Defect detection and avoidance method based on gypsum board production full data collection

Abstract

Disclosed in the present invention is a defect detection and avoidance method based on gypsum board production full data collection, comprising the following steps: performing timing analysis on adjustment amounts of gypsum board production device adjustment items to extract companion relationships between the gypsum board production device adjustment items, and abstracting the companion relationships into a companion network among the gypsum board production device adjustment items; and on the basis of the number of real-time gypsum board production device adjustment items, determining real-time settings of production conditions of the gypsum board production device. According to the present invention, learning and training is performed on characterization data of gypsum board defect items and adjustment amounts of gypsum board production device adjustment items to obtain a defect adjustment calculation model, timing analysis is performed on the adjustment amounts of the gypsum board production device adjustment items to extract companion relationships between the gypsum board production device adjustment items, and the companion relationships are abstracted into a companion network among the gypsum board production device adjustment items, thereby achieving full detection and full avoidance of potential defects of gypsum boards, and improving the proactiveness of defect detection and avoidance.

Description

A defect detection and avoidance method based on full data collection of gypsum board production Technical Field

[0001] The present invention relates to the technical field of defect detection, and in particular to a defect detection avoidance method based on full data collection of gypsum board production. Background Art

[0002] During the production, molding, cutting, and internal transportation of gypsum boards in the dryer, various problems such as board breakage and broken board corners may occur. The appearance of problematic boards is mainly determined by visual observation by on-the-job personnel. Sometimes, they are busy with other work at their posts and fail to discover the external defects of individual products in time, causing the boards to be stacked, which is not conducive to product quality control. In addition, the entire rack of stacked boards needs to be unloaded by forklift for quality selection and re-stacking, which is a large workload.

[0003] Therefore, it is necessary to adopt an intelligent automatic defect monitoring and defect avoidance method to monitor and avoid defects in the surface shape, temperature and humidity of the gypsum board. The existing intelligent automatic defect monitoring method generally performs defect detection and defect avoidance on one surface of the gypsum board first, then flips the gypsum board over, and then performs defect monitoring and defect avoidance on the other surface of the gypsum board. This detection method still uses passive defect detection and defect avoidance, and requires the gypsum board defects to be detected and avoided one by one, which makes it difficult to predictively monitor and avoid potential defects.

[0004] Summary of the Invention

[0005] The purpose of the present invention is to provide a defect detection and avoidance method based on full data collection of gypsum board production, so as to solve the technical problem that the existing technology adopts passive defect detection and defect avoidance, which requires gypsum board defects to be detected and avoided one by one, and it is difficult to predictively monitor and avoid potential defects.

[0006] In order to solve the above technical problems, the present invention specifically provides the following technical solutions:

[0007] A defect detection and avoidance method based on full data collection of gypsum board production includes the following steps:

[0008] Step S1: Targetedly matching gypsum board defect items with gypsum board production equipment adjustment items;

[0009] Step S2: Based on the targeted matching relationship between the gypsum board defect items and the adjustment items of the gypsum board production equipment in step S1, the characterization data of the gypsum board defect items and the adjustment amounts of the gypsum board production equipment adjustment items in the gypsum board production full data collection log are trained to obtain a defect adjustment calculation model;

[0010] Step S3: In the gypsum board production full data collection log, perform a time series analysis of the adjustment amounts of the gypsum board production equipment adjustment items to extract the associated relationships between the gypsum board production equipment adjustment items, and abstract the associated relationships into an associated network between the gypsum board production equipment adjustment items;

[0011] Step S4: extracting characterization data of real-time defect items of the gypsum board from the full real-time gypsum board production data obtained through real-time monitoring, and inputting the characterization data of the real-time defect items of the gypsum board into a defect adjustment calculation model to obtain an adjustment amount of the real-time gypsum board production equipment adjustment item;

[0012] Step S5: Count the number of real-time gypsum board production equipment adjustment items, and determine the real-time setting of the production working conditions of the gypsum board production equipment based on the number of real-time gypsum board production equipment adjustment items, so as to achieve full detection and full avoidance of potential defects of the gypsum board.

[0013] As a preferred solution of the present invention, the targeted matching of gypsum board defect items with gypsum board production equipment adjustment items includes:

[0014] The gypsum board defect items include: gypsum board width, gypsum board standing edge angle, gypsum board thickness, gypsum board wet board weight, gypsum board dry board weight, gypsum board temperature after drying, and gypsum board humidity after drying;

[0015] The adjustment items of the gypsum board production equipment include: gypsum board forming knife distance, gypsum board forming knife angle, gypsum board forming board height, gypsum board dry material feeding amount, gypsum board wet material feeding amount, gypsum board dryer gate temperature;

[0016] The adjustment of the distance between the gypsum board forming knives and the angle of the gypsum board forming knives can specifically avoid the defects of the gypsum board width and the gypsum board vertical edge angle;

[0017] The height adjustment of the gypsum board forming board can specifically avoid the defects of the gypsum board thickness;

[0018] The adjustment of the amount of dry gypsum board material and the amount of wet gypsum board material is targeted to avoid the defects of the single weight of the wet gypsum board and the single weight of the dry gypsum board;

[0019] The adjustment of the gate temperature of the gypsum board dryer specifically avoids the defects of the temperature and humidity of the gypsum board after drying.

