Factory data intelligent analysis and management system based on digital-intelligent integration

Abstract

The invention discloses an intelligent factory data analysis and management system based on digital-intelligent integration, which belongs to the technical field of digital-intelligent integration and comprises a data acquisition and preprocessing module, a situation awareness module, an analog simulation model module, an intelligent analysis module, an intelligent prediction and diagnosis module, a dynamic decision module and a self-adaptive optimization scheduling module. According to the invention, data of the data acquisition and preprocessing module and the context awareness module are organically combined, the time-varying feature matrix is extracted, dynamic evaluation and update selection are carried out on features, and deep influence factors can be mined out when association of production data and environment data is analyzed, so that the production efficiency is improved. Through an abnormal score, a trend prediction result and a virtual simulation model state of an intelligent analysis module, and through calculating a risk factor and a composite risk score, when an equipment fault is predicted, real-time operation data and historical fault data of the equipment and performance simulation in virtual simulation are comprehensively considered, and a potential fault hidden danger is found in advance.

Description

Technical Field

[0001] The present invention belongs to the technical field of digital and intelligent integration, and specifically refers to a factory data intelligent analysis and management system based on digital and intelligent integration. Background Art

[0002] With the rapid development of industry, traditional manufacturing is transforming towards intelligence. Among them, data-driven decision support systems have become key factors in improving production efficiency and service quality. However, existing factory data analysis systems often have limitations such as serious data island phenomena, slow model updates, and difficulty in adapting to rapidly changing market demands.

[0003] However, there are still certain defects in the existing factory data intelligent analysis and management. The existing factory data intelligent analysis and management rely on simple statistical analysis or predefined rules to process and analyze data, lacking the ability to deeply analyze real-time data and context-aware results, not comprehensively discovering hidden patterns and trends, difficult to timely detect abnormal situations in the production process, lacking the ability to comprehensively consider real-time operation data of equipment, historical fault data, and virtual simulation performance simulation, resulting in low accuracy in predicting equipment failures. Therefore, a factory data intelligent analysis and management system based on digital and intelligent integration is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a factory data intelligent analysis and management system based on digital and intelligent integration to solve the problems raised in the above background art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: A factory data intelligent analysis and management system based on digital and intelligent integration, including a data acquisition and preprocessing module, a context awareness module, a simulation model module, an intelligent analysis module, an intelligent prediction and diagnosis module, a dynamic decision-making module, and an adaptive optimization and scheduling module;

[0006] The data acquisition and preprocessing module is used to obtain factory data from sensors and equipment and perform preprocessing;

[0007] The context awareness module is used to obtain real-time environmental information in the factory according to integrated sensors and monitoring devices;

[0008] The simulation model module is used to construct a virtual simulation model to simulate the factory production process according to the results of the data acquisition and preprocessing module and the context awareness module;

[0009] The intelligent analysis module is used to deeply analyze the data information of the data acquisition and preprocessing module and the context awareness module to discover hidden patterns and trends;

[0010] The intelligent prediction and diagnosis module is used to predict potential problems by combining the results of the intelligent analysis module and the information of the virtual simulation model;

[0011] The dynamic decision-making module is used to make decisions based on the predicted diagnosis results combined with the simulation information of the simulation model;

[0012] The adaptive optimization scheduling module is used to automatically adjust the resource configuration and operation parameter configuration according to the made decisions.

[0013] The data acquisition and preprocessing module is wirelessly connected to the simulation model module and the intelligent analysis module. The intelligent analysis module is wirelessly connected to the situation awareness module and the intelligent prediction and diagnosis module. The situation awareness module is wirelessly connected to the simulation model module. The intelligent prediction and diagnosis module is wirelessly connected to the dynamic decision-making and simulation model modules. The dynamic decision-making module is wirelessly connected to the simulation model module and the adaptive optimization scheduling module. The adaptive optimization scheduling module is wirelessly connected to the real-time feedback control module.

