bottled mineral water in the previous time window; AC pre Indicates the pass rate of bottled mineral water in the previous time window.

[0035] An industrial manufacturing enterprise data resource evaluation and management system, the system comprising:

[0036] A data acquisition module is used to acquire first data, second data and third data of a time window on a production line of a bottled mineral water enterprise; the first data includes a bottle body image, a label image and a bottle cap image of the bottled mineral water taken by a camera; the second data includes the bottle body weight, water weight and total weight of the bottled mineral water obtained by a weighing sensor; the third data includes the pH value, mineral content and conductivity of the bottled mineral water obtained by a water quality monitoring device; by preprocessing the three types of data, first detection data, second detection data and third detection data are obtained, and the three types of detection data are integrated; comprehensive detection data is obtained and input into a mineral water quality detection model;

[0037] A quality assessment module, used to assess the quality of bottled water through the feature extraction layer, feature analysis layer and output layer of the model; the feature extraction layer extracts features from the comprehensive detection data to obtain image features, weight features and water quality features; the feature analysis layer analyzes the three features in combination with the attention mechanism to obtain a first quality score value, a second quality score value and a third quality score value, and obtains a comprehensive quality score value based on the three quality score values;

[0038] The output layer is used to determine the quality level of the bottled mineral water according to the comparison between the comprehensive quality score value and a preset threshold value; the quality level includes qualified and unqualified;

[0039] The output determination module is used to determine the production volume of bottled mineral water in the next time window based on the order volume, return volume, qualified rate of bottled mineral water of the enterprise in the previous time window and the order volume in the next time window.

[0040] Preferably, the first quality score is:

[0041]

[0042] Among them, R 1,i represents the first score value of the i-th bottled mineral water; LeakyReLU represents the activation function; n 1 Indicates the number of image features; m j represents the dimension of the jth image feature; f image,jh represents the hth feature value of the jth image feature; Attention(f image,jh ) represents the attention mechanism weight of the hth feature value of the jth image feature of the i-th bottle of mineral water;

[0043] The second quality score is:

[0044]

[0045] Among them, R 2,i represents the second score of the i-th bottled mineral water; n 2 represents the number of weight features; f weight,k represents the k-th eigenvalue; LeakyReLU represents the activation function; Attention(f weight,k ) represents the attention mechanism weight of the k-th weight feature value;

[0046] The third quality score value is:

[0047]

[0048] Among them, R 2,i represents the third score value of the i-th bottled mineral water; n 3 Indicates the quantity of water quality characteristics; f water,m represents the mth water quality feature value; LeakyReLU represents the activation function; Attention(f weight,m ) represents the attention mechanism weight of the m-th water quality feature value.

[0049] Preferably, the comprehensive quality score is:

[0050] AR i =λ 1 *R 1,i +λ 2 *R 2,i +λ3 *R 3,i +γ*ln(1+R 1,i *R 2,i *R 3,i );

[0051] Among them, AR i represents the comprehensive quality score of the i-th bottled mineral water; 1 represents the weight of the first quality score; 2 represents the second quality score weight; 3 represents the weight of the third quality score; ln is the logarithm with base e.

[0052] Preferably, the production volume of bottled mineral water in the next time window is:

[0053]

[0054] Among them, MF next Indicates the production volume of bottled mineral water in the next time window; OR next Indicates the order volume of the next time window; BQ pre represents the qualified number of bottled mineral water in the previous time window; ξ represents the compensation coefficient for non-qualified products; ROR pre represents the number of returned orders for bottled mineral water in the previous time window; ψ represents the adjustment coefficient for returned orders; θ represents the feedback coefficient; OR pre represents the order volume of bottled mineral water in the previous time window; AC pre Indicates the pass rate of bottled mineral water in the previous time window.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] 1. The present invention preprocesses and integrates multi-source data to construct the first data (bottle body image, label image and bottle cap image of bottled mineral water), the second data (bottle body weight, water weight and total weight of bottled mineral water) and the third data (pH value, mineral content and conductivity of bottled mineral water) of a time window on the production line of a bottled mineral water enterprise into comprehensive detection data, providing accurate and rich input data for the model, ensuring the accuracy and effectiveness of bottled mineral water quality detection, and facilitating the enterprise to adjust the production plan according to the qualified rate, thereby improving the product quality and production management efficiency of the enterprise.

