Dynamic data analysis method and system based on intelligent operation platform

Abstract

The invention discloses a dynamic data analysis method and system based on an intelligent operation platform, and relates to the technical field of data detection, and the method comprises the steps: collecting the operation data of the intelligent operation platform, carrying out the preprocessing, and constructing a causal network based on the operation data; constructing an intelligent operation platform digital twin model according to the causal network, analyzing the operation state of the intelligent operation platform through the digital twin model, and adjusting the causal network based on the operation state; and displaying the operation data and the operation state of the intelligent operation platform in real time and storing the data and the state in a database. According to the invention, the operation data of the intelligent operation platform is collected to construct the causal network, and the digital twinborn model is constructed according to the causal network, so that the relevance between the digital twinborn model and the intelligent operation platform is effectively improved; the operation data and the operation state of the intelligent operation platform are subjected to anomaly detection through the digital twinborn model, the anomaly detection accuracy and flexibility of the intelligent operation platform are greatly improved, and safe operation of the intelligent operation platform is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of data detection technology, and in particular to a dynamic data analysis method and system based on a smart operation platform. Background Art

[0002] In recent years, with the rapid development of information technology, the Internet of Things and big data analysis technology, smart operation platforms have gradually become core tools for improving the efficiency of urban management, industrial production and commercial services. Smart operation platforms integrate distributed multi-source data on a unified platform through sensor networks, data acquisition equipment and smart terminals, providing real-time monitoring, resource optimization and intelligent decision support for complex scenarios. In existing technologies, smart operation platforms usually use traditional statistical analysis or single-point prediction models based on machine learning to analyze operation data. These technologies have high analysis accuracy in static environments, but their adaptability in dynamic and complex environments is poor. It is difficult to capture the causal relationship within the system and cannot accurately reveal the mechanism of action between variables, resulting in insufficient performance of the model in anomaly detection, real-time optimization and global decision support. In addition, they lack the ability to comprehensively model the overall status of the platform, making it difficult to achieve real-time monitoring and dynamic feedback adjustment of the system's operating status. Summary of the invention

[0003] In view of the problems existing in the above-mentioned existing dynamic data analysis method and system based on the intelligent operation platform, the present invention is proposed.

[0004] Therefore, the problem to be solved by the present invention is that the existing technology is difficult to capture the causal relationship within the system, cannot accurately reveal the mechanism of action between various variables, and is difficult to achieve real-time monitoring and dynamic feedback adjustment of the system operation status.

[0005] To solve the above technical problems, the present invention provides the following technical solutions: a dynamic data analysis method based on a smart operation platform, which includes collecting the operation data of the smart operation platform for preprocessing, and constructing a causal network based on the operation data; constructing a digital twin model of the smart operation platform according to the causal network, and analyzing the operation status of the smart operation platform through the digital twin model, and adjusting the causal network based on the operation status; displaying the operation data and operation status of the smart operation platform in real time and storing them in a database.

[0006] As a preferred solution of the dynamic data analysis method based on the smart operation platform described in the present invention, the pre-processing of the collected operation data of the smart operation platform refers to obtaining the operation data based on the smart operation platform, using wavelet transform and Kalman filtering to jointly perform noise reduction on the operation data, and performing data cleaning on the noise-reduced operation data, and converting the operation data into JSON (JavaScript Object Notation, a lightweight data exchange format) format after time alignment.

[0007] As a preferred solution of the dynamic data analysis method based on the smart operation platform of the present invention, wherein: the construction of a causal network based on the operation data refers to defining each type of data in the operation data as a variable, extracting variables from the pre-processed smart operation platform operation data and forming a variable set X = {x 1 , x 2 , …, x n}, x n For the nth variable, take each variable as a node to form an undirected association graph, and initialize all nodes so that there are connecting edges between nodes;

[0008] Get adjacent nodes x i and x j All common neighbor nodes of are exhaustively enumerated to form a combination Z. For each combination Z, check x separately. i and x j Conditional independence of:

