Artificial Intelligence-Based Intelligent Manufacturing Quality Inspection Methods and Systems

Abstract

This invention relates to the field of intelligent manufacturing technology and discloses an intelligent manufacturing quality inspection method and system based on artificial intelligence, including: S1, performing quality inspection on continuously produced products; S2, generating a single-piece quality state to characterize the quality features of the individual product; S3, organizing multiple single-piece quality states according to manufacturing sequence information or time information; S4, constraining the continuity of the single-piece quality states in the time dimension; S5, based on a multi-level quality state expression structure; and S6, outputting the process quality state. By constructing a multi-level quality state expression structure containing a single-piece quality state layer and a process quality state layer, and introducing a process evaluation constraint mechanism based on time continuity in the process quality state construction and judgment process, the manufacturing process quality inspection is expanded from discrete single-piece judgment to overall quality state judgment of a continuous production process, thereby improving the ability to characterize quality changes in the manufacturing process.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, specifically to an intelligent manufacturing quality inspection method and system based on artificial intelligence. Background Technology

[0002] Intelligent manufacturing refers to a manufacturing model that uses information technology, automation, and intelligent technology to monitor, analyze, and manage the product production process. Its core objective is to improve the stability of the manufacturing process and the level of product quality. Quality inspection, as an important component of intelligent manufacturing, is usually used to detect and evaluate the quality status of products formed during the production process. With the development of artificial intelligence technology, the introduction of artificial intelligence into intelligent manufacturing quality inspection can achieve automatic identification and judgment of product quality status through the analysis of inspection data. Currently, AI-based intelligent manufacturing quality inspection usually takes a single product as the inspection object, and determines whether the product meets the quality requirements by analyzing the quality inspection data of each product.

[0003] However, in current technology, the quality status of a product is assessed by independently judging the quality of a single product, while the production process is continuous and the quality status is continuous over time. Relying solely on the quality judgment of a single product is insufficient to reflect the overall quality status of the manufacturing process. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent manufacturing quality inspection method and system based on artificial intelligence, which solves the problem that existing technologies, which rely solely on the quality judgment of a single product, cannot reflect the overall quality status during continuous manufacturing.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent manufacturing quality inspection method based on artificial intelligence, comprising:

[0006] S1. During the manufacturing process, quality inspection is carried out on continuously produced products to obtain quality inspection data related to the quality of the products, and the corresponding manufacturing sequence information or time information is associated with the quality inspection data.

[0007] S2. Based on the quality inspection data, use an artificial intelligence model to perform quality analysis on a single product and generate a single-piece quality status to characterize the quality features of the single product.

[0008] S3. According to the manufacturing sequence information or time information, organize the quality states of multiple individual pieces and construct a multi-level quality state expression structure that includes an individual piece quality state layer and a process quality state layer. The process quality state layer is used to characterize the quality states of multiple products in a continuous manufacturing process.

[0009] S4. In the process of constructing the process quality status layer, a process evaluation constraint mechanism based on time continuity is introduced to constrain the continuity of the individual quality status in the time dimension.

[0010] S5. Based on the multi-level quality status expression structure and combined with the time-continuous process evaluation constraint mechanism, use an artificial intelligence model to determine the process quality status of the manufacturing process.

[0011] S6. Output the process quality status, wherein the process quality status corresponds to the process quality status layer in the multi-level quality status expression structure.

[0012] Preferably, the step of performing quality inspection on continuously produced products and obtaining quality inspection data includes:

[0013] The product is tested at the manufacturing line or workstation to obtain test data used to characterize the product quality;

[0014] The test data is associated with the manufacturing sequence information or time information of the corresponding product to form quality test data corresponding to the manufacturing process.

[0015] Preferably, the step of using an artificial intelligence model to perform quality analysis on a single product and generate a single-piece quality status includes:

[0016] The quality inspection data is input into an artificial intelligence model to analyze and process the quality characteristics of a single product and obtain the corresponding quality analysis results.

[0017] The artificial intelligence model includes a model structure for performing feature representation and quality discrimination on the quality inspection data;

[0018] Based on the quality analysis results, a single-piece quality status corresponding to the individual product is generated.