[0020] As a preferred solution of the present invention, the defect adjustment calculation model is obtained by learning and training the characterization data of the gypsum board defect items and the adjustment amount of the gypsum board production equipment adjustment items, including:

[0021] Extract the characterization data of gypsum board defects and the adjustment amount of gypsum board production equipment adjustment items from the gypsum board production full data collection log;

[0022] Using the characterization data of the gypsum board defect item as the input item of the BP neural network, and using the adjustment amount of the gypsum board production equipment as the output item of the BP neural network;

[0023] The BP neural network is used to perform convolution learning on the input and output items of the BP neural network to obtain a defect adjustment measurement model;

[0024] The model expression of the defect adjustment calculation model is:

[0025] Wr=BP(Ds);

[0026] Where Wr is the adjustment value of the gypsum board production equipment adjustment item r, Ds is the characterization data of the gypsum board defect item s, BP is the BP neural network, r is the identifier of the gypsum board production equipment adjustment item, and s is the identifier of the gypsum board defect item;

[0027] The complete gypsum board production data consists of a combination of characterization data of gypsum board width, gypsum board standing edge angle, gypsum board thickness, gypsum board wet board weight, gypsum board dry board weight, gypsum board drying temperature, and gypsum board drying humidity;

[0028] The gypsum board production full data collection log is a log consisting of all gypsum board production data collected in time series;

[0029] Extracting the characterization data of the gypsum board defect item includes:

[0030] In the full data collection log for gypsum board production, the characterization data for gypsum board width, gypsum board edge angle, gypsum board thickness, gypsum board wet board weight, gypsum board dry board weight, gypsum board temperature after drying, and gypsum board humidity after drying at each time series are compared with the standard data for gypsum board width, gypsum board edge angle, gypsum board thickness, gypsum board wet board weight, gypsum board dry board weight, gypsum board temperature after drying, and gypsum board humidity after drying. Items whose similarity between the characterization data at each time series and the standard data is less than the allowable threshold are identified as gypsum board defects at that time series.

[0031] Extract the characterization data of gypsum board defects at each time series from the full data collection log of gypsum board production;

[0032] Extracting the adjustment amount of the gypsum board production equipment adjustment item includes:

[0033] In the gypsum board production full data collection log, the gypsum board defect items at each time sequence are determined using the targeted matching relationship to determine the gypsum board production equipment adjustment items that are targeted and matched with the gypsum board defect items at each time sequence;

[0034] In the full data collection log of gypsum board production, the adjustment amounts of the gypsum board production equipment adjustment items that specifically match the gypsum board defect items at each time sequence are extracted.

[0035] As a preferred solution of the present invention, the time series analysis of the adjustment amounts of the gypsum board production equipment adjustment items to extract the associated relationships between the gypsum board production equipment adjustment items includes:

[0036] Sorting the adjustment amounts of the various gypsum board production equipment adjustment items according to the time sequence to obtain a time sequence of the adjustment amounts of the gypsum board production equipment adjustment items;

[0037] Performing sequence similarity analysis on the time series of the adjustment amounts of each gypsum board production equipment adjustment item to quantify the associated rate that characterizes the associated relationship between each gypsum board production equipment adjustment item. The associated relationship characterizes the degree of mutual influence between the gypsum board production equipment adjustment items.

[0038] The quantitative formula of the associated rate is: ij =|W i,T -W j,T |;

[0039] Where H ij is the concomitant rate between gypsum board production equipment adjustment item i and gypsum board production equipment adjustment item j, W i,T is the time series of adjustment amount of gypsum board production equipment adjustment item i, W j,T is the time series of adjustment value of gypsum board production equipment adjustment item j, |W i,T -W j,T | for W i,T and W j,T where i and j are the identifiers of the gypsum board production equipment adjustment items, and T is the time series identifier.

[0040] As a preferred solution of the present invention, the associated relationship is abstracted into an associated network between gypsum board production equipment adjustment items, including:

[0041] The gypsum board production equipment adjustment items are abstracted into topological nodes, and topological edges are set between the topological nodes for topological connection;

[0042] The associated rate between the gypsum board production equipment adjustment items is used as the edge weight of the topological edge between the corresponding topological nodes to abstract the associated network.

[0043] As a preferred solution of the present invention, the method of determining the real-time setting of the production working conditions of the gypsum board production equipment according to the number of real-time adjustment items of the gypsum board production equipment includes:

[0044] If the number of real-time gypsum board production equipment adjustment items is 0, the current production condition setting of the gypsum board production equipment is maintained;

[0045] If the number of real-time gypsum board production equipment adjustment items is 1, the adjustment amount of the real-time gypsum board production equipment adjustment item is mapped to the associated network to obtain a single real-time associated network, and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment is obtained in the single real-time associated network. The production working conditions of the gypsum board production equipment are set in real time according to the adjustment amount of the real-time gypsum board production equipment adjustment item and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment, so as to achieve full detection and full avoidance of potential defects of the gypsum board;

[0046] If the number of gypsum board production equipment adjustment items is greater than 1, the adjustment amount of each gypsum board real-time production equipment adjustment item is mapped to the associated network in turn to obtain multiple real-time associated networks, and the multiple real-time associated networks are fused to obtain the adjustment amount of the gypsum board production equipment real-time associated adjustment item. The production working conditions of the gypsum board production equipment are set in real time according to the adjustment amount of the gypsum board real-time production equipment adjustment item and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment, so as to achieve full detection and full avoidance of potential defects of the gypsum board.

