Architecture system for intelligent production of digital factory and implementation method thereof

Abstract

The invention relates to a digital factory intelligent production oriented architecture system and an implementation method thereof, and the architecture system comprises a data sensing layer, an edge computing layer, a cloud computing layer and an application layer. The data sensing layer sends equipment state information of corresponding production equipment acquired in real time to the edge computing layer to perform abnormal signal analysis processing, production plan scheduling optimization in a workshop and workshop environment analysis, and the cloud computing layer performs multi-modal data fusion on data provided by the edge computing layer and different downstream tasks. The method comprises the following steps: carrying out model node communication training on internal characteristics extracted from each modal data after multi-modal data fusion, carrying out incremental updating processing on different downstream task models according to different extracted internal characteristics after training is finished, calling a service interface provided by a cloud computing layer by an application layer, and realizing that the architecture is applied to different application scenes, so as to realize multi-modal data fusion. Quick response to changes in a factory is met, and the overall production efficiency of a digital factory is improved.

Description

Technical Field

[0001] The present invention relates to the field of digital factory intelligent production, and in particular to an architecture system oriented to digital factory intelligent production and an implementation method thereof. Background Art

[0002] With the rapid development of intelligent manufacturing, digital factories have become the core of modern industrial production. In order to improve production efficiency, reduce costs and ensure product quality, digital factories need an efficient and reliable production management platform, hoping to use the production management platform to dispatch and process data from various departments in the factory. However, existing production management platforms often have problems such as information islands, difficulties in system integration and low data processing efficiency, which cannot meet the needs of rapid response and real-time monitoring of modern industrial production.

[0003] Chinese invention patent CN116880412B discloses a cloud-based visual production management platform, which uses an anomaly detection method based on the mean variance method to perform anomaly monitoring and early warning. According to historical data and actual production conditions, a reasonable anomaly threshold is set to make anomaly detection more targeted, avoid false alarms and missed alarms, and ensure that real anomalies can be discovered and resolved in a timely manner. It ensures that the anomaly monitoring and early warning module can monitor data in real time, and immediately trigger an early warning notification to relevant staff when an anomaly occurs or exceeds the threshold. The exception handling mechanism is implemented in the data acquisition module. When an anomaly occurs, an alarm is promptly issued to the staff to prompt the abnormal situation of data collection, ensuring the stable collection and transmission of data. The platform can discover abnormal situations in production more timely, provide more accurate data analysis and decision support, and make production management more efficient and intelligent.

[0004] However, the cloud-based visual production management platform disclosed in the above-mentioned invention patent CN116880412B has shortcomings: because it sets an anomaly detection model based on historical data and adopts a fixed production scheduling algorithm, it lacks flexibility and scalability, and cannot respond quickly to changes in factory production, making it difficult to improve the overall production efficiency of the digital factory. Summary of the invention

[0005] The first technical problem to be solved by the present invention is to provide an architecture system for intelligent production in digital factories that can monitor and analyze the equipment data of each workshop in real time to optimize production scheduling efficiency in response to the above-mentioned existing technologies.

[0006] The second technical problem to be solved by the present invention is to provide a method for realizing the above-mentioned architecture system.

[0007] The technical solution adopted by the present invention to solve the first technical problem is: an architecture system for intelligent production in a digital factory, characterized in that it includes:

[0008] The data perception layer is formed by a number of industrial data acquisition modules deployed on multiple production equipment in the digital factory. The industrial data acquisition module is configured to collect the equipment status information of the corresponding production equipment in real time, and adjust the display content of the control equipment of the digital factory in real time according to the equipment status information; wherein the equipment status information of the production equipment includes equipment operation mode, equipment operation status, spindle speed, spindle load and spindle temperature, and the equipment operation status includes operation status, standby status and abnormal status. At least one industrial data acquisition module is deployed corresponding to each production equipment;

[0009] The edge computing layer is formed by multiple edge computing devices located at the edge of the network, and is configured to perform abnormal signal analysis and processing on the device status information sent by the data perception layer, perform scheduling optimization on the production plan in the workshop, and analyze the workshop environment; wherein the edge computing device is located in the vicinity of the industrial data generation source in a close distance, and the edge computing layer includes an abnormal detection and analysis module and a workshop scheduling and analysis module, the abnormal detection and analysis module is configured to receive the device status information sent by the data perception layer, and perform abnormal signal analysis and processing on the device status information; the workshop scheduling and analysis module is configured to perform real-time analysis and processing on the production information in the workshop to complete the scheduling optimization of the production plan; the production information in the workshop includes the production equipment resources currently available in the workshop and the production tasks in the production queue;

[0010] The cloud computing layer, including the cloud computing center and multiple model nodes, is configured to perform multimodal data fusion according to the device status information provided by the edge computing layer and different downstream tasks, extract the intrinsic features of each modal data after multimodal data fusion, and then communicate the extracted intrinsic features with the model nodes for training. After the training is completed, the different downstream task models are incrementally updated according to the different extracted intrinsic features;

[0011] The application layer applies the architecture to different application scenarios by calling the service interface provided by the cloud computing layer.

