Production line data exception management and control system and method based on big data

Through the production line data abnormality control system based on big data, the distance measuring camera is used to capture product shapes and cluster algorithms to identify abnormal products, which solves the problems of inaccurate prediction and high cost in the existing technology, and achieves rapid discovery of quality problems and flexibility Maintenance reduces defective rate and production costs.

CN119027044APending Publication Date: 2024-11-26CHANGSHU INSTITUTE OF TECHNOLOGY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411046187.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing production line abnormality supervision methods are difficult to eliminate problem products in a timely manner in terms of inaccurate prediction and high cost, and regular maintenance cannot flexibly adapt to the production line operation conditions, resulting in high production costs and equipment losses.

Method used

Adopt a production line data abnormality control system based on big data, including model comparison module, quality evaluation module, product clustering module, abnormality identification module and production offset module, and use the distance measuring camera to capture the shape of the product surface and establish a standard slice model. Calculate completion rate and flow vectors, identify abnormal products and perform real-time supervision and maintenance.

Benefits of technology

It has achieved the evaluation of product quality in a three-dimensional dimension, quickly discovered quality problems in the production process, reduced defective rates, improved production efficiency, timely adjusted production lines, and ensured product stability and consistency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119027044A_ABST
    Figure CN119027044A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of production line management, in particular to a production line data exception management and control system and method based on big data, and the system comprises a model comparison module, a quality evaluation module, a product clustering module, an exception recognition module and a production offset module. The quality evaluation module is used for outputting the completion rate of a product; the product clustering module is used for generating a flow vector and carrying out clustering; the abnormity identification module is used for removing abnormal products from a production line; the production deviation module is used for calculating the abnormal risk of each process and judging the maintenance time; the method can help to improve production efficiency, optimize resource utilization and improve quality management, can rapidly find quality problems in the production process, can timely adjust and improve a production line, can reduce production pauses caused by equipment faults, and can realize accuracy and flexibility of a production plan.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of production line management, and specifically to a production line data anomaly control system and method based on big data. Background Art

[0002] The production line refers to connecting each production link in the production process in sequence to form an assembly line-style operation system. On the production line, the product will start from raw materials, go through multiple processes, and finally be assembled. Each process may have production anomalies. Some obvious production anomalies can be discovered through product sampling, but some difficult-to-detect anomalies are difficult to supervise, resulting in a high rate of defective products.

[0003] The existing patent CN108090905A discloses a method for monitoring production line anomalies. It uses the images generated by the network prediction process to find anomalies in the production line by calculating the prediction differences. However, this method requires the monitoring equipment to have strong prediction computing capabilities, and the difference calculation method of the plane image is not suitable for the production system of three-dimensional products. Therefore, in actual use, there are problems such as high production costs and inaccurate predictions, and it is impossible to eliminate problematic products in time.

[0004] In addition, as the production line operates, the equipment on the entire production line is in a state of constant wear and tear. Therefore, it is necessary to overhaul processes with abnormal risks in advance. However, there are many processes on the production line, and regular maintenance cannot flexibly adapt to the operation of the production line, which can easily lead to unnecessary losses. Summary of the invention

[0005] The purpose of the present invention is to provide a production line data anomaly control system and method based on big data to solve the problems raised in the above background technology.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a production line data anomaly management and control system based on big data, comprising: a model comparison module, a quality evaluation module, a product clustering module, an anomaly identification module and a production offset module;

[0007] The model comparison module is used to process a standard product before the production line is put into operation. A standard slicing model of the product is established according to the state of the standard product after each process. After the production line is put into operation, the products on the production line are numbered, and the shape of the product surface is measured by a rangefinder camera after each process.

[0008] The quality evaluation module is used to number the products after the production line is running, and to photograph the surface shape of the products after the products go through a process, to construct a product model using the surface shape, to calculate the residual between the product model and the slice model, and to output the ratio of the residual to the model volume as the product completion rate.

[0009] The product clustering module is used to generate flow vectors for products in each process, with the change in the completion rate of the product before and after the process as the first coordinate and the time it takes to pass the process as the second coordinate. The flow vectors of all products constitute the flow vector set of the process. The flow vector set is clustered using a clustering algorithm to generate a standard flow vector, which represents the directional trend of the flow vector set.

