Industrial internet of things discrete manufacturing multi-source data real-time collaborative optimization method and system

Abstract

The invention discloses an industrial Internet of Things discrete manufacturing multi-source data real-time collaborative optimization method and system, and the method comprises the following steps: collecting equipment state data, process parameter data, material information data and order demand data in a discrete manufacturing process in real time through collection equipment in an industrial Internet of Things; performing cleaning, missing data complementation and format conversion on the acquired data to complete data preprocessing; the system has the beneficial effects that real-time acquisition and preprocessing of discrete manufacturing multi-source data are realized based on the industrial Internet of Things, various data in the production process can be quickly acquired and effectively processed, and a foundation is laid for subsequent collaborative optimization analysis; through data feature extraction and classification and establishment of the data association model, the internal relationship among the multi-source data is deeply mined, seamless joint and intelligent analysis of the multi-source data are realized, and compared with the prior art, the data relationship in the production process can be more accurately grasped.

Description

Technical Field

[0001] The present invention belongs to the field of discrete manufacturing technology, and specifically relates to a method and system for real-time collaborative optimization of multi-source data in discrete manufacturing of the industrial Internet of Things. Background Art

[0002] The Industrial Internet of Things is the in-depth application of Internet of Things technology in the industrial field. By deeply integrating physical equipment, sensors, network communications and data analysis technologies, it realizes the intelligence, automation and efficient collaboration of industrial production processes. Discrete manufacturing is an important branch of the manufacturing industry. Its production process involves multi-source heterogeneous data such as equipment status, process parameters, material information, order requirements, etc.; moreover, discrete manufacturing often faces small-batch, multi-variety production models, and order changes are frequent, which makes production process management extremely complex. In this context, it is crucial to achieve real-time collaborative optimization of multi-source data. Through real-time collaborative optimization, enterprises can dynamically adjust production plans and respond to changes in market demand in a timely manner. At the same time, real-time collaboration of multi-source data can also achieve full-process quality traceability and improve product quality.

[0003] In the prior art, for example, a multimodal industrial data fusion method and system for discrete manufacturing is disclosed in patent publication number CN118940220B. The patent discloses the following steps: real-time collection of discrete manufacturing industrial data and feature extraction to obtain feature data, inputting the feature data into a differentiable computer network to obtain a data fusion result, and the differentiable computer is also used to complete the prediction target based on the data fusion result; the discrete manufacturing industrial data includes video, text, image and sound data; the differentiable computer network includes a data fusion module and a memory system, the data fusion module fuses the feature data in the intelligent body, and the memory system is used to store and read and write the fused data of the intelligent body; the data in the memory system is input into the evaluation system to calculate the reward value, and the data fusion of the intelligent body is guided and the memory system is updated according to the reward value.

[0004] Although this patent can integrate discrete manufacturing industrial data and achieve prediction goals, it has shortcomings in real-time collaborative optimization of data. It cannot fully meet the needs of discrete manufacturing for seamless connection and intelligent analysis of multi-source data in complex and changeable production environments. It lacks the adaptive ability to automatically adjust analysis strategies according to different data characteristics. Data compatibility needs to be improved, and it is difficult to effectively improve the overall production efficiency of the enterprise. Summary of the Invention

[0005] The purpose of the present invention is to provide a method and system for real-time collaborative optimization of multi-source data in discrete manufacturing of the Industrial Internet of Things, to solve the problem in the existing technology that it is difficult to achieve real-time collaborative optimization of multi-source data in discrete manufacturing and cannot meet the needs of complex production environments, to achieve seamless docking and intelligent analysis of multi-source data, and to improve the overall production efficiency of the enterprise.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for real-time collaborative optimization of multi-source data in discrete manufacturing of industrial Internet of Things, comprising the following steps:

[0007] Through the collection equipment in the Industrial Internet of Things, the equipment status data, process parameter data, material information data and order demand data in the discrete manufacturing process are collected in real time; the collected data is cleaned, missing data is supplemented and format converted to complete data preprocessing;

[0008] A feature extraction algorithm is used to extract features from the preprocessed data to obtain feature vectors. Based on the data type and application scenario, the data is divided into equipment status, process parameter, material information, and order requirement categories.