[0019] Preferably, the construction of the multi-level quality state expression structure includes:

[0020] According to the manufacturing sequence information or time information, the quality status of multiple individual parts is sorted and organized to form a sequence of individual part quality statuses;

[0021] Based on the single-piece quality state sequence, a process quality state layer is constructed to characterize the quality states of multiple products in a continuous process, forming a multi-level quality state expression structure that includes the single-piece quality state layer and the process quality state layer.

[0022] Preferably, the construction of the process quality state layer based on the single-piece quality state sequence includes:

[0023] Within a predetermined number of individual quality states or a predetermined time range, the sequence of individual quality states is grouped.

[0024] Based on the individual quality status within each group, corresponding process quality statuses are generated to form the process quality status layer.

[0025] Preferably, the generation of the process quality status includes:

[0026] The quality status of individual items within each group is uniformly represented to form a group quality status representation for the corresponding group.

[0027] Based on the group quality status representation, the process quality status of the corresponding group is determined, and the process quality status of each group is used to form the process quality status layer.

[0028] Preferably, the process evaluation constraint mechanism based on time continuity includes:

[0029] Based on the manufacturing sequence information or time information, sort the quality status of the individual piece or the process quality status by time.

[0030] Within a predetermined time window or a predetermined number of consecutive individual quality states, the individual quality state or process quality state is jointly processed.

[0031] Based on the joint processing results, constraints are applied to the updating of the process quality status.

[0032] Preferably, the updating of the process quality status is constrained, including:

[0033] Based on the predetermined time window or a predetermined number of consecutive unit quality states, determine whether the unit quality state or process quality state meets the continuity condition.

[0034] Under the condition of continuity, the process quality status may be updated;

[0035] If the continuity condition is not met, the process quality status remains unchanged.

[0036] Preferably, the output process quality status includes:

[0037] The process quality status is associated with the corresponding manufacturing sequence information or time information to form process quality status data;

[0038] The process quality status data is output in a data format that maintains the correspondence between the process quality status and the process quality status layer in the multi-level quality status expression structure.

[0039] An AI-based intelligent manufacturing quality inspection system, the system comprising:

[0040] The quality inspection module performs quality inspection on continuously produced products during the manufacturing process, acquires quality inspection data related to product quality, and associates the quality inspection data with manufacturing sequence information or time information.

[0041] The single-piece quality status analysis module inputs the quality inspection data into the artificial intelligence model, analyzes and processes the quality characteristics of a single product, and generates the single-piece quality status corresponding to that single product.

[0042] The multi-level quality status modeling module organizes multiple individual quality statuses according to the manufacturing sequence information or time information, and constructs a multi-level quality status expression structure that includes an individual quality status layer and a process quality status layer.

[0043] The process quality status construction module groups the individual quality statuses based on the individual quality status sequence, and generates corresponding process quality statuses based on the individual quality statuses within each group, thereby forming the process quality status layer.

[0044] The time continuity constraint evaluation module constrains the updating of the process quality status within a predetermined time window or a predetermined number of continuous states.

[0045] The process quality status output module associates the process quality status with the corresponding manufacturing sequence information or time information, and outputs it in data form that maintains its correspondence with the process quality status layer in the multi-level quality status expression structure.

[0046] This invention provides an intelligent manufacturing quality inspection method and system based on artificial intelligence. It has the following beneficial effects:

[0047] 1. This invention constructs a multi-level quality state expression structure that includes a single-piece quality state layer and a process quality state layer, and introduces a process evaluation constraint mechanism based on time continuity in the process quality state construction and judgment process. This expands the manufacturing process quality detection from discrete single-piece judgment to the overall quality state judgment of a continuous production process, thereby improving the ability to characterize the quality changes in the manufacturing process.

[0048] 2. This invention utilizes artificial intelligence models to analyze and process the quality inspection data of individual products, generating a unified form of single-piece quality status. Based on this, multiple single-piece quality statuses are organized and modeled, realizing a structured expression of quality status at different stages of the manufacturing process, which facilitates centralized management and analysis of quality information of continuously produced products.

