Operation fault automatic detection method for intelligent manufacturing production line

Abstract

The invention discloses an operation fault automatic detection method for an intelligent manufacturing production line, and relates to the technical field of production equipment fault detection, and the method comprises the steps: decomposing the production line into at least one processing link, and obtaining at least one product classification; obtaining a target feature product processed by the feature product in a normal operation production line; collecting a processing part in a processing link in the production line; processing attributes of products in the production line are obtained; processing shapes of products in the production line are obtained; the block chain sets access verification at each distributed database node; product quality abnormity detection is carried out in the production line, products with quality abnormity are obtained, and the products with the same processing attributes in the production line are verified; and determining a processing link with a fault in the production line according to the verification result. By collecting the machining part, the machining attribute and the machining shape, it can be determined that the position of the corresponding machining link in the production line has a fault according to the detection result.

Description

Technical Field

[0001] This invention relates to the field of production equipment fault detection technology, specifically to an automated fault detection method for intelligent manufacturing production lines. Background Technology

[0002] An automated production line refers to a production organization form in which the product manufacturing process is realized by an automated machine system. It is formed on the basis of further development of continuous production lines. Its characteristics are: the workpiece is automatically transferred from one machine tool to another, and the machine tools automatically perform processing, loading and unloading, inspection, etc.; the workers' tasks are only to adjust, supervise and manage the automated line, and they do not participate in direct operation; all machines and equipment operate at a uniform rhythm, and the production process is highly continuous.

[0003] However, traditional production line processing presents several unresolved issues. Different methods can be used to monitor electrical equipment malfunctions during production. Sometimes manual inspection is employed, but this method is typically ineffective, expensive, and time-consuming. On the other hand, using specialized quality testing equipment may require significant adjustments to the production line and substantial upfront investment. Summary of the Invention

[0004] To address the aforementioned technical problems, this paper provides an automated fault detection method for intelligent manufacturing production lines. This solution addresses the issue raised in the background section regarding the various methods used to monitor electrical equipment faults during production. While manual inspection is sometimes employed, these methods are typically ineffective, expensive, and time-consuming. However, using specialized quality testing equipment may require significant adjustments to the production line and substantial upfront investment.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] An automated fault detection method for intelligent manufacturing production lines includes:

[0007] The process involves acquiring products that have entered the production line, creating identification codes, matching the identification codes with the products, and printing the identification codes onto the products. The identification codes can be either serial numbers or QR codes.

[0008] The production line is broken down into at least one processing stage, and the products in the production line are classified to obtain at least one product category, in which the products are the same type of product;

[0009] In each product category, one product is selected as the feature product, and the target feature product is obtained after the feature product is processed on a normally operating production line.

[0010] In the processing stage of the production line, the processing part is collected, and the processing stage is matched with the processing part. The processing part includes the part of the product that is processed in the processing stage, and the processing part is matched with the corresponding identification code of the product.

[0011] Obtain the processing attributes of products in the production line, including the product category and matching the processing attributes with the corresponding identification code of the product.

[0012] The processing shape of the product in the production line is obtained. The processing shape is the product form after real-time processing. The processing shape is matched with the identification code corresponding to the product.

[0013] Track down at least one product with the same processing attributes, set up a distributed database node on the blockchain, and use distributed storage technology to store the identification code, processing part, processing attributes, and processing shape of at least one product with the same processing attributes;

[0014] The blockchain sets up access verification on each distributed database node. Visitors who pass the access verification can obtain the processing part, processing attributes and processing shape of the product in the distributed database node.

[0015] Product quality anomaly detection is carried out on the production line to obtain products with quality anomalies and to verify products with the same processing attributes on the production line.

[0016] The verification results will determine the processing steps in the production line where the fault exists.

[0017] Preferably, the step of breaking down the production line into at least one processing stage includes the following steps:

[0018] Set up an input layer, a hidden layer, and an output layer to obtain a standardized production line sample set;

[0019] Input neurons are created in the input layer, hidden neurons are created in the hidden layer, and output neurons are created in the output layer.