[0009] H 0 :Q(x i ⊥x j |Z);

[0010] H 0 is a hypothesis, Q is a joint probability distribution, and statistical test methods are used to determine H 0 Is it true that if there is a combination Z such that H 0 If true, delete x i and x j The connecting edges between them, otherwise the connecting edges are retained;

[0011] Through the conditional independence test, the connection edges between all nodes in the undirected association graph are optimized and the optimized node x is extracted. i The adjacent nodes form the adjacent node set P a (x i );

[0012] Calculate the node x through Gaussian distribution i The conditional probability distribution P(x i |P a (x i )):

[0013]

[0014] Where N is a normal distribution, β 0 is the bias term, β j is the adjacent node x j For node x i The influence coefficient of 2 is the conditional variance;

[0015] Based on the conditional probability distribution P(x i |P a (x i ))Define the joint probability distribution P:

[0016]

[0017] Where n is the number of nodes;

[0018] Define the objective function L as:

[0019] L=-logP+λ‖β‖ 1 ;

[0020] where λ is the regularization strength, ‖β‖ 1 is the absolute value and of the influence coefficient;

[0021] Based on the objective function, the Bayesian Information Criterion is used to calculate the BIC (Bayesian Information Criterion) value of each connection edge direction:

[0022]

[0023] For every connecting edge there exists x i →x j and x j →x i In two directions, the BIC values ​​of the two directions are calculated and compared, and the direction with the larger BIC value is selected as the optimal direction. The directions of the connecting edges between all nodes are counted, and the causal relationship matrix M is constructed based on the causal relationship between nodes. If the node x i →x j If it holds, M(i, j) is 1, otherwise it is 0. The undirected association graph is optimized based on the causal relationship matrix to form a causal network G, and the average of the BIC values ​​of the two connection edge directions is used as the weight of each connection edge in the causal network.

[0024] As a preferred solution of the dynamic data analysis method based on the smart operation platform of the present invention, wherein: the construction of the digital twin model of the smart operation platform according to the causal network includes:

[0025] According to the edge weights in the causal network G, the key variable set H is screened:

[0026] H={x i ∈G|∑ j ω ij >τ};

[0027] where ω ij Output variable x for the node i Its adjacent nodes output variable x j The connection edge weight, τ is the weight threshold;

[0028] Take the key variable set H as the input variable set I, and define the output variable set O based on the input variable set I:

[0029]

[0030] where ω jk Output variable x for the node k Its adjacent node output variable x j The connection edge weights of

[0031] For each output variable x in the output variable set O k , extract its adjacent node set P from the causal network a (x k ), through the adjacent node set P a (x k ) for the output variable x k Make adjustments:

[0032]

[0033] where x′ k To adjust the output variable;

[0034] Combined with the time series characteristics, the output variable state update formula is defined as:

[0035] x′ k (t+1) = ReLU(W*x′ k (t)+b);

[0036] where x′ k (t+1) is the updated adjusted output variable at time t+1, W is the updated weight, b is the updated bias, and x′ k (t) is the adjusted output variable at time t, and ReLU is the activation function;

[0037] The mean square error (MSE) is defined as the loss function, and the output variable state update formula is trained using the training set. The weight W and the bias b are iteratively updated based on the gradient descent method. The training is stopped after the loss converges, and the trained output variable state update formula is used as the digital twin model of the smart operation platform.

[0038] As a preferred solution of the dynamic data analysis method based on the smart operation platform of the present invention, wherein: the operation status of the smart operation platform is analyzed by the digital twin model, and the causal network is adjusted based on the operation status, which means that the variable x is obtained by the digital twin model of the smart operation platform. k The adjusted output variable at time t+1 The adjacent node set P extracted through the causal network a (x k ) for causal verification:

[0039]

[0040] in To adjust the output variable Check value of

[0041] calculate and The absolute difference between k ), if the anomaly detection index D(x k ) is greater than the preset anomaly detection threshold, then the variable x k is an abnormal variable, and the operation status of the smart operation platform is judged to be abnormal, and the abnormal detection indicator D(x k ) Calculate the abnormal variable x k Strength of influence on adjacent variables:

[0042] Im(x j )=ω jk *D(x k );

[0043] Where Im(x j ) is the abnormal variable x k For adjacent variables x j The influence intensity is based on Im(x j ) for adjacent variables x j Make adjustments:

[0044]

[0045] in is the adjusted variable, γ is the feedback propagation limitation coefficient;

[0046] Substitute the adjusted variables into the causal network to form an adjusted causal network.