[0009] Analyze the intrinsic connections between different categories of data, establish data association models based on machine learning algorithms, and train the models using historical data;

[0010] Input the real-time collected and pre-processed data into the data association model for real-time collaborative analysis; combine the preset production goals and constraints, and use optimization algorithms to dynamically adjust the production plan and process parameters;

[0011] Feedback the optimized results to the production execution link to monitor the actual execution of the production process; collect the difference information between the actual production data and the optimization results, and update and optimize the data association model.

[0012] As a preferred technical solution of the present invention, the missing data completion includes

[0013] Numerical data completion: A time series-based prediction model is used to divide time series data containing missing values ​​into training and test sets. The time step is used as a window, and the historical data in the window is used as the input of the LSTM model to predict the data at the missing location.

[0014] As a preferred technical solution of the present invention, categorical data completion: using a decision tree algorithm, with complete features as input variables and the categorical features where the missing data are located as output variables, a decision tree model is constructed; by training the model, the missing categorical data is predicted based on the values ​​of other features.

[0015] As a preferred technical solution of the present invention, the feature extraction includes equipment status data feature extraction, process parameter data feature extraction, material information data feature extraction, and order demand data feature extraction.

[0016] As a preferred technical solution of the present invention, the specific implementation method of establishing a data association model is as follows:

[0017] Use association rule mining algorithm to analyze the association relationship between different categories of data;

[0018] Build a graph neural network model, using equipment status data, process parameter data, material information data, and order demand data as nodes in the graph, and the associations between different types of data nodes as edges. The edge weights are initialized based on the confidence or correlation coefficient of the association rules.

[0019] Using historical production data as a training set, the data is input into the graph neural network model. The model parameters are updated through the backpropagation algorithm and stochastic gradient descent method. The objective function is to predict the correlation between different categories of data. The model is continuously optimized so that the model can accurately learn and predict the intrinsic connections between various types of data.

[0020] As a preferred technical solution of the present invention, an optimization algorithm is used to dynamically adjust the production plan and process parameters in combination with preset production goals and constraints. The specific implementation method is as follows:

[0021] Construct a mixed integer programming model with the objective functions of minimizing production costs and maximizing on-time delivery rates; the constraints include equipment capacity constraints, material supply constraints, and order priority constraints;

[0022] Use commercial solvers or heuristic algorithms to solve mixed integer programming models to obtain the optimal production plan adjustment plan, including the allocation of production tasks, the arrangement of production sequences, and the determination of production time. When an order changes, re-enter the changed order demand data and real-time equipment and material status data to quickly solve the model and update the production plan.

[0023] Using a reinforcement learning algorithm, the adjustable range of process parameters is used as the action space, and product quality indicators and production efficiency indicators are used as reward functions. The intelligent agent adjusts the process parameters during the production process, optimizes the strategy based on the reward feedback obtained, and learns the optimal process parameter adjustment plan.

[0024] The present invention also discloses an industrial Internet of Things discrete manufacturing multi-source data real-time collaborative optimization system, including

[0025] The data acquisition module is used to collect equipment status data, process parameter data, material information data, and order demand data in the discrete manufacturing process in real time through the acquisition devices in the industrial Internet of Things, and transmit the collected data to the data preprocessing module;

[0026] The data preprocessing module is used to receive the data transmitted by the data acquisition module, clean it, fill in missing data and convert the format. After completing the data preprocessing, the data is transmitted to the data feature extraction module;

[0027] The data feature extraction module is used to extract features from the data processed by the data preprocessing module using a feature extraction algorithm to obtain feature vectors, and transmit the feature vectors to the data classification module;

[0028] The data classification module is used to classify the data corresponding to the feature vector transmitted by the data feature extraction module into equipment status, process parameter, material information and order requirement categories according to the data type and application scenario, and transmit the classified data to the data association model establishment module;

[0029] The data association model building module is used to analyze the intrinsic connections between different types of data, build a data association model based on machine learning algorithms, and train the model using historical data; the trained data association model is transmitted to the real-time collaborative optimization analysis module;

[0030] The real-time collaborative optimization and analysis module is used to receive the real-time data transmitted by the data classification module and the data association model transmitted by the data association model establishment module, input the real-time data into the data association model for real-time collaborative analysis; based on the preset production goals and constraints, it uses the optimization algorithm to dynamically adjust the production plan and process parameters, and outputs the optimization results to the result feedback module;