[0014] Among them, the data acquisition and preprocessing module sets the acquisition parameters of each data source through sensors and devices in the factory, including the acquisition frequency, acquisition time range, and data accuracy. It grabs data in real time from each sensor and device according to the set acquisition strategy, and performs data cleaning, data conversion, and data screening on the real-time grabbed data. Data cleaning includes removing outliers, filling missing values, and deduplicating. The preprocessed data is stored in the database.

[0015] Among them, the situation awareness module obtains the environmental information in the factory in real time according to the integrated sensors and monitoring devices; it collects the factory environmental data information in real time through the integrated sensors and monitoring devices, defines the rules and trigger conditions of different environmental parameters, analyzes the relationships between different environmental parameters, constructs a situation model, identifies and classifies different working situations according to the real-time environmental parameters through the situation model, and automatically adjusts the device parameter information according to the result information of the situation model.

[0016] Among them, the simulation model module constructs a virtual simulation model to simulate the factory production process according to the results of the data acquisition and preprocessing module and the situation awareness module; according to the data information obtained in real time by the data acquisition and preprocessing and the environmental information monitored in real time in the situation awareness, according to the factory layout and production process, constructs a virtual simulation model in the virtual simulation software, imports the device models in the factory into the virtual simulation software, and performs configuration and calibration. The virtual simulation model dynamically updates the state of the virtual simulation model in real time according to the data information of the data acquisition and preprocessing module and the situation awareness module. The virtual simulation model simulates according to the set scenarios and records various index information generated during the simulation process, including factory data changes and device status information.

[0017] Among them, the intelligent analysis module conducts in-depth analysis based on the data information of the data acquisition and preprocessing module and the context awareness module to discover hidden patterns and trends. Let the data of the data preprocessing module be d pc (T(t)), and the data of the context awareness module be d cx (T(t)). The data of the data acquisition and preprocessing module and the context awareness module are fused, and the implementation formula is:

[0018] D fn (t) = λ1d pc (T(t)) + λ2d cx (T(t)),

[0019] In the formula, D fn (t) represents the fused data at time t, and λ1 and λ2 represent weight coefficients;

[0020] Let the time-varying feature matrix be F(t) = [f1(t), f2(t),..., f i (t)], F(t) represents the feature matrix extracted at time t, and f i (t) represents the i-th feature extracted at time t, and f i (t) is extracted from the fused data D fn (t), and the implementation formula is:

[0021] f i (t) = g i (D fn (t)),

[0022] In the formula, g i represents the feature extraction function, and D fn (t) represents the fused data at time t;

[0023] The importance of the extracted features is dynamically evaluated, and the implementation formula is:

[0024]

[0025] In the formula, I(f i , t) represents the importance of feature f i at time t, L(F(t)) represents the loss function, and the evaluated features are updated and selected, which is implemented as:

[0026] S(t) = {f i |I(f i , t) > θ(t)},

[0027] In the formula, S(t) represents the features selected at time t, and θ(t) represents the updated features.

[0028] Among them, anomaly detection is performed according to the selected feature S(t), and the implementation formula is:

[0029] h t = σ(W h x t + U h h t-1 + b n ),

[0030] In the formula, h t represents the hidden state at time t, x t represents the feature S(t) input at time t, W h and U h represent weight matrices, representing the input feature and the hidden state at the previous moment respectively, b h represents the bias term, and σ represents the activation function; anomaly scoring is performed according to the hidden state, and the implementation is:

[0031]

[0032] In the formula, A t represents the anomaly score at time t, represents the predicted hidden state at time t.

[0033] Among them, it is assumed that comprehensive prediction is performed according to multiple-dimensional trends. Let the predicted value of the d-th dimension at time t be The implementation formula is:

[0034]

[0035] In the formula, represents the observed value of the d-th dimension in the past p steps, and represent the autoregressive coefficient and the moving average coefficient of the d-th dimension, represents the error term of the d-th dimension, and combining the prediction results of multiple dimensions generates the final trend prediction result, and the implementation formula is:

[0036]

[0037] In the formula, represents the final trend prediction result, D represents the number of dimensions, w d (A t ) represents the weights of each dimension, represents the predicted value of the d-th dimension at time t.