[0047] As a preferred solution of the present invention, mapping the adjustment amount of the real-time gypsum board production equipment adjustment item to the associated network to obtain a single real-time associated network includes:

[0048] The adjustment amount of the real-time gypsum board production equipment adjustment item is mapped to the node value of the corresponding topological node in the companion network, and the node values ​​of the remaining topological nodes in the companion network are calculated one by one according to the edge weights of the topological nodes and topological edges to obtain a single real-time companion network.

[0049] As a preferred solution of the present invention, obtaining the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment in a single real-time associated network includes:

[0050] The topological nodes in the single real-time associated network except the topological nodes corresponding to the gypsum board real-time production equipment adjustment items are used as the gypsum board production equipment real-time associated adjustment items;

[0051] The node value of the topological node corresponding to the real-time associated adjustment item of the gypsum board production equipment in the single real-time associated network is used as the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment.

[0052] As a preferred solution of the present invention, the adjustment amount of each gypsum board real-time production equipment adjustment item is sequentially mapped to the associated network to obtain multiple real-time associated networks, including:

[0053] The adjustment amount of each gypsum board real-time production equipment adjustment item is sequentially mapped to the node value of the corresponding topological node of each associated network, and the node values ​​of the remaining topological nodes in each associated network are calculated one by one based on the edge weights of the topological nodes and topological edges to obtain multiple real-time associated networks;

[0054] The node values ​​of multiple real-time companion networks are superimposed and fused to obtain a single fused real-time companion network.

[0055] As a preferred solution of the present invention, the method of fusing multiple real-time associated networks to obtain the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment includes:

[0056] The topology nodes in the single fusion real-time associated network except for all the topology nodes corresponding to the gypsum board real-time production equipment adjustment items are used as the gypsum board production equipment real-time associated adjustment items;

[0057] The node value of the topological node corresponding to the real-time associated adjustment item of the gypsum board production equipment in the single fused real-time associated network is used as the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] The present invention conducts learning and training on the characterization data of gypsum board defect items and the adjustment amounts of gypsum board production equipment adjustment items in the full data collection log to obtain a defect adjustment calculation model, and performs time series analysis on the adjustment amounts of gypsum board production equipment adjustment items to extract the associated relationships between the gypsum board production equipment adjustment items, and abstracts the associated relationships into an associated network between the gypsum board production equipment adjustment items, thereby achieving full detection and full avoidance of potential defects in gypsum boards and improving the initiative of defect detection and avoidance. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] To more clearly illustrate the embodiments of the present invention or technical solutions in the prior art, the following briefly describes the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are merely illustrative, and those skilled in the art can, without inventive effort, derive other implementation drawings from the provided drawings.

[0061] FIG1 is a flow chart of a defect detection avoidance method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0063] Existing intelligent automatic defect monitoring methods generally first perform defect detection and defect avoidance on one surface of the gypsum board, then flip the gypsum board over, and then perform defect detection and defect avoidance on the other surface of the gypsum board. This detection method still uses passive defect detection and defect avoidance, requiring each gypsum board defect to be detected and avoided one by one, making it difficult to predictively monitor and avoid potential defects. Therefore, the present invention provides a defect detection and avoidance method based on full data collection of gypsum board production. The defect adjustment calculation model is used to determine the adjustment items of gypsum board production equipment that avoid gypsum board defects based on full data collection of gypsum board production. The adjustment items of all gypsum board production equipment that avoid potential gypsum board defects are obtained based on the constructed associated network, thereby achieving full detection and avoidance of potential gypsum board defects.

[0064] As shown in FIG1 , the present invention provides a defect detection and avoidance method based on full data collection of gypsum board production, comprising the following steps:

[0065] Step S1: Targetedly matching gypsum board defect items with gypsum board production equipment adjustment items;

[0066] Step S2: Based on the targeted matching relationship between the gypsum board defect items and the adjustment items of the gypsum board production equipment in step S1, the characterization data of the gypsum board defect items and the adjustment amounts of the gypsum board production equipment adjustment items in the gypsum board production full data collection log are trained to obtain a defect adjustment calculation model;

[0067] Step S3: In the gypsum board production full data collection log, perform a time series analysis of the adjustment amounts of the gypsum board production equipment adjustment items to extract the associated relationships between the gypsum board production equipment adjustment items, and abstract the associated relationships into an associated network between the gypsum board production equipment adjustment items;

[0068] Step S4: extracting characterization data of real-time defect items of the gypsum board from the full real-time gypsum board production data obtained through real-time monitoring, and inputting the characterization data of the real-time defect items of the gypsum board into a defect adjustment calculation model to obtain an adjustment amount of the real-time gypsum board production equipment adjustment item;

[0069] Step S5: Count the number of real-time gypsum board production equipment adjustment items, and determine the real-time setting of the production working conditions of the gypsum board production equipment based on the number of real-time gypsum board production equipment adjustment items, so as to achieve full detection and full avoidance of potential defects of the gypsum board.