[0049] 3. This invention links process quality status with manufacturing sequence information or time information and outputs it in a data form that maintains hierarchical correspondence, so that process quality status can directly reflect the stage quality situation in the manufacturing process, which is convenient for integration with subsequent quality analysis, process monitoring or system interface applications. Attached Figure Description

[0050] Figure 1 This is a flowchart of the AI-based intelligent manufacturing quality inspection method of the present invention;

[0051] Figure 2 This is an architecture diagram of the AI-based intelligent manufacturing quality inspection system of the present invention. Detailed Implementation

[0052] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Please see the appendix Figure 1 This invention provides an intelligent manufacturing quality inspection method based on artificial intelligence, comprising:

[0054] S1. During the manufacturing process, quality inspection is carried out on continuously produced products to obtain quality inspection data related to product quality, and the corresponding manufacturing sequence information or time information is associated with the quality inspection data.

[0055] Furthermore, quality inspection is conducted on continuously produced products to obtain quality inspection data, including:

[0056] At the manufacturing line or workstation, products are tested to obtain test data used to characterize product quality;

[0057] The test data is associated with the manufacturing sequence or time information of the corresponding product to form quality test data corresponding to the manufacturing process.

[0058] Specifically, when products pass through the manufacturing line or workstation, quality inspection devices installed at the line or workstation are used to inspect the products and obtain inspection data to characterize the product quality. The quality inspection devices can be selected according to the product type and inspection requirements. For example, industrial cameras can be used to acquire product appearance images, dimensional measurement sensors can be used to acquire product size data, or process parameter data reflecting the manufacturing process status can be collected. Quality inspection data can be acquired for each continuously produced product without affecting the normal operation of the production line, so that the collected data truly reflects the quality status of the product during the manufacturing process. For example, on an assembly line, when products pass through the inspection station in sequence, industrial cameras will take pictures of each product and collect corresponding inspection data, thereby forming a quality inspection record corresponding to each product.

[0059] After obtaining the aforementioned quality inspection data, the inspection data is associated with the manufacturing sequence information or time information of the corresponding product. The manufacturing sequence information can be obtained through the production line serial number, workstation sequence number, or product serial number in the production line. The time information can be obtained by attaching a timestamp to the inspection data. That is, when a product completes quality inspection, the sequential position or inspection time of the product in the manufacturing process is recorded simultaneously, and the manufacturing sequence information or time information is bound with the corresponding quality inspection data, thereby forming quality inspection data corresponding to the manufacturing process sequence. This ensures that each piece of quality inspection data not only contains the product's quality characteristics but also its sequential relationship in the manufacturing process. For example, on an electronic product assembly line, the inspection data of each product carries a corresponding production line serial number or inspection time, thus providing basic data support for subsequent organization and analysis of product quality status according to the manufacturing sequence.

[0060] S2. Based on quality inspection data, use artificial intelligence models to perform quality analysis and processing on individual products, and generate a single-piece quality status to characterize the quality features of the individual product.

[0061] Furthermore, artificial intelligence models are used to perform quality analysis on individual products, generating individual product quality status, including:

[0062] The quality inspection data is input into the artificial intelligence model to analyze and process the quality characteristics of individual products and obtain the corresponding quality analysis results.

[0063] Among them, the artificial intelligence model includes a model structure for feature representation and quality judgment of quality inspection data;

[0064] Based on the quality analysis results, the individual quality status corresponding to the single product is generated.

[0065] Specifically, in this embodiment, after obtaining the quality inspection data corresponding to a single product, the quality inspection data is input into an artificial intelligence model to analyze and process the quality characteristics of the single product to obtain the corresponding quality analysis results. The artificial intelligence model includes a model structure for feature representation and quality discrimination of the quality inspection data. The feature representation structure maps the original quality inspection data into intermediate feature representations that can characterize the product's quality characteristics, and the quality discrimination structure outputs the quality analysis results based on the intermediate feature representations. In a specific implementation, the first... The quality inspection data for each product is recorded as follows: The artificial intelligence model performs feature representation processing on the quality inspection data to obtain quality feature representation. The process can be represented as follows:

[0066] ;

