Intelligent manufacturing quality monitoring method and system based on digital twinning

Abstract

The application provides an intelligent manufacturing quality monitoring method and system based on digital twinning, which comprises: constructing traceable manufacturing data for a manufacturing entity; creating a virtual mapping body containing parameter correlation based on the data chain; comparing the quality prediction value of the virtual mapping body with the measured value of the corresponding node in real time to identify quality deviation; when there is deviation, locating abnormal process parameters by reverse tracing using the virtual mapping body; based on the contribution of the parameters, matching and executing device adjustment instructions from the process knowledge base. The application realizes accurate and automatic root cause positioning of quality problems by constructing a full-process traceable data chain and a virtual mapping body, solves the problem of low efficiency of traditional methods relying on manual experience and traceability, uses the virtual mapping body for real-time prediction and comparison, realizes early warning and preventive control of quality risks, changes quality control from passive detection to active intervention, and effectively improves process stability and quality consistency.

Description

Technical Field

[0001] This application relates to the field of intelligent manufacturing technology, and in particular to an intelligent manufacturing quality monitoring method and system based on digital twins. Background Technology

[0002] In modern manufacturing, product quality monitoring mainly falls into two categories: one is offline sampling inspection of final products, which is inherently slow and cannot prevent batch defects; the other is deploying online inspection equipment in key processes, which can detect defects in real time, but typically only provides a "pass / fail" judgment, unable to deeply trace or predict the causes of defects, and even less able to proactively adjust the processes that generate defects. Existing technologies include systems that attempt to correlate production and quality data, but these systems often suffer from severe data silos, with data from different stages lacking precise correlation in terms of time sequence and objects, leading to incomplete analysis results and difficulties in traceability. Summary of the Invention

[0003] This application provides a digital twin-based intelligent manufacturing quality monitoring method and system to solve the problem in related technologies that it is impossible to deeply trace and predict the causes of defects generated during product manufacturing.

[0004] The first aspect of this application provides a method for intelligent manufacturing quality monitoring based on digital twins, the method comprising: The manufacturing entity's identification code is identified based on a preset quality monitoring node, and the manufacturing entity's quality inspection data at the corresponding quality monitoring node is collected to obtain the manufacturing entity's manufacturing data. Construct a virtual mapping body corresponding to the manufacturing entity based on the manufacturing data; The first predicted value of the virtual mapping body at the quality monitoring node is compared with the measured value of the manufacturing entity to determine whether there is a quality deviation between the virtual mapping body and the manufacturing entity. When there is a quality deviation between the virtual mapping body and the manufacturing entity, a reverse retrieval is performed starting from the current quality monitoring node to locate the abnormal process parameters. Based on the contribution of the abnormal process parameters to the quality deviation, the corresponding equipment setting adjustment instructions are matched from the preset process knowledge base for dynamic adjustment.

[0005] Optionally, in the first implementation of the first aspect of this application, the step of identifying the identification code of the manufacturing entity based on a preset quality monitoring node, and collecting the quality inspection data of the manufacturing entity at the corresponding quality monitoring node to obtain the manufacturing data of the manufacturing entity includes: The identification codes carried by the manufacturing entities are identified based on all quality monitoring nodes on the same production line, and an association record between the identification codes and node identifiers is generated. During the processing of the manufacturing entity at the quality monitoring node, the equipment process parameters and quality inspection data of the quality monitoring node are acquired to generate single-node data; Based on the flow sequence of the manufacturing entity along the production path, the single-node data generated by the manufacturing entity at each quality monitoring node is arranged and aggregated according to the node time sequence, and the aggregated single-node data is bound by the identification code to generate manufacturing data.

[0006] Optionally, in a second implementation of the first aspect of this application, the step of constructing a virtual mapping body corresponding to the manufacturing entity based on the manufacturing data includes: Separate a set of quality characteristic parameters representing the quality attributes of the manufactured entity and a set of process state parameters directly related to the production and processing process from the single-node data of the manufacturing data. Based on preset physical constraint rules, a first correlation analysis is performed on the set of quality characteristic parameters and the set of process state parameters to establish the direct influence relationship between the parameters. A second correlation analysis is performed on the set of quality characteristic parameters and the set of process state parameters based on historical manufacturing data to determine the implicit coupling relationship between the parameters; By fusing the direct influence relationship with the implicit coupling relationship, an association network reflecting the interaction path between parameters is constructed; Using the identification code of the manufactured entity as an index, a structured data object is created in the digital space; The relationship network, the set of quality characteristic parameters, and the set of process state parameters are stored as the core data fields of the structured data object to generate a virtual mapping body.

[0007] Optionally, in a third implementation of the first aspect of this application, the step of comparing the first predicted value of the virtual mapping body at the quality monitoring node with the measured value of the manufactured entity to determine whether there is a quality deviation between the virtual mapping body and the manufactured entity includes: Extract the set of process status parameters corresponding to all preceding nodes of the manufacturing entity at the current quality monitoring node from the manufacturing data; The process state parameter set is input into the association network in the virtual mapping body, and the first predicted value of the quality characteristic parameter corresponding to the manufacturing entity at the current quality monitoring node is output. Extract single-node data corresponding to the current quality monitoring node from the manufacturing data, parse the actual acquired quality inspection data from the single-node data, and obtain the measured value of the quality characteristic parameter; The difference between the first predicted value and the measured value is calculated to obtain the single-point deviation value of each quality characteristic parameter; The single-point deviation value is compared with a first deviation threshold. If the single-point deviation value exceeds the first deviation threshold, it is determined that there is a quality deviation between the virtual mapping body and the manufacturing entity.

[0008] Optionally, in the fourth implementation of the first aspect of this application, the method further includes: Aggregate the single-point deviation values ​​of all manufacturing entities in the current production batch to determine the statistical distribution characteristics of the deviation values ​​in the current production batch; The statistical distribution characteristics are compared with the second deviation threshold; If the statistical distribution characteristics exceed the second deviation threshold, then it is determined that there is a quality deviation between the virtual mapping body and the manufactured entity.

[0009] Optionally, in the fifth implementation of the first aspect of this application, the step of performing a reverse search starting from the current quality monitoring node to locate abnormal process parameters includes: Using the quality characteristic parameter whose single-point deviation value exceeds the first deviation threshold as the retrieval target, the upstream process status parameter node directly connected to the retrieval target is located in the association network; Extract the target measured values ​​of the upstream process status parameter nodes at each preceding quality monitoring node from the manufacturing data; The measured target value is compared with the preset standard process parameters to determine the abnormal process parameters of the upstream process status parameter node.