[0047] As a preferred solution of the dynamic data analysis method based on the smart operation platform described in the present invention, the real-time display of the operation data and operation status of the smart operation platform refers to collecting the operation data of the smart operation platform and obtaining the abnormal variables and operation status of the smart operation platform through the operation data, and then visually displaying them, and highlighting the obtained abnormal variables in the operation data for display.

[0048] As a preferred solution of the dynamic data analysis method based on the smart operation platform described in the present invention, the storage in the database refers to the smart operation platform storing the collected operation data and operation status into operation records according to the timestamp in the database, and the database regularly performs integrity and security checks on the stored data, and uploads the stored data to the cloud for backup.

[0049] Another object of the present invention is to provide a dynamic data analysis system based on a smart operation platform, which includes a data collection module for collecting and preprocessing the operation data of the smart operation platform;

[0050] The model building module is used to build a causal network based on the operation data and build a digital twin model of the smart operation platform based on the causal network;

[0051] The analysis and adjustment module is used to analyze abnormal variables in the operation data according to the digital twin model, determine the operation status of the smart operation platform based on the abnormal variables, and adjust the causal network through the abnormal variables;

[0052] The display storage module is used to display the operation data and operation status of the smart operation platform in real time and store them in the database.

[0053] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above-mentioned dynamic data analysis method based on a smart operation platform when executing the computer program.

[0054] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned dynamic data analysis method based on a smart operation platform.

[0055] The beneficial effects of the present invention are as follows: the present invention constructs a causal network by collecting the operating data of the smart operation platform, and constructs a digital twin model based on the causal network, thereby effectively improving the correlation between the digital twin model and the smart operation platform. By performing anomaly detection on the operating data and operating status of the smart operation platform through the digital twin model, the accuracy and flexibility of anomaly detection of the smart operation platform are greatly improved, thereby ensuring the safe operation of the smart operation platform. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0057] Figure 1 The figure is a flowchart of the dynamic data analysis method based on the smart operation platform.

[0058] Figure 2 This is a structural diagram of the dynamic data analysis system based on the smart operation platform. DETAILED DESCRIPTION

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0060] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0061] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0062] Example 1

[0063] Reference Figure 1 , which is the first embodiment of the present invention, and this embodiment provides a dynamic data analysis method based on a smart operation platform. The dynamic data analysis method based on a smart operation platform includes:

[0064] S1. Collect the operation data of the smart operation platform for preprocessing, and build a causal network based on the operation data;

[0065] Specifically, collecting the operation data of the smart operation platform for preprocessing refers to obtaining the operation data based on the smart operation platform. The operation data includes equipment data, user behavior data, system operation logs, etc., and using wavelet transform and Kalman filtering to jointly perform noise reduction on the operation data, and then performing data cleaning on the noise-reduced operation data, and converting the operation data into JSON format after time alignment. JSON is a lightweight data exchange format that is easy to read and efficient to parse.

[0066] By integrating multi-source heterogeneous data (device data, user behavior data, system operation logs), the operating status of the smart operation platform can be fully covered, providing rich contextual information for subsequent analysis. Multi-dimensional data collection can support global analysis of complex systems and improve the depth and breadth of operating status analysis. The joint noise reduction method can maintain the integrity and key features of the data in complex scenarios (such as high-noise environments or non-stationary data streams), improve the quality of operating data, reduce the interference of noise on subsequent analysis, and improve the accuracy and robustness of data processing. Data cleaning can effectively identify and repair erroneous values, missing values, and duplicate data in operating data, and improve data quality. By time-aligning the operating data, the time base of multi-source data can be unified, and the problem of asynchronous data flow can be solved. The use of JSON format for standardized storage and transmission of data can improve data compatibility and parsing efficiency.