[0067] in, This represents the feature representation mapping structure in an artificial intelligence model. For characterizing the first The feature representation of each product's quality characteristics is then used by the artificial intelligence model based on these feature representations. Perform quality discrimination processing to obtain the corresponding quality analysis results. ,For example:

[0068] ;

[0069] in, Indicates the quality discrimination structure, To reflect the analysis results of the product's quality status. Using the above method, while maintaining the abstract nature of the model structure, automatic analysis and processing of individual product quality features can be achieved. For example, in an electronic product testing scenario, the product's appearance image and size inspection data are input into the artificial intelligence model as quality inspection data. The model first extracts feature representations reflecting appearance and size characteristics, and then outputs the corresponding quality analysis results.

[0070] After obtaining the above quality analysis results, a single-piece quality status corresponding to the individual product is generated based on the quality analysis results. This status characterizes the product's quality condition at the current inspection moment. The single-piece quality status can be represented in numerical, vector, or categorical form based on the quality analysis results; for example, the quality analysis results can be mapped to single-piece quality status. Its generation process can be represented as:

[0071] ;

[0072] in, This represents the mapping relationship used to convert quality analysis results into the quality status of individual components. For the first The individual quality status of each product is represented in a unified form by converting the quality analysis results into individual quality status. This facilitates the organization and processing of multiple individual quality statuses according to the manufacturing sequence. For example, in the above-mentioned electronic product assembly line, each product generates a corresponding individual quality status after being analyzed by the artificial intelligence model, which is used to construct the quality status expression structure of the manufacturing process.

[0073] S3. Organize the quality states of multiple individual pieces according to manufacturing sequence information or time information, and construct a multi-level quality state expression structure that includes an individual piece quality state layer and a process quality state layer. The process quality state layer is used to characterize the quality states of multiple products in a continuous manufacturing process.

[0074] Furthermore, a multi-level quality state representation structure is constructed, including:

[0075] Based on manufacturing sequence information or time information, multiple individual piece quality states are sorted and organized to form a sequence of individual piece quality states;

[0076] Based on the single-piece quality state sequence, a process quality state layer is constructed to characterize the quality states of multiple products in a continuous process, forming a multi-level quality state expression structure that includes the single-piece quality state layer and the process quality state layer.

[0077] Specifically, after obtaining multiple individual quality states, the individual quality states are first sorted and organized according to the corresponding manufacturing sequence information or time information to form an individual quality state sequence. Specifically, for multiple products produced consecutively, their individual quality states are denoted as follows: The subscripts are numbered according to the manufacturing sequence or time sequence. By sorting the quality status of each individual piece according to the manufacturing sequence information or time information, a sequence of individual piece quality statuses is formed.

[0078] ;

[0079] This single-piece quality state sequence is used to reflect the changes in product quality state with manufacturing sequence or time during the manufacturing process. By organizing discrete single-piece quality states according to the manufacturing sequence, subsequent quality analysis can no longer be limited to isolated products, but can be carried out based on the sequential relationship of products in the manufacturing process. For example, on an assembly line, products pass through the inspection station in sequence according to a fixed rhythm. After each product completes the quality analysis, a corresponding single-piece quality state is generated. The system arranges these single-piece quality states in sequence according to the product's production line number to form a single-piece quality state sequence consistent with the production line sequence.

[0080] After forming the above-mentioned single-piece quality state sequence, a process quality state layer is further constructed based on the single-piece quality state sequence to characterize the quality states of multiple consecutive products in the manufacturing process. Several consecutive single-piece quality states in the single-piece quality state sequence can be modeled as a whole to generate the corresponding process quality state. Let the first single-piece quality state in the single-piece quality state sequence be... The set of individual component quality states contained within a continuous interval is ,in This represents the set of indices corresponding to the continuous interval. The set of individual component quality states can then be processed using a state mapping function to generate the corresponding process quality states. Its representation can be as follows:

[0081] ;

[0082] in, This represents a state modeling function used to map multiple individual component quality states to process quality states. To characterize the overall quality status of multiple products within this continuous interval, a multi-level quality status expression structure including a single-piece quality status layer and a process quality status layer is formed in the above manner, so that the quality information of the manufacturing process can be described at both the single-piece level and the process level.