[0010] Optionally, in the sixth implementation of the first aspect of this application, the step of dynamically adjusting the equipment settings by matching the corresponding adjustment instructions from a preset process knowledge base based on the contribution of the abnormal process parameters to the quality deviation includes: Based on the parameter type and abnormal change direction of the abnormal process parameters, the corresponding baseline adjustment amount is matched from the process knowledge base; Based on the contribution of the abnormal process parameters to the quality deviation, the baseline adjustment amount is weighted and corrected to generate a candidate adjustment instruction set; The device setting value adjustment amount corresponding to the candidate adjustment instruction set is input into the correlation network to determine the second predicted value of the quality characteristic parameter corresponding to the device setting value adjustment amount; Based on the difference between the second predicted value and the measured value, the adjustment instruction that reduces the quality deviation the most is selected from the candidate adjustment instruction set and dynamically adjusted.

[0011] A second aspect of this application provides a digital twin-based intelligent manufacturing quality monitoring system, which is used to implement a digital twin-based intelligent manufacturing quality monitoring method. The digital twin-based intelligent manufacturing quality monitoring system includes: The acquisition module is used to identify the identification code of the manufacturing entity based on the preset quality monitoring node, and to acquire the quality inspection data of the manufacturing entity at the corresponding quality monitoring node to obtain the manufacturing data of the manufacturing entity. The construction module is used to construct a virtual mapping body corresponding to the manufacturing entity based on the manufacturing data; The comparison module is used to compare the first predicted value of the virtual mapping body at the quality monitoring node with the measured value of the manufacturing entity to determine whether there is a quality deviation between the virtual mapping body and the manufacturing entity. The positioning module is used to locate abnormal process parameters by performing a reverse retrieval starting from the current quality monitoring node when there is a quality deviation between the virtual mapping body and the manufacturing entity. The adjustment module is used to dynamically adjust the equipment settings by matching the corresponding adjustment instructions from a preset process knowledge base based on the contribution of the abnormal process parameters to the quality deviation.

[0012] A third aspect of this application provides an electronic device, including a memory and a processor, wherein the processor is configured to execute a computer program stored in the memory, and when the processor executes the computer program, it implements the steps of the intelligent manufacturing quality monitoring method based on digital twins provided in the first aspect of this application.

[0013] The fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements the steps of the intelligent manufacturing quality monitoring method based on digital twins provided in the first aspect of this application.

[0014] In summary, the intelligent manufacturing quality monitoring method and system based on digital twins provided in this application constructs traceable manufacturing data spanning multiple quality monitoring nodes for the manufacturing entity; a virtual mapping entity containing parameter correlations is created based on this data chain; the quality prediction value of the virtual mapping entity is compared with the measured value of the corresponding node in real time to identify quality deviations; when deviations exist, the abnormal process parameters are located by reverse tracing using the virtual mapping entity; and equipment adjustment instructions are matched and executed from the process knowledge base based on the contribution of the parameter. This application, by constructing a fully traceable data chain and virtual mapping entity, achieves precise and automated root cause localization of quality problems, solving the problems of traditional methods relying on manual experience and inefficient traceability. The use of the virtual mapping entity for real-time prediction and comparison effectively improves process stability and quality consistency. Attached Figure Description

[0015] Figure 1 A flowchart illustrating the intelligent manufacturing quality monitoring method based on digital twins provided in this application embodiment; Figure 2 A schematic diagram illustrating the quality deviation between the virtual mapping object and the manufactured entity provided in this application embodiment; Figure 3 A schematic diagram of the program modules of the intelligent manufacturing quality monitoring system based on digital twin provided in the embodiments of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0016] To make the inventive objectives, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] To address the problem in related technologies that cannot deeply trace and predict the causes of defects during product manufacturing, embodiments of this application provide a smart manufacturing quality monitoring method based on digital twins, such as... Figure 1 This is a flowchart illustrating the intelligent manufacturing quality monitoring method based on digital twins provided in this embodiment. The intelligent manufacturing quality monitoring method based on digital twins includes the following steps: Step 110: Identify the identification code of the manufacturing entity based on the preset quality monitoring nodes, and collect the quality inspection data of the manufacturing entity at the corresponding quality monitoring nodes to obtain the manufacturing data of the manufacturing entity.

[0018] Specifically, each entity entering the production line is assigned a globally unique identification code, and data acquisition terminals with synchronized clocks are deployed at multiple pre-set quality monitoring nodes to achieve real-time binding with the identification code. The data acquisition terminals automatically read the identification code as the entity passes by and record equipment operating parameters, sensor outputs, visual inspection results, and precise timestamps. The acquired information is then packaged and stored in units of nodes and linked together through the identification code to form manufacturing data arranged in the flow sequence, thereby ensuring that each piece of data can be traced back to a specific entity, a specific point in time, and a specific process location. At the same time, to improve data availability, a data integrity verification and time synchronization correction mechanism is adopted to ensure the temporal consistency and quality reliability of cross-node data.

[0019] Step 120: Construct a virtual mapping body corresponding to the manufacturing entity based on the manufacturing data.

[0020] Specifically, the process of constructing a virtual mapping entity corresponding to the manufacturing entity based on manufacturing data involves systematically organizing the quality characteristic parameters and process state parameters contained in the manufacturing data, and clarifying the direct relationship between quality characteristics and process influencing factors according to established physical constraints. Subsequently, by combining statistical correlation patterns extracted from historical manufacturing records, potential parameter coupling relationships are incorporated into the correlation structure. By integrating direct and implicit coupling relationships, a correlation structure network reflecting the inherent production path characteristics and key quality influencing factors of the manufacturing entity is formed. Then, using the manufacturing entity's identification information as an index, an independent data object is created, storing the aforementioned correlation structure, parameter set, and production path information as core elements within this object. This enables the object to dynamically respond to the entity's manufacturing process, thus completing the construction of the virtual mapping entity.

[0021] Step 130: Compare the first predicted value of the virtual mapping body at the quality monitoring node with the measured value of the manufacturing entity to determine whether there is a quality deviation between the virtual mapping body and the manufacturing entity.

[0022] Specifically, the process of comparing the first predicted value of the virtual mapping entity at the quality monitoring node with the measured value of the manufacturing entity first requires extracting the process state parameters formed by the manufacturing entity in all processes before the current node from the manufacturing data. These parameters are then input into the association structure contained in the virtual mapping entity to deduce the predicted value of the key quality attribute corresponding to the current node. Simultaneously, the actual inspection records collected at the current node are parsed from the manufacturing data to obtain the measured value of the quality attribute. The difference between the predicted value and the measured value is calculated, and the difference is judged according to a predetermined deviation threshold. If the difference exceeds a preset range, it is determined that there is a deviation between the predicted performance and the actual performance, thus revealing a non-negligible quality change trend in the production process.