[0012] Improved, in the architecture system for digital factory intelligent production, the process of the industrial data acquisition module in the data perception layer collecting the equipment status information of the corresponding production equipment in real time includes the following steps:

[0013] The industrial data acquisition module is networked when it is normally connected to the production equipment in the workshop to obtain the equipment status information of the corresponding production equipment;

[0014] The device operating status is determined based on the acquired device status information, and the data transmission frequency is adjusted accordingly based on the device operating status. The device status information is sent to the edge server for analysis and processing according to the adjusted data transmission frequency corresponding to its operating status; wherein each device operating status has a corresponding data transmission frequency.

[0015] Further improved, in the architecture system for digital factory intelligent production, the abnormal detection and analysis module in the edge computing layer performs abnormal signal analysis and processing in the following manner:

[0016] Step a1: The anomaly detection and analysis module makes an anomaly analysis and judgment on the device status information sent by the data perception layer according to a preset anomaly detection model:

[0017] When the abnormal detection and analysis module determines that an abnormal signal is obtained, the production equipment corresponding to the abnormal signal is made to enter an abnormal state, stop production work, and go to step a2; otherwise, the production equipment is made to continue production work;

[0018] Step a2, the abnormality detection and analysis module analyzes the abnormality cause and abnormality type of the abnormal signal to obtain an abnormality analysis result; wherein the abnormality analysis result includes the abnormality cause and abnormality type of the abnormal signal;

[0019] Step a3, the anomaly detection and analysis module takes corresponding anomaly handling measures according to the anomaly analysis results, and performs multi-index testing on the production equipment after the anomaly handling measures are taken;

[0020] Step a4: The anomaly detection and analysis module makes a judgment based on the multi-index test results:

[0021] When the production equipment successfully passes the multi-index test, the production equipment is restored to a normal state and the operating state of the production equipment is updated; otherwise, the multi-index test is re-executed on the production equipment until the production equipment successfully passes the multi-index test.

[0022] Furthermore, in the architecture system for intelligent production in digital factories, the workshop scheduling analysis module in the edge computing layer performs real-time analysis and processing on the production information in the workshop to complete the scheduling optimization process of the production plan, which includes the following steps:

[0023] When the workshop scheduling analysis module determines that a new production task has appeared or the production equipment is in an abnormal state, it analyzes the current workshop environment and counts the available production equipment resources in the current workshop and the production tasks in the production queue;

[0024] The workshop scheduling analysis module outputs a scheduling plan suitable for the current workshop conditions based on the statistically available production equipment resources in the current workshop and the production tasks in the production queue;

[0025] The workshop scheduling analysis module verifies the feasibility of the scheduling plan based on the task priority of each production task in the scheduling plan and the production constraints corresponding to each production task:

[0026] When the scheduling plan is feasible, each production equipment in the current workshop is ordered to execute the scheduling plan; otherwise, the scheduling plan is readjusted to obtain a new scheduling plan that is feasible.

[0027] Furthermore, in the architecture system for digital factory intelligent production, the process of analyzing the workshop environment by the edge computing layer includes the following steps:

[0028] Step b1, traverse the operating status of all production equipment in the workshop, and calculate the distribution number of production equipment corresponding to each operating status; wherein the number of operating statuses of all production equipment in the workshop is marked as N, and the i-th operating status among the N operating statuses is marked as S i , in running state S i The distribution quantity of the corresponding production equipment is marked as

[0029]

[0030]

[0031] Among them, J represents the total number of production equipment in the workshop, s j represents the operating status of the jth production equipment in the workshop, δ(s j =S i ) is about the running state s j With running status S i The indicator function of

[0032] Step b2, according to the running status of each production equipment obtained by traversal, determine the total number of production equipment in a faulty state, and calculate the equipment failure rate in the current workshop; wherein, the equipment failure rate in the workshop is calculated as follows:

[0033]

[0034]

[0035] Among them, λ represents the equipment failure rate in the current workshop, S b Indicates that the production equipment is in a faulty state;

[0036] Step b3, obtaining the operating time of each production equipment in different operating states, and calculating the operating state time proportion of each operating state corresponding to each production equipment; wherein the operating state time proportion is calculated as follows:

[0037]

[0038] Among them, f ji Indicates that the jth production equipment is in operating state S i The duration corresponding to the time, T j,i Indicates that the jth production equipment is in operating state S i The length of time taken;

[0039] Step b4, counting the production quality of each type of workpiece produced in the factory, calculating the number of high-quality workpieces of each type, and calculating the high-quality rate of each type of workpiece; wherein the high-quality rate of each type of workpiece is calculated as follows:

[0040]

[0041] Among them, e u represents the quality rate of the u-th type of workpiece, P u represents the total number of workpieces of type u produced in the workshop, p uv It represents the quality of the vth workpiece among the uth type of workpieces produced in the workshop;

[0042] Step b5, count the production quality of each type of workpieces required for each production order, obtain the number of high-quality workpieces of each type, and calculate the production progress of the production tasks corresponding to each production order; the total number of production tasks received in the workshop is marked as K, and the production progress of the kth production task among the K production tasks is marked as r k :

[0043]

[0044] Among them, Q k represents the target production quantity of the kth production task, P u represents the total number of workpieces of type u produced in the workshop, p uv It represents the quality of the vth workpiece among the uth type of workpieces produced in the workshop;

[0045] Step b6, counting the number of workpieces produced by the workshop for all types of workpieces every day, and calculating the workshop capacity of the workshop; where the workshop capacity of the workshop on the dth day is marked as c d :

[0046]

[0047] Among them, c d,u It represents the total number of workpieces of the uth type produced by the workshop on the dth day, and U represents the total number of types of workpieces produced by the workshop.