[0083] Furthermore, a process quality state layer is constructed based on the individual component quality state sequence, including:

[0084] Within a predetermined number of individual quality states or within a predetermined time range, the sequence of individual quality states is grouped.

[0085] Based on the individual quality status within each group, corresponding process quality statuses are generated to form a process quality status layer.

[0086] Specifically, based on the aforementioned sequence of individual quality states, the sequence of individual quality states is grouped within a predetermined number of individual quality states or a predetermined time range. Specifically, the number of individual quality states included in each group or the corresponding time range can be pre-set according to the cycle characteristics of the manufacturing process or the quality assessment requirements. Without changing the original order of the individual quality states, the continuous individual quality states are divided into multiple groups of individual quality states. This allows the individual quality states corresponding to adjacent products in the continuous production process to be organized into several groups with clear boundaries, providing a basis for the subsequent generation of process quality states that reflect the quality status of different production intervals. For example, on a certain assembly line, the corresponding sequence of individual quality states can be grouped according to a group of several consecutive products, so that each group of individual quality states corresponds to a specific production interval.

[0087] After completing the above grouping process, based on the individual quality status within each group, and following the generation method of process quality status in the aforementioned multi-level quality status expression structure, a corresponding process quality status is generated for each group. Specifically, multiple individual quality statuses within the same group are taken as a whole input, and the overall quality status of the products within the group is modeled using the aforementioned state mapping relationship. This generates a process quality status that characterizes the overall quality level of the group. By generating process quality statuses for different groups and combining them to form a process quality status layer, the quality status of the manufacturing process can be segmented according to the production interval, facilitating further analysis and processing of the process quality status based on the manufacturing sequence or time dimension.

[0088] Furthermore, process quality status generation includes:

[0089] The quality status of individual items within each group is uniformly represented to form a group quality status representation for the corresponding group.

[0090] Based on the grouped quality status representation, the process quality status of the corresponding group is determined, and the process quality status of each group is used to form a process quality status layer.

[0091] Specifically, after grouping the individual quality status sequence, the individual quality status within each group is uniformly represented to form a group quality status representation for the corresponding group. That is, multiple individual quality statuses within the same group are first organized according to the aforementioned definition of individual quality status to ensure consistency in structure and form, thereby forming a group quality status representation that can reflect the overall product quality status within the group. This avoids the impact of differences in the representation of individual quality status within the group on the generation of subsequent process quality status.

[0092] After forming the above-mentioned group quality status representation, the process quality status of the corresponding group is determined based on the group quality status representation. Specifically, the group quality status representation is used as the overall input, and the process quality status is generated according to the aforementioned process quality status generation method to represent the overall quality status of multiple products in the group. The process quality status corresponding to each group is used to form a process quality status layer, so that the quality status of the manufacturing process can be described in the form of multiple continuous intervals.

[0093] S4. In the process of constructing the process quality status layer, a process evaluation constraint mechanism based on time continuity is introduced to constrain the continuity of the quality status of a single item in the time dimension.

[0094] Furthermore, process evaluation constraint mechanisms based on time continuity include:

[0095] Based on manufacturing sequence information or time information, sort the quality status of individual pieces or process quality status by time.

[0096] Within a predetermined time window or a predetermined number of consecutive individual quality statuses, the individual quality status or process quality status is jointly processed.

[0097] Based on the joint processing results, constraints are applied to the updating of the process quality status.

[0098] Specifically, after sorting by completion time, the individual quality status or process quality status is jointly processed within a predetermined time window or a predetermined number of consecutive quality statuses. In practice, the current process quality status is used as a reference, and its corresponding set of consecutive quality states is selected for processing. This set of consecutive quality states can be represented as:

[0099] ;

[0100] in, This indicates the process quality status at the current moment. Indicates the length of the time window or the number of consecutive states. For a set of continuous quality states participating in joint processing, the evaluation of process quality states is subject to time continuity constraints by jointly processing multiple continuous quality states within a predetermined time range.