[0010] The abnormality identification module is used to calculate the dot product of each vector in the flow vector set and the standard flow vector, mark the products whose dot product is not within the preset range as abnormal products, and remove the abnormal products from the production line after the abnormal products pass through the process;

[0011] The production deviation module is used to calculate the overall flow vector set of the production line, calculate the deviation speed of the angle of the overall flow vector over time, and when the deviation speed exceeds the threshold, further calculate the deviation speed of the standard flow vector of each process within the supervision cycle, obtain the abnormal risk of each process, and display it on the control terminal.

[0012] Further, the model comparison module includes: a model shaping unit and a stereoscopic shooting unit;

[0013] The model shaping unit is used to construct a product model according to the curved surface morphology of the product surface before and after the product passes through the process;

[0014] The stereoscopic shooting unit is used to shoot various side views of the product through a range-finding camera and analyze the range and depth of field of the product surface.

[0015] Furthermore, the quality evaluation module includes: a model slicing unit, a residual comparison unit and a completion quality unit;

[0016] The model slicing unit is used to capture views of products passing through the production line and generate surface functions of each surface of the product;

[0017] The residual comparison unit is used to compare the state function of each surface of the product with the model of each surface of the standard product to generate the residual of the two models;

[0018] The completion quality unit is used to calculate the proportion of the residual to the entire model as the completion rate output of the product.

[0019] Further, the product clustering module includes: a flow vector unit and a standard clustering unit;

[0020] The flow vector unit is used to generate a flow vector according to the change of the completion rate of the product before and after passing the process and the time of passing the process. All flow vectors in the same process constitute a flow vector set of the process.

[0021] The standard clustering unit is used to cluster the flow vector sets of each process and output the clustering results as standard flow vectors.

[0022] Further, the anomaly identification module includes: a group analysis unit and an anomaly marking unit;

[0023] The group analysis unit is used to analyze the dot product of the flow vector corresponding to each product passing through the process and the standard flow vector;

[0024] The abnormal marking unit is used to mark products whose flow vectors have a deviation from the standard flow vector that is higher than a threshold value, and remove them from the production line as abnormal products.

[0025] Further, the production offset module includes: a production line pointer unit, a pointing offset unit and a risk accumulation unit;

[0026] The production line pointer unit is used to calculate the relationship between the completion rate and production time of the entire production line and generate an overall flow vector;

[0027] The directional deviation unit is used to calculate the deviation speed of the overall flow vector and to alarm when the deviation speed is higher than a predetermined value;

[0028] The risk accumulation unit is used to calculate the dot product of the standard flow vector of each process and the overall flow vector after the production line alarms, and output it to the maintenance personnel as the failure risk of the process.

[0029] The production line data anomaly control method based on big data includes the following steps:

[0030] Step S1. A standard product is processed before the production line is put into operation. After each process, a camera capable of identifying depth of field is used to shoot orthographic projection views of each surface of the standard product, and a surface function is generated for each view. All the surface functions obtained constitute a standard model of the product;

[0031] Step S2. The production line is running. After the product to be processed has gone through the process, the orthographic projection views of each surface of the product are photographed and the surface functions are generated. All the surface functions obtained constitute the verification model. The verification model is subtracted from the product standard model, and the total residual is used as the completion rate of the product;

[0032] Step S3. Generate a flow vector by taking the change in the completion rate of the product before and after the process as the first coordinate and the duration of the process as the second coordinate, cluster the flow vectors of the processed products in the same process, and use the clustering result as the standard flow vector;

[0033] Step S4. Calculate the product deviation according to the product flow vector and the standard flow vector, mark the product whose product deviation is outside the preset range as a failed product of the process, and remove it from the production line;

[0034] Step S5. At the end of the production line, the total completion rate of the product is used as the first coordinate, and the total processing time is used as the second coordinate to generate a total flow vector. The total flow vectors of the processed products are clustered to obtain a production offset vector. The angular change rate of the production offset vector within a preset period is used as the abnormal risk of the production line. When the risk exceeds the threshold, the production line is inspected.