[0070] The present invention constructs a defect adjustment calculation model, which can calculate the adjustment amount of the gypsum board production equipment adjustment item based on the characterization data of the gypsum board defect item. The adjustment amount of the gypsum board production equipment adjustment item is used to adjust the characterization data of the gypsum board defect item to the standard data. For example, if the characterization data of the gypsum board width and the gypsum board vertical edge angle deviate from the standard data of the gypsum board width and the gypsum board vertical edge angle, the corresponding adjustment amounts of the gypsum board forming knife distance and the gypsum board forming knife angle are such that the gypsum board width and the gypsum board vertical edge angle can be restored from the characterization data to the standard data, thereby avoiding the defects of the gypsum board width and the gypsum board vertical edge angle. The defects of the gypsum board thickness, the unit weight of the wet board of the gypsum board, the unit weight of the dry board of the gypsum board, the temperature of the gypsum board after drying, and the humidity of the gypsum board after drying are the same as the defects of the gypsum board width and the gypsum board vertical edge angle, which will not be elaborated here. Therefore, the defect adjustment calculation model realizes the targeted matching calculation of the adjustment amount of the gypsum board production equipment adjustment item, improves the degree of measurement automation, and thereby improves the efficiency of avoiding gypsum board defects.

[0071] Targeted matching of gypsum board defects with gypsum board production equipment adjustments, including:

[0072] Gypsum board defect items include: gypsum board width, gypsum board standing edge angle, gypsum board thickness, gypsum board wet board weight, gypsum board dry board weight, gypsum board drying temperature, gypsum board drying humidity;

[0073] Adjustment items for gypsum board production equipment include: gypsum board forming knife distance, gypsum board forming knife angle, gypsum board forming board height, gypsum board dry material feeding amount, gypsum board wet material feeding amount, gypsum board dryer gate temperature;

[0074] Among them, the adjustment of the distance and angle of the gypsum board forming knife can specifically avoid the defects of the gypsum board width and the gypsum board vertical edge angle;

[0075] Adjustment of the height of the gypsum board molding board can specifically avoid the defects of the gypsum board thickness;

[0076] Adjustment of dry material feeding amount and wet material feeding amount of gypsum board can avoid the defects of wet board weight and dry board weight of gypsum board;

[0077] The adjustment of the gate temperature of the gypsum board dryer can specifically avoid the defects of temperature and humidity after the gypsum board is dried.

[0078] The defect adjustment calculation model is obtained by learning and training the characterization data of gypsum board defects and the adjustment amount of gypsum board production equipment adjustment items, including:

[0079] Extract the characterization data of gypsum board defects and the adjustment amount of gypsum board production equipment adjustment items from the gypsum board production full data collection log;

[0080] The characterization data of the gypsum board defect items are used as the input items of the BP neural network, and the adjustment amount of the gypsum board production equipment is used as the output item of the BP neural network;

[0081] The BP neural network is used to perform convolution learning on the input and output items of the BP neural network to obtain a defect adjustment measurement model;

[0082] The model expression of the defect adjustment calculation model is: Wr=BP(Ds);

[0083] Where Wr is the adjustment value of the gypsum board production equipment adjustment item r, Ds is the characterization data of the gypsum board defect item s, BP is the BP neural network, r is the identifier of the gypsum board production equipment adjustment item, and s is the identifier of the gypsum board defect item;

[0084] The complete gypsum board production data consists of a combination of characterization data of gypsum board width, gypsum board standing edge angle, gypsum board thickness, gypsum board wet board weight, gypsum board dry board weight, gypsum board drying temperature, and gypsum board drying humidity;

[0085] The gypsum board production full data collection log is a log consisting of all gypsum board production data collected in time series;

[0086] Extraction of characterization data for gypsum board defects, including:

[0087] In the full data collection log for gypsum board production, the characterization data for gypsum board width, gypsum board edge angle, gypsum board thickness, gypsum board wet board weight, gypsum board dry board weight, gypsum board temperature after drying, and gypsum board humidity after drying at each time series are compared with the standard data for gypsum board width, gypsum board edge angle, gypsum board thickness, gypsum board wet board weight, gypsum board dry board weight, gypsum board temperature after drying, and gypsum board humidity after drying. Items whose similarity between the characterization data at each time series and the standard data is less than the allowable threshold are identified as gypsum board defects at that time series.

[0088] Extract the characterization data of gypsum board defects at each time series from the full data collection log of gypsum board production;

[0089] The adjustment amount of the gypsum board production equipment adjustment item is extracted, including:

[0090] In the gypsum board production full data collection log, the gypsum board defect items at each time sequence are used to determine the gypsum board production equipment adjustment items that specifically match the gypsum board defect items at each time sequence;

[0091] In the full data collection log of gypsum board production, the adjustment amounts of the gypsum board production equipment adjustment items that specifically match the gypsum board defect items at each time sequence are extracted.