[0020] Data from the standardized production line sample set is transmitted to the input layer to obtain the output results from the output layer;

[0021] Determine whether the output of the output layer is reasonable. If it is, do not process it. If not, obtain the flow path of the standardized production line sample set, adjust the neurons in the path, and retransmit the data of the standardized production line sample set to the input layer until the data output of the standardized production line sample set is reasonable.

[0022] By combining the input neurons of the input layer, the hidden neurons of the hidden layer, and the output neurons of the output layer, we obtain the production line decomposition model.

[0023] By inputting the production line into the production line decomposition model, at least one processing step can be obtained.

[0024] Preferably, classifying the products in the production line to obtain at least one product category includes the following steps:

[0025] Perform contour recognition on the products in the production line to obtain at least one product contour;

[0026] At least one product profile is compared with each other, and product profiles with differences less than a preset difference are grouped into the same product category to obtain at least one product category.

[0027] Preferably, the tracking to obtain at least one product with the same processing attributes includes the following steps:

[0028] In the production line, it is determined whether the product categories of two products are consistent. If not, no action is taken.

[0029] If so, then the two products have the same processing attributes.

[0030] Preferably, storing the identification code, processing location, processing attribute, and processing shape of at least one product with the same processing attributes using distributed storage technology includes the following steps:

[0031] The identification code, processing part, processing attribute and processing shape of at least one product with the same processing attributes are divided and segmented, and the data is evenly stored on multiple distributed database nodes.

[0032] Configure data replication and redundancy backup strategies in the distributed database, and use master-slave replication or multi-master replication to replicate data to multiple distributed database nodes;

[0033] Distributed transaction processing technology is used to achieve data consistency and synchronization in a distributed database;

[0034] A data sharding routing method is adopted to implement load balancing and performance optimization strategies in a distributed database;

[0035] In a distributed database, disaster recovery and fault restoration mechanisms, including fault detection and automatic failover, should be configured.

[0036] Preferably, the blockchain sets up access verification on each distributed database node, including the following steps:

[0037] Set up a verification question and a verification answer on each distributed database node;

[0038] Send a verification question to the visitor and obtain the visitor's actual answer;

[0039] Determine whether the actual answer matches the verification answer. If yes, the verification passes; otherwise, the verification fails.

[0040] Preferably, the step of detecting product quality anomalies on the production line and obtaining products with quality anomalies includes the following steps:

[0041] Obtain the product's processed shape, identify target feature products in the same product category as the product, and pair them with the product;

[0042] Compare the processed parts of the product with the target feature product that it is paired with;

[0043] Obtain the first surface contour of the processed part of the product, and obtain the second surface contour of the processed part of the target feature product paired with the product.

[0044] When the first surface profile is inconsistent with the second surface profile, the product quality is abnormal.

[0045] Record any abnormal processing parts as abnormal processing parts.

[0046] Preferably, the verification of products with the same processing attributes in the production line includes the following steps:

[0047] Obtain the product category to which the product with quality issues belongs, and use it as the category of abnormal products.

[0048] The abnormal processing parts of the products in the abnormal product category are compared with the abnormal processing parts of the target feature products that are paired with the products, and the products in the abnormal product category with inconsistencies are counted.

[0049] If the number of products in an inconsistent abnormal product category exceeds the preset number, then the production line is faulty.

[0050] If the number of products in an inconsistent abnormal product category is not greater than the preset number, then there is no fault in the production line.

[0051] Preferably, determining the faulty processing step in the production line based on the verification results includes the following steps:

[0052] When a production line malfunctions, the processing step corresponding to the abnormal processing part is obtained and identified as the malfunctioning processing step.