[0038] Among them, the intelligent prediction and diagnosis module predicts potential problems according to the results of the intelligent analysis module and the virtual simulation model information; let the state of the virtual simulation model be Ssim (t), fuse the anomaly score A of the intelligent analysis module t , the trend prediction result and the virtual simulation model state to obtain the multi-dimensional feature matrix F dw (t). The implementation formula is as follows:

[0039]

[0040] According to the confidence level C t of the anomaly score A t (A t ) and the reliability R t (S sim (t)) of the simulation state, dynamically adjust the weights of each feature. The implementation formula is as follows:

[0041]

[0042] In the formula, w i (t) represents the weight of the i-th feature at time t, and λ and μ respectively represent the sensitivity parameters that control the influence of confidence level and reliability on the weight;

[0043] Perform feature weighting according to the updated weights. The implementation formula is as follows:

[0044] F w (t) = W(t) · F dw (t),

[0045] In the formula, F w (t) represents the weighted weight, and W(t) represents the diagonal matrix composed of w i (t).

[0046] The intelligent prediction and diagnosis module calculates multiple risk factors according to the weighted feature matrix F w (t). The implementation formula is as follows:

[0047]

[0048] In the formula, R f (t) represents the risk factor, represents the weighted trend prediction value, extracted from F w (t), represents the weighted trend prediction value The change rate within the unit time Δt, S sim (t) w represents the weighted simulation state value, extracted from F w (t), S real (t) represents the actual observed state, and perform a composite risk score R c(t), set the warning threshold, and trigger a warning when the composite risk score R c (t) exceeds the threshold.

[0049] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0050] 1. Through the intelligent analysis module of the present invention, by using data analysis technology, deeply analyze the real-time data and the results of context awareness, discover hidden patterns and trends, and can perform anomaly detection to timely discover abnormal situations in the production process, which not only improves the decision-making efficiency of the enterprise and reduces risks;

[0051] 2. Through anomaly detection of the present invention, timely discover abnormal situations in the production process, provide timely and effective warnings for the decision-making of the enterprise. Through multi-dimensional feature fusion and dynamic weight adjustment, it can accurately identify potential risk points in the production process, organically combine the data of the data acquisition and preprocessing module and the context awareness module, extract the time-varying feature matrix, and dynamically evaluate and update the selection of features. When analyzing the correlation between production data and environmental data, it can dig out deep influencing factors;

[0052] 3. Through the anomaly score, trend prediction results of the intelligent analysis module and the status of the virtual simulation model of the present invention, by calculating the risk factor and the composite risk score, when predicting equipment failures, comprehensively consider the real-time operation data of the equipment, historical failure data and performance simulation in the virtual simulation to discover potential failure hazards in advance;

[0053] 4. Through the dynamic decision-making module of the present invention, make decisions according to the results of the intelligent prediction and diagnosis module and the information of the simulation model to ensure the scientificity and timeliness of the decisions. The adaptive optimization scheduling module can automatically adjust the resource allocation and operation parameters according to the decisions to realize the real-time optimization of the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 It is a schematic structural diagram of the factory data intelligent analysis and management system based on digital and intelligent integration of the present invention;

[0055] Figure 2 It is the operation process of the factory data intelligent analysis and management system based on digital and intelligent integration of the present invention Figure 1 ;

[0056] Figure 3 It is the operation process of the factory data intelligent analysis and management system based on digital and intelligent integration of the present invention Figure 2 ;

[0057] Figure 4 It is the operation process of the factory data intelligent analysis and management system based on digital and intelligent integration of the present invention Figure 3 . DETAILED DESCRIPTION OF THE INVENTION

[0058] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0059] Embodiment

[0060] Please refer to Figures 1 - 4 As shown, the present invention provides a technical solution: including a data acquisition and preprocessing module, a context awareness module, a simulation model module, an intelligent analysis module, an intelligent prediction and diagnosis module, a dynamic decision-making module, and an adaptive optimization and scheduling module;

[0061] The data acquisition and preprocessing module is used to obtain factory data from sensors and devices and perform preprocessing;

[0062] The context awareness module is used to obtain the environmental information in the factory in real time according to the integrated sensors and monitoring devices;