[0092] Characterization data of real-time defect items of gypsum boards are extracted from the real-time gypsum board production full data obtained by real-time monitoring, and the real-time defect item characterization data of gypsum boards are input into the defect adjustment calculation model to obtain the adjustment amount of the real-time production equipment adjustment item of gypsum boards.

[0093] Extraction of characterization data of real-time defects of gypsum boards, including:

[0094] In the full gypsum board production data collected in real time, the real-time characterization data of gypsum board width, gypsum board edge angle, gypsum board thickness, gypsum board wet board unit weight, gypsum board dry board unit weight, gypsum board temperature after drying, and gypsum board humidity after drying are compared with the standard data of gypsum board width, gypsum board edge angle, gypsum board thickness, gypsum board wet board unit weight, gypsum board dry board unit weight, gypsum board temperature after drying, and gypsum board humidity after drying for corresponding item similarity. Items whose similarity between the real-time characterization data and the standard data is less than the allowable threshold are identified as real-time defects of the gypsum board.

[0095] Extract the representation data of real-time defects of gypsum boards from the full gypsum board production data collected in real time;

[0096] Extraction of adjustment amounts for real-time gypsum board production equipment adjustment items, including:

[0097] In the real-time collected gypsum board production data, the real-time gypsum board defect items are matched with the gypsum board production equipment adjustment items according to the targeted matching relationship.

[0098] From the full gypsum board production data collected in real time, the adjustment amount of the gypsum board real-time production equipment adjustment item that specifically matches the real-time defect item of the gypsum board is extracted.

[0099] The present invention mines the degree of mutual influence between production equipment adjustment items. Since production equipment adjustment items and gypsum board defect items have a targeted matching relationship, mining the degree of mutual influence between production equipment adjustment items is also mining the degree of mutual influence between gypsum board defect items. Reflected in the production process, mining the influence relationship between gypsum board defects can, when a certain gypsum board defect is detected, perceive in advance other potential gypsum board defects that will be caused by the gypsum board defect. That is, a certain gypsum board defect A has a mutual influence relationship with another gypsum board defect B. Then, after the occurrence of gypsum board defect A is detected, it can be directly grasped that gypsum board defect B will occur. Therefore, the present invention mines the degree of mutual influence between production equipment adjustment items to construct an associated network, which can be used for active and predictive perception of potential gypsum board defects and active and predictive avoidance of potential gypsum board defects, as follows:

[0100] Perform time series analysis on the adjustment items of gypsum board production equipment to extract the associated relationships between the adjustment items of gypsum board production equipment, including:

[0101] Sorting the adjustment amounts of the various gypsum board production equipment adjustment items according to the time sequence to obtain a time sequence of the adjustment amounts of the gypsum board production equipment adjustment items;

[0102] The sequence similarity analysis of the time series of the adjustment amount of each gypsum board production equipment adjustment item is performed to quantify the association rate that represents the association relationship between each gypsum board production equipment adjustment item. The association relationship represents the degree of mutual influence between the gypsum board production equipment adjustment items.

[0103] The quantitative formula of the accompanying rate is: H ij =|W i,T -W j,T |;

[0104] Where H ij is the concomitant rate between gypsum board production equipment adjustment item i and gypsum board production equipment adjustment item j, W i,T is the time series of adjustment amount of gypsum board production equipment adjustment item i, W j,T is the time series of adjustment value of gypsum board production equipment adjustment item j, |W i,T -W j,T | for W i,T and W j,T where i and j are the identifiers of the gypsum board production equipment adjustment items, and T is the time series identifier.

[0105] The sequence similarity analysis of the adjustment amount time series of the gypsum board production equipment adjustment items is used to quantify the association rate that characterizes the association relationship between each gypsum board production equipment adjustment item. Among them, the higher the similarity of the adjustment amount time series between the gypsum board production equipment adjustment items, the higher the association rate between the gypsum board production equipment adjustment items. Reflected in the production process, it shows that when the corresponding production equipment is adjusted to avoid gypsum board defect A, the corresponding production equipment is usually adjusted to avoid gypsum board defect B. Moreover, the more consistent the adjustment rhythm of the corresponding production equipment to avoid gypsum board defect A is with the adjustment rhythm of the corresponding production equipment to avoid gypsum board defect B, the closer the relationship between gypsum board defect B and gypsum board defect A is. The more consistent the rhythm of the appearance of gypsum board defects A, the higher the degree of mutual influence between gypsum board defects B and gypsum board defects A, and the highest phenomenon can be formed as a companion phenomenon, that is, when gypsum board defect A appears, gypsum board defect B will definitely appear. Therefore, the present invention uses the adjustment amount time series sequence of the gypsum board production equipment adjustment items to perform sequence similarity analysis to quantify the companion relationship between each gypsum board production equipment adjustment item. This is a full mining of the data information between the production equipment adjustment items, thereby obtaining the companion relationship between each gypsum board production equipment adjustment item, and ultimately realizing active predictive perception of potential gypsum board defects and active predictive avoidance of potential gypsum board defects.