[0048] Improved, in the architecture system for digital factory intelligent production, the cloud computing layer performs multimodal data fusion according to the data provided by the edge computing layer and different downstream tasks, including the following steps c1 to c5:

[0049] Step c1, according to the modal data type after the fusion of each modal data, a modal feature extractor corresponding to the modal data type is established through a pre-training method;

[0050] Step c2, retraining the established modal feature extractor according to the production task, so that the trained modal feature extractor meets the specific downstream task requirements in the application layer;

[0051] Step c3, selecting one or more model data according to the downstream task requirements in the application layer and performing feature extraction processing using the trained modal feature extractor corresponding to the model data to obtain extracted modal features, and training the output dimensions of each trained modal feature extractor using a fully automatic encoder to obtain modal features after dimensionality reduction processing;

[0052] Step c4, putting each modal feature after dimensionality reduction processing into a contrastive learning network to learn the corresponding relationship between different modal features;

[0053] Step c5: According to different downstream tasks, the corresponding mode is selected through the multimodal selection gate to select the multimodal features, and the multimodal features are transmitted to the downstream model in the application layer that needs to update the weight for training processing;

[0054] Step c6, after the downstream model in the application layer is trained, the updated parameters and weights are returned to the modal fusion model to train the multimodal fusion model.

[0055] The technical solution adopted by the present invention to solve the second technical problem is: a method for realizing an architecture for intelligent production in a digital factory, realizing an architecture system for intelligent production in a digital factory as described in any one of the items, characterized in that it comprises the following steps:

[0056] Step 1: The data perception layer collects the equipment status information of the corresponding production equipment in the workshop in real time, and adjusts the display content of the control equipment of the digital factory in real time according to the equipment status information; wherein the equipment status information of the production equipment includes the equipment operation mode, equipment operation status, spindle speed, spindle load and spindle temperature, and the equipment operation status includes operation status, standby status and abnormal status, and each production equipment is deployed with at least one industrial data acquisition module;

[0057] Step 2: The edge computing layer analyzes and processes abnormal signals of the equipment status information sent by the data perception layer, optimizes the production plan in the workshop, and analyzes the workshop environment;

[0058] Step 3: The cloud computing layer performs multimodal data fusion according to the data provided by the edge computing layer and different downstream tasks, and extracts the intrinsic features of each modal data after multimodal data fusion, and then communicates the extracted intrinsic features with the model nodes for training. After the training is completed, the different downstream task models are incrementally updated according to the different extracted intrinsic features.

[0059] Step 4: The application layer calls the service interface provided by the cloud computing layer to apply the architecture to different application scenarios.

[0060] Compared with the prior art, the advantages of the present invention are: the architecture system for intelligent production of digital factories of the present invention sets a data perception layer, an edge computing layer, a cloud computing layer and an application layer. The data perception layer collects the equipment status information of the corresponding production equipment in real time and sends it to the edge computing layer, and the data perception layer adjusts the display content of the control equipment of the digital factory according to the equipment status information in real time, and then the edge computing layer performs abnormal signal analysis and processing on the equipment status information sent by the data perception layer, optimizes the production plan in the workshop and analyzes the workshop environment, and the cloud computing layer and the edge computing layer provide data and different downstream tasks for multimodal data fusion, and extracts the intrinsic features of each modal data after multimodal data fusion, and then communicates the extracted intrinsic features with the model node for training, and after the training is completed, according to the different extracted intrinsic features, the different downstream task models are incrementally updated, and the application layer is enabled to apply the architecture to different application scenarios by calling the service interface provided by the cloud computing layer. In this way, the architecture system can not only obtain the equipment status information of the production equipment in the workshop in real time, but also analyze abnormal signals and workshop environment according to the equipment status information, and dynamically schedule and optimize the production plan in the workshop, so as to meet the needs of rapid response to changes in factory production and improve the overall production efficiency of the digital factory. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 A schematic diagram of an architecture system for intelligent production in a digital factory according to an embodiment of the present invention;

[0062] Figure 2 for Figure 1 A flowchart of the implementation method of the architecture system shown. DETAILED DESCRIPTION

[0063] The present invention is further described in detail below with reference to the accompanying drawings.

[0064] This embodiment provides an architecture system for intelligent production in a digital factory. Figure 1 As shown, the architecture system for digital factory intelligent production in this embodiment includes a data perception layer 1, an edge computing layer 2, a cloud computing layer 3 and an application layer 4. Among them:

[0065] The data perception layer is formed by a number of industrial data acquisition modules deployed on multiple production equipment in the digital factory. The industrial data acquisition module is configured to collect the equipment status information of the corresponding production equipment in real time, and adjust the display content of the control equipment of the digital factory in real time according to the equipment status information; wherein the equipment status information of the production equipment includes the equipment operation mode, equipment operation status, spindle speed, spindle load and spindle temperature, and the equipment operation status includes operation status, standby status and abnormal status. At least one industrial data acquisition module is deployed corresponding to each production equipment; for example, in the data perception layer of this embodiment, the industrial data acquisition module includes a camera, a laser radar, a temperature sensor and an air pressure sensor, etc.;