[0057] 2. The present invention realizes a comprehensive and detailed evaluation of the quality of mineral water through a mineral water quality detection model. The feature extraction layer extracts the image features, weight features and water quality features of bottled mineral water, and calculates the first quality score value, the second quality score value and the third quality score value through the feature analysis layer. The comprehensive quality score value is obtained by comprehensive calculation and compared with the preset threshold value to determine the quality level of the mineral water. The comprehensive and accurate analysis of the mineral water quality detection model is conducive to the accurate detection of the quality of bottled mineral water and lays a good foundation for the production plan of the next cycle, thereby improving the product quality and production management efficiency of the enterprise.

[0058] 3. The present invention calculates the qualified rate of bottled mineral water in the previous time window, and determines the production volume in the next time window in combination with the order volume, order return volume and order volume in the next cycle in the previous time window, so that the enterprise can flexibly adjust the production plan, optimize resource allocation, reduce production costs, and thus improve the production management efficiency of the enterprise. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A flow chart of a data resource evaluation management method for an industrial manufacturing enterprise provided by an embodiment of the present invention;

[0060] Figure 2 A module diagram of an industrial manufacturing enterprise data resource evaluation management system provided by an embodiment of the present invention;

[0061] Figure 3 Schematic diagram of a mineral water quality detection model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0063] Embodiment 1

[0064] Company A is an industrial manufacturing company, mainly responsible for the production of bottled mineral water. In order to improve the production management efficiency of bottled mineral water, Company A has applied a data resource assessment and management method for industrial manufacturing companies.

[0065] A data resource evaluation and management method for an industrial manufacturing enterprise comprises the following steps:

[0066] Reference Figure 1In S10, the first data, the second data and the third data of a time window on the production line of the bottled mineral water enterprise are obtained, and the first detection data, the second detection data and the third detection data are obtained by preprocessing the three kinds of data, and the three kinds of detection data are integrated; the comprehensive detection data is obtained and input into the mineral water quality detection model; the schematic diagram of the mineral water quality detection model is shown in Figure 3 As shown;

[0067] In this embodiment, the previous time window is a time window of one month;

[0068] The first data includes the bottle body image, label image and bottle cap image of the bottled mineral water taken by a camera; the second data includes the bottle body weight, water weight and total weight of the bottled mineral water obtained by a weighing sensor; the third data includes the pH value, mineral content and conductivity of the bottled mineral water obtained by a water quality monitoring device; the preprocessing process is to crop, denoise, grayscale and normalize the first data; and perform missing value processing, standardization and normalization on the second data and the third data.

[0069] Furthermore, the comprehensive detection data is:

[0070] AF i =[I i ,WE i ,WQ i ];

[0071] I i =[I bottle,i ,I label,i ,I cap,i ];

[0072] WE i =[WE b,i ,WE w,i ,WE wb,i ];

[0073] WQ i =[WQ p,i ,WQ m,i ,WQ d,i ];

[0074] WQ m,i =[Q ca,i ,Q mg,i ,Q k,i ,Q na,i ,Q hco3,i ,Q so4,i ,Q cl,i ];

[0075] Among them, AF irepresents the comprehensive test data of the i-th bottled mineral water; I i represents the first detection data of the i-th bottled mineral water; WE i represents the second detection data of the i-th bottled mineral water; WQ i represents the third detection data of the i-th bottled mineral water; bottle,i I represents the bottle image data of the i-th bottle of mineral water; label,i Represents the label image data of the i-th bottled mineral water; cap,i represents the bottle cap image data of the i-th bottled mineral water; WE b,i represents the weight of the bottle of the i-th bottle of mineral water; WE w,i represents the water weight of the i-th bottle of mineral water; WE wb,i represents the total weight of the i-th bottled mineral water; WQ p,i represents the pH value of the i-th bottled mineral water; WQ m,i represents the mineral content data of the i-th bottled mineral water; WQ d,i represents the conductivity data of the i-th bottled mineral water; Q ca,i , Q mg,i , Q k,i , Q na,i , Q hco3,i , Q so4,i , Q cl,i Represents the content of calcium, magnesium, potassium, sodium, bicarbonate, sulfate and chloride in the i-th bottled mineral water.