[0106] The present invention constructs an associated network, which can topologically visualize the associated relationship between gypsum board production equipment adjustment items, fully visualize the associated relationship between gypsum board production equipment adjustment items, and can obtain the adjustment amount of the gypsum board production equipment real-time associated adjustment item that avoids the potential gypsum board defect item relative to the detected gypsum board defect item based on the adjustment amount of the gypsum board real-time production equipment adjustment item for avoiding the detected gypsum board defect item, thereby achieving full avoidance of all potential gypsum board defect items and improving the comprehensiveness of defect avoidance. There is a targeted match between the adjustment amount of the gypsum board production equipment real-time associated adjustment item and the potential gypsum board defect item, so the full avoidance of all potential gypsum board defect items is equivalent to the full detection of all potential gypsum board defect items, thereby improving the comprehensiveness of defect detection.

[0107] The associated relationship is abstracted into an associated network between the adjustment items of gypsum board production equipment, including:

[0108] The gypsum board production equipment adjustment items are abstracted into topological nodes, and topological edges are set between the topological nodes for topological connection;

[0109] The association rate between the adjustment items of gypsum board production equipment is used as the edge weight of the topological edge between the corresponding topological nodes, which is abstracted as an associated network.

[0110] The real-time setting of the production working conditions of the gypsum board production equipment is determined based on the number of real-time gypsum board production equipment adjustment items, including:

[0111] If the number of real-time gypsum board production equipment adjustment items is 0, the current production condition setting of the gypsum board production equipment is maintained;

[0112] If the number of real-time gypsum board production equipment adjustment items is 1, the adjustment amount of the real-time gypsum board production equipment adjustment item is mapped to the associated network to obtain a single real-time associated network, and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment is obtained in the single real-time associated network. The production working conditions of the gypsum board production equipment are set in real time based on the adjustment amount of the real-time gypsum board production equipment adjustment item and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment, so as to achieve full detection and full avoidance of potential defects of the gypsum board;

[0113] If the number of gypsum board production equipment adjustment items is greater than 1, the adjustment amount of each gypsum board real-time production equipment adjustment item is mapped to the associated network in turn to obtain multiple real-time associated networks, and the multiple real-time associated networks are fused to obtain the adjustment amount of the gypsum board production equipment real-time associated adjustment item. The production working conditions of the gypsum board production equipment are set in real time according to the adjustment amount of the gypsum board real-time production equipment adjustment item and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment, so as to achieve full detection and full avoidance of potential defects of the gypsum board.

[0114] Map the adjustment values ​​of the real-time gypsum board production equipment adjustment items to the associated network to obtain a single real-time associated network, including:

[0115] The adjustment amount of the real-time gypsum board production equipment adjustment item is mapped to the node value of the corresponding topological node in the companion network, and the node values ​​of the remaining topological nodes in the companion network are calculated one by one according to the edge weights of the topological nodes and topological edges to obtain a single real-time companion network.

[0116] The adjustment amount of the real-time associated adjustment item of the gypsum board production equipment is obtained in a single real-time associated network, including:

[0117] The topological nodes in the single real-time associated network except the topological nodes corresponding to the gypsum board real-time production equipment adjustment items are used as the gypsum board production equipment real-time associated adjustment items;

[0118] The node value of the topological node corresponding to the real-time associated adjustment item of the gypsum board production equipment in the single real-time associated network is used as the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment.

[0119] If multiple gypsum board defects are detected at the same time, the associated network is mapped for each defect item, and the mapped associated networks are fused to generate a simultaneous combined adjustment amount to avoid all potential defects at one time, improve defect avoidance efficiency, and reduce avoidance interference. The details are as follows:

[0120] The adjustment amount of each gypsum board real-time production equipment adjustment item is mapped to the associated network in turn to obtain multiple real-time associated networks, including:

[0121] The adjustment amount of each gypsum board real-time production equipment adjustment item is sequentially mapped to the node value of the corresponding topological node of each associated network, and the node values ​​of the remaining topological nodes in each associated network are calculated one by one based on the edge weights of the topological nodes and topological edges to obtain multiple real-time associated networks;

[0122] The node values ​​of multiple real-time companion networks are superimposed and fused to obtain a single fused real-time companion network.

[0123] Multiple real-time associated networks are integrated to obtain the adjustment values ​​of the real-time associated adjustment items of the gypsum board production equipment, including:

[0124] The topology nodes in the single fusion real-time associated network except for all the topology nodes corresponding to the gypsum board real-time production equipment adjustment items are used as the gypsum board production equipment real-time associated adjustment items;

[0125] The node value of the topological node corresponding to the real-time associated adjustment item of the gypsum board production equipment in the single fused real-time associated network is used as the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment.

[0126] The present invention conducts learning and training on the characterization data of gypsum board defect items and the adjustment amounts of gypsum board production equipment adjustment items in the full data collection log to obtain a defect adjustment calculation model, and performs time series analysis on the adjustment amounts of gypsum board production equipment adjustment items to extract the associated relationships between the gypsum board production equipment adjustment items, and abstracts the associated relationships into an associated network between the gypsum board production equipment adjustment items, thereby achieving full detection and full avoidance of potential defects in gypsum boards and improving the initiative of defect detection and avoidance.

[0127] The above embodiments are merely exemplary embodiments of the present application and are not intended to limit the scope of the present application. The scope of protection of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and scope of protection of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the scope of protection of the present application.