[0066] The edge computing layer is formed by multiple edge computing devices located at the edge of the network, and is configured to perform abnormal signal analysis and processing on the device status information sent by the data perception layer, perform scheduling optimization on the production plan in the workshop, and analyze the workshop environment; wherein the edge computing device is located in the vicinity of the industrial data generation source in a close distance, and the edge computing layer includes an abnormal detection and analysis module and a workshop scheduling and analysis module, the abnormal detection and analysis module is configured to receive the device status information sent by the data perception layer, and perform abnormal signal analysis and processing on the device status information; the workshop scheduling and analysis module is configured to perform real-time analysis and processing on the production information in the workshop to complete the scheduling optimization of the production plan; the production information in the workshop includes the production equipment resources currently available in the workshop and the production tasks in the production queue;

[0067] The cloud computing layer, including the cloud computing center and multiple model nodes, is configured to perform multimodal data fusion according to the device status information provided by the edge computing layer and different downstream tasks, extract the intrinsic features of each modal data after multimodal data fusion, and then communicate the extracted intrinsic features with the model nodes for training. After the training is completed, the different downstream task models are incrementally updated according to the different extracted intrinsic features;

[0068] The application layer calls the service interface provided by the cloud computing layer to apply the architecture to different application scenarios.

[0069] Specifically in this embodiment, the process of the industrial data collection module in the data perception layer collecting the equipment status information of the corresponding production equipment in real time includes the following steps:

[0070] The industrial data acquisition module is networked when it is normally connected to the production equipment in the workshop to obtain the equipment status information of the corresponding production equipment;

[0071] The device operating status is determined based on the acquired device status information, and the data transmission frequency is adjusted accordingly based on the device operating status. The device status information is sent to the edge server for analysis and processing according to the adjusted data transmission frequency corresponding to its operating status; wherein each device operating status has a corresponding data transmission frequency.

[0072] More specifically, in this embodiment, the anomaly detection and analysis module in the edge computing layer performs abnormal signal analysis processing in the following steps a1 to a4:

[0073] Step a1: The anomaly detection and analysis module makes an anomaly analysis and judgment on the device status information sent by the data perception layer according to a preset anomaly detection model:

[0074] When the abnormal detection and analysis module determines that an abnormal signal is obtained, the production equipment corresponding to the abnormal signal is made to enter an abnormal state, stop production work, and go to step a2; otherwise, the production equipment is made to continue production work;

[0075] Step a2, the abnormality detection and analysis module analyzes the abnormality cause and abnormality type of the abnormal signal to obtain an abnormality analysis result; wherein the abnormality analysis result includes the abnormality cause and abnormality type of the abnormal signal;

[0076] Step a3, the anomaly detection and analysis module takes corresponding anomaly handling measures according to the anomaly analysis results, and performs multi-index testing on the production equipment after the anomaly handling measures are taken;

[0077] Step a4: The anomaly detection and analysis module makes a judgment based on the multi-index test results:

[0078] When the production equipment successfully passes the multi-index test, the production equipment is restored to a normal state and the operating state of the production equipment is updated; otherwise, the multi-index test is re-executed on the production equipment until the production equipment successfully passes the multi-index test.

[0079] For the workshop scheduling analysis module in this embodiment, the workshop scheduling analysis module in the edge computing layer performs real-time analysis and processing on the production information in the workshop to complete the scheduling optimization process of the production plan, including the following steps:

[0080] When the workshop scheduling analysis module determines that a new production task has appeared or the production equipment is in an abnormal state, it analyzes the current workshop environment and counts the available production equipment resources in the current workshop and the production tasks in the production queue;

[0081] The workshop scheduling analysis module outputs a scheduling plan suitable for the current workshop conditions based on the statistically available production equipment resources in the current workshop and the production tasks in the production queue;

[0082] The workshop scheduling analysis module verifies the feasibility of the scheduling plan based on the task priority of each production task in the scheduling plan and the production constraints corresponding to each production task:

[0083] When the scheduling plan is feasible, each production equipment in the current workshop is ordered to execute the scheduling plan; otherwise, the scheduling plan is readjusted to obtain a new scheduling plan that is feasible.

[0084] As a way of analyzing the workshop environment by the edge computing layer, in this embodiment, the process of analyzing the workshop environment by the edge computing layer includes the following steps b1 to b6:

[0085] Step b1, traverse the operating status of all production equipment in the workshop, and calculate the distribution number of production equipment corresponding to each operating status; wherein the number of operating statuses of all production equipment in the workshop is marked as N, and the i-th operating status among the N operating statuses is marked as S i , in running state S i The distribution quantity of the corresponding production equipment is marked as

[0086]

[0087]

[0088] Among them, J represents the total number of production equipment in the workshop, s j represents the operating status of the jth production equipment in the workshop, δ(s j =S i ) is about the running state s j With running status S i The indicator function of

[0089] Step b2, according to the running status of each production equipment obtained by traversal, determine the total number of production equipment in a faulty state, and calculate the equipment failure rate in the current workshop; wherein, the equipment failure rate in the workshop is calculated as follows:

[0090]

[0091]

[0092] Among them, λ represents the equipment failure rate in the current workshop, S b Indicates that the production equipment is in a faulty state;