[0031] The result feedback module is used to feed back the optimization results output by the real-time collaborative optimization analysis module to the production execution link, while monitoring the actual execution of the production process, collecting the difference information between the actual production data and the optimization results, and transmitting the difference information to the model update module;

[0032] The model updating module is used to update and optimize the data association model established by the data association model establishing module according to the difference information transmitted by the result feedback module.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The present invention realizes the real-time collection and preprocessing of multi-source data in discrete manufacturing based on the Industrial Internet of Things. It can quickly obtain various types of data in the production process and effectively process them, laying the foundation for subsequent collaborative optimization analysis.

[0035] By extracting and classifying data features and establishing data association models, we can deeply explore the inherent connections between multi-source data, achieve seamless integration and intelligent analysis of multi-source data, and more accurately grasp the data relationships in the production process compared to existing technologies.

[0036] The real-time collaborative optimization analysis link dynamically adjusts production plans and process parameters based on production goals and constraints, enabling enterprises to respond promptly to production changes such as order changes, effectively improving the overall production efficiency of the enterprise;

[0037] The result feedback and model update mechanism ensures that the model can continuously adapt to changes in the production process, continuously improve the accuracy and effectiveness of data collaborative optimization, and enhance the system's adaptability and data compatibility. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a flow chart of the method for real-time collaborative optimization of multi-source data in discrete manufacturing of the industrial Internet of Things of the present invention;

[0039] Figure 2 This is a schematic diagram of the industrial Internet of Things discrete manufacturing multi-source data real-time collaborative optimization system of the present invention. DETAILED DESCRIPTION

[0040] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0041] Example 1

[0042] See also Figure 1 , which is the first embodiment of the present invention, provides a method for real-time collaborative optimization of multi-source data in discrete manufacturing of the industrial Internet of Things, including the following steps:

[0043] Through the collection equipment in the industrial Internet of Things, multi-source heterogeneous data such as equipment status data, process parameter data, material information data and order demand data in the discrete manufacturing process are collected in real time; the collected data are cleaned, missing data are supplemented and format converted to complete data preprocessing; for structured data, such as equipment status data tables, the hash value of each data record is calculated by the hash algorithm, and the data records with the same hash value are extracted, and the field content is further compared. If they are completely consistent, they are regarded as duplicate data and deleted; for unstructured data, such as material information text, a text similarity calculation algorithm (such as cosine similarity) is used, and a similarity threshold (such as 0.9) is set. If the similarity between two texts exceeds the threshold, the one with more complete information is retained and the remaining duplicate texts are deleted; using statistical analysis methods, for numerical data, statistical quantities such as the mean and standard deviation of the data are calculated, and data outside the range of ±3 times the standard deviation of the mean is regarded as abnormal error data; for equipment For Boolean data of operating status, if a state combination that does not conform to the normal operation logic of the equipment occurs (such as the equipment being in both the operating and closed states at the same time) is detected, it will be judged as erroneous data and will be directly discarded or corrected according to the equipment operation log. A data mapping table is used to define the correspondence between fields of different data sources and fields of a unified format. For example, the "equipment operating speed" field collected by sensors from different manufacturers in the equipment status data is uniformly mapped to the "operating rate" field in the standard format, and unit conversion is performed (such as converting km / h to m / s). For text data of material information, natural language processing technology is used to extract text information into structured key information, such as extracting material names, specifications, models, and other information through named entity recognition technology, and then storing and transmitting it in a unified JSON format. For production drawing data in image format, image recognition algorithms are used to extract key dimensions, technical parameters, and other information in the drawings and convert them into a structured data format.

[0044] A feature extraction algorithm is used to extract features from the preprocessed data to obtain feature vectors. Based on the data type and application scenario, the data is divided into different categories such as equipment status, process parameters, material information, and order requirements.

[0045] Analyze the intrinsic connections between different categories of data, establish data association models based on machine learning algorithms, and train the models using historical data;

[0046] Input the real-time collected and pre-processed data into the data association model for real-time collaborative analysis; combine the preset production goals and constraints, and use optimization algorithms to dynamically adjust production plans, process parameters, etc.