[0023] Step 140: When there is a quality deviation between the virtual mapping body and the manufacturing entity, perform a reverse search starting from the current quality monitoring node to locate the abnormal process parameters.

[0024] Specifically, when a quality deviation exists between the virtual mapping entity and the manufactured entity, the process involves reverse retrieval and locating abnormal process parameters, starting from the current quality monitoring node. The core objective is the quality characteristic with the largest deviation, and the process traces upstream process influencing factors step by step based on the correlation structure within the virtual mapping entity. By tracking process state parameter nodes directly related to this quality characteristic, the actual records of the corresponding processes in the manufacturing data are further extracted, and key data are compared with standard process requirements. If the actual value of a certain process state deviates from the normal range in terms of fluctuation amplitude, trend of change, or stability, it can be identified as a potential abnormal process parameter.

[0025] Step 150: Based on the contribution of abnormal process parameters to quality deviation, dynamically adjust the corresponding equipment setting values ​​by matching the adjustment instructions from the preset process knowledge base.

[0026] Specifically, based on the attribute characteristics, direction of change, and correlation with quality deviation of abnormal process parameters, corresponding basic adjustment suggestions are extracted from the process knowledge base and weighted to reflect the actual impact of abnormal factors on deviation. Then, the adjusted equipment settings are input into the correlation structure to deduce the possible quality change effect after adjustment, thereby obtaining the second predicted performance corresponding to the adjustment amount. This second predicted performance is compared with the current actual performance to select the adjustment scheme that can improve the deviation to the greatest extent from multiple candidate adjustment amounts. This scheme is then sent to the equipment execution layer for implementation, enabling the production process to be flexibly adjusted according to the real-time quality status, thereby achieving dynamic control of manufacturing quality.

[0027] In one optional implementation of this embodiment, the steps of identifying the identification code of the manufacturing entity based on preset quality monitoring nodes and collecting the quality inspection data of the manufacturing entity at the corresponding quality monitoring nodes to obtain the manufacturing data of the manufacturing entity include: identifying the identification code carried by the manufacturing entity according to all quality monitoring nodes on the same production line, and generating an association record between the identification code and the node identification; during the processing of the manufacturing entity at the quality monitoring node, acquiring the equipment process parameters and quality inspection data of the quality monitoring node to generate single-node data; and according to the flow sequence of the manufacturing entity along the production path, arranging and aggregating the single-node data generated by the manufacturing entity at each quality monitoring node according to the node time sequence, and binding the aggregated single-node data through the identification code to generate manufacturing data.

[0028] In this embodiment, during production line operation, the identification device corresponding to the quality monitoring node reads the identification code carried by the manufacturing entity when it enters its scope of operation. This identification code is a unique code that identifies a specific manufacturing entity and can be attached to the surface of the entity or a carrier accompanying the workpiece. The identification device maintains time synchronization with the production line control system, recording the node identifier of the current quality monitoring node and the time information of the reading while completing the identification code reading. This establishes a one-to-one correspondence between the identification code and the node identifier, and writes this correspondence into a data record to indicate that the manufacturing entity has arrived at the designated quality monitoring node at a specific time. For example, in an automotive parts assembly scenario, when an engine block enters a hole machining station, its identification code is read by the identification device and associated with the node identifier of the hole machining station, forming a traceable arrival record. During the process of the manufacturing entity performing processing or inspection operations at the quality monitoring node, information about the operating status around that node is continuously collected. Equipment process parameters refer to numerical information reflecting the operating status of the processing process, such as the rotational speed, feed rate, temperature, or pressure of the processing equipment, used to describe the actual working conditions of the equipment during operation. Quality inspection data refers to the results obtained from measuring or inspecting the manufactured entity at a specific node, reflecting the quality characteristics exhibited by the manufactured entity at that stage. After the manufactured entity completes its corresponding operation within the node's scope, the equipment process parameters, quality inspection data, and the aforementioned identification codes and node identifier association information collected during that time period are integrated to form a structurally complete single-node data set. This single-node data set reflects the complete processing and inspection process experienced by the manufactured entity at that node. For example, in the aforementioned engine cylinder block hole machining station, the spindle speed, coolant temperature, and hole diameter measurements collected during machining are integrated into the same single-node data set and stored along with the cylinder block identification code and station identifier. As the manufactured entity flows sequentially between multiple quality monitoring nodes along a predetermined production path, the aforementioned single-node data is continuously generated at different nodes. Based on the actual flow sequence of the manufactured entity in the production line, the single-node data generated by each quality monitoring node is sorted chronologically. This chronological sorting is based on the chronological order of events recorded in the single-node data set, thereby reconstructing the actual production experience of the manufactured entity. After collecting single-node data from different quality monitoring nodes, the identification code of the manufacturing entity is used as an index to bind single-node data belonging to the same identification code, forming a coherent record sequence at the data level. This record sequence is the manufacturing data, used to completely describe the entire process state changes of the manufacturing entity from the first quality monitoring node to the last node. For example, after the same engine block completes hole machining, it continues to flow to the cleaning station and the inspection station. The single-node data generated at each station is appended to the original record in chronological order and maintained in a unified association through the cylinder block identification code, thus forming manufacturing data covering multiple processes.

[0029] In one optional implementation of this embodiment, the step of constructing a virtual mapping body corresponding to a manufacturing entity based on manufacturing data includes: separating a set of quality characteristic parameters representing the quality attributes of the manufacturing entity and a set of process state parameters directly related to the production and processing process from the single-node data of the manufacturing data; performing a first association analysis on the set of quality characteristic parameters and the set of process state parameters based on preset physical constraint rules to establish a direct influence relationship between the parameters; performing a second association analysis on the set of quality characteristic parameters and the set of process state parameters based on historical manufacturing data to determine the implicit coupling relationship between the parameters; fusing the direct influence relationship and the implicit coupling relationship to construct an association relationship network reflecting the action path between the parameters; creating a structured data object in the digital space using the identification code of the manufacturing entity as an index; and storing the association relationship network, the set of quality characteristic parameters, and the set of process state parameters as the core data fields of the structured data object to generate a virtual mapping body.