[0093] Step b3, obtaining the operating time of each production equipment in different operating states, and calculating the operating state time proportion of each operating state corresponding to each production equipment; wherein the operating state time proportion is calculated as follows:

[0094]

[0095] Among them, f ji Indicates that the jth production equipment is in operating state S i The duration corresponding to the time, T j,i Indicates that the jth production equipment is in operating state S i The length of time taken;

[0096] Step b4, counting the production quality of each type of workpiece produced in the factory, calculating the number of high-quality workpieces of each type, and calculating the high-quality rate of each type of workpiece; wherein the high-quality rate of each type of workpiece is calculated as follows:

[0097]

[0098] Among them, e u represents the quality rate of the u-th type of workpiece, P u represents the total number of workpieces of type u produced in the workshop, p uv It represents the quality of the vth workpiece among the uth type of workpieces produced in the workshop;

[0099] Step b5, count the production quality of each type of workpieces required for each production order, obtain the number of high-quality workpieces of each type, and calculate the production progress of the production tasks corresponding to each production order; the total number of production tasks received in the workshop is marked as K, and the production progress of the kth production task among the K production tasks is marked as r k :

[0100]

[0101] Among them, Q k represents the target production quantity of the kth production task, P u represents the total number of workpieces of type u produced in the workshop, p uv It represents the quality of the vth workpiece among the uth type of workpieces produced in the workshop;

[0102] Step b6, counting the number of workpieces produced by the workshop for all types of workpieces every day, and calculating the workshop capacity of the workshop; where the workshop capacity of the workshop on the dth day is marked as c d :

[0103]

[0104] Among them, c d,u It represents the total number of workpieces of the uth type produced by the workshop on the dth day, and U represents the total number of types of workpieces produced by the workshop.

[0105] It should be noted that, in this embodiment, the cloud computing layer performs multimodal data fusion according to the data provided by the edge computing layer and different downstream tasks, including the following steps c1 to c5:

[0106] Step c1, according to the modal data type after the fusion of each modal data, a modal feature extractor corresponding to the modal data type is established through a pre-training method; wherein, in this embodiment, the model of the modal feature extractor is as follows:

[0107] F c =M c (X c );

[0108] Among them, X c Indicates modal data of type c, M c represents the modal feature extractor for this type of modal data of type c, F c Represents the modal data X of type c c Extracted modal features;

[0109] Step c2, retraining the established modal feature extractor according to the production task so that the trained modal feature extractor meets the specific downstream task requirements in the application layer; wherein the trained modal feature extraction model that meets the specific downstream task requirements in the application layer is as follows:

[0110] F task =M task (X task ,Y task );

[0111] Among them, Y task Indicates the label of the specific downstream task, X task Represents the training data of specific downstream tasks, M task represents the trained modality feature extractor corresponding to the specific downstream task, F task It represents the modal features extracted by the trained modal feature extractor and meeting the specific downstream task requirements in the application layer;

[0112] Step c3, according to the downstream task requirements in the application layer, one or more model data are selected and feature extraction processing is performed using the trained modal feature extractor corresponding to the model data to obtain the extracted modal features, and the output dimensions of each trained modal feature extractor are trained using a fully automatic encoder to obtain the modal features after dimensionality reduction processing; wherein the model for obtaining the modal features after dimensionality reduction processing is as follows:

[0113] X encode =A(X,D);

[0114] Among them, X encode represents the modal features after dimensionality reduction, A represents the autoencoder, X represents the input data of the autoencoder, and D represents the dimension of the input data X;

[0115] Step c4, putting each modal feature after dimensionality reduction processing into a contrastive learning network to learn the corresponding relationship between different modal features; wherein the model of the contrastive learning network is as follows:

[0116]

[0117] Among them, z i and z j They refer to the modal characteristics after dimensionality reduction, τ is the temperature parameter, sim(z i ,z j ) represents the modal characteristic z i and z j The cosine similarity of

[0118] Step c5, according to the different downstream tasks, the corresponding mode is selected through the multimodal selection gate to select the multimodal features, and the multimodal features are transmitted to the downstream model in the application layer that needs to update the weight for training processing; wherein, the selection model of the multimodal features is as follows:

[0119] F select =C(F i ,Task;θ i ), F i ∈{F1,F2,…,F n}, 1≤i≤n;

[0120] Among them, F select is the selected modal feature, C is the multimodal selection gate, and F i represents the features of the i-th modality, n is the total number of modalities, Task is the type of downstream task, θ i is the weight of the i-th mode;

[0121] Step c6, after the downstream model in the application layer is trained, the updated parameters and weights are returned to the modal fusion model to train the multimodal fusion model. The model for training the multimodal fusion model is as follows:

[0122]

[0123] Among them, W fusion represents the weight of the multimodal fusion model, L fusionRepresents the loss function of the multimodal fusion model, W i Represents the weight from the i-th downstream model. Wherein, in this embodiment, the modal fusion model is located in the cloud computing layer. It is responsible for fusing data from different modalities to generate a more comprehensive and accurate feature representation. Specifically, the modal fusion model receives the updated parameters and weights from the downstream model of the application layer after training, and uses this information (i.e., the updated parameters and weights after training) to train the multimodal fusion model. The existence and function of this modal fusion model is to improve the performance of downstream tasks by fusing information from different modalities to enhance the learning ability and generalization ability of the model.