[0047] Feedback the optimized results to the production execution link to monitor the actual execution of the production process; collect the difference information between the actual production data and the optimization results, and update and optimize the data association model.

[0048] In this embodiment, preferably, missing data completion includes

[0049] Numerical data completion: A time series-based prediction model is used to divide time series data containing missing values ​​into training and test sets. The time step is used as a window, and the historical data in the window is used as the input of the LSTM model to predict the data at the missing location.

[0050] Categorical data completion: Use the decision tree algorithm to build a decision tree model with complete features as input variables and the categorical features where the missing data is located as output variables; by training the model, predict the missing categorical data based on the values ​​of other features.

[0051] In this embodiment, preferably, feature extraction includes

[0052] Device status data feature extraction:

[0053] Time Domain Feature Extraction: Calculates statistics such as the mean, variance, maximum, minimum, and peak-to-peak values ​​of equipment operating parameters (such as current and voltage) to reflect the basic operating status of the equipment; extracts waveform features such as rise time, fall time, and overshoot to analyze the dynamic performance of the equipment during startup and shutdown.

[0054] Frequency domain feature extraction: Perform fast Fourier transform (FFT) on time domain data such as equipment vibration signals to obtain the frequency components of the signal, calculate the amplitude and phase of each frequency component, extract key frequency features such as the primary frequency and secondary frequency, and determine whether the equipment has abnormal vibration frequencies, such as specific frequency vibrations caused by bearing failures.

[0055] Time-frequency domain feature extraction: Using wavelet transform, the device status data is converted from the time domain to the time-frequency domain, the signal energy distribution at different times and frequencies is analyzed, and the wavelet coefficients are obtained as feature vectors, which can more clearly capture the transient changes in the device operating status;

[0056] Process parameter data feature extraction:

[0057] The principal component analysis (PCA) method is used to reduce the dimensionality of high-dimensional data composed of multiple process parameters (such as temperature, pressure, flow rate, etc.). The covariance matrix of the data is calculated, the eigenvalues ​​and eigenvectors are solved, and the first few principal components with the largest variance contribution are selected as eigenvectors. This method reduces the data dimension while retaining the main information, facilitating subsequent analysis and processing.

[0058] Material information data feature extraction:

[0059] For text information such as material names and specifications, word embedding technology, such as the Word2Vec model, is used to map each word into a low-dimensional vector. The feature vector of the material text is then obtained by summing or averaging the vectors. For categorical information such as material batches and suppliers, one-hot encoding or label encoding is used to convert the categorical information into a numerical feature vector.

[0060] Order demand data feature extraction:

[0061] Time-related feature extraction: Extract timestamp information such as order placement time and delivery time, calculate features such as the remaining delivery time of the order and the waiting time of the order in the production queue, and reflect the urgency and production priority of the order;

[0062] Quantity-related feature extraction: Calculate features such as the number of ordered products and order batch size to analyze the impact of order size on production plans; extract trend characteristics of order quantity changes, such as the growth or decline rate of recent order quantities, to assist in predicting future order demand.

[0063] In this embodiment, preferably, the specific implementation method of establishing the data association model is as follows:

[0064] Use association rule mining algorithm to analyze the association relationship between different categories of data;

[0065] Build a graph neural network model, using equipment status data, process parameter data, material information data, and order demand data as nodes in the graph, and the associations between different types of data nodes as edges. The edge weights are initialized based on the confidence or correlation coefficient of the association rules.

[0066] Using historical production data as a training set, the data is input into the graph neural network model. The model parameters are updated through the backpropagation algorithm and stochastic gradient descent method. The objective function is to predict the correlation between different categories of data. The model is continuously optimized so that the model can accurately learn and predict the intrinsic connections between various types of data.

[0067] The association rule mining algorithm is the Apriori algorithm.

[0068] In this embodiment, preferably, the production plan and process parameters are dynamically adjusted using an optimization algorithm in combination with preset production goals and constraints. The specific implementation method is as follows:

[0069] Construct a mixed integer programming model with the objective functions of minimizing production costs and maximizing on-time delivery rates; the constraints include equipment capacity constraints, material supply constraints, and order priority constraints;

[0070] Use commercial solvers or heuristic algorithms to solve mixed integer programming models to obtain the optimal production plan adjustment plan, including the allocation of production tasks, the arrangement of production sequences, and the determination of production time. When an order changes, re-enter the changed order demand data and real-time equipment and material status data to quickly solve the model and update the production plan.