[0035] Further, step S1 includes:

[0036] Step S11. Build a shooting space at the end of each process of the production line, preset the origin position, and establish a spatial rectangular coordinate system with the length, width and height of the shooting space as the X-axis, Y-axis and Z-axis respectively;

[0037] Step S12. Processing a standard product in the production line, after each process, placing the standard product in the shooting space, and using a camera capable of identifying depth of field to shoot n-side projection views of the product, where n is a preset value, and n≥6;

[0038] Step S13. Segment the area where the product is located from the projection view, and establish a surface function z=F(x,y) according to the position and depth of field of the segmentation result, where x, y and z represent the horizontal coordinate, vertical coordinate and vertical coordinate of the product surface respectively;

[0039] Step S14. Execute step S13 for all n views of the standard product, and all the surface functions obtained constitute the product standard model L, where L = {F1(x,y), F2(x,y), ..., Fn(x,y)}, where Fn(x,y) represents the surface function corresponding to the nth surface of the standard product.

[0040] This step is used to model the standard state of the product. Different from the direct comparison of the pictures taken of the product, this step comprehensively draws the precise structure of the product in all directions by combining depth of field with picture segmentation, which can achieve more accurate identification of product quality.

[0041] Further, step S2 includes:

[0042] Step S21. The production line is operated to start production. After the product to be processed passes the process, the regular six-sided projection view of the product is photographed. The surface function corresponding to the n-side view of the product is calculated according to the method of step S13-step S14 to obtain the verification model B = {G1(x,y), G2(x,y), ..., Gn(x,y)}, where Gn(x,y) represents the surface function corresponding to the n-th side of the product to be processed;

[0043] Step S22. Calculate the completion rate of the product to be processed according to the product standard model and the verification model according to the following formula:

[0044]

[0045] Among them, E represents the completion rate of the current process of the product to be processed, E0 represents the completion rate of the product to be processed in the previous process, the initial value is 0, i is the view number, Di is the plane Fi(x, y) = 0, representing the orthographic projection area, Fi(x, y) and Gi(x, y) represent the surface functions of the standard product and the i-th surface of the product to be processed respectively, and A is the area unit.

[0046] This step uses the surface model comparison method to calculate the residual between the actual product appearance and the standard product appearance, thereby identifying the quality differences of the products. The quality of the products can be evaluated in a three-dimensional dimension, helping to improve the efficiency of production supervision.

[0047] Further, step S3 includes:

[0048] Step S31. After the product passes through the process, record the time it takes for the product to pass through the process, use the product completion rate as the first coordinate and the time it takes to pass through the process as the second coordinate to construct the product flow vector Said Where t represents the time it takes for the product to be processed to pass through the process;

[0049] Step S32. According to the flow vectors of all processed products in the process, a flow vector set {v1, v2, ..., vm} is formed, where m is the number of processed products in the process, and vm represents the flow vector of the mth processed product;

[0050] Step S33. Cluster the flow vector set using a vector clustering algorithm. The clustering result is recorded as a standard flow vector. The standard flow vector represents the overall trend of the vectors in the flow vector set. The standard flow vector is updated once each time an element is added to the flow vector set.

[0051] Further, step S4 includes:

[0052] Step S41. After the product passes through the process and the product flow vector is constructed, the deviation between the product flow vector and the standard flow vector is calculated according to the following formula:

[0053]

[0054] Where P represents the vector deviation of the product, represents the standard flow vector;

[0055] Step S42. Preset the product qualification range. If the vector deviation of the product is not within the qualification range, the product will be removed from the production line as a processing failure. If the vector deviation of the product is within the qualification range, the product will be sent to the next process and the flow vector of the product will be added to the flow vector set of the current process.

[0056] This step can eliminate abnormal workpieces on the production line so that they do not flow into other processes, saving the supervision process of subsequent processes. At the same time, it can reduce the defective rate of the final product, quickly discover quality problems in the production process, and improve production efficiency.

[0057] Further, step S5 includes:

[0058] Step S51. After the product passes the last process of the production line, the final completion rate of the product is recorded as Ed, and the total time the product passes through the production line is recorded as T. Ed and T are used as vector coordinates to generate the total flow vector of the product.

[0059] Step S52. A total flow vector set is formed according to the flow vectors of all processed products of the production line. The total flow vector set is clustered according to the method of step S33. The clustering result is recorded as the production offset vector.