[0063] The simulation model module is used to build a virtual simulation model according to the results of the data acquisition and preprocessing module and the context awareness module to simulate the factory production process;

[0064] The intelligent analysis module is used to perform in-depth analysis on the data information of the data acquisition and preprocessing module and the context awareness module to discover hidden patterns and trends;

[0065] The intelligent prediction and diagnosis module is used to predict potential problems by combining the results of the intelligent analysis module and the information of the virtual simulation model;

[0066] The dynamic decision-making module is used to make decisions according to the prediction and diagnosis results combined with the simulation information of the simulation model;

[0067] The adaptive optimization and scheduling module is used to automatically adjust the resource configuration and operation parameter configuration according to the made decisions.

[0068] The data acquisition and preprocessing module is wirelessly connected to the simulation model module and the intelligent analysis module. The intelligent analysis module is wirelessly connected to the context awareness module and the intelligent prediction and diagnosis module. The context awareness module is wirelessly connected to the simulation model module. The intelligent prediction and diagnosis module is wirelessly connected to the dynamic decision-making and simulation model modules. The dynamic decision-making module is wirelessly connected to the simulation model module and the adaptive optimization and scheduling module. The adaptive optimization and scheduling module is wirelessly connected to the real-time feedback control module.

[0069] Among them, the data acquisition and preprocessing module sets the acquisition parameters of each data source through sensors and devices in the factory, including the acquisition frequency, acquisition time range, and data accuracy. It grabs data in real time from each sensor and device according to the set acquisition strategy, and performs data cleaning, data conversion, and data screening on the real-time grabbed data. Data cleaning includes removing outliers, filling in missing values, and deduplicating. The preprocessed data is stored in a database.

[0070] Among them, the situation awareness module obtains the environmental information in the factory in real time according to the integrated sensors and monitoring devices; collects the factory environmental data information in real time through the integrated sensors and monitoring devices, defines the rules and trigger conditions of different environmental parameters, analyzes the relationships between different environmental parameters, constructs a situation model, identifies and classifies different working situations according to the real-time environmental parameters through the situation model, and automatically adjusts the device parameter information according to the result information of the situation model.

[0071] Among them, the simulation model module constructs a virtual simulation model to simulate the factory production process according to the results of the data acquisition and preprocessing module and the situation awareness module; according to the data information obtained in real time by the data acquisition and preprocessing and the environmental information monitored in real time in the situation awareness, constructs a virtual simulation model in the virtual simulation software according to the factory layout and production process, imports the device models in the factory into the virtual simulation software, and performs configuration and calibration. The virtual simulation model dynamically updates the state of the virtual simulation model in real time according to the data information of the data acquisition and preprocessing module and the situation awareness module. The virtual simulation model simulates according to the set scenarios and records various index information generated during the simulation process, including factory data changes and device status information.

[0072] Among them, the intelligent analysis module conducts in-depth analysis according to the data information of the data acquisition and preprocessing module and the situation awareness module to discover hidden patterns and trends; let the data of the data preprocessing module be d pc (T(t)), and the data of the situation awareness module be d cx (T(t)). The data of the data acquisition and preprocessing module and the situation awareness module are fused, and the implementation formula is:

[0073] D fn (t) = λ1d pc (T(t)) + λ2d cx (T(t)),

[0074] In the formula, D fn (t) represents the fused data at time t, and λ1 and λ2 represent weight coefficients;

[0075] Let the time-varying feature matrix be F(t) = [f1(t), f2(t),..., f i(t)], where F(t) represents the feature matrix extracted at time t, and f i (t) represents the i-th feature extracted at time t, and f i (t) extracts features from the fused data D fn (t). The implementation formula is:

[0076] f i (t) = g i (D fn (t)),

[0077] In the formula, g i represents the feature extraction function, and D fn (t) represents the fused data at time t;

[0078] Based on the extracted features, dynamic importance assessment is performed. The implementation formula is:

[0079]

[0080] In the formula, I(f i , t) represents the importance of feature f i at time t, L(F(t)) represents the loss function. After evaluating the features, update and selection are performed, and the implementation is:

[0081] S(t) = {f i | I(f i , t) > θ(t)},

[0082] In the formula, S(t) represents the features selected at time t, and θ(t) represents the updated features.