[0076] This embodiment preprocesses and integrates multi-source data to construct the first data (bottle body image, label image and bottle cap image of bottled mineral water), the second data (bottle body weight, water weight and total weight of bottled mineral water) and the third data (pH value, mineral content and conductivity of bottled mineral water) of a time window on the production line of a bottled mineral water enterprise into comprehensive detection data, providing accurate and rich input data for the model, ensuring the accuracy and effectiveness of bottled mineral water quality detection, and facilitating the enterprise to adjust the production plan according to the qualified rate, thereby improving the product quality and production management efficiency of the enterprise.

[0077] Reference Figure 1 In S20, the feature extraction layer of the model extracts features from the comprehensive detection data to obtain image features, weight features, and water quality features; and the feature analysis layer combines the attention mechanism to analyze the three features to obtain a first quality score value, a second quality score value, and a third quality score value, and obtains a comprehensive quality score value based on the three quality score values;

[0078] Furthermore, the first quality score value is:

[0079]

[0080] Among them, R 1,i represents the first score value of the i-th bottled mineral water; LeakyReLU represents the activation function; n 1 Indicates the number of image features; m j represents the dimension of the jth image feature; f image,jh represents the hth feature value of the jth image feature; Attention(f image,jh ) represents the attention mechanism weight of the hth feature value of the jth image feature of the i-th bottle of mineral water;

[0081] The image features include: the length of the crack on the bottle surface, the width of the crack and the direction of the crack, the relative position and rotation angle between the edge of the bottle cap and the bottle body, the position offset of the label, the printing clarity and color difference;

[0082] The second quality score is:

[0083]

[0084] Among them, R 2,i represents the second score of the i-th bottled mineral water; n 2 represents the number of weight features; f weight,k represents the k-th eigenvalue; LeakyReLU represents the activation function; Attention(f weight,k ) represents the attention mechanism weight of the k-th weight feature value;

[0085] The weight characteristics include weight distribution deviation and filling uniformity; the weight distribution deviation indicates the deviation between the total weight of each bottle of mineral water and the target weight; the filling uniformity indicates the ratio between the water weight and the bottle weight during the filling process;

[0086] The third quality score value is:

[0087]

[0088] Among them, R 2,i represents the third score value of the i-th bottled mineral water; n 3 Indicates the quantity of water quality characteristics; f water,m represents the mth water quality feature value; LeakyReLU represents the activation function; Attention(f weight,m ) represents the attention mechanism weight of the m-th water quality feature value.

[0089] The water quality characteristics include pH deviation, mineral ratio and the correlation between conductivity and mineral content; wherein pH deviation indicates the deviation between the actual measured pH value and the standard pH value; mineral ratio includes the ratio of calcium, magnesium, potassium, sodium, bicarbonate, sulfate and chloride content in bottled mineral water to the total mineral content; the correlation between conductivity and mineral content is used to analyze the correlation between conductivity and mineral content to evaluate the relationship between the total amount of dissolved ions in water and the mineral concentration. The correlation is calculated by the Pearson correlation coefficient;

[0090] Furthermore, the comprehensive quality score is:

[0091] AR i =λ 1 *R 1,i +λ 2 *R 2,i +λ 3 *R 3,i +γ*ln(1+R 1,i *R 2,i *R 3,i );

[0092] Among them, AR i represents the comprehensive quality score of the i-th bottled mineral water; 1 represents the weight of the first quality score; 2 represents the second quality score weight; 3 represents the weight of the third quality score; ln is the logarithm with base e.