[0030] In this embodiment, the manufacturing data consists of multiple single-node data arranged chronologically. Each single-node data contains multiple types of parameter information reflecting the state of the manufactured entity. To form a clear understanding of the quality state of the manufactured entity, the parameters in the single-node data need to be functionally distinguished. Quality characteristic parameters refer to numerical information directly used to describe the quality performance of the manufactured entity, such as dimensions, form and position deviations, or surface conditions, reflecting whether the manufactured entity meets design requirements. Process state parameters refer to operational state information generated during production and processing that affects the quality results, such as equipment speed, feed rate, or processing environment conditions. By performing semantic recognition and classification on the fields in the single-node data, the parameters used for quality evaluation are extracted into a set of quality characteristic parameters, and the parameters describing the operational state of the processing process are extracted into a set of process state parameters, thereby achieving the separation of information of different natures. After completing the parameter separation, the direct relationship between the set of quality characteristic parameters and the set of process state parameters is sorted out based on the known physical constraints in the manufacturing process. Physical constraint rules refer to deterministic relationships determined by process principles, material properties, or equipment structure, used to describe the action path that a change in a certain process state will inevitably cause a change in a specific quality characteristic. For example, in the machining of engine block bores, there is a clear relationship between spindle speed and bore surface roughness, stemming from the physical mechanisms of the cutting process. By incorporating such rules into the parameter relationship construction process, a direct influence relationship can be established between quality characteristic parameters and corresponding process state parameters, indicating the clear direction of a process factor's impact on the quality outcome. Building upon this, it is necessary to further explore potential relationships not covered by explicit rules in the production process. By performing statistical correlation calculations on historical manufacturing data chains, a stable trend of correspondence between quality characteristic parameters and process state parameters can be identified in a large number of production records. Statistical correlation calculation refers to quantitatively analyzing the synchronicity of parameter changes across multiple manufacturing entities and production processes to determine whether a stable correlation exists between the two. For instance, in the cylinder block machining process, there may not be a direct physical formula describing the relationship between coolant temperature and bore diameter, but a consistent trend is observed in long-term production records. This trend reflects a potential coupling relationship. Through such analysis, implicit coupling relationships beyond direct influence relationships can be supplemented, making the connections between parameters more complete. Once both direct influence relationships and implicit coupling relationships are identified, they are integrated to form a correlation network that reflects the interaction paths between parameters. A correlation network is a structural representation used to describe multiple parameter nodes and their mutual influence relationships. Each node corresponds to a quality characteristic parameter or process state parameter, and the connections between nodes represent the direction of influence and the strength of the correlation.This network can present the multi-path propagation relationship from changes in process state to changes in quality result. For example, the spindle speed affects the overall process by influencing the cutting state, which in turn affects the hole diameter and surface condition. Then, using the identifier of the manufactured entity as a unique index, a structured data object is created in the digital space. This structured data object is a data carrier used to centrally store information related to a specific manufactured entity. It employs a fixed field structure to ensure the integrity and parsability of the information. The aforementioned constructed relationship network, quality characteristic parameter set, and process state parameter set are stored as the core data fields of this data object. This allows the object to not only contain the static attributes of the manufactured entity but also reflect its dynamic quality evolution characteristics during the manufacturing process, forming a virtual mapping synchronized with the physical production process.

[0031] In one optional implementation of this embodiment, the step of comparing the first predicted value of the virtual mapping body at the quality monitoring node with the measured value of the manufacturing entity to determine whether there is a quality deviation between the virtual mapping body and the manufacturing entity includes: extracting the set of process state parameters corresponding to all preceding nodes of the manufacturing entity at the current quality monitoring node from the manufacturing data; inputting the set of process state parameters into the association network in the virtual mapping body and outputting the first predicted value of the quality characteristic parameter corresponding to the manufacturing entity at the current quality monitoring node; extracting single-node data corresponding to the current quality monitoring node from the manufacturing data, parsing the actually acquired quality inspection data from the single-node data, and obtaining the measured value of the quality characteristic parameter; calculating the difference between the first predicted value and the measured value to obtain the single-point deviation value of each quality characteristic parameter; comparing the single-point deviation value with a first deviation threshold, and if the single-point deviation value exceeds the first deviation threshold, it is determined that there is a quality deviation between the virtual mapping body and the manufacturing entity.

[0032] In this embodiment, the manufacturing data has completely recorded the processing and inspection information experienced by the manufacturing entity at each quality monitoring node in chronological order. When the manufacturing entity reaches a certain current quality monitoring node, the position corresponding to that node can be located in the manufacturing data based on the node identifier and time information. The set of process state parameters corresponding to the preceding node refers to the set of parameter information collected and recorded in all quality monitoring nodes before the current node, reflecting the operating status of the processing process. By searching forward along the manufacturing data, the data of single nodes belonging to the same manufacturing entity and occurring earlier than the current node are read one by one, and the process state parameter fields are extracted from them to form a parameter set arranged in chronological order. This set can completely reflect the changes in processing conditions that the manufacturing entity has experienced before entering the current node. For example, in the engine block processing scenario, when the cylinder block enters the final inspection station, the set of process state parameters corresponding to the preceding node will include information such as the spindle speed of the hole processing station and the medium temperature of the cleaning station. After obtaining the set of preceding process state parameters, this set needs to be transferred to the established relational network within the virtual mapping entity. Once the set of process state parameters is loaded, the corresponding nodes of each parameter in the network are assigned actual values, triggering the internal relational transfer process within the network. This allows the quality characteristic parameters associated with the current quality monitoring node to gradually form a deduced result. Through this transfer process, the potential quality state of the manufactured entity at the current node before actual inspection can be obtained. This quality state is represented numerically as a first predicted value, reflecting the quality performance deduced based on historical processing. For example, before the final inspection station, the expected values ​​of the cylinder bore diameter and surface condition at this node can be deduced from the status information of the preceding hole machining and cleaning processes. Simultaneously, single-node data directly corresponding to the current quality monitoring node is extracted from the manufacturing data. This single-node data includes the quality inspection data obtained at that node. Quality inspection data refers to the result information obtained after measuring devices or testing equipment directly act on the manufactured entity, accurately reflecting the quality state of the manufactured entity at that node. By parsing the relevant fields in the single-node data, the measured values ​​of the quality characteristic parameters corresponding to the first predicted value can be extracted, such as the actual aperture measurement results obtained in the final inspection station. Figure 2As shown, the curve represents the predicted value of the virtual mapping entity, and the bar chart represents the measured value of the manufactured entity. After obtaining both the first predicted value and the measured value, the difference between the two is calculated to obtain the single-point deviation value. The single-point deviation value describes the difference between the quality state reflected by the virtual mapping entity and the actual state of the manufactured entity at the current quality monitoring node. For example, if the predicted aperture is 50.008 mm and the measured aperture is 50.01 mm, a clear single-point deviation value is formed. The single-point deviation value is compared with a pre-set first deviation threshold. The first deviation threshold is a criterion used to define the allowable deviation range, derived from design requirements or quality control targets. When the single-point deviation value exceeds this threshold, it indicates that the difference between the predicted result and the actual result has exceeded the acceptable range, thus confirming the existence of a quality deviation between the virtual mapping entity and the manufactured entity. Taking the engine cylinder block as an example, when the aperture deviation exceeds the design allowable range, an anomaly in the quality of the current manufactured entity can be identified.