[0083] Among them, based on the selected features S(t), anomaly detection is performed. The implementation formula is:

[0084] h t = σ(W h x t + U h h t-1 + b n ),

[0085] In the formula, h t represents the hidden state at time t, x t represents the feature S(t) input at time t, W h and U h represent weight matrices, representing the input feature and the hidden state at the previous moment respectively, b h represents the bias term, and σ represents the activation function; based on the hidden state, anomaly scoring is performed, and the implementation is:

[0086]

[0087] In the formula, A t represents the anomaly score at time t, and represents the predicted hidden state at time t.

[0088] Among them, assuming comprehensive prediction is performed based on multiple-dimensional trends, and the predicted value of the d-th dimension at time t is The implementation formula is:

[0089]

[0090] In the formula, represents the observed value of the d-th dimension in the past p steps, and represent the autoregressive coefficient and the moving average coefficient of the d-th dimension, represents the error term of the d-th dimension. Combining the prediction results of multiple dimensions generates the final trend prediction result. The implementation formula is:

[0091]

[0092] In the formula, represents the final trend prediction result, D represents the number of dimensions, and w d (A t ) represents the weight of each dimension, and represents the predicted value of the d-th dimension at time t.

[0093] Among them, the intelligent prediction and diagnosis module predicts potential problems based on the results of the intelligent analysis module and the virtual simulation model information; assuming the state of the virtual simulation model is S sim (t), the anomaly score A t of the intelligent analysis module, the trend prediction result and the state of the virtual simulation model are fused. The multi-dimensional feature matrix F dw (t) implementation formula is:

[0094]

[0095] According to the confidence level C t of the anomaly score A t (A t ) and the reliability R t (S sim (t)) of the simulation state, the weights of each feature are dynamically adjusted. The implementation formula is:

[0096]

[0097] In the formula, w i(t) represents the weight of the i-th feature at time t, and λ and μ respectively represent the sensitivity parameters that control the influence of confidence and reliability on the weight;

[0098] Feature weighting is performed according to the updated weights, and the implementation formula is:

[0099] F w (t) = W(t)·F dw (t),

[0100] In the formula, F w (t) represents the weighted weight, and W(t) represents the diagonal matrix composed of w i (t).

[0101] The intelligent prediction and diagnosis module calculates multiple risk factors according to the weighted feature matrix F w (t), and the implementation formula is:

[0102]

[0103] In the formula, R f (t) represents the risk factor, represents the weighted trend prediction value, extracted from F w (t), represents the weighted trend prediction value The change rate within the unit time Δt, S sim (t) w represents the weighted simulation state value, extracted from F w (t), S real (t) represents the actual observed state, and the composite risk score R c (t) is set according to the risk factor, and the warning threshold is set. When the composite risk score R c (t) exceeds the threshold, a warning is triggered.

[0104] Working principle: Collect parameters such as the acquisition frequency, time range, and data accuracy through sensors and device settings in the factory. According to the set acquisition strategy, grab data from sensors and devices in real time, perform data cleaning, data conversion, and data screening on the grabbed data, and the preprocessed data is stored in the database;

[0105] Collect the environmental information in the factory in real time according to the integrated sensors and monitoring devices. Define the rules and trigger conditions for different environmental parameters based on the real-time collected environmental data, construct a situation model, identify and classify different working situations through the situation model, automatically adjust the device parameters according to the result information of the situation model, construct a virtual simulation model to simulate the production process of the factory based on the results of the data acquisition preprocessing module and the situation awareness module, import the device models in the factory into the virtual simulation software, and perform configuration and calibration. The virtual simulation model is dynamically updated in real time according to the data information of the data acquisition preprocessing module and the situation awareness module, simulate according to the set scenarios, and record various index information generated during the simulation process. The intelligent analysis module fuses the data of the data acquisition preprocessing module and the situation awareness module, performs feature extraction, dynamically evaluates the importance of the extracted features, and updates the selection. Perform anomaly detection and trend prediction based on the selected features. The intelligent prediction and diagnosis module fuses the anomaly score of the intelligent analysis module, the trend prediction result and the state of the virtual simulation model, dynamically adjusts the weights of each feature according to the confidence of the anomaly score and the reliability of the simulation state, calculates multiple risk factors based on the weighted feature matrix, and performs a composite risk score. When the composite risk score exceeds the set warning threshold, trigger a warning. The dynamic decision-making module formulates decisions based on the results of the intelligent prediction and diagnosis module, combined with the simulation information of the simulation model, including adjusting the production process, optimizing resource allocation or changing operation parameters. The adaptive optimization scheduling module automatically adjusts the resource allocation and operation parameter configuration according to the decisions made by the dynamic decision-making module.