[0093] This embodiment uses a mineral water quality detection model to achieve a comprehensive and detailed assessment of the quality of mineral water. The feature extraction layer extracts the image features, weight features and water quality features of bottled mineral water, and calculates the first quality score value, the second quality score value and the third quality score value through the feature analysis layer. The comprehensive quality score value is calculated comprehensively and compared with the preset threshold to determine the quality level of the mineral water. The comprehensive and accurate analysis of the mineral water quality detection model is conducive to the accurate detection of the mineral water quality and lays a good foundation for the production plan of the next cycle, thereby improving the product quality and production management efficiency of the enterprise.

[0094] Reference Figure 1 S30 in which the comprehensive quality score value is compared with a preset threshold value through an output layer to determine the quality level of the bottled mineral water; the quality level includes qualified and unqualified;

[0095] Reference Figure 1 S40, determining the production volume of bottled mineral water in the next time window according to the order volume, order return volume, qualified rate of bottled mineral water of the enterprise in the previous time window and the order volume in the next time window;

[0096] Furthermore, the production volume of bottled mineral water in the next time window is:

[0097]

[0098] Among them, MF next Indicates the production volume of bottled mineral water in the next time window; OR next Indicates the order volume of the next time window; BQ pre represents the qualified number of bottled mineral water in the previous time window; ξ represents the compensation coefficient for non-qualified products; ROR pre represents the number of returned orders for bottled mineral water in the previous time window; ψ represents the adjustment coefficient for returned orders; θ represents the feedback coefficient; OR pre represents the order volume of bottled mineral water in the previous time window; AC pre Indicates the pass rate of bottled mineral water in the previous time window.

[0099] Among them, the non-conforming product compensation coefficient is used to control the impact of non-conforming products in the previous cycle on the production plan of the next time window; the return order adjustment coefficient is used to control the impact of the return quantity in the previous cycle on the production plan of the next time window; the feedback coefficient is used to control the impact of the orders and returns in the previous time window on the production plan of the next time window;

[0100] Part of the data of Company A in this embodiment is shown in Table 1;

[0101] Table 1A Partial data table of enterprises

[0102] parameter Data Value <![CDATA[OR next ]]> 10000 <![CDATA[BQ pre ]]> 8500 <![CDATA[ROR pre ]]> 200 <![CDATA[OR pre ]]> 9500 <![CDATA[AC pre ]]> 0.85

[0103] This embodiment calculates the qualified rate of bottled mineral water in the previous time window, and determines the production volume in the next time window in combination with the order volume, order return volume and order volume in the next cycle in the previous time window, so that the enterprise can flexibly adjust the production plan, optimize resource allocation, reduce production costs, and thus improve the production management efficiency of the enterprise.

[0104] This embodiment preprocesses and integrates multi-source data to construct comprehensive detection data, providing accurate and rich input data for the mineral water quality detection model, thereby achieving a comprehensive and detailed evaluation of the mineral water quality. The feature extraction layer extracts the image features, weight features and water quality features of bottled mineral water, and the feature analysis layer calculates the first quality score value, the second quality score value and the third quality score value, and comprehensively calculates the comprehensive quality score value, which is compared with the preset threshold to determine the quality level of the mineral water, providing data support for the enterprise's production plan adjustment, thereby improving the enterprise's product quality and production management efficiency.

[0105] Embodiment 2

[0106] Company B is an industrial manufacturing company responsible for bottling mineral water. In order to optimize production management efficiency, Company B has applied an industrial manufacturing enterprise data resource evaluation and management system.

[0107] An industrial manufacturing enterprise data resource evaluation management system, the system module diagram is as follows Figure 2 As shown; the module includes:

[0108] The data acquisition module is used to acquire the first data, the second data and the third data of a time window on the production line of Company B; the first data includes the bottle body image, the label image and the bottle cap image of the bottled mineral water taken by the camera; the second data includes the bottle body weight, the water weight and the total weight of the bottled mineral water obtained by the weighing sensor; the third data includes the pH value, the mineral content and the conductivity of the bottled mineral water obtained by the water quality monitoring equipment; by preprocessing these three kinds of data, the first detection data, the second detection data and the third detection data are obtained, and the three kinds of detection data are integrated; the comprehensive detection data is obtained and input into the mineral water quality detection model; the schematic diagram of the mineral water quality detection model is shown as follows Figure 3 As shown;

[0109] In this embodiment, the previous time window is a time window of one month;

[0110] The preprocessing process is to perform cropping, denoising, grayscale and normalization on the first data; and to perform missing value processing, standardization and normalization on the second data and the third data.