Claims

1. A defect detection avoidance method based on full data collection of gypsum board production, characterized by: The following steps are involved: Step S1, matching gypsum board defect items with gypsum board production equipment adjustment items in a targeted manner; Step S2: Based on the targeted matching relationship between the gypsum board defect item and the adjustment item of the gypsum board production equipment in step S1, in the gypsum board production full data collection log, the characterization data of the gypsum board defect item and the adjustment amount of the gypsum board production equipment adjustment item are trained to obtain a defect adjustment calculation model; Step S3: In the gypsum board production full data collection log, the time series analysis of the adjustment amount of the gypsum board production equipment adjustment items is performed to extract the associated relationship between the gypsum board production equipment adjustment items, and the associated relationship is abstracted into an associated network between the gypsum board production equipment adjustment items; Step S4, extracting the representation data of the real-time defect items of the gypsum board from the real-time gypsum board production full data obtained by real-time monitoring, and inputting the representation data of the real-time defect items of the gypsum board into the defect adjustment calculation model to obtain the adjustment amount of the real-time production equipment adjustment item of the gypsum board; Step S5, counting the number of real-time gypsum board production equipment adjustment items, and determining the real-time setting of the production working conditions of the gypsum board production equipment according to the number of real-time gypsum board production equipment adjustment items, so as to achieve full detection and full avoidance of potential defects of the gypsum board.

2. The defect detection and avoidance method based on full data collection of gypsum board production according to claim 1 is characterized by: The targeted matching of gypsum board defect items with gypsum board production equipment adjustment items includes: The gypsum board defect items include: gypsum board width, gypsum board vertical edge angle, gypsum board thickness, gypsum board wet board unit weight, gypsum board dry board unit weight, gypsum board temperature after drying, gypsum board humidity after drying; The gypsum board production equipment adjustment items include: gypsum board forming knife distance, gypsum board forming knife angle, gypsum board forming board height, gypsum board dry material feeding amount, gypsum board wet material feeding amount, gypsum board dryer gate temperature; The adjustment of the distance between the gypsum board forming knife and the angle of the gypsum board forming knife can specifically avoid the defects of the gypsum board width and the gypsum board vertical edge angle; The height adjustment of the gypsum board forming board can specifically avoid the defects of the thickness of the gypsum board; The adjustment of the amount of dry material and wet material of the gypsum board is targeted to avoid the defects of the single weight of the wet board and the single weight of the dry board; The adjustment of the gate temperature of the gypsum board dryer can specifically avoid the defects of the temperature and humidity of the gypsum board after drying.

3. The defect detection and avoidance method based on full data collection of gypsum board production according to claim 2 is characterized by: The defect adjustment calculation model is obtained by learning and training the characterization data of the gypsum board defect items and the adjustment amount of the gypsum board production equipment adjustment items, including: Extract the characterization data of gypsum board defect items and the adjustment amount of gypsum board production equipment adjustment items from the gypsum board production full data collection log; Using the characterization data of the gypsum board defect item as the input item of the BP neural network, and using the adjustment amount of the gypsum board production equipment as the output item of the BP neural network; The BP neural network is used to perform convolution learning on the input items and output items of the BP neural network to obtain a defect adjustment measurement model; The model expression of the defect adjustment calculation model is: Wr=BP(Ds); Wherein, Wr is the adjustment amount of the gypsum board production equipment adjustment item r, Ds is the characterization data of the gypsum board defect item s, BP is the BP neural network, r is the identifier of the gypsum board production equipment adjustment item, and s is the identifier of the gypsum board defect item; The full data of gypsum board production consists of the characterization data combination of gypsum board width, gypsum board vertical edge angle, gypsum board thickness, gypsum board wet board unit weight, gypsum board dry board unit weight, gypsum board drying temperature and gypsum board drying humidity; The gypsum board production full data collection log is a log consisting of the gypsum board production full data collected in time sequence; The extraction of the characterization data of the gypsum board defect item includes: In the gypsum board production full data collection log, the characterization data of gypsum board width, gypsum board vertical edge angle, gypsum board thickness, gypsum board wet board unit weight, gypsum board dry board unit weight, gypsum board drying temperature and gypsum board drying humidity at each time series are combined with the gypsum board width, gypsum board vertical edge angle, gypsum board thickness, gypsum board wet board unit weight, gypsum board dry board unit weight, gypsum board drying temperature and gypsum board drying humidity at each time series. The standard data of board thickness, wet board weight of gypsum board, dry board weight of gypsum board, temperature of gypsum board after drying and humidity of gypsum board after drying are compared for the similarity of corresponding items, and the corresponding items whose similarity between the characterization data at each time sequence and the standard data is less than the allowable threshold are taken as the defective items of gypsum board at each time sequence; Extract the characterization data of gypsum board defect items at each time series from the full data collection log of gypsum board production; The extraction of the adjustment amount of the gypsum board production equipment adjustment item includes: In the gypsum board production full data collection log, the gypsum board defect items at each time sequence use the targeted matching relationship to determine the gypsum board production equipment adjustment items that are targetedly matched with the gypsum board defect items at each time sequence; In the full data collection log of gypsum board production, the adjustment amounts of the gypsum board production equipment adjustment items that specifically match the gypsum board defect items at each time sequence are extracted.