[0033] In one optional implementation of this embodiment, the single-point deviation values ​​of all manufacturing entities in the current production batch are aggregated to determine the statistical distribution characteristics of the deviation values ​​in the current production batch; the statistical distribution characteristics are compared with a second deviation threshold; if the statistical distribution characteristics exceed the second deviation threshold, it is determined that there is a quality deviation between the virtual mapping body and the manufacturing entity.

[0034] In this embodiment, after calculating the single-point deviation value of a single manufacturing entity, the perspective needs to be expanded from the individual entity to the entire current production batch to identify whether there are systematic quality changes. The current production batch refers to a group of manufacturing entities manufactured continuously under the same production conditions and within a similar time window, characterized by continuous process environment and similar equipment status. Aggregating the single-point deviation values ​​of all manufacturing entities within this batch involves grouping single-point deviation values ​​belonging to the same quality characteristic but from different manufacturing entities, using quality characteristic parameters as the dimension, to form a dataset suitable for overall analysis. In this process, aggregation is not simply stacking values, but rather uniformly collecting deviation values ​​based on the identification information and time sequence of the manufacturing entities, enabling it to reflect the overall performance of the quality status within the batch. After completing the deviation value aggregation, it is necessary to further extract the statistical distribution characteristics of the batch's deviation values. Statistical distribution characteristics are quantitative indicators used to describe the overall shape of a set of values, reflecting the concentration and trend of deviations within the batch, such as the concentration range, dispersion, or overall offset direction of the deviation values. By analyzing the aggregated set of deviation values, it can be determined whether the quality deviation is concentrated in a few manufacturing entities or exhibits a general change within the batch. For example, in the final inspection scenario of engine blocks, if the bore diameter deviation values ​​of most cylinder blocks in a batch all shift in the same direction, the statistical distribution characteristics will show a significant shift in the overall distribution center, rather than a scattered distribution of individual anomalies. After obtaining the statistical distribution characteristics, it is necessary to compare them with a second deviation threshold. The second deviation threshold is a control limit used for batch-level quality judgment. It focuses on whether the overall deviation level exceeds the allowable range, rather than the individual performance of a single manufacturing entity. When the statistical distribution characteristics reflect that the deviation concentration range, overall offset amplitude, or dispersion exceeds this threshold, it indicates that the quality status of the current production batch has undergone a systematic change. Even if some individual units are still within the acceptable range, the overall quality stability has been affected. If the comparison result exceeds the second deviation threshold, it can be determined that there is a quality deviation between the virtual mapping and the manufacturing entity. This judgment emphasizes the identification of deviations at the batch level, used to reveal the inconsistency between the overall state of the production process and the ideal state reflected by the virtual mapping. For example, in engine block production, if the bore diameter deviation of a single cylinder block has not yet exceeded the limit, but the bore diameter deviation distribution of the entire batch of cylinder blocks has shifted upward and exceeded the second deviation threshold, it can be determined that the production process of this batch has deviated from a stable state.

[0035] In one optional implementation of this embodiment, the step of performing a reverse retrieval starting from the current quality monitoring node to locate abnormal process parameters includes: using a quality characteristic parameter whose single-point deviation value exceeds a first deviation threshold as the retrieval target, locating the upstream process status parameter node directly connected to the retrieval target in the association network; extracting the target measured value of the upstream process status parameter node at each preceding quality monitoring node from the manufacturing data; comparing the target measured value with preset standard process parameters to determine the abnormal process parameters of the upstream process status parameter node.

[0036] In this embodiment, when a manufacturing entity is determined to have a single-point deviation exceeding a first deviation threshold at a certain quality monitoring node, the exceeding quality characteristic parameter becomes the retrieval target for subsequent traceability analysis. The quality characteristic parameter directly describes the quality status of the manufacturing entity; exceeding the deviation threshold means the manufacturing result has significantly deviated from expectations. Analysis is conducted around this retrieval target using the established relational network within the virtual mapping entity. In the relational network, centered on the quality characteristic parameter node where the deviation occurred, tracing back along the connection direction pointing to that node allows the location of all upstream process status parameter nodes that have a direct relationship with it. After locating the upstream process status parameter nodes, it is necessary to further return to the manufacturing data to verify the performance of the relevant parameters in the actual production process. The manufacturing data completely records the status information of the manufacturing entity at each preceding quality monitoring node. Through the manufacturing entity's identification code, the single-node data generated by its corresponding preceding node can be located one by one in the data chain, and the measured values ​​corresponding to the upstream process status parameter nodes can be extracted from it. The target measured value refers to the actual parameter value recorded by equipment or sensors during actual processing or operation, reflecting the actual performance of the process at a specific point in time. By extracting the target measured values ​​of multiple preceding nodes in chronological order, the trajectory of the process state parameters throughout the entire production process can be reconstructed. After obtaining the target measured value, it needs to be compared with the preset standard process parameter values. The standard process parameter values ​​refer to the benchmark values ​​or allowable ranges determined in process design or production control to ensure quality stability and to determine whether the process state is within a reasonable range. The comparison process focuses not only on the deviation of the value at a certain moment but also on the continuous changes of the parameter in multiple preceding nodes, such as whether there are continuous offsets, abrupt changes, or increased fluctuations. If the target measured value deviates significantly from the standard process parameter value in terms of magnitude, direction of change, or stability, it can be determined that the upstream process state parameter node has an anomaly in the actual production process. Taking the application scenario of out-of-tolerance bore diameter in engine cylinder blocks as an example, when the final inspection station finds that the single-point deviation value of the bore diameter quality characteristic parameter exceeds the first deviation threshold, the bore diameter becomes the retrieval target. By tracing back in the correlation network, the upstream process state parameter nodes directly related to the bore diameter can be located, such as the spindle speed and feed rate of the bore machining station. The measured spindle speed recorded at the bore machining station and its predecessor nodes in the manufacturing data is extracted and compared with the standard speed range specified in the process document. If it is found that the measured spindle speed is consistently higher than the benchmark value in multiple machining cycles, the process state parameter corresponding to the spindle speed can be determined to be an abnormal process parameter, thereby realizing the reverse location of specific process abnormalities from quality deviation.