[0111] Furthermore, the comprehensive detection data is:

[0112] AF i =[I i ,WE i ,WQ i ];

[0113] I i =[I bottle,i ,I label,i ,I cap,i ];

[0114] WE i =[WE b,i ,WE w,i ,WE wb,i ];

[0115] WQ i =[WQ p,i ,WQ m,i ,WQ d,i ];

[0116] WQ m,i =[Q ca,i ,Q mg,i ,Q k,i ,Q na,i ,Q hco3,i ,Q so4,i ,Q cl,i ];

[0117] Among them, AF i represents the comprehensive test data of the i-th bottled mineral water; I i represents the first detection data of the i-th bottled mineral water; WE i represents the second detection data of the i-th bottled mineral water; WQ i represents the third detection data of the i-th bottled mineral water; bottle,i I represents the bottle image data of the i-th bottle of mineral water; label,i Represents the label image data of the i-th bottled mineral water; cap,i represents the bottle cap image data of the i-th bottled mineral water; WE b,i represents the weight of the bottle of the i-th bottle of mineral water; WE w,i represents the water weight of the i-th bottle of mineral water; WE wb,i represents the total weight of the i-th bottled mineral water; WQ p,i represents the pH value of the i-th bottled mineral water; WQ m,i represents the mineral content data of the i-th bottled mineral water; WQ d,i represents the conductivity data of the i-th bottled mineral water; Q ca,i , Q mg,i , Q k,i , Q na,i , Q hco3,i , Q so4,i , Q cl,i Represents the content of calcium, magnesium, potassium, sodium, bicarbonate, sulfate and chloride in the i-th bottled mineral water.

[0118] A quality assessment module, used to assess the quality of bottled water through the feature extraction layer, feature analysis layer and output layer of the model; the feature extraction layer extracts features from the comprehensive detection data to obtain image features, weight features and water quality features; the feature analysis layer analyzes the three features in combination with the attention mechanism to obtain a first quality score value, a second quality score value and a third quality score value, and obtains a comprehensive quality score value based on the three quality score values;

[0119] The output layer is used to determine the quality level of the bottled mineral water according to the comparison between the comprehensive quality score value and a preset threshold value; the quality level includes qualified and unqualified;

[0120] Furthermore, the first quality score value is:

[0121]

[0122] Among them, R 1,i represents the first score value of the i-th bottled mineral water; LeakyReLU represents the activation function; n 1 Indicates the number of image features; m j represents the dimension of the jth image feature; f image,jh represents the hth feature value of the jth image feature; Attention(f image,jh ) represents the attention mechanism weight of the hth feature value of the jth image feature of the i-th bottle of mineral water;

[0123] The image features include: the length of the crack on the bottle surface, the width of the crack and the direction of the crack, the relative position and rotation angle between the edge of the bottle cap and the bottle body, the position offset of the label, the printing clarity and color difference;

[0124] The second quality score is:

[0125]

[0126] Among them, R 2,i represents the second score of the i-th bottled mineral water; n 2 represents the number of weight features; f weight,k represents the k-th eigenvalue; LeakyReLU represents the activation function; Attention(f weight,k ) represents the attention mechanism weight of the k-th weight feature value;

[0127] The weight characteristics include weight distribution deviation and filling uniformity; the weight distribution deviation indicates the deviation between the total weight of each bottle of mineral water and the target weight; the filling uniformity indicates the ratio between the water weight and the bottle weight during the filling process;

[0128] The third quality score value is:

[0129]

[0130] Among them, R 2,i represents the third score value of the i-th bottled mineral water; n 3 Indicates the quantity of water quality characteristics; f water,m represents the mth water quality feature value; LeakyReLU represents the activation function; Attention(fweight,m ) represents the attention mechanism weight of the m-th water quality feature value.