[0124] In addition, this embodiment also provides an architecture implementation method for digital factory intelligent production to implement the above-mentioned architecture system for digital factory intelligent production. Specifically, the architecture implementation method for digital factory intelligent production includes the following steps 1 to 4:

[0125] Step 1: The data perception layer collects the equipment status information of the corresponding production equipment in the workshop in real time, and adjusts the display content of the control equipment of the digital factory in real time according to the equipment status information; wherein the equipment status information of the production equipment includes the equipment operation mode, equipment operation status, spindle speed, spindle load and spindle temperature, and the equipment operation status includes operation status, standby status and abnormal status, and each production equipment is deployed with at least one industrial data acquisition module;

[0126] Step 2: The edge computing layer analyzes and processes abnormal signals of the equipment status information sent by the data perception layer, optimizes the production plan in the workshop, and analyzes the workshop environment;

[0127] Step 3: The cloud computing layer performs multimodal data fusion according to the data provided by the edge computing layer and different downstream tasks, and extracts the intrinsic features of each modal data after multimodal data fusion, and then communicates the extracted intrinsic features with the model nodes for training. After the training is completed, the different downstream task models are incrementally updated according to the different extracted intrinsic features.

[0128] Step 4: The application layer calls the service interface provided by the cloud computing layer to apply the architecture to different application scenarios.

[0129] Among them, the process of the industrial data acquisition module in the data perception layer collecting the equipment status information of the corresponding production equipment in real time, the way in which the anomaly detection and analysis module in the edge computing layer performs abnormal signal analysis and processing, the workshop scheduling and analysis module in the edge computing layer performs real-time analysis and processing of the production information in the workshop to complete the scheduling optimization process of the production plan, and the process of the edge computing layer analyzing the workshop environment can all be referred to the aforementioned content recorded in the architecture system for intelligent production in digital factories, and will not be repeated here.

[0130] Although the preferred embodiments of the present invention are described in detail above, it should be clearly understood that various modifications and variations are possible for those skilled in the art. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. The architecture system for intelligent production in digital factories is characterized by: include: The data perception layer is formed by a number of industrial data acquisition modules deployed on multiple production equipment in the digital factory. The industrial data acquisition module is configured to collect the equipment status information of the corresponding production equipment in real time, and adjust the display content of the control equipment of the digital factory in real time according to the equipment status information; wherein the equipment status information of the production equipment includes equipment operation mode, equipment operation status, spindle speed, spindle load and spindle temperature, and the equipment operation status includes operation status, standby status and abnormal status. At least one industrial data acquisition module is deployed corresponding to each production equipment; The edge computing layer is formed by multiple edge computing devices located at the edge of the network, and is configured to perform abnormal signal analysis and processing on the device status information sent by the data perception layer, perform scheduling optimization on the production plan in the workshop, and analyze the workshop environment; wherein the edge computing device is located in the vicinity of the industrial data generation source in a close distance, and the edge computing layer includes an abnormal detection and analysis module and a workshop scheduling and analysis module, the abnormal detection and analysis module is configured to receive the device status information sent by the data perception layer, and perform abnormal signal analysis and processing on the device status information; the workshop scheduling and analysis module is configured to perform real-time analysis and processing on the production information in the workshop to complete the scheduling optimization of the production plan; the production information in the workshop includes the production equipment resources currently available in the workshop and the production tasks in the production queue; The cloud computing layer, including the cloud computing center and multiple model nodes, is configured to perform multimodal data fusion according to the device status information provided by the edge computing layer and different downstream tasks, extract the intrinsic features of each modal data after multimodal data fusion, and then communicate the extracted intrinsic features with the model nodes for training. After the training is completed, the different downstream task models are incrementally updated according to the different extracted intrinsic features; The application layer implements the application of this architecture to different application scenarios by calling the service interface provided by the cloud computing layer.

2. The architecture system for digital factory intelligent production according to claim 1 is characterized in that: The process of real-time acquisition of equipment status information of corresponding production equipment by the industrial data acquisition module in the data perception layer includes the following steps: The industrial data acquisition module is networked when it is normally connected to the production equipment in the workshop to obtain the equipment status information of the corresponding production equipment; The device operating status is determined based on the acquired device status information, and the data transmission frequency is adjusted accordingly based on the device operating status. The device status information is sent to the edge server for analysis and processing according to the adjusted data transmission frequency corresponding to its operating status; wherein each device operating status has a corresponding data transmission frequency.

3. The architecture system for digital factory intelligent production according to claim 1 is characterized in that: The anomaly detection and analysis module in the edge computing layer performs abnormal signal analysis and processing in the following manner: Step a1: The anomaly detection and analysis module makes an anomaly analysis and judgment on the device status information sent by the data perception layer according to a preset anomaly detection model: When the abnormal detection and analysis module determines that an abnormal signal is obtained, the production equipment corresponding to the abnormal signal is made to enter an abnormal state, stop production work, and go to step a2; otherwise, the production equipment is made to continue production work; Step a2, the abnormality detection and analysis module analyzes the abnormality cause and abnormality type of the abnormal signal to obtain an abnormality analysis result; wherein the abnormality analysis result includes the abnormality cause and abnormality type of the abnormal signal; Step a3, the anomaly detection and analysis module takes corresponding anomaly handling measures according to the anomaly analysis results, and performs multi-index testing on the production equipment after the anomaly handling measures are taken; Step a4: The anomaly detection and analysis module makes a judgment based on the multi-index test results: When the production equipment successfully passes the multi-index test, the production equipment is restored to a normal state and the operating state of the production equipment is updated; otherwise, the multi-index test is re-executed on the production equipment until the production equipment successfully passes the multi-index test.