[0037] In one optional implementation of this embodiment, the step of dynamically adjusting the equipment setpoints by matching the corresponding adjustment instructions from a preset process knowledge base based on the contribution of abnormal process parameters to quality deviation includes: matching the corresponding baseline adjustment amount from the process knowledge base according to the parameter type and abnormal change direction of the abnormal process parameters; weighting and correcting the baseline adjustment amount according to the contribution of abnormal process parameters to quality deviation to generate a candidate adjustment instruction set; inputting the corresponding equipment setpoint adjustment amounts from the candidate adjustment instruction set into the correlation network to determine the second predicted value of the quality characteristic parameter corresponding to the equipment setpoint adjustment amount; and selecting the adjustment instruction that reduces the quality deviation the most from the candidate adjustment instruction set for dynamic adjustment based on the difference between the second predicted value and the measured value.

[0038] In this embodiment, after the abnormal process parameter is clearly located, the abnormal information needs to be transformed into an actionable control basis. The parameter type of the abnormal process parameter refers to the category of process attribute it represents during production, such as speed, temperature, or pressure, used to distinguish its impact on equipment operation. The abnormal change direction refers to the deviation trend of the parameter relative to the standard process parameter value, such as being too high or too low. The process knowledge base is an information collection formed by systematically organizing existing production experience, process specifications, and equipment adjustment rules. It records the basic adjustment suggestions corresponding to different types of process parameters when a specific deviation direction occurs. By using the type and abnormal change direction of the abnormal process parameter as index conditions, and retrieving matching records in the process knowledge base, a baseline adjustment amount for correcting the abnormal state can be obtained. The baseline adjustment amount describes the basic range of adjustment to the equipment setpoint under ideal conditions. After obtaining the baseline adjustment amount, further correction is needed based on the contribution of the abnormal process parameter to the quality deviation. The contribution is a numerical indicator used to quantify the degree of influence of a certain process parameter on the quality deviation; its magnitude reflects the weight of the parameter change in the formation of the quality deviation. By incorporating contribution into the adjustment calculation process, the baseline adjustment can be weighted and corrected to match the adjustment magnitude with the actual quality impact intensity, thus generating multiple candidate adjustment instructions with different adjustment magnitudes. The candidate adjustment instruction set is a collection of several sets of equipment setpoint adjustments, representing various possible control schemes to avoid the uncertainty risks associated with a single adjustment. After forming the candidate adjustment instruction set, the potential quality improvement effect of each adjustment instruction needs to be pre-evaluated. To this end, the corresponding equipment setpoint adjustments in the candidate adjustment instruction set are applied one by one to the correlation network in the virtual mapping entity. The correlation network reflects the transmission path from process state changes to quality characteristic changes. When an adjustment is assigned to a corresponding process state parameter node, the influence relationship within the network changes accordingly, driving the related quality characteristic parameters to re-form the deduced results. Through this process, the possible state of the quality characteristic parameters at the current quality monitoring node after executing the adjustment instruction can be obtained, and this state is represented numerically as a second predicted value. After obtaining the second predicted value, the difference between it and the measured value of the current quality characteristic parameter needs to be calculated to quantify the degree of improvement of the quality deviation by the adjustment instruction. By comparing the changes in the differences corresponding to different candidate adjustment instructions, it is possible to identify which set of adjustment instructions can reduce the quality deviation by the greatest extent without introducing new adverse effects. Finally, the set of instructions with the most significant improvement effect is selected from the candidate adjustment instruction set and used as the basis for dynamic adjustment execution, which is then sent to the equipment control layer. Taking the machining scenario of engine cylinder bores as an example, when the spindle speed is identified as an abnormal process parameter and shows an upward trend, a baseline downward adjustment amount for the high speed can be matched in the process knowledge base.By considering the contribution of the rotational speed to the orifice diameter deviation, multiple sets of rotational speed adjustment commands are generated after correcting the reduction range. Subsequently, different rotational speed reduction amounts are input into a correlation network to deduce the possible second predicted value of the orifice diameter, and compared with the current measured orifice diameter. Through comparison, it can be found that a certain adjustment amount can maximize the reduction in orifice diameter deviation. This adjustment command is then determined and implemented, thereby achieving precise dynamic control based on quality feedback.

[0039] This application provides a digital twin-based intelligent manufacturing quality monitoring method that constructs traceable manufacturing data spanning multiple quality monitoring nodes for a manufacturing entity. Based on this data chain, a virtual mapping entity containing parameter relationships is created. The predicted quality values ​​of the virtual mapping entity are compared in real-time with the measured values ​​of the corresponding nodes to identify quality deviations. When deviations exist, the virtual mapping entity is used to trace and locate abnormal process parameters. Based on the contribution of these parameters, equipment adjustment instructions are matched from the process knowledge base and executed. This application, by constructing a fully traceable data chain and virtual mapping entity, achieves precise and automated root cause localization of quality problems, solving the problems of traditional methods relying on manual experience and inefficient traceability. By using the virtual mapping entity for real-time prediction and comparison, early warning and preventative control of quality risks are achieved, shifting quality control from passive detection to proactive intervention, effectively improving process stability and quality consistency.

[0040] Figure 3 This application provides an embodiment of a digital twin-based intelligent manufacturing quality monitoring system, which can be used to implement the digital twin-based intelligent manufacturing quality monitoring method described in the foregoing embodiments. Figure 3 As shown, this intelligent manufacturing quality monitoring system based on digital twins mainly includes: The data acquisition module is used to identify the identification code of the manufacturing entity based on the preset quality monitoring nodes, and to collect the quality inspection data of the manufacturing entity at the corresponding quality monitoring nodes to obtain the manufacturing data of the manufacturing entity. The building module is used to construct a virtual mapping body corresponding to the manufacturing entity based on the manufacturing data; The comparison module is used to compare the first predicted value of the virtual mapping body at the quality monitoring node with the measured value of the manufacturing entity to determine whether there is a quality deviation between the virtual mapping body and the manufacturing entity. The positioning module is used to perform reverse retrieval starting from the current quality monitoring node to locate abnormal process parameters when there is a quality deviation between the virtual mapping body and the manufacturing entity. The adjustment module is used to dynamically adjust the equipment settings by matching the corresponding adjustment instructions from the preset process knowledge base based on the contribution of abnormal process parameters to the quality deviation.