[0053] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0054] By collecting data on processing location, processing attributes, and processing shape, products with quality abnormalities are identified. Products with the same processing attributes in the production line are then verified. Based on the verification results, the faulty processing stage in the production line is determined. This allows for direct product inspection using algorithms. Furthermore, based on the product inspection, the fault in the production line can be derived. Since all inspections are performed after product processing is complete, extensive adjustments to the production line are unnecessary, reducing upfront investment. By decomposing the production line into at least one processing stage and identifying the processing location of each stage, the faulty processing stage can be determined based on the inspection results. This allows for pinpointing the location of the fault in the corresponding processing stage in the production line. Because the fault range is significantly narrowed, the cause of the fault can be quickly identified. Attached Figure Description

[0055] Figure 1 This is a schematic diagram of the automated fault detection method for intelligent manufacturing production lines according to the present invention.

[0056] Figure 2 This is a schematic diagram illustrating the process of breaking down the production line into at least one processing stage according to the present invention;

[0057] Figure 3 To classify products in a production line according to the present invention, at least one product classification process diagram is obtained;

[0058] Figure 4 This invention provides a schematic diagram of the process flow for tracking at least one product with the same processing attributes.

[0059] Figure 5 This invention provides a schematic diagram illustrating the process of using distributed storage technology to store the identification code, processing location, processing attributes, and processing shape of at least one product with the same processing attributes.

[0060] Figure 6 This is a schematic diagram illustrating the access verification process set up on each distributed database node in the blockchain of this invention.

[0061] Figure 7 The present invention provides a schematic diagram of the process for detecting product quality anomalies on a production line and obtaining products with quality anomalies.

[0062] Figure 8 This is a schematic diagram of the verification process for products with the same processing attributes in the production line according to the present invention. Detailed Implementation

[0063] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0064] Reference Figure 1 As shown, an automated fault detection method for intelligent manufacturing production lines includes:

[0065] The process involves acquiring products that have entered the production line, creating identification codes, matching the identification codes with the products, and printing the identification codes onto the products. The identification codes can be either serial numbers or QR codes.

[0066] The production line is broken down into at least one processing stage, and the products in the production line are classified to obtain at least one product category, in which the products are the same type of product;

[0067] In each product category, one product is selected as the feature product, and the target feature product is obtained after the feature product is processed on a normally operating production line.

[0068] In the processing stage of the production line, the processing part is collected, and the processing stage is matched with the processing part. The processing part includes the part of the product that is processed in the processing stage, and the processing part is matched with the corresponding identification code of the product.

[0069] Obtain the processing attributes of products in the production line, including the product category and matching the processing attributes with the corresponding identification code of the product.

[0070] The processing shape of the product in the production line is obtained. The processing shape is the product form after real-time processing. The processing shape is matched with the identification code corresponding to the product.

[0071] Track down at least one product with the same processing attributes, set up a distributed database node on the blockchain, and use distributed storage technology to store the identification code, processing part, processing attributes, and processing shape of at least one product with the same processing attributes;

[0072] The blockchain sets up access verification on each distributed database node. Visitors who pass the access verification can obtain the processing part, processing attributes and processing shape of the product in the distributed database node.

[0073] Product quality anomaly detection is carried out on the production line to obtain products with quality anomalies and to verify products with the same processing attributes on the production line.

[0074] The verification results will determine the processing steps in the production line where the fault exists.

[0075] Reference Figure 2 As shown, breaking down the production line into at least one processing stage includes the following steps:

[0076] Set up an input layer, a hidden layer, and an output layer to obtain a standardized production line sample set;

[0077] Input neurons are created in the input layer, hidden neurons are created in the hidden layer, and output neurons are created in the output layer.

[0078] Data from the standardized production line sample set is transmitted to the input layer to obtain the output results from the output layer;

[0079] Determine whether the output of the output layer is reasonable. If it is, do not process it. If not, obtain the flow path of the standardized production line sample set, adjust the neurons in the path, and retransmit the data of the standardized production line sample set to the input layer until the data output of the standardized production line sample set is reasonable.

[0080] By combining the input neurons of the input layer, the hidden neurons of the hidden layer, and the output neurons of the output layer, we obtain the production line decomposition model.