[0101] After obtaining the joint processing result of the above continuous quality state set, the update of the process quality state is constrained based on the joint processing result. That is, according to the time continuity characteristics reflected by the continuous quality state set, the current process quality state is controlled to participate in the subsequent update processing, so that the change of process quality state is consistent with the time continuity relationship in the manufacturing process.

[0102] Furthermore, the updating of process quality status is constrained, including:

[0103] Based on the quality status of a predetermined number of consecutive units within a predetermined time window, determine whether the quality status of a unit or the quality status of the process meets the continuity condition.

[0104] The process quality status may be updated if the continuity condition is met.

[0105] If the continuity condition is not met, the process quality status is not updated.

[0106] Specifically, after completing the joint processing based on time continuity, it is determined whether the quality status of a single piece or the quality status of a process meets the continuity condition. First, based on the set of continuous quality statuses selected within the aforementioned time window or continuous quantity range, it is determined whether the set meets the preset continuity requirements in the time dimension, such as whether it contains a sufficient number of continuous quality statuses or whether it covers the predetermined continuous time range. By determining the continuity condition, it is distinguished whether the current process quality status is in a stable continuous manufacturing interval, thereby providing a basis for whether to update the process quality status in the future.

[0107] After completing the continuity condition determination, the update of the process quality status is controlled according to the determination result. When the determination result shows that the current quality status meets the continuity condition, the process quality status is allowed to be updated. When the determination result shows that the current quality status does not meet the continuity condition, the process quality status is kept unchanged, ensuring that the change of the process quality status corresponds to the time continuity characteristics of the manufacturing process, thereby ensuring the consistency of the process quality status in the manufacturing sequence or time dimension.

[0108] S5. Based on a multi-level quality status expression structure and combined with a time-continuous process evaluation constraint mechanism, an artificial intelligence model is used to determine the process quality status of the manufacturing process.

[0109] Specifically, after constructing a multi-level quality status expression structure and introducing a time-continuous process evaluation constraint mechanism, an artificial intelligence model is used to determine the process quality status of the manufacturing process. The artificial intelligence model can analyze and determine the process quality status that reflects the overall quality status of the manufacturing process. The model type can be a judgment model based on machine learning or deep learning, such as a classification model, regression model, or state discrimination model, to output the judgment results that characterize the quality status of the manufacturing process.

[0110] First, the process quality states formed in the multi-level quality state expression structure are used as model inputs. Under the constraint of time continuity, only process quality states that meet the continuity requirements are selected for judgment. The artificial intelligence model uses the same model structure as the aforementioned artificial intelligence model for single-piece quality analysis, or it can be a model structure built separately for process quality states. Its input dimensions and judgment targets are set according to the expression form of the process quality states. Based on the above inputs, the artificial intelligence model judges the process quality states. The judgment process can be abstractly represented as follows:

[0111]

[0112] in, This indicates the process quality status involved in the judgment under the constraint of time continuity. This represents the mapping relationship of the artificial intelligence model used to determine the process quality status. This provides the process quality assessment result corresponding to the manufacturing process at the current moment. Through the above method, the artificial intelligence model can make a unified judgment on the quality status of the manufacturing process based on the multi-level quality status expression structure and the time continuity constraint mechanism, thereby providing a judgment basis for subsequent process quality status output.

[0113] S6. Output process quality status, wherein the process quality status corresponds to the process quality status layer in the multi-level quality status expression structure.

[0114] Furthermore, the output process quality status includes:

[0115] The process quality status is associated with the corresponding manufacturing sequence information or time information to form process quality status data;

[0116] The process quality status data is output in a data format that maintains the correspondence between the process quality status and the process quality status layers in the multi-level quality status expression structure.

[0117] Specifically, after determining the process quality status of the manufacturing process, the process quality status is output and processed. Specifically, the determined process quality status is associated with the corresponding manufacturing sequence information or time information to form process quality status data that can reflect the quality status of different stages of the manufacturing process. When the process quality status data is output, it maintains its correspondence with the process quality status layer in the multi-level quality status expression structure, so that the output data can directly reflect the quality status corresponding to each continuous interval in the manufacturing process.