4. The defect detection and avoidance method based on full data collection of gypsum board production according to claim 3 is characterized by: The method of performing time series analysis on the adjustment amount of the gypsum board production equipment adjustment items to extract the associated relationship between the gypsum board production equipment adjustment items includes: Sorting the adjustment amounts of each gypsum board production equipment adjustment item according to the time sequence to obtain the adjustment amount time sequence of the gypsum board production equipment adjustment item; Perform sequence similarity analysis on the time series of adjustment amounts of each gypsum board production equipment adjustment item to quantify the associated rate that characterizes the associated relationship between each gypsum board production equipment adjustment item, wherein the associated relationship is a representation of the degree of mutual influence between the gypsum board production equipment adjustment items; The quantitative formula of the associated rate is: H ij =|W i,T -W j,T |; In the formula, H ij is the concomitant rate between gypsum board production equipment adjustment item i and gypsum board production equipment adjustment item j, W i,T is the time series of adjustment items i of gypsum board production equipment, W j,T is the time series of adjustment items j of gypsum board production equipment, |W i,T -W j,T | for W i,T and W j,T The Euclidean distance of , i, j are the identifiers of the gypsum board production equipment adjustment items, and T is the time series identifier.

5. The defect detection and avoidance method based on full data collection of gypsum board production according to claim 4 is characterized in that: The associated relationship is abstracted into an associated network between the adjustment items of gypsum board production equipment, including: The gypsum board production equipment adjustment items are abstracted into topological nodes, and topological edges are set between topological nodes for topological connection; The association rate between the gypsum board production equipment adjustment items is used as the edge weight of the topological edge between the corresponding topological nodes to abstract the associated network.

6. The defect detection and avoidance method based on full data collection of gypsum board production according to claim 5 is characterized by: The method of determining the real-time setting of the production working conditions of the gypsum board production equipment according to the number of real-time gypsum board production equipment adjustment items includes: If the number of real-time gypsum board production equipment adjustment items is 0, the current production condition setting of the gypsum board production equipment is maintained; If the number of real-time gypsum board production equipment adjustment items is 1, the adjustment amount of the real-time gypsum board production equipment adjustment item is mapped to the associated network to obtain a single real-time associated network, and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment is obtained in the single real-time associated network. The production working condition of the gypsum board production equipment is set in real time according to the adjustment amount of the real-time gypsum board production equipment adjustment item and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment, so as to achieve full detection and full avoidance of potential defects of the gypsum board; If the number of gypsum board production equipment adjustment items is greater than 1, the adjustment amount of each gypsum board real-time production equipment adjustment item is mapped to the associated network in turn to obtain multiple real-time associated networks, and the multiple real-time associated networks are fused to obtain the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment. The production conditions of the gypsum board production equipment are set in real time according to the adjustment amount of the real-time production equipment adjustment item of the gypsum board and the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment, so as to achieve full detection and full avoidance of potential defects of the gypsum board.

7. A defect detection and avoidance method based on full data collection of gypsum board production according to claim 6, characterized in that: The step of mapping the adjustment amount of the real-time gypsum board production equipment adjustment item to the associated network to obtain a single real-time associated network includes: The adjustment amount of the real-time gypsum board production equipment adjustment item is mapped to the node value of the corresponding topological node of the associated network, and the node values ​​of the remaining topological nodes in the associated network are calculated one by one according to the edge weights of the topological nodes and topological edges to obtain a single real-time associated network.

8. The defect detection and avoidance method based on full data collection of gypsum board production according to claim 6 is characterized in that: The step of obtaining the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment in a single real-time associated network includes: The topological nodes other than the topological nodes corresponding to the real-time gypsum board production equipment adjustment items in the single real-time associated network are used as the real-time associated adjustment items of the gypsum board production equipment; The node value of the topological node corresponding to the real-time associated adjustment item of the gypsum board production equipment in the single real-time associated network is used as the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment.

9. The defect detection and avoidance method based on full data collection of gypsum board production according to claim 6 is characterized in that: The step of sequentially mapping the adjustment amount of each gypsum board real-time production equipment adjustment item to the associated network to obtain multiple real-time associated networks includes: The adjustment amount of each gypsum board real-time production equipment adjustment item is mapped to the node value of the corresponding topological node of each associated network in turn, and the node values ​​of the remaining topological nodes in each associated network are calculated one by one according to the edge weights of the topological nodes and the topological edges, so as to obtain multiple real-time associated networks; The node values ​​of multiple real-time companion networks are superimposed and fused to obtain a single fused real-time companion network.

10. The defect detection and avoidance method based on full data collection of gypsum board production according to claim 6 is characterized in that: The method of fusing multiple real-time associated networks to obtain the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment includes: The topological nodes except all the topological nodes corresponding to the real-time gypsum board production equipment adjustment items in the single fusion real-time associated network are used as the real-time associated adjustment items of the gypsum board production equipment; The node value of the topological node corresponding to the real-time associated adjustment item of the gypsum board production equipment in the single fused real-time associated network is used as the adjustment amount of the real-time associated adjustment item of the gypsum board production equipment.

Images