[0060] Step S53. Every time an element is added to the total flow vector set, the production offset vector is updated once, and the production offset vector before the ath update is recorded as The production offset vector after the ath update is recorded as The time interval of the ath update is recorded as t a , calculate the abnormal risk of the production line according to the following method:

[0061]

[0062] Where Y represents the abnormal risk of the production line, a represents the current update number, and u represents the total update number of the production offset vector;

[0063] Step S54. When the abnormal risk Y of the production line exceeds the preset threshold, the production line is shut down for maintenance, and according to the method of step S41, the vector deviation of the standard flow vector of each process relative to the production offset vector during the shutdown is calculated, and the maintenance order of the process is determined according to the descending order of the vector deviation.

[0064] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0065] 1. The present invention can establish a standard slicing model of a product according to the state of the standard product after each process. After each process, the marked product is compared with the slicing model, and the model residual is used to evaluate the completion rate of the product. The quality of the product can be evaluated in a three-dimensional dimension, which helps to improve production efficiency, optimize resource utilization, improve quality management, and achieve accuracy and flexibility in production planning.

[0066] 2. The present invention can calculate the quality flow vector of each product, cluster the flow vectors of all products, calculate the vector product of the flow vector of each product and the standard flow vector, identify abnormal products after each process is executed, quickly discover quality problems in the production process, reduce product defective rate, and improve production efficiency.

[0067] 3. The present invention can calculate the offset speed of the completion rate pointer according to the full-cycle flow vector set of the production line, thereby obtaining the abnormal risk of each process, and inspecting the process according to the accumulation of abnormal risks. The production line can be adjusted and improved in time to ensure the stability and consistency of the product production process and reduce production stoppages caused by equipment failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0069] Figure 1 It is a structural schematic diagram of the production line data abnormality control system based on big data of the present invention;

[0070] Figure 2 It is a schematic diagram of the steps of the production line data anomaly control method based on big data of the present invention. DETAILED DESCRIPTION

[0071] 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.

[0072] See also Figure 1 , the present invention provides a technical solution: a production line data anomaly management and control system based on big data, including: a model comparison module, a quality evaluation module, a product clustering module, an anomaly identification module and a production offset module;

[0073] The model comparison module is used to process a standard product before the production line is put into operation. A standard slicing model of the product is established according to the state of the standard product after each process. After the production line is put into operation, the products on the production line are numbered, and the shape of the product surface is measured by a rangefinder camera after each process.

[0074] The model comparison module includes: a model shaping unit and a stereoscopic shooting unit;

[0075] The model shaping unit is used to construct a product model according to the curved surface morphology of the product surface before and after the product passes through the process;

[0076] The stereoscopic shooting unit is used to shoot various side views of the product through a range-finding camera and analyze the range and depth of field of the product surface.

[0077] The quality evaluation module is used to number the products after the production line is running, and to photograph the surface shape of the products after the products go through a process, to construct a product model using the surface shape, to calculate the residual between the product model and the slice model, and to output the ratio of the residual to the model volume as the product completion rate.

[0078] The quality evaluation module includes: a model slicing unit, a residual comparison unit and a completion quality unit;

[0079] The model slicing unit is used to capture views of products passing through the production line and generate surface functions of each surface of the product;

[0080] The residual comparison unit is used to compare the state function of each surface of the product with the model of each surface of the standard product to generate the residual of the two models;

[0081] The completion quality unit is used to calculate the proportion of the residual to the entire model as the completion rate output of the product.

[0082] The product clustering module is used to generate flow vectors for products in each process, with the change in the completion rate of the product before and after the process as the first coordinate and the time it takes to pass the process as the second coordinate. The flow vectors of all products constitute the flow vector set of the process. The flow vector set is clustered using a clustering algorithm to generate a standard flow vector, which represents the directional trend of the flow vector set.

[0083] The product clustering module includes: a flow vector unit and a standard clustering unit;

[0084] The flow vector unit is used to generate a flow vector according to the change of the completion rate of the product before and after passing the process and the time of passing the process. All flow vectors in the same process constitute a flow vector set of the process.

[0085] The standard clustering unit is used to cluster the flow vector sets of each process and output the clustering results as standard flow vectors.