[0071] Using a reinforcement learning algorithm, the adjustable range of process parameters is used as the action space, and product quality indicators and production efficiency indicators are used as reward functions. The intelligent agent adjusts the process parameters during the production process, optimizes the strategy based on the reward feedback obtained, and learns the optimal process parameter adjustment plan.

[0072] Heuristic algorithms include genetic algorithms and particle swarm optimization algorithms.

[0073] In this embodiment, preferably,.

[0074] Example 2

[0075] See also Figure 2 , which is the second embodiment of the present invention, and provides an industrial Internet of Things discrete manufacturing multi-source data real-time collaborative optimization system, including

[0076] The data acquisition module is used to collect equipment status data, process parameter data, material information data, and order demand data in the discrete manufacturing process in real time through the acquisition devices in the industrial Internet of Things, and transmit the collected data to the data preprocessing module;

[0077] The data preprocessing module is used to receive the data transmitted by the data acquisition module, clean it, fill in missing data and convert the format. After completing the data preprocessing, the data is transmitted to the data feature extraction module;

[0078] The data feature extraction module is used to extract features from the data processed by the data preprocessing module using a feature extraction algorithm to obtain feature vectors, and transmit the feature vectors to the data classification module;

[0079] The data classification module is used to classify the data corresponding to the feature vector transmitted by the data feature extraction module into equipment status, process parameter, material information and order requirement categories according to the data type and application scenario, and transmit the classified data to the data association model establishment module;

[0080] The data association model building module is used to analyze the intrinsic connections between different types of data, build a data association model based on machine learning algorithms, and train the model using historical data; the trained data association model is transmitted to the real-time collaborative optimization analysis module;

[0081] The real-time collaborative optimization and analysis module is used to receive the real-time data transmitted by the data classification module and the data association model transmitted by the data association model establishment module, input the real-time data into the data association model for real-time collaborative analysis; based on the preset production goals and constraints, it uses the optimization algorithm to dynamically adjust the production plan and process parameters, and outputs the optimization results to the result feedback module;

[0082] The result feedback module is used to feed back the optimization results output by the real-time collaborative optimization analysis module to the production execution link, while monitoring the actual execution of the production process, collecting the difference information between the actual production data and the optimization results, and transmitting the difference information to the model update module;

[0083] The model updating module is used to update and optimize the data association model established by the data association model establishing module according to the difference information transmitted by the result feedback module.

[0084] Although the embodiments of the present invention have been shown and described, as detailed above, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the invention, and the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A real-time collaborative optimization method for multi-source data in discrete manufacturing of the Industrial Internet of Things, characterized by: The steps include: Through the collection equipment in the Industrial Internet of Things, the equipment status data, process parameter data, material information data and order demand data in the discrete manufacturing process are collected in real time; the collected data is cleaned, missing data is supplemented and format converted to complete data preprocessing; A feature extraction algorithm is used to extract features from the preprocessed data to obtain feature vectors. Based on the data type and application scenario, the data is divided into equipment status, process parameter, material information, and order requirement categories. Analyze the intrinsic connections between different categories of data, establish data association models based on machine learning algorithms, and train the models using historical data; Input the real-time collected and pre-processed data into the data association model for real-time collaborative analysis; combine the preset production goals and constraints, and use optimization algorithms to dynamically adjust the production plan and process parameters; Feedback the optimized results to the production execution link to monitor the actual execution of the production process; Collect the difference information between actual production data and optimization results, and update and optimize the data association model.

2. The method for real-time collaborative optimization of multi-source data in discrete manufacturing of the Industrial Internet of Things according to claim 1 is characterized by: The missing data completion includes Numerical data completion: A time series-based prediction model is used to divide time series data containing missing values ​​into training and test sets. The time step is used as a window, and the historical data in the window is used as the input of the LSTM model to predict the data at the missing location. Categorical data completion: Use the decision tree algorithm to build a decision tree model with complete features as input variables and the categorical features where the missing data is located as output variables; by training the model, predict the missing categorical data based on the values ​​of other features.