[0118] Process quality status data can be organized according to manufacturing sequence or time sequence and output in the form of structured data, such as stored or transmitted in a record manner, so that each piece of process quality status data corresponds to a continuous interval in the manufacturing process. Thus, the process quality status can be directly used in subsequent quality analysis, process monitoring or system integration without the need for additional parsing of the correspondence between quality statuses.

[0119] Please see the appendix Figure 2 An AI-based intelligent manufacturing quality inspection system, comprising:

[0120] The quality inspection module performs quality inspections on continuously produced products during the manufacturing process, acquires quality inspection data related to product quality, and associates manufacturing sequence information or time information with the quality inspection data.

[0121] The single-piece quality status analysis module inputs quality inspection data into an artificial intelligence model to analyze and process the quality characteristics of a single product and generate the single-piece quality status corresponding to that single product.

[0122] The multi-level quality status modeling module organizes multiple individual quality statuses according to manufacturing sequence information or time information, and constructs a multi-level quality status expression structure that includes an individual quality status layer and a process quality status layer.

[0123] The process quality status construction module groups the individual quality statuses based on the individual quality status sequence, and generates corresponding process quality statuses based on the individual quality statuses within each group to form a process quality status layer.

[0124] The time continuity constraint evaluation module constrains the updating of process quality status within a predetermined time window or a predetermined number of continuous states.

[0125] The process quality status output module associates the process quality status with the corresponding manufacturing sequence information or time information, and outputs it in a data format that maintains its correspondence with the process quality status layer in the multi-level quality status expression structure.

[0126] Specifically, the quality inspection module is set up at the manufacturing line or workstation to inspect continuously produced products, acquire quality inspection data related to product quality, and associate the quality inspection data with manufacturing sequence information or time information. The single-piece quality status analysis module inputs the quality inspection data into an artificial intelligence model to analyze and process the quality characteristics of individual products and generate corresponding single-piece quality status. The multi-level quality status modeling module organizes multiple single-piece quality statuses according to manufacturing sequence information or time information, constructing a multi-level quality status expression structure that includes a single-piece quality status layer and a process quality status layer. The process quality status construction module groups single-piece quality statuses based on the single-piece quality status sequence and generates corresponding process quality statuses based on the single-piece quality statuses within each group to form a process quality status layer. The time continuity constraint evaluation module constrains the update of the process quality status within a predetermined time window or a predetermined number of continuous states. The process quality status output module associates the process quality status with the corresponding manufacturing sequence information or time information and outputs it in a data form that maintains its correspondence with the process quality status layer in the multi-level quality status expression structure, thereby realizing the systematic detection and management of the manufacturing process quality status.

[0127] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An intelligent manufacturing quality inspection method based on artificial intelligence, characterized in that, include: S1. During the manufacturing process, quality inspection is carried out on continuously produced products to obtain quality inspection data related to the quality of the products, and the corresponding manufacturing sequence information or time information is associated with the quality inspection data. S2. Based on the quality inspection data, use an artificial intelligence model to perform quality analysis on a single product and generate a single-piece quality status to characterize the quality features of the single product. S3. According to the manufacturing sequence information or time information, organize the quality states of multiple individual pieces and construct a multi-level quality state expression structure that includes an individual piece quality state layer and a process quality state layer. The process quality state layer is used to characterize the quality states of multiple products in a continuous manufacturing process. S4. In the process of constructing the process quality status layer, a process evaluation constraint mechanism based on time continuity is introduced to constrain the continuity of the individual quality status in the time dimension. S5. Based on the multi-level quality status expression structure and combined with the time-continuous process evaluation constraint mechanism, use an artificial intelligence model to determine the process quality status of the manufacturing process. S6. Output the process quality status, wherein the process quality status corresponds to the process quality status layer in the multi-level quality status expression structure.

2. The intelligent manufacturing quality inspection method based on artificial intelligence according to claim 1, characterized in that, The process of conducting quality inspections on continuously produced products and obtaining quality inspection data includes: The product is tested at the manufacturing line or workstation to obtain test data used to characterize the product quality; The test data is associated with the manufacturing sequence information or time information of the corresponding product to form quality test data corresponding to the manufacturing process.