[0086] The abnormality identification module is used to calculate the dot product of each vector in the flow vector set and the standard flow vector, mark the products whose dot product is not within the preset range as abnormal products, and remove the abnormal products from the production line after the abnormal products pass through the process;

[0087] The anomaly identification module includes: a group analysis unit and an anomaly marking unit;

[0088] The group analysis unit is used to analyze the dot product of the flow vector corresponding to each product passing through the process and the standard flow vector;

[0089] The abnormal marking unit is used to mark products whose flow vectors have a deviation from the standard flow vector that is higher than a threshold value, and remove them from the production line as abnormal products.

[0090] The production deviation module is used to calculate the overall flow vector set of the production line, calculate the deviation speed of the angle of the overall flow vector over time, and when the deviation speed exceeds the threshold, further calculate the deviation speed of the standard flow vector of each process within the supervision cycle, obtain the abnormal risk of each process, and display it on the control terminal.

[0091] The production offset module includes: a production line pointer unit, a pointing offset unit and a risk accumulation unit;

[0092] The production line pointer unit is used to calculate the relationship between the completion rate and production time of the entire production line and generate an overall flow vector;

[0093] The directional deviation unit is used to calculate the deviation speed of the overall flow vector and to alarm when the deviation speed is higher than a predetermined value;

[0094] The risk accumulation unit is used to calculate the dot product of the standard flow vector of each process and the overall flow vector after the production line alarms, and output it to the maintenance personnel as the failure risk of the process.

[0095] The production line data anomaly control method based on big data includes the following steps:

[0096] Step S1. A standard product is processed before the production line is put into operation. After each process, a camera capable of identifying depth of field is used to shoot orthographic projection views of each surface of the standard product, and a surface function is generated for each view. All the surface functions obtained constitute a standard model of the product;

[0097] Step S1 includes:

[0098] Step S11. Build a shooting space at the end of each process of the production line, preset the origin position, and establish a spatial rectangular coordinate system with the length, width and height of the shooting space as the X-axis, Y-axis and Z-axis respectively;

[0099] Step S12. Processing a standard product in the production line, after each process, placing the standard product in the shooting space, and using a camera capable of identifying depth of field to shoot n-side projection views of the product, where n is a preset value, and n≥6;

[0100] Step S13. Segment the area where the product is located from the projection view, and establish a surface function z=F(x,y) according to the position and depth of field of the segmentation result, where x, y and z represent the horizontal coordinate, vertical coordinate and vertical coordinate of the product surface respectively;

[0101] Step S14. Execute step S13 for all n views of the standard product, and all the surface functions obtained constitute the product standard model L, where L = {F1(x,y), F2(x,y), ..., Fn(x,y)}, where Fn(x,y) represents the surface function corresponding to the nth surface of the standard product.

[0102] Step S2. The production line is running. After the product to be processed has gone through the process, the orthographic projection views of each surface of the product are photographed and the surface functions are generated. All the surface functions obtained constitute the verification model. The verification model is subtracted from the product standard model, and the total residual is used as the completion rate of the product;

[0103] Step S2 includes:

[0104] Step S21. The production line is operated to start production. After the product to be processed passes the process, the regular six-sided projection view of the product is photographed. The surface function corresponding to the n-side view of the product is calculated according to the method of step S13-step S14 to obtain the verification model B = {G1(x,y), G2(x,y), ..., Gn(x,y)}, where Gn(x,y) represents the surface function corresponding to the n-th side of the product to be processed;

[0105] Step S22. Calculate the completion rate of the product to be processed according to the product standard model and the verification model according to the following formula:

[0106]

[0107] Among them, E represents the completion rate of the current process of the product to be processed, E0 represents the completion rate of the product to be processed in the previous process, the initial value is 0, i is the view number, Di is the plane Fi(x, y) = 0, representing the orthographic projection area, Fi(x, y) and Gi(x, y) represent the surface functions of the standard product and the i-th surface of the product to be processed respectively, and A is the area unit.