[0106] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

[0107] The above describes the present invention and its implementation manners. Such a description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual structure is not limited thereto. Generally speaking, if those of ordinary skill in the art are inspired by it and design similar structural manners and embodiments without creative efforts without departing from the purpose of the present invention, they shall fall within the protection scope of the present invention.

Claims

1. The intelligent analysis and management system for factory data based on digital and intelligent integration is characterized in that: It includes a data acquisition and preprocessing module, a context awareness module, a simulation model module, an intelligent analysis module, an intelligent prediction and diagnosis module, a dynamic decision-making module, and an adaptive optimization and scheduling module; The data acquisition and preprocessing module is used to obtain factory data from sensors and devices and perform preprocessing; The context awareness module is used to obtain the environmental information inside the factory in real time based on integrated sensors and monitoring devices; The simulation model module is used to construct a virtual simulation model to simulate the factory production process according to the results of the data acquisition and preprocessing module and the context awareness module; The intelligent analysis module is used to conduct in-depth analysis on the data information of the data acquisition and preprocessing module and the context awareness module to discover hidden patterns and trends; The intelligent prediction and diagnosis module is used to predict potential problems by combining the results of the intelligent analysis module and the information of the virtual simulation model; The dynamic decision-making module is used to make decisions according to the prediction and diagnosis results combined with the simulation information of the simulation model; The adaptive optimization and scheduling module is used to automatically adjust the resource configuration and operation parameter configuration according to the made decisions.

2. The factory data intelligent analysis and management system based on digital and intelligent integration according to claim 1, wherein: The data acquisition and preprocessing module sets the acquisition parameters of each data source through sensors and devices in the factory, including the acquisition frequency, acquisition time range, and data accuracy, grabs data in real time from each sensor and device according to the set acquisition strategy, and performs data cleaning, data conversion, and data screening on the real-time grabbed data. Data cleaning includes removing outliers, filling missing values, and de-duplication, and stores the preprocessed data in the database.

3. The factory data intelligent analysis and management system based on digital and intelligent integration according to claim 1, wherein: The context awareness module obtains the environmental information inside the factory in real time based on integrated sensors and monitoring devices; collects the factory environmental data information in real time through integrated sensors and monitoring devices, defines the rules and trigger conditions of different environmental parameters, analyzes the relationships between different environmental parameters, constructs a context model, identifies and classifies different working contexts according to the real-time environmental parameters through the context model, and automatically adjusts the device parameter information according to the result information of the context model.

4. The factory data intelligent analysis and management system based on digital and intelligent integration according to claim 1, characterized in that: The simulation model module constructs a virtual simulation model to simulate the factory production process according to the results of the data acquisition and preprocessing module and the context awareness module; constructs a virtual simulation model in the virtual simulation software according to the data information obtained in real time by the data acquisition and preprocessing and the environmental information monitored in real time in the context awareness, imports the device models in the factory into the virtual simulation software, and performs configuration and calibration. The virtual simulation model dynamically updates the state of the virtual simulation model in real time according to the data information of the data acquisition and preprocessing module and the context awareness module, and the virtual simulation model simulates according to the set scenarios and records various index information generated during the simulation process, including factory data changes and device status information.