3. The method for real-time collaborative optimization of multi-source data in discrete manufacturing of the Industrial Internet of Things according to claim 1 is characterized by: The feature extraction includes equipment status data feature extraction, process parameter data feature extraction, material information data feature extraction, and order demand data feature extraction.

4. The method for real-time collaborative optimization of multi-source data in discrete manufacturing of the Industrial Internet of Things according to claim 1 is characterized by: The specific implementation method of establishing a data association model is as follows: Use association rule mining algorithm to analyze the association relationship between different categories of data; Build a graph neural network model, using equipment status data, process parameter data, material information data, and order demand data as nodes in the graph, and the associations between different types of data nodes as edges. The edge weights are initialized based on the confidence or correlation coefficient of the association rules. Using historical production data as a training set, the data is input into the graph neural network model. The model parameters are updated through the backpropagation algorithm and stochastic gradient descent method. The objective function is to predict the correlation between different categories of data. The model is continuously optimized so that the model can accurately learn and predict the intrinsic connections between various types of data.

5. The method for real-time collaborative optimization of multi-source data in discrete manufacturing of industrial Internet of Things according to claim 1 is characterized by: The association rule mining algorithm is the Apriori algorithm.

6. The method for real-time collaborative optimization of multi-source data in discrete manufacturing of the Industrial Internet of Things according to claim 1 is characterized by: Combined with the preset production goals and constraints, the optimization algorithm is used to dynamically adjust the production plan and process parameters. The specific implementation method is as follows: Construct a mixed integer programming model with the objective functions of minimizing production costs and maximizing on-time delivery rates; the constraints include equipment capacity constraints, material supply constraints, and order priority constraints; Use commercial solvers or heuristic algorithms to solve mixed integer programming models to obtain the optimal production plan adjustment plan, including the allocation of production tasks, the arrangement of production sequences, and the determination of production time. When an order changes, re-enter the changed order demand data and real-time equipment and material status data to quickly solve the model and update the production plan. Using a reinforcement learning algorithm, the adjustable range of process parameters is used as the action space, and product quality indicators and production efficiency indicators are used as reward functions. The intelligent agent adjusts the process parameters during the production process, optimizes the strategy based on the reward feedback obtained, and learns the optimal process parameter adjustment plan.

7. The method for real-time collaborative optimization of multi-source data in discrete manufacturing of the Industrial Internet of Things according to claim 6 is characterized by: The heuristic algorithm includes a genetic algorithm and a particle swarm optimization algorithm.

8. An industrial internet of things (IIoT) discrete manufacturing multi-source data real-time collaborative optimization system applied to the IIoT discrete manufacturing multi-source data real-time collaborative optimization method according to any one of claims 1 to 7, characterized in that: include The data acquisition module is used to collect equipment status data, process parameter data, material information data, and order demand data in the discrete manufacturing process in real time through the acquisition devices in the industrial Internet of Things, and transmit the collected data to the data preprocessing module; The data preprocessing module is used to receive the data transmitted by the data acquisition module, clean it, fill in missing data and convert the format. After completing the data preprocessing, the data is transmitted to the data feature extraction module; The data feature extraction module is used to extract features from the data processed by the data preprocessing module using a feature extraction algorithm to obtain feature vectors, and transmit the feature vectors to the data classification module; The data classification module is used to classify the data corresponding to the feature vector transmitted by the data feature extraction module into equipment status, process parameter, material information and order requirement categories according to the data type and application scenario, and transmit the classified data to the data association model establishment module; The data association model building module is used to analyze the intrinsic connections between different types of data, build a data association model based on machine learning algorithms, and train the model using historical data; the trained data association model is transmitted to the real-time collaborative optimization analysis module; The real-time collaborative optimization and analysis module is used to receive the real-time data transmitted by the data classification module and the data association model transmitted by the data association model establishment module, input the real-time data into the data association model for real-time collaborative analysis; based on the preset production goals and constraints, it uses the optimization algorithm to dynamically adjust the production plan and process parameters, and outputs the optimization results to the result feedback module; The result feedback module is used to feed back the optimization results output by the real-time collaborative optimization analysis module to the production execution link, while monitoring the actual execution of the production process, collecting the difference information between the actual production data and the optimization results, and transmitting the difference information to the model update module; The model updating module is used to update and optimize the data association model established by the data association model establishing module according to the difference information transmitted by the result feedback module.

Images