[0108] Step S3. Generate a flow vector by taking the change in the completion rate of the product before and after the process as the first coordinate and the duration of the process as the second coordinate, cluster the flow vectors of the processed products in the same process, and use the clustering result as the standard flow vector;

[0109] Step S3 includes:

[0110] Step S31. After the product passes through the process, record the time it takes for the product to pass through the process, use the product completion rate as the first coordinate and the time it takes to pass through the process as the second coordinate to construct the product flow vector Said Where t represents the time it takes for the product to be processed to pass through the process;

[0111] Step S32. According to the flow vectors of all processed products in the process, a flow vector set {v1, v2, ..., vm} is formed, where m is the number of processed products in the process, and vm represents the flow vector of the mth processed product;

[0112] Step S33. Cluster the flow vector set using a vector clustering algorithm. The clustering result is recorded as a standard flow vector. The standard flow vector represents the overall trend of the vectors in the flow vector set. The standard flow vector is updated once each time an element is added to the flow vector set.

[0113] Step S4. Calculate the product deviation according to the product flow vector and the standard flow vector, mark the product whose product deviation is outside the preset range as a failed product of the process, and remove it from the production line;

[0114] Step S4 includes:

[0115] Step S41. After the product passes through the process and the product flow vector is constructed, the deviation between the product flow vector and the standard flow vector is calculated according to the following formula:

[0116]

[0117] Where P represents the vector deviation of the product, represents the standard flow vector;

[0118] Step S42. Preset the product qualification range. If the vector deviation of the product is not within the qualification range, the product will be removed from the production line as a processing failure. If the vector deviation of the product is within the qualification range, the product will be sent to the next process and the flow vector of the product will be added to the flow vector set of the current process.

[0119] Step S5. At the end of the production line, the total completion rate of the product is used as the first coordinate, and the total processing time is used as the second coordinate to generate a total flow vector. The total flow vectors of the processed products are clustered to obtain a production offset vector. The angular change rate of the production offset vector within a preset period is used as the abnormal risk of the production line. When the risk exceeds the threshold, the production line is inspected.

[0120] Step S5 includes:

[0121] Step S51. After the product passes the last process of the production line, the final completion rate of the product is recorded as Ed, and the total time the product passes through the production line is recorded as T. Ed and T are used as vector coordinates to generate the total flow vector of the product.

[0122] Step S52. A total flow vector set is formed according to the flow vectors of all processed products of the production line. The total flow vector set is clustered according to the method of step S33. The clustering result is recorded as the production offset vector.

[0123] Step S53. Every time an element is added to the total flow vector set, the production offset vector is updated once, and the production offset vector before the ath update is recorded as The production offset vector after the ath update is recorded as The time interval of the ath update is recorded as t a , calculate the abnormal risk of the production line according to the following method:

[0124]

[0125] Where Y represents the abnormal risk of the production line, a represents the current update number, and u represents the total update number of the production offset vector;

[0126] Step S54. When the abnormal risk Y of the production line exceeds the preset threshold, the production line is shut down for maintenance, and according to the method of step S41, the vector deviation of the standard flow vector of each process relative to the production offset vector during the shutdown is calculated, and the maintenance order of the process is determined according to the descending order of the vector deviation.

[0127] Example: There are 3 processes in a production line, and 6 products need to be produced in a certain period of time. After the product passes the first process, the camera takes a picture of the product and draws the product surface function. Compared with the model of the standard product after passing the first process, the completion rates of the 6 products are 0.75, 0.60, 0.70, 0.80, 0.65 and 0.50 respectively, and the time it takes to pass the process is 1 minute. The flow vector set of the process is {(0.75, 1), (0.60, 1), (0.70, 1), (0.80, 1), (0.65, 1), (0.50, 1)}. After clustering, the standard flow vector (0.67, 1) is obtained. The calculated vector deviations are 0.04, 0.04, 0.02, 0.06, 0.01 and 0.11 respectively. The vector deviation of the 6th product is greater than the preset value 0.1, so the 6th product is removed from the production line.

[0128] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0129] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A production line data anomaly control method based on big data, characterized in that: The method comprises the following steps: Step S1. A standard product is processed before the production line is put into operation. After each process, a camera capable of identifying depth of field is used to shoot orthographic projection views of each surface of the standard product, and a surface function is generated for each view. All the surface functions obtained constitute a standard model of the product; Step S2. After the product to be processed has gone through the production process, the orthographic projection views of each surface of the product are photographed and the surface functions are generated. All the surface functions obtained constitute the verification model. The verification model is subtracted from the product standard model, and the total residual is used as the completion rate of the product; Step S3. Generate a flow vector by taking the change in the completion rate of the product before and after the process as the first coordinate and the duration of the process as the second coordinate, cluster the flow vectors of the processed products in the same process, and use the clustering result as the standard flow vector; Step S4. Calculate the product deviation according to the product flow vector and the standard flow vector, mark the product whose product deviation is outside the preset range as a failed product of the process, and remove it from the production line; Step S5. At the end of the production line, the total completion rate of the product is used as the first coordinate, and the total processing time is used as the second coordinate to generate a total flow vector. The total flow vectors of the processed products are clustered to obtain a production offset vector. The angular change rate of the production offset vector within a preset period is used as the abnormal risk of the production line. When the risk exceeds the threshold, the production line is inspected.