[0131] The water quality characteristics include pH deviation, mineral ratio and the correlation between conductivity and mineral content; wherein pH deviation indicates the deviation between the actual measured pH value and the standard pH value; mineral ratio includes the ratio of calcium, magnesium, potassium, sodium, bicarbonate, sulfate and chloride content in bottled mineral water to the total mineral content; the correlation between conductivity and mineral content is used to analyze the correlation between conductivity and mineral content to evaluate the relationship between the total amount of dissolved ions in water and the mineral concentration. The correlation is calculated by the Pearson correlation coefficient;

[0132] Furthermore, the comprehensive quality score is:

[0133] AR i =λ 1 *R 1,i +λ 2 *R 2,i +λ 3 *R 3,i +γ*ln(1+R 1,i *R 2,i *R 3,i );

[0134] Among them, AR i represents the comprehensive quality score of the i-th bottled mineral water; 1 represents the weight of the first quality score; 2 represents the second quality score weight; 3 represents the weight of the third quality score; ln is the logarithm with base e.

[0135] The output determination module is used to determine the production volume of bottled mineral water in the next time window based on the order volume, return volume, qualified rate of bottled mineral water of the enterprise in the previous time window and the order volume in the next time window.

[0136] Furthermore, the production volume of bottled mineral water in the next time window is:

[0137]

[0138] Among them, MF next Indicates the production volume of bottled mineral water in the next time window; OR next Indicates the order volume of the next time window; BQ pre represents the qualified number of bottled mineral water in the previous time window; ξ represents the compensation coefficient for non-qualified products; ROR prerepresents the number of returned orders for bottled mineral water in the previous time window; ψ represents the adjustment coefficient for returned orders; θ represents the feedback coefficient; OR pre represents the order volume of bottled mineral water in the previous time window; AC pre Indicates the pass rate of bottled mineral water in the previous time window.

[0139] Among them, the non-conforming product compensation coefficient is used to control the impact of non-conforming products in the previous cycle on the production plan of the next time window; the return order adjustment coefficient is used to control the impact of the return quantity in the previous cycle on the production plan of the next time window; the feedback coefficient is used to control the impact of the orders and returns in the previous time window on the production plan of the next time window;

[0140] Part of the data of Company B in this example is shown in Table 2;

[0141] Table 2B Partial data of enterprises

[0142] parameter Data Value <![CDATA[OR next ]]> 20000 <![CDATA[BQ pre ]]> 18000 <![CDATA[ROR pre ]]> 500 <![CDATA[OR pre ]]> 19500 <![CDATA[AC pre ]]> 0.9

[0143] Although 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 appended claims and their equivalents.

Claims

1. A data resource evaluation and management method for an industrial manufacturing enterprise, characterized in that: The first data, the second data and the third data of a time window on the production line of a bottled mineral water enterprise are obtained; the first data includes the bottle body image, the label image and the bottle cap image of the bottled mineral water taken by the camera; the second data includes the bottle body weight, the water weight and the total weight of the bottled mineral water obtained by the weighing sensor; the third data includes the pH value, the mineral content and the conductivity of the bottled mineral water obtained by the water quality monitoring equipment; the first detection data, the second detection data and the third detection data are obtained by preprocessing the three kinds of data, and the three kinds of detection data are integrated; the comprehensive detection data is obtained and input into the mineral water quality detection model; The model passes through the feature extraction layer, feature analysis layer and output layer in sequence; the feature extraction layer extracts features from the comprehensive detection data to obtain image features, weight features and water quality features; the feature analysis layer analyzes the three features in combination with the attention mechanism to obtain a first quality score value, a second quality score value and a third quality score value, and analyzes the three quality score values ​​to obtain a comprehensive quality score value; the output layer is used to compare the comprehensive quality score value with a preset threshold value to determine the quality level of the bottled mineral water; the quality level includes qualified and unqualified; The production volume of bottled mineral water in the next time window is determined based on the company's order volume, order return volume, and pass rate for bottled mineral water in the previous time window and the order volume in the next time window.