[0081] By inputting the production line into the production line decomposition model, at least one processing step can be obtained.

[0082] The purpose of breaking down a production line into at least one processing stage is to associate each processing stage with a specific product. Thus, each processing stage uniquely corresponds to a single processing part of the product. Therefore, when it's necessary to determine if a machine in a processing stage is faulty, the corresponding processing part of the product can be inspected. If it's fault-free, the corresponding processing part of the product matches the normal processing shape. If there's a discrepancy, a fault exists. Subdividing the processing stage reduces the number of corresponding machines, while coarser subdivisions increase the number. This allows the number of machines in each processing stage to be reduced according to demand, enabling rapid identification of the malfunctioning machine when a processing stage is found to be faulty.

[0083] Reference Figure 3 As shown, classifying products in the production line to obtain at least one product category includes the following steps:

[0084] Perform contour recognition on the products in the production line to obtain at least one product contour;

[0085] At least one product profile is compared with each other, and product profiles with differences less than a preset difference are grouped into the same product category to obtain at least one product category.

[0086] Products within the same product category are identical, but products processed on the production line can differ. Therefore, classification is necessary to facilitate management during inspection.

[0087] Reference Figure 4 As shown, tracking down at least one product with the same processing attributes includes the following steps:

[0088] In the production line, it is determined whether the product categories of two products are consistent. If not, no action is taken.

[0089] If so, then the two products have the same processing attributes.

[0090] The purpose of tracking is to inspect products with the same processing attributes to avoid errors in the inspection results of individual products, which could lead to misjudgments. When problems are found in the inspection of products with the same processing attributes, it indicates that there is indeed a problem in the corresponding processing stage, thereby improving the accuracy of the inspection.

[0091] Reference Figure 5 As shown, storing the identification code, processing location, processing attribute, and processing shape of at least one product with the same processing attributes using distributed storage technology includes the following steps:

[0092] The identification code, processing part, processing attribute and processing shape of at least one product with the same processing attributes are divided and segmented, and the data is evenly stored on multiple distributed database nodes.

[0093] Configure data replication and redundancy backup strategies in the distributed database, and use master-slave replication or multi-master replication to replicate data to multiple distributed database nodes;

[0094] Distributed transaction processing technology is used to achieve data consistency and synchronization in a distributed database;

[0095] A data sharding routing method is adopted to implement load balancing and performance optimization strategies in a distributed database;

[0096] In a distributed database, disaster recovery and fault restoration mechanisms, including fault detection and automatic failover, should be configured.

[0097] The purpose of distributed storage is to improve the fault tolerance of data storage, so that data can be reconstructed in a timely manner after data corruption, and fault detection can continue without having to reacquire the data used for detection.

[0098] Reference Figure 6 As shown, the blockchain sets up access verification on each distributed database node, including the following steps:

[0099] Set up a verification question and a verification answer on each distributed database node;

[0100] Send a verification question to the visitor and obtain the visitor's actual answer;

[0101] Determine whether the actual answer matches the verification answer. If yes, the verification passes; otherwise, the verification fails.

[0102] The purpose of access verification is to ensure the security of production line data, prevent external personnel from obtaining production line data, and thus avoid data leakage.

[0103] Reference Figure 7 As shown, the process of detecting product quality defects on the production line and identifying products with defects includes the following steps:

[0104] Obtain the product's processed shape, identify target feature products in the same product category as the product, and pair them with the product;

[0105] Compare the processed parts of the product with the target feature product that it is paired with;

[0106] Obtain the first surface contour of the processed part of the product, and obtain the second surface contour of the processed part of the target feature product paired with the product.

[0107] When the first surface profile is inconsistent with the second surface profile, the product quality is abnormal.

[0108] Record any abnormal processing parts as abnormal processing parts.