5. The intelligent analysis and management system for factory data based on digital and intelligent integration according to claim 1, wherein: The intelligent analysis module performs in-depth analysis based on the data information of the data acquisition and preprocessing module and the context awareness module to discover hidden patterns and trends. Let the data of the data preprocessing module be d pc (T(t)), and the data of the context awareness module be d cx (T(t)). The data of the data acquisition and preprocessing module and the context awareness module are fused, and the implementation formula is: D fn d(t) = λ1d pc + λ2d(T(t)) cx (T(t)), In the formula, D fn (t) represents the data after fusion at time t, and λ1 and λ2 represent the weight coefficients; Let the time-varying feature matrix be F(t) = [f1(t), f2(t),..., f i (t)], where F(t) represents the feature matrix extracted at time t, and f i (t) represents the i-th feature extracted at time t. f i (t) extracts features from the fused data D fn (t). The implementation formula is as follows: f i f(t) = g i (D fn (t)), In the formula, g i represents the feature extraction function, and D fn (t) represents the data after fusion at time t; Dynamically evaluate the importance according to the extracted features, and the implementation formula is: In the formula, I(f i , t) represents the importance of feature f i at time t, and L(F(t)) represents the loss function. The updated selection of the evaluated features is implemented as: S(t) = {f i |I(f i , t) > θ(t)}, In the formula, S(t) represents the feature selected at time t, and θ(t) represents the updated feature.

6. The factory data intelligent analysis and management system based on digital and intelligent integration according to claim 5, wherein: Perform anomaly detection according to the selected feature S(t), and the implementation formula is: h t = σ(W h x t + U h h t-1 + b n ), In the formula, h t represents the hidden state at time t, x t represents the feature S(t) input at time t, W h and U h represent weight matrices, representing the input feature and the hidden state at the previous moment respectively, b h represents the bias term, and σ represents the activation function; the anomaly score is calculated based on the hidden state and is implemented as: In the formula, A t represents the anomaly score at time t, represents the predicted hidden state at time t.

7. The intelligent analysis and management system for factory data based on digital and intelligent integration according to claim 6, characterized in that: Suppose a comprehensive prediction is made based on trends in multiple dimensions. Let the predicted value of the d-th dimension at time t be The implementation formula is: In the formula, represents the d - dimensional observation value in the past p steps, and represent the autoregressive coefficient and the moving - average coefficient of the d - dimension, represents the error term of the d - dimension. Combining the prediction results of multiple dimensions generates the final trend prediction result. The implementation formula is: In the formula, represents the final trend prediction result, D represents the number of dimensions, and w d (A t ) represents the weight of each dimension, represents the predicted value of the d-th dimension at time t.

8. The intelligent analysis and management system for factory data based on digital and intelligent integration according to claim 1, wherein: The intelligent prediction and diagnosis module predicts potential problems based on the results of the intelligent analysis module and the information of the virtual simulation model; assuming the state of the virtual simulation model is S sim (t), and fusing the anomaly score A t of the intelligent analysis module, the trend prediction result and the state of the virtual simulation model, the multi-dimensional feature matrix F dw (t) is implemented by the formula: According to the anomaly score A t with confidence C t (A t ) and the reliability R of the simulation state t (S sim (t)), dynamically adjust the weights of each feature, and the implementation formula is: In the formula, w i (t) represents the weight of the i-th feature at time t, and λ and μ respectively represent the sensitivity parameters that control the influence of confidence and reliability on the weight; Perform feature weighting according to the updated weights, and the implementation formula is: F w (t) = W(t)·F dw (t), In the formula, F w (t) represents the weighted weight, and W(t) represents the diagonal matrix composed of w i (t).

9. The factory data intelligent analysis and management system based on digital and intelligent integration according to claim 8, characterized in that: The intelligent prediction and diagnosis module calculates multiple risk factors based on the weighted feature matrix F w (t), and the implementation formula is as follows: In the formula, R f (t) represents the risk factor, represents the weighted trend prediction value, extracted from F w (t), represents the weighted trend prediction value The change rate within the unit time Δt, S sim (t) w represents the weighted simulation state value, extracted from F w (t), S real (t) represents the actual observed state, and the composite risk score R c (t) is calculated according to the risk factor. Set the warning threshold. When the composite risk score R c (t) exceeds the threshold, a warning is triggered.

Images