2. The method for evaluating and managing data resources of an industrial manufacturing enterprise according to claim 1, characterized in that: The preprocessing process is to perform cropping, denoising, grayscale and normalization on the first data; and to perform missing value processing, standardization and normalization on the second data and the third data.

3. The method for evaluating and managing data resources of an industrial manufacturing enterprise according to claim 1, characterized in that: The comprehensive test data are: AF i =[I i ,WE i ,WQ i ]; I i =[I bottle,i ,I label,i ,I cap,i ]; WE i =[WE b,i ,WE w,i ,WE wb,i ]; WQ i =[WQ p,i ,WQ m,i ,WQ d,i ]; WQ m,i =[Q ca,i ,Q mg,i ,Q k,i ,Q na,i ,Q hco3,i ,Q so4,i ,Q cl,i ]; Among them, AF i represents the comprehensive test data of the i-th bottled mineral water; I i represents the first detection data of the i-th bottled mineral water; WE i represents the second detection data of the i-th bottled mineral water; WQ i represents the third detection data of the i-th bottled mineral water; bottle,i I represents the bottle image data of the i-th bottle of mineral water; label,i I represents the label image data of the i-th bottled mineral water; cap,i represents the bottle cap image data of the i-th bottled mineral water; WE b,i represents the weight of the bottle of the i-th bottle of mineral water; WE w,i represents the water weight of the i-th bottle of mineral water; WE wb,i represents the total weight of the i-th bottled mineral water; WQ p,i represents the pH value of the i-th bottled mineral water; WQ m,i represents the mineral content data of the i-th bottled mineral water; WQ d,i represents the conductivity data of the i-th bottled mineral water; Q ca,i , Q mg,i , Q k,i , Q na,i , Q hco3,i , Q so4,i , Q cl,i Represents the content of calcium, magnesium, potassium, sodium, bicarbonate, sulfate and chloride in the i-th bottled mineral water.

4. The method for evaluating and managing data resources of an industrial manufacturing enterprise according to claim 1, characterized in that: The first quality score value is: Among them, R 1,i represents the first score of the i-th bottle of mineral water; LeakyReLU represents the activation function; n1 represents the number of image features; m j represents the dimension of the jth image feature; f image,jh represents the hth feature value of the jth image feature; Attention(f image,jh ) represents the attention mechanism weight of the hth feature value of the jth image feature of the i-th bottle of mineral water; The second quality score is: Among them, R 2,i represents the second score value of the i-th bottled mineral water; n2 represents the number of weight features; f weight,k represents the k-th eigenvalue; LeakyReLU represents the activation function; Attention(f weight,k ) represents the attention mechanism weight of the k-th weight feature value; The third quality score value is: Among them, R 2,i represents the third score value of the i-th bottled mineral water; n3 represents the number of water quality characteristics; f water,m represents the mth water quality feature value; LeakyReLU represents the activation function; Attention(f weight,m ) represents the attention mechanism weight of the m-th water quality feature value.

5. The method for data resource evaluation and management of industrial manufacturing enterprises according to claim 1 is characterized in that: The comprehensive quality rating is: AR i =λ1*R 1,i +λ2*R 2,i +λ3*R 3,i +γ*ln(1+R 1,i *R 2,i *R 3,i ); Among them, AR i represents the comprehensive quality score of the i-th bottled mineral water; λ1 represents the weight of the first quality score; λ2 represents the weight of the second quality score; λ3 represents the weight of the third quality score; ln is the logarithm with base e.