[0109] Production line anomalies inevitably lead to product processing anomalies. Therefore, product inspection is necessary. Since each processing part of the product corresponds to a processing stage in the production line, if the processing part of the product differs from the processing part of the target product, it indicates an anomaly in the processing of that part. Consequently, the corresponding processing stage in the production line may also be anomaly. Because an anomaly in a single product can be caused by other reasons, it is necessary to verify products with the same processing attributes in the production line. When many identical products have the same problem, it indicates that the corresponding processing stage in the production line is definitely anomaly. Therefore, investigating that processing stage can quickly identify the fault, as the processing stage is very small and contains very few or only one machine.

[0110] Reference Figure 8 As shown, verifying products with the same processing attributes on the production line includes the following steps:

[0111] Obtain the product category to which the product with quality issues belongs, and use it as the category of abnormal products.

[0112] The abnormal processing parts of the products in the abnormal product category are compared with the abnormal processing parts of the target feature products that are paired with the products, and the products in the abnormal product category with inconsistencies are counted.

[0113] If the number of products in an inconsistent abnormal product category exceeds the preset number, then the production line is faulty.

[0114] If the number of products in an inconsistent abnormal product category is not greater than the preset number, then there is no fault in the production line.

[0115] The steps to identify faulty processing stages in the production line based on the verification results are as follows:

[0116] When a production line malfunctions, the processing step corresponding to the abnormal processing part is obtained and identified as the malfunctioning processing step.

[0117] Furthermore, this solution also proposes a storage medium on which a computer-readable program is stored. When the computer-readable program is invoked, it executes the aforementioned automated fault detection method for intelligent manufacturing production lines.

[0118] It is understandable that the storage medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid-state drive (SSD).

[0119] In summary, the advantages of this invention are as follows: By collecting data on processing location, processing attributes, and processing shape, products with quality abnormalities are identified. Products with the same processing attributes in the production line are verified, and the faulty processing stage in the production line is determined based on the verification results. This allows for direct product detection using algorithms, and the fault in the production line can be derived from the product detection. Furthermore, all detection is performed after the product has been processed, eliminating the need for extensive adjustments to the production line and reducing initial investment. By decomposing the production line into at least one processing stage and determining the processing location of each stage, the faulty processing stage can be identified based on the detection results, thus pinpointing the location of the fault in the corresponding processing stage in the production line. Because the fault range is significantly narrowed, the cause of the fault can be quickly determined.

[0120] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. An automated fault detection method for intelligent manufacturing production lines, characterized in that, include: The process involves acquiring products that have entered the production line, creating identification codes, matching the identification codes with the products, and printing the identification codes onto the products. The identification codes can be either serial numbers or QR codes. The production line is broken down into at least one processing stage, and the products in the production line are classified to obtain at least one product category, in which the products are the same type of product; In each product category, one product is selected as the feature product, and the target feature product is obtained after the feature product is processed on a normally operating production line. In the processing stage of the production line, the processing part is collected, and the processing stage is matched with the processing part. The processing part includes the part of the product that is processed in the processing stage, and the processing part is matched with the corresponding identification code of the product. Obtain the processing attributes of products in the production line, including the product category and matching the processing attributes with the corresponding identification code of the product. The processing shape of the product in the production line is obtained. The processing shape is the product form after real-time processing. The processing shape is matched with the identification code corresponding to the product. Track down at least one product with the same processing attributes, set up a distributed database node on the blockchain, and use distributed storage technology to store the identification code, processing part, processing attributes, and processing shape of at least one product with the same processing attributes; The blockchain sets up access verification on each distributed database node. Visitors who pass the access verification can obtain the processing part, processing attributes and processing shape of the product in the distributed database node. Product quality anomaly detection is carried out on the production line to obtain products with quality anomalies and to verify products with the same processing attributes on the production line. The verification results will determine the processing steps in the production line where the fault exists.