2. The method for controlling production line data anomalies based on big data according to claim 1, characterized in that: Step S1 includes: Step S11. Build a shooting space at the end of each process of the production line, preset the origin position, and establish a spatial rectangular coordinate system with the length, width and height of the shooting space as the X-axis, Y-axis and Z-axis respectively; Step S12. Processing a standard product in the production line, after each process, placing the standard product in the shooting space, and using a camera capable of identifying depth of field to shoot n-side projection views of the product, where n is a preset value, and n≥6; Step S13. Segment the area where the product is located from the projection view, and establish a surface function z=F(x,y) according to the position and depth of field of the segmentation result, where x, y and z represent the horizontal coordinate, vertical coordinate and vertical coordinate of the product surface respectively; Step S14. Execute step S13 for all n views of the standard product, and all the surface functions obtained constitute the product standard model L, where L = {F1(x,y), F2(x,y), ..., Fn(x,y)}, where Fn(x,y) represents the surface function corresponding to the nth surface of the standard product.

3. The method for controlling production line data anomalies based on big data according to claim 2 is characterized in that: Step S2 includes: Step S21. The production line is operated to start production. After the product to be processed passes the process, the regular six-sided projection view of the product is photographed. The surface function corresponding to the n-side view of the product is calculated according to the method of step S13-step S14 to obtain the verification model B = {G1(x,y), G2(x,y), ..., Gn(x,y)}, where Gn(x,y) represents the surface function corresponding to the n-th side of the product to be processed; Step S22. Calculate the completion rate of the product to be processed according to the product standard model and the verification model according to the following formula: Among them, E represents the completion rate of the current process of the product to be processed, E0 represents the completion rate of the product to be processed in the previous process, the initial value is 0, i is the view number, Di is the plane Fi(x, y) = 0, representing the orthographic projection area, Fi(x, y) and Gi(x, y) represent the surface functions of the standard product and the i-th surface of the product to be processed respectively, and A is the area unit.

4. The method for controlling production line data anomalies based on big data according to claim 3 is characterized in that: Step S3 includes: Step S31. After the product passes through the process, record the time it takes for the product to pass through the process, use the product completion rate as the first coordinate and the time it takes to pass through the process as the second coordinate to construct the product flow vector Said Where t represents the time it takes for the product to be processed to pass through the process; Step S32. According to the flow vectors of all processed products in the process, a flow vector set {v1, v2, ..., vm} is formed, where m is the number of processed products in the process, and vm represents the flow vector of the mth processed product; Step S33. Cluster the flow vector set using a vector clustering algorithm, and record the clustering result as a standard flow vector. The standard flow vector represents the overall trend of the vectors in the flow vector set. The standard flow vector is updated once each time an element is added to the flow vector set. Step S4 includes: Step S41. After the product passes through the process and the product flow vector is constructed, the deviation between the product flow vector and the standard flow vector is calculated according to the following formula: Where P represents the vector deviation of the product, represents the standard flow vector; Step S42. Preset the product qualification range. If the vector deviation of the product is not within the qualification range, the product will be removed from the production line as a processing failure. If the vector deviation of the product is within the qualification range, the product will be sent to the next process and the flow vector of the product will be added to the flow vector set of the current process.