4. The architecture system for digital factory intelligent production according to claim 3 is characterized in that: The workshop scheduling analysis module in the edge computing layer performs real-time analysis and processing of production information in the workshop to complete the scheduling optimization process of the production plan, which includes the following steps: When the workshop scheduling analysis module determines that a new production task has appeared or the production equipment is in an abnormal state, it analyzes the current workshop environment and counts the available production equipment resources in the current workshop and the production tasks in the production queue; The workshop scheduling analysis module outputs a scheduling plan suitable for the current workshop conditions based on the statistically available production equipment resources in the current workshop and the production tasks in the production queue; The workshop scheduling analysis module verifies the feasibility of the scheduling plan based on the task priority of each production task in the scheduling plan and the production constraints corresponding to each production task: When the scheduling plan is feasible, each production equipment in the current workshop is ordered to execute the scheduling plan; otherwise, the scheduling plan is readjusted to obtain a new scheduling plan that is feasible.

5. The architecture system for digital factory intelligent production according to claim 4 is characterized in that: The process of analyzing the workshop environment by the edge computing layer includes the following steps: Step b1, traverse the operating status of all production equipment in the workshop, and calculate the distribution number of production equipment corresponding to each operating status; wherein the number of operating statuses of all production equipment in the workshop is marked as N, and the i-th operating status among the N operating statuses is marked as S i , in running state S i The corresponding distribution quantity of production equipment is marked as M Si : Among them, J represents the total number of production equipment in the workshop, s j represents the operating status of the jth production equipment in the workshop, δ(s j =S i ) is about the running state s j With running status S i The indicator function of Step b2, according to the running status of each production equipment obtained by traversal, determine the total number of production equipment in a faulty state, and calculate the equipment failure rate in the current workshop; wherein, the equipment failure rate in the workshop is calculated as follows: Among them, λ represents the equipment failure rate in the current workshop, S b Indicates that the production equipment is in a faulty state; Step b3, obtaining the operating time of each production equipment in different operating states, and calculating the operating state time proportion of each operating state corresponding to each production equipment; wherein the operating state time proportion is calculated as follows: Among them, f ji Indicates that the jth production equipment is in operating state S i The duration corresponding to the time, T j,i Indicates that the jth production equipment is in operating state S i The duration of time taken; Step b4, counting the production quality of each type of workpiece produced in the factory, calculating the number of high-quality workpieces of each type, and calculating the high-quality rate of each type of workpiece; wherein the high-quality rate of each type of workpiece is calculated as follows: Among them, e u represents the quality rate of the u-th type of workpiece, P u represents the total number of workpieces of type u produced in the workshop, p uv It represents the quality of the vth workpiece among the uth type of workpieces produced in the workshop; Step b5, count the production quality of each type of workpieces required for each production order, obtain the number of high-quality workpieces of each type, and calculate the production progress of the production tasks corresponding to each production order; the total number of production tasks received in the workshop is marked as K, and the production progress of the kth production task among the K production tasks is marked as r k : Among them, Q k represents the target production quantity of the kth production task, P u represents the total number of workpieces of type u produced in the workshop, p uv It represents the quality of the vth workpiece among the uth type of workpieces produced in the workshop; Step b6, counting the number of workpieces produced by the workshop for all types of workpieces every day, and calculating the workshop capacity of the workshop; where the workshop capacity of the workshop on the dth day is marked as c d : Among them, c d,u It represents the total number of workpieces of the uth type produced by the workshop on the dth day, and U represents the total number of types of workpieces produced by the workshop.

6. A method for implementing an architecture for intelligent production in a digital factory, which implements the architecture system for intelligent production in a digital factory as claimed in any one of claims 1 to 5, characterized in that: The steps include: Step 1: The data perception layer collects the equipment status information of the corresponding production equipment in the workshop in real time, and adjusts the display content of the control equipment of the digital factory in real time according to the equipment status information; wherein the equipment status information of the production equipment includes the equipment operation mode, equipment operation status, spindle speed, spindle load and spindle temperature, and the equipment operation status includes operation status, standby status and abnormal status, and each production equipment is deployed with at least one industrial data acquisition module; Step 2: The edge computing layer analyzes and processes abnormal signals of the equipment status information sent by the data perception layer, optimizes the production plan in the workshop, and analyzes the workshop environment; Step 3: The cloud computing layer performs multimodal data fusion according to the data provided by the edge computing layer and different downstream tasks, and extracts the intrinsic features of each modal data after multimodal data fusion, and then communicates the extracted intrinsic features with the model nodes for training. After the training is completed, the different downstream task models are incrementally updated according to the different extracted intrinsic features. Step 4: The application layer calls the service interface provided by the cloud computing layer to apply the architecture to different application scenarios.