6. The method for data resource evaluation and management of industrial manufacturing enterprises according to claim 1, characterized in that: The production volume of bottled mineral water in the next time window is: Among them, MF next Indicates the production volume of bottled mineral water in the next time window; OR next Indicates the order volume of the next time window; BQ pre represents the qualified number of bottled mineral water in the previous time window; ξ represents the compensation coefficient for non-qualified products; ROR pre represents the number of returned orders for bottled mineral water in the previous time window; ψ represents the adjustment coefficient for returned orders; θ represents the feedback coefficient; OR pre represents the order volume of bottled mineral water in the previous time window; AC pre Indicates the pass rate of bottled mineral water in the previous time window.

7. An industrial manufacturing enterprise data resource evaluation and management system, characterized in that: The system comprises: A data acquisition module is used to acquire first data, second data and third data of a time window on a production line of a bottled mineral water enterprise; the first data includes a bottle body image, a label image and a bottle cap image of the bottled mineral water taken by a camera; the second data includes the bottle body weight, water weight and total weight of the bottled mineral water obtained by a weighing sensor; the third data includes the pH value, mineral content and conductivity of the bottled mineral water obtained by a water quality monitoring device; by preprocessing the three types of data, first detection data, second detection data and third detection data are obtained, and the three types of detection data are integrated; comprehensive detection data is obtained and input into a mineral water quality detection model; A quality assessment module, used to assess the quality of bottled water through the feature extraction layer, feature analysis layer and output layer of the model; the feature extraction layer extracts features from the comprehensive detection data to obtain image features, weight features and water quality features; the feature analysis layer analyzes the three features in combination with the attention mechanism to obtain a first quality score value, a second quality score value and a third quality score value, and obtains a comprehensive quality score value based on the three quality score values; The output layer is used to determine the quality level of the bottled mineral water according to the comparison between the comprehensive quality score value and a preset threshold value; the quality level includes qualified and unqualified; The output determination module is used to determine the production volume of bottled mineral water in the next time window based on the order volume, return volume, qualified rate of bottled mineral water of the enterprise in the previous time window and the order volume in the next time window.

8. The industrial manufacturing enterprise data resource evaluation and management system according to claim 7, characterized in that: The first quality score value is: Among them, R 1,i represents the first score of the i-th bottle of mineral water; LeakyReLU represents the activation function; n1 represents the number of image features; m j represents the dimension of the jth image feature; f image,jh represents the hth feature value of the jth image feature; Attention(f image,jh ) represents the attention mechanism weight of the hth feature value of the jth image feature of the i-th bottle of mineral water; The second quality score is: Among them, R 2,i represents the second score value of the i-th bottled mineral water; n2 represents the number of weight features; f weight,k represents the k-th eigenvalue; LeakyReLU represents the activation function; Attention(f weight,k ) represents the attention mechanism weight of the k-th weight feature value; The third quality score value is: Among them, R 2,i represents the third score value of the i-th bottled mineral water; n3 represents the number of water quality characteristics; f water,m represents the mth water quality feature value; LeakyReLU represents the activation function; Attention(f weight,m ) represents the attention mechanism weight of the m-th water quality feature value.

9. The industrial manufacturing enterprise data resource evaluation and management system according to claim 7, characterized in that: The comprehensive quality rating is: AR i =λ1*R 1,i +λ2*R 2,i +λ3*R 3,i +γ*ln(1+R 1,i *R 2,i *R 3,i ); Among them, AR i represents the comprehensive quality score of the i-th bottled mineral water; λ1 represents the weight of the first quality score; λ2 represents the weight of the second quality score; λ3 represents the weight of the third quality score; ln is the logarithm with base e.

10. The industrial manufacturing enterprise data resource evaluation and management system according to claim 7, characterized in that: The production volume of bottled mineral water in the next time window is: Among them, MF next Indicates the production volume of bottled mineral water in the next time window; OR next Indicates the order volume of the next time window; BQ pre represents the qualified number of bottled mineral water in the previous time window; ξ represents the compensation coefficient for non-qualified products; ROR pre represents the number of returned orders for bottled mineral water in the previous time window; ψ represents the adjustment coefficient for returned orders; θ represents the feedback coefficient; OR pre represents the order volume of bottled mineral water in the previous time window; AC pre Indicates the pass rate of bottled mineral water in the previous time window.

Images