2. The automated fault detection method for intelligent manufacturing production lines according to claim 1, characterized in that, The process of breaking down the production line into at least one processing stage includes the following steps: Set up an input layer, a hidden layer, and an output layer to obtain a standardized production line sample set; Input neurons are created in the input layer, hidden neurons are created in the hidden layer, and output neurons are created in the output layer. Data from the standardized production line sample set is transmitted to the input layer to obtain the output results from the output layer; Determine whether the output of the output layer is reasonable. If it is, do not process it. If not, obtain the flow path of the standardized production line sample set, adjust the neurons in the path, and retransmit the data of the standardized production line sample set to the input layer until the data output of the standardized production line sample set is reasonable. By combining the input neurons of the input layer, the hidden neurons of the hidden layer, and the output neurons of the output layer, we obtain the production line decomposition model. By inputting the production line into the production line decomposition model, at least one processing step can be obtained.

3. The automated fault detection method for intelligent manufacturing production lines according to claim 2, characterized in that, The process of classifying products in the production line to obtain at least one product category includes the following steps: Perform contour recognition on the products in the production line to obtain at least one product contour; At least one product profile is compared with each other, and product profiles with differences less than a preset difference are grouped into the same product category to obtain at least one product category.

4. The automated fault detection method for intelligent manufacturing production lines according to claim 3, characterized in that, The process of tracking to obtain at least one product with the same processing attributes includes the following steps: In the production line, it is determined whether the product categories of two products are consistent. If not, no action is taken. If so, then the two products have the same processing attributes.

5. The automated fault detection method for intelligent manufacturing production lines according to claim 4, characterized in that, The process of storing the identification code, processing location, processing attribute, and processing shape of at least one product with the same processing attributes using distributed storage technology includes the following steps: The identification code, processing part, processing attribute and processing shape of at least one product with the same processing attributes are divided and segmented, and the data is evenly stored on multiple distributed database nodes. Configure data replication and redundancy backup strategies in the distributed database, and use master-slave replication or multi-master replication to replicate data to multiple distributed database nodes; Distributed transaction processing technology is used to achieve data consistency and synchronization in a distributed database; A data sharding routing method is adopted to implement load balancing and performance optimization strategies in a distributed database; In a distributed database, disaster recovery and fault restoration mechanisms, including fault detection and automatic failover, should be configured.

6. The automated fault detection method for intelligent manufacturing production lines according to claim 5, characterized in that, The blockchain sets up access verification on each distributed database node, including the following steps: Set up a verification question and a verification answer on each distributed database node; Send a verification question to the visitor and obtain the visitor's actual answer; Determine whether the actual answer matches the verification answer. If yes, the verification passes; otherwise, the verification fails.

7. The automated fault detection method for intelligent manufacturing production lines according to claim 6, characterized in that, The process of detecting product quality anomalies on the production line and identifying products with quality anomalies includes the following steps: Obtain the product's processed shape, identify target feature products in the same product category as the product, and pair them with the product; Compare the processed parts of the product with the target feature product that it is paired with; Obtain the first surface contour of the processed part of the product, and obtain the second surface contour of the processed part of the target feature product paired with the product. When the first surface profile is inconsistent with the second surface profile, the product quality is abnormal. Record any abnormal processing parts as abnormal processing parts.

8. The automated fault detection method for intelligent manufacturing production lines according to claim 7, characterized in that, The verification of products with the same processing attributes in the production line includes the following steps: Obtain the product category to which the product with quality issues belongs, and use it as the category of abnormal products. The abnormal processing parts of the products in the abnormal product category are compared with the abnormal processing parts of the target feature products that are paired with the products, and the products in the abnormal product category with inconsistencies are counted. If the number of products in an inconsistent abnormal product category exceeds the preset number, then the production line is faulty. If the number of products in an inconsistent abnormal product category is not greater than the preset number, then there is no fault in the production line.

9. The automated fault detection method for an intelligent manufacturing production line according to claim 8, characterized in that, The process of identifying faulty processing steps in the production line based on the verification results includes the following steps: When a production line malfunctions, the processing step corresponding to the abnormal processing part is obtained and identified as the malfunctioning processing step.

Images