7. The method for implementing the architecture for intelligent production in a digital factory according to claim 6, characterized in that: The process of the industrial data acquisition module in the data perception layer collecting the equipment status information of the corresponding production equipment in real time includes the following steps: The industrial data acquisition module is networked when it is normally connected to the production equipment in the workshop to obtain the equipment status information of the corresponding production equipment; The device operating status is determined based on the acquired device status information, and the data transmission frequency is adjusted accordingly based on the device operating status. The device status information is sent to the edge server for analysis and processing according to the adjusted data transmission frequency corresponding to its operating status; wherein each device operating status has a corresponding data transmission frequency.

8. The method for implementing the architecture for intelligent production in a digital factory according to claim 7, characterized in that: The anomaly detection and analysis module in the edge computing layer performs abnormal signal analysis and processing in the following manner: Step a1: The anomaly detection and analysis module makes an anomaly analysis and judgment on the device status information sent by the data perception layer according to a preset anomaly detection model: When the abnormal detection and analysis module determines that an abnormal signal is obtained, the production equipment corresponding to the abnormal signal is made to enter an abnormal state, stop production work, and go to step a2; otherwise, the production equipment is made to continue production work; Step a2, the abnormality detection and analysis module analyzes the abnormality cause and abnormality type of the abnormal signal to obtain an abnormality analysis result; wherein the abnormality analysis result includes the abnormality cause and abnormality type of the abnormal signal; Step a3, the anomaly detection and analysis module takes corresponding anomaly handling measures according to the anomaly analysis results, and performs multi-index testing on the production equipment after the anomaly handling measures are taken; Step a4: The anomaly detection and analysis module makes a judgment based on the multi-index test results: When the production equipment successfully passes the multi-index test, the production equipment is restored to a normal state and the operating state of the production equipment is updated; otherwise, the multi-index test is re-executed on the production equipment until the production equipment successfully passes the multi-index test; The workshop scheduling analysis module in the edge computing layer performs real-time analysis and processing of production information in the workshop to complete the scheduling optimization process of the production plan, which includes the following steps: When the workshop scheduling analysis module determines that a new production task has appeared or the production equipment is in an abnormal state, it analyzes the current workshop environment and counts the available production equipment resources in the current workshop and the production tasks in the production queue; The workshop scheduling analysis module outputs a scheduling plan suitable for the current workshop conditions based on the statistically available production equipment resources in the current workshop and the production tasks in the production queue; The workshop scheduling analysis module verifies the feasibility of the scheduling plan based on the task priority of each production task in the scheduling plan and the production constraints corresponding to each production task: When the scheduling plan is feasible, each production equipment in the current workshop is ordered to execute the scheduling plan; otherwise, the scheduling plan is readjusted to obtain a new scheduling plan that is feasible.

9. The method for implementing the architecture for intelligent production in a digital factory according to claim 8, characterized in that: The process of analyzing the workshop environment by the edge computing layer includes the following steps: Step b1, traverse the operating status of all production equipment in the workshop, and calculate the distribution number of production equipment corresponding to each operating status; wherein the number of operating statuses of all production equipment in the workshop is marked as N, and the i-th operating status among the N operating statuses is marked as S i , in running state S i The corresponding distribution quantity of production equipment is marked as M Si : Among them, J represents the total number of production equipment in the workshop, s j represents the operating status of the jth production equipment in the workshop, δ(s j =S i ) is about the running state s j With running status S i The indicator function of Step b2, according to the running status of each production equipment obtained by traversal, determine the total number of production equipment in a faulty state, and calculate the equipment failure rate in the current workshop; wherein, the equipment failure rate in the workshop is calculated as follows: Among them, λ represents the equipment failure rate in the current workshop, S b Indicates that the production equipment is in a faulty state; Step b3, obtaining the operating time of each production equipment in different operating states, and calculating the operating state time proportion of each operating state corresponding to each production equipment; wherein the operating state time proportion is calculated as follows: Among them, f ji Indicates that the jth production equipment is in operating state S i The duration corresponding to the time, T j,i Indicates that the jth production equipment is in operating state S i The duration of time taken; Step b4, counting the production quality of each type of workpiece produced in the factory, calculating the number of high-quality workpieces of each type, and calculating the high-quality rate of each type of workpiece; wherein the high-quality rate of each type of workpiece is calculated as follows: Among them, e u represents the quality rate of the u-th type of workpiece, P u represents the total number of workpieces of type u produced in the workshop, p uv It represents the quality of the vth workpiece among the uth type of workpieces produced in the workshop; Step b5, count the production quality of each type of workpieces required for each production order, obtain the number of high-quality workpieces of each type, and calculate the production progress of the production tasks corresponding to each production order; the total number of production tasks received in the workshop is marked as K, and the production progress of the kth production task among the K production tasks is marked as r k : Among them, Q k represents the target production quantity of the kth production task, P u represents the total number of workpieces of type u produced in the workshop, p uv It represents the quality of the vth workpiece among the uth type of workpieces produced in the workshop; Step b6, counting the number of workpieces produced by the workshop for all types of workpieces every day, and calculating the workshop capacity of the workshop; where the workshop capacity of the workshop on the dth day is marked as c d : Among them, c d,u It represents the total number of workpieces of the uth type produced by the workshop on the dth day, and U represents the total number of types of workpieces produced by the workshop.

Images