5. The method for controlling production line data anomalies based on big data according to claim 4 is characterized in that: Step S5 includes: Step S51. After the product passes the last process of the production line, the final completion rate of the product is recorded as Ed, and the total time the product passes through the production line is recorded as T. Ed and T are used as vector coordinates to generate the total flow vector of the product. Step S52. A total flow vector set is formed according to the flow vectors of all processed products of the production line. The total flow vector set is clustered according to the method of step S33. The clustering result is recorded as the production offset vector. Step S53. Every time an element is added to the total flow vector set, the production offset vector is updated once, and the production offset vector before the ath update is recorded as The production offset vector after the ath update is recorded as The time interval of the ath update is recorded as t a , calculate the abnormal risk of the production line according to the following method: Where Y represents the abnormal risk of the production line, a represents the current update number, and u represents the total update number; Step S54: When the abnormal risk Y of the production line exceeds the preset threshold, the production line is shut down for maintenance.

6. The production line data abnormality control system based on big data is characterized by: The system includes the following modules: a model comparison module, a quality evaluation module, a product clustering module, an anomaly identification module, and a production shift module; The model comparison module is used to process a standard product before the production line is put into operation, and to establish a product standard slice model according to the state of the standard product after each process. After the production line is put into operation, the products on the production line are numbered, and the shape of the product surface is measured by a rangefinder camera after each process of the product. The quality evaluation module is used to number the products after the production line is running, and after the products go through a process, take a picture of the surface shape of the products, construct a product model using the surface shape, calculate the residual between the product model and the slice model, and output the ratio of the residual to the model volume as the product completion rate; The product clustering module is used to generate flow vectors for products in each process, with the change in the completion rate of the product before and after the process as the first coordinate and the time it takes to pass the process as the second coordinate. The flow vectors of all products constitute the flow vector set of the process. The flow vector set is clustered using a clustering algorithm to generate a standard flow vector, which represents the directional trend of the flow vector set. The abnormality identification module is used to calculate the dot product of each vector in the flow vector set and the standard flow vector, mark the products whose dot product is not within the preset range as abnormal products, and remove the abnormal products from the production line after the abnormal products pass through the process; The production deviation module is used to calculate the overall flow vector set of the production line, calculate the deviation speed of the angle of the overall flow vector over time, and when the deviation speed exceeds the threshold, further calculate the deviation speed of the standard flow vector of each process within the supervision cycle, obtain the abnormal risk of each process, and display it on the control terminal.

7. The production line data anomaly management and control system based on big data according to claim 6 is characterized by: The model comparison module includes: a model shaping unit and a stereoscopic shooting unit; The model shaping unit is used to construct a product model according to the curved surface morphology of the product surface before and after the product passes through the process; The stereoscopic shooting unit is used to shoot various side views of the product through a range-finding camera and analyze the range and depth of field of the product surface.

8. The production line data anomaly management and control system based on big data according to claim 7 is characterized by: The quality evaluation module includes: a model slicing unit, a residual comparison unit and a completion quality unit; The model slicing unit is used to capture views of products passing through the production line and generate surface functions of each surface of the product; The residual comparison unit is used to compare the state function of each surface of the product with the model of each surface of the standard product to generate the residual of the two models; The completion quality unit is used to calculate the proportion of the residual to the entire model as the completion rate output of the product.

9. The production line data anomaly management and control system based on big data according to claim 8 is characterized by: The product clustering module includes: a flow vector unit and a standard clustering unit; The flow vector unit is used to generate a flow vector according to the change of the completion rate of the product before and after passing the process and the time of passing the process. All flow vectors in the same process constitute a flow vector set of the process. The standard clustering unit is used to cluster the flow vector sets of each process and output the clustering results as standard flow vectors; The anomaly identification module includes: a group analysis unit and an anomaly marking unit; The group analysis unit is used to analyze the dot product of the flow vector corresponding to each product passing through the process and the standard flow vector; The abnormal marking unit is used to mark products whose flow vectors have a deviation from the standard flow vector that is higher than a threshold value, and remove them from the production line as abnormal products.

10. The production line data anomaly management and control system based on big data according to claim 9 is characterized in that: The production offset module includes: a production line pointer unit, a pointing offset unit and a risk accumulation unit; The production line pointer unit is used to calculate the relationship between the completion rate and production time of the entire production line and generate an overall flow vector; The directional deviation unit is used to calculate the deviation speed of the overall flow vector and to alarm when the deviation speed is higher than a predetermined value; The risk accumulation unit is used to calculate the dot product of the standard flow vector of each process and the overall flow vector after the production line alarms, and output it to the maintenance personnel as the failure risk of the process.