[0041] In one optional implementation of this embodiment, the acquisition module is specifically used to: identify the identification code carried by the manufacturing entity according to all quality monitoring nodes on the same production line, and generate an association record between the identification code and the node identification; during the processing of the manufacturing entity at the quality monitoring node, acquire the equipment process parameters and quality inspection data of the quality monitoring node, and generate single-node data; according to the flow sequence of the manufacturing entity along the production path, arrange and aggregate the single-node data generated by the manufacturing entity at each quality monitoring node according to the node time sequence, and bind the aggregated single-node data through the identification code to generate manufacturing data.

[0042] In one optional implementation of this embodiment, the construction module is specifically used to: separate a set of quality characteristic parameters representing the quality attributes of the manufactured entity and a set of process state parameters directly related to the production and processing process from the single-node data of the manufacturing data; perform a first association analysis on the set of quality characteristic parameters and the set of process state parameters based on preset physical constraint rules to establish a direct influence relationship between the parameters; perform a second association analysis on the set of quality characteristic parameters and the set of process state parameters based on historical manufacturing data to determine the implicit coupling relationship between the parameters; integrate the direct influence relationship and the implicit coupling relationship to construct an association relationship network reflecting the action path between the parameters; create a structured data object in the digital space using the identification code of the manufactured entity as an index; and store the association relationship network, the set of quality characteristic parameters, and the set of process state parameters as the core data fields of the structured data object to generate a virtual mapping body.

[0043] In an optional implementation of this embodiment, the comparison module is specifically used for: extracting the set of process state parameters corresponding to all preceding nodes of the manufacturing entity at the current quality monitoring node from the manufacturing data; inputting the set of process state parameters into the association network in the virtual mapping body, and outputting the first predicted value of the quality characteristic parameter corresponding to the manufacturing entity at the current quality monitoring node; extracting single-node data corresponding to the current quality monitoring node from the manufacturing data, parsing the actually acquired quality inspection data from the single-node data, and obtaining the measured value of the quality characteristic parameter; calculating the difference between the first predicted value and the measured value to obtain the single-point deviation value of each quality characteristic parameter; comparing the single-point deviation value with the first deviation threshold, and if the single-point deviation value exceeds the first deviation threshold, it is determined that there is a quality deviation between the virtual mapping body and the manufacturing entity.

[0044] In an optional implementation of this embodiment, the comparison module is further configured to: aggregate the single-point deviation values ​​of all manufacturing entities in the current production batch to determine the statistical distribution characteristics of the deviation values ​​in the current production batch; compare the statistical distribution characteristics with a second deviation threshold; and if the statistical distribution characteristics exceed the second deviation threshold, determine that there is a quality deviation between the virtual mapping body and the manufacturing entity.

[0045] In one optional implementation of this embodiment, the positioning module is specifically used to: locate the upstream process status parameter node directly connected to the search target in the association network, using the quality characteristic parameter whose single-point deviation value exceeds the first deviation threshold as the search target; extract the target measured value of the upstream process status parameter node at each preceding quality monitoring node in the manufacturing data; compare the target measured value with the preset standard process parameters to determine the abnormal process parameters of the upstream process status parameter node.

[0046] In one optional implementation of this embodiment, the adjustment module is specifically used for: matching the corresponding baseline adjustment amount from the process knowledge base according to the parameter type and abnormal change direction of the abnormal process parameters; weighting and correcting the baseline adjustment amount according to the contribution of the abnormal process parameters to the quality deviation, and generating a candidate adjustment instruction set; inputting the equipment setpoint adjustment amount corresponding to the candidate adjustment instruction set into the correlation network to determine the second predicted value of the quality characteristic parameter corresponding to the equipment setpoint adjustment amount; and dynamically adjusting the adjustment instruction that reduces the quality deviation the most from the candidate adjustment instruction set based on the difference between the second predicted value and the measured value.

[0047] The intelligent manufacturing quality monitoring system based on digital twins provided in this application constructs traceable manufacturing data spanning multiple quality monitoring nodes for the manufacturing entity; a virtual mapping body containing parameter correlations is created based on this data chain; the quality prediction value of the virtual mapping body is compared with the measured value of the corresponding node in real time to identify quality deviations; when deviations exist, the abnormal process parameters are located by reverse tracing using the virtual mapping body; and equipment adjustment instructions are matched and executed from the process knowledge base based on the contribution of the parameter. This application, by constructing a fully traceable data chain and virtual mapping body, achieves precise and automated root cause localization of quality problems, solving the problems of traditional methods relying on manual experience and inefficient traceability. By using the virtual mapping body for real-time prediction and comparison, early warning and preventative control of quality risks are achieved, shifting quality control from passive detection to proactive intervention, effectively improving process stability and quality consistency.

[0048] According to the scheme provided in this application Figure 4 An electronic device is provided as an embodiment of this application. This electronic device can be used to implement the intelligent manufacturing quality monitoring method based on digital twins in the foregoing embodiments, mainly including: The system includes a memory 401, a processor 402, and a computer program 403 stored on the memory 401 and executable on the processor 402. The memory 401 and the processor 402 are connected via communication. When the processor 402 executes the computer program 403, it implements the intelligent manufacturing quality monitoring method based on digital twins as described in the foregoing embodiments. The number of processors can be one or more.

[0049] The memory 401 can be a high-speed random access memory (RAM) or a non-volatile memory, such as a disk storage device. The memory 401 is used to store executable program code, and the processor 402 is coupled to the memory 401.

[0050] Furthermore, embodiments of this application also provide a computer-readable storage medium, which may be disposed in the electronic device described in the above embodiments, and the computer-readable storage medium may be as described above. Figure 4 The memory in the illustrated embodiment.

[0051] The computer-readable storage medium stores a computer program that, when executed by a processor, implements the digital twin-based intelligent manufacturing quality monitoring method described in the foregoing embodiments. Furthermore, the computer-readable storage medium can also be a USB flash drive, external hard drive, read-only memory (ROM), RAM, magnetic disk, or optical disk, or any other medium capable of storing program code.