3. The intelligent manufacturing quality inspection method based on artificial intelligence according to claim 1, characterized in that, The process of using artificial intelligence models to perform quality analysis on individual products and generate a single-item quality status includes: The quality inspection data is input into an artificial intelligence model to analyze and process the quality characteristics of a single product and obtain the corresponding quality analysis results. The artificial intelligence model includes a model structure for performing feature representation and quality discrimination on the quality inspection data; Based on the quality analysis results, a single-piece quality status corresponding to the individual product is generated.

4. The intelligent manufacturing quality inspection method based on artificial intelligence according to claim 1, characterized in that, The construction of the multi-level quality state expression structure includes: According to the manufacturing sequence information or time information, the quality status of multiple individual parts is sorted and organized to form a sequence of individual part quality statuses; Based on the single-piece quality state sequence, a process quality state layer is constructed to characterize the quality states of multiple products in a continuous process, forming a multi-level quality state expression structure that includes the single-piece quality state layer and the process quality state layer.

5. The intelligent manufacturing quality inspection method based on artificial intelligence according to claim 4, characterized in that, The process quality state layer constructed based on the single-piece quality state sequence includes: Within a predetermined number of individual quality states or a predetermined time range, the sequence of individual quality states is grouped. Based on the individual quality status within each group, corresponding process quality statuses are generated to form the process quality status layer.

6. The intelligent manufacturing quality inspection method based on artificial intelligence according to claim 5, characterized in that, The generation of the process quality status includes: The quality status of individual items within each group is uniformly represented to form a group quality status representation for the corresponding group. Based on the group quality status representation, the process quality status of the corresponding group is determined, and the process quality status of each group is used to form the process quality status layer.

7. The intelligent manufacturing quality inspection method based on artificial intelligence according to claim 1, characterized in that, The time-continuity-based process evaluation constraint mechanism includes: Based on the manufacturing sequence information or time information, sort the quality status of the individual piece or the process quality status by time. Within a predetermined time window or a predetermined number of consecutive individual quality states, the individual quality state or process quality state is jointly processed. Based on the joint processing results, constraints are applied to the updating of the process quality status.

8. The intelligent manufacturing quality inspection method based on artificial intelligence according to claim 7, characterized in that, The updating of the process quality status is constrained, including: Based on the predetermined time window or a predetermined number of consecutive unit quality states, determine whether the unit quality state or process quality state meets the continuity condition. Under the condition of continuity, the process quality status may be updated; If the continuity condition is not met, the process quality status remains unchanged.

9. The intelligent manufacturing quality inspection method based on artificial intelligence according to claim 1, characterized in that, The output process quality status includes: The process quality status is associated with the corresponding manufacturing sequence information or time information to form process quality status data; The process quality status data is output in a data format that maintains the correspondence between the process quality status and the process quality status layer in the multi-level quality status expression structure.

10. An intelligent manufacturing quality inspection system based on artificial intelligence, characterized in that, The system for the AI-based intelligent manufacturing quality inspection method according to any one of claims 1-9 comprises: The quality inspection module performs quality inspection on continuously produced products during the manufacturing process, acquires quality inspection data related to product quality, and associates the quality inspection data with manufacturing sequence information or time information. The single-piece quality status analysis module inputs the quality inspection data into the artificial intelligence model, analyzes and processes the quality characteristics of a single product, and generates the single-piece quality status corresponding to that single product. The multi-level quality status modeling module organizes multiple individual quality statuses according to the manufacturing sequence information or time information, and constructs a multi-level quality status expression structure that includes an individual quality status layer and a process quality status layer. The process quality status construction module groups the individual quality statuses based on the individual quality status sequence, and generates corresponding process quality statuses based on the individual quality statuses within each group, thereby forming the process quality status layer. The time continuity constraint evaluation module constrains the updating of the process quality status within a predetermined time window or a predetermined number of continuous states. The process quality status output module associates the process quality status with the corresponding manufacturing sequence information or time information, and outputs it in data form that maintains its correspondence with the process quality status layer in the multi-level quality status expression structure.

Images