[0052] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0053] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0054] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for intelligent manufacturing quality monitoring based on digital twins, characterized in that, include: The manufacturing entity's identification code is identified based on a preset quality monitoring node, and the manufacturing entity's quality inspection data at the corresponding quality monitoring node is collected to obtain the manufacturing entity's manufacturing data. Constructing a virtual mapping body corresponding to the manufacturing entity based on the manufacturing data; specifically including: separating a set of quality characteristic parameters representing the quality attributes of the manufacturing entity and a set of process state parameters directly related to the production and processing process from the single-node data of the manufacturing data; performing a first association analysis on the set of quality characteristic parameters and the set of process state parameters based on preset physical constraint rules to establish a direct influence relationship between the parameters; performing a second association analysis on the set of quality characteristic parameters and the set of process state parameters based on historical manufacturing data to determine the implicit coupling relationship between the parameters; fusing the direct influence relationship and the implicit coupling relationship to construct an association relationship network reflecting the action path between the parameters; creating a structured data object in the digital space using the identification code of the manufacturing entity as an index; storing the association relationship network, the set of quality characteristic parameters, and the set of process state parameters as the core data fields of the structured data object to generate a virtual mapping body; The first predicted value of the virtual mapping body at the quality monitoring node is compared with the measured value of the manufacturing entity to determine whether there is a quality deviation between the virtual mapping body and the manufacturing entity. When there is a quality deviation between the virtual mapping body and the manufacturing entity, a reverse retrieval is performed starting from the current quality monitoring node to locate the abnormal process parameters. Based on the contribution of the abnormal process parameters to the quality deviation, the corresponding equipment setting adjustment instructions are matched from the preset process knowledge base for dynamic adjustment.

2. The intelligent manufacturing quality monitoring method based on digital twins according to claim 1, characterized in that, The step of identifying the identification code of the manufacturing entity based on the preset quality monitoring node, and collecting the quality inspection data of the manufacturing entity at the corresponding quality monitoring node to obtain the manufacturing data of the manufacturing entity includes: The identification codes carried by the manufacturing entities are identified based on all quality monitoring nodes on the same production line, and an association record between the identification codes and node identifiers is generated. During the processing of the manufacturing entity at the quality monitoring node, the equipment process parameters and quality inspection data of the quality monitoring node are acquired to generate single-node data; Based on the flow sequence of the manufacturing entity along the production path, the single-node data generated by the manufacturing entity at each quality monitoring node is arranged and aggregated according to the node time sequence, and the aggregated single-node data is bound by the identification code to generate manufacturing data.

3. The intelligent manufacturing quality monitoring method based on digital twins according to claim 2, characterized in that, The step of comparing the first predicted value of the virtual mapping body at the quality monitoring node with the measured value of the manufactured entity to determine whether there is a quality deviation between the virtual mapping body and the manufactured entity includes: Extract the set of process status parameters corresponding to all preceding nodes of the manufacturing entity at the current quality monitoring node from the manufacturing data; The process state parameter set is input into the association network in the virtual mapping body, and the first predicted value of the quality characteristic parameter corresponding to the manufacturing entity at the current quality monitoring node is output. Extract single-node data corresponding to the current quality monitoring node from the manufacturing data, parse the actual acquired quality inspection data from the single-node data, and obtain the measured value of the quality characteristic parameter; The difference between the first predicted value and the measured value is calculated to obtain the single-point deviation value of each quality characteristic parameter; The single-point deviation value is compared with a first deviation threshold. If the single-point deviation value exceeds the first deviation threshold, it is determined that there is a quality deviation between the virtual mapping body and the manufacturing entity.

4. The intelligent manufacturing quality monitoring method based on digital twins according to claim 3, characterized in that, The method further includes: Aggregate the single-point deviation values ​​of all manufacturing entities in the current production batch to determine the statistical distribution characteristics of the deviation values ​​in the current production batch; The statistical distribution characteristics are compared with the second deviation threshold; If the statistical distribution characteristics exceed the second deviation threshold, then it is determined that there is a quality deviation between the virtual mapping body and the manufactured entity.

5. The intelligent manufacturing quality monitoring method based on digital twins according to claim 3, characterized in that, The step of performing a reverse search starting from the current quality monitoring node to locate abnormal process parameters includes: Using the quality characteristic parameter whose single-point deviation value exceeds the first deviation threshold as the retrieval target, the upstream process status parameter node directly connected to the retrieval target is located in the association network; Extract the target measured values ​​of the upstream process status parameter nodes at each preceding quality monitoring node from the manufacturing data; The measured target value is compared with the preset standard process parameters to determine the abnormal process parameters of the upstream process status parameter node.

6. The intelligent manufacturing quality monitoring method based on digital twins according to claim 1, characterized in that, The step of dynamically adjusting the equipment settings based on the contribution of the abnormal process parameters to the quality deviation by matching the corresponding adjustment instructions from a preset process knowledge base includes: Based on the parameter type and abnormal change direction of the abnormal process parameters, the corresponding baseline adjustment amount is matched from the process knowledge base; Based on the contribution of the abnormal process parameters to the quality deviation, the baseline adjustment amount is weighted and corrected to generate a candidate adjustment instruction set; The device setting value adjustment amount corresponding to the candidate adjustment instruction set is input into the correlation network to determine the second predicted value of the quality characteristic parameter corresponding to the device setting value adjustment amount; Based on the difference between the second predicted value and the measured value, the adjustment instruction that reduces the quality deviation the most is selected from the candidate adjustment instruction set and dynamically adjusted.

7. A digital twin-based intelligent manufacturing quality monitoring system, characterized in that, The digital twin-based intelligent manufacturing quality monitoring system is used to implement the digital twin-based intelligent manufacturing quality monitoring method of claim 1, wherein the digital twin-based intelligent manufacturing quality monitoring system comprises: The acquisition module is used to identify the identification code of the manufacturing entity based on the preset quality monitoring node, and to acquire the quality inspection data of the manufacturing entity at the corresponding quality monitoring node to obtain the manufacturing data of the manufacturing entity. The construction module is used to construct a virtual mapping body corresponding to the manufacturing entity based on the manufacturing data; The comparison module is used to compare the first predicted value of the virtual mapping body at the quality monitoring node with the measured value of the manufacturing entity to determine whether there is a quality deviation between the virtual mapping body and the manufacturing entity. The positioning module is used to locate abnormal process parameters by performing a reverse retrieval starting from the current quality monitoring node when there is a quality deviation between the virtual mapping body and the manufacturing entity. The adjustment module is used to dynamically adjust the equipment settings by matching the corresponding adjustment instructions from a preset process knowledge base based on the contribution of the abnormal process parameters to the quality deviation.

8. An electronic device, characterized in that, Includes memory and processor, of which: The processor is used to execute computer programs stored in the memory; When the processor executes the computer program, it implements the steps in the intelligent manufacturing quality monitoring method based on digital twins as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps in the intelligent manufacturing quality monitoring method based on digital twins as described in any one of claims 1 to 6.

Images