[0067] Furthermore, constructing a causal network based on operation data means defining each type of data in the operation data as a variable, extracting variables from the preprocessed operation data of the smart operation platform and forming a variable set X = {x 1 , x 2 ,…,x n}, x n For the nth variable, take each variable as a node to form an undirected association graph, and initialize all nodes so that there are connecting edges between nodes;

[0068] Get adjacent nodes x i and x j All common neighbor nodes of are exhaustively enumerated to form a combination Z. For example, if the common neighbor nodes include 1 and 2, their combinations include (1), (2) and (1,2). For each combination Z, check x separately. i and x j Conditional independence of:

[0069] H 0 :Q(x i ⊥x j |Z);

[0070] H 0 is a hypothesis, Q is a joint probability distribution, and statistical test methods are used to determine H 0Is it true, such as mutual information or Pearson correlation, if there is a combination Z such that H 0 If true, delete x i and x j The connecting edges between them, otherwise the connecting edges are retained;

[0071] Through the conditional independence test, the connection edges between all nodes in the undirected association graph are optimized and the optimized node x is extracted. i The adjacent nodes form the adjacent node set P a (x i );

[0072] Calculate the node x through Gaussian distribution i The conditional probability distribution P(x i |P a (x i )):

[0073]

[0074] Where N is a normal distribution, β 0 is the bias term, β j is the adjacent node x j For node x i The influence coefficient of 2 is the conditional variance;

[0075] Based on the conditional probability distribution P(x i |P a (x i ))Define the joint probability distribution P:

[0076]

[0077] Where n is the number of nodes;

[0078] Define the objective function L as:

[0079] L=-log P+λ‖β‖ 1 ;

[0080] where λ is the regularization strength, ‖β‖ 1 is the absolute value and of the influence coefficient;

[0081] The BIC value of each connection edge direction is calculated based on the objective function using the Bayesian Information Criterion:

[0082]

[0083] For every connecting edge there exists x i →x j and x j →x iIn two directions, the BIC values ​​of the two directions are calculated and compared, and the direction with the larger BIC value is selected as the optimal direction. The directions of the connecting edges between all nodes are counted, and the causal relationship matrix M is constructed based on the causal relationship between nodes. If the node x i →x j If it holds, M(i, j) is 1, otherwise it is 0. The undirected association graph is optimized based on the causal relationship matrix to form a causal network G, and the average of the BIC values ​​of the two connection edge directions is used as the weight of each connection edge in the causal network.

[0084] The construction of the undirected association graph provides an initial structural framework for subsequent causal network modeling, covering all potential related variables and their relationships. By defining the variable set in the running data and initializing the connection edges, the association graph is ensured to be complete, providing comprehensive input for subsequent optimization. The conditional independence test removes redundant connection edges through statistical methods, thereby improving the simplicity and interpretability of the association graph. The edge optimization process significantly reduces the complexity of the network and retains key association information, providing a guarantee for building a more accurate causal network. The conditional probability distribution of the Gaussian distribution modeling node can intuitively describe the statistical relationship between variables and provide a basis for the calculation of the joint probability distribution. The parameterized form of the Gaussian distribution (such as the bias term and conditional variance) is convenient for variables. The linear or nonlinear dependencies between variables are modeled and analyzed. The definition of joint probability distribution can comprehensively describe the global statistical characteristics between variables and provide an overall perspective for causal relationship modeling. The objective function introduces regularization strength to effectively avoid overfitting problems. At the same time, the interpretability and predictive performance of the model are improved by minimizing the objective function. The introduction of BIC comprehensively considers the goodness of fit and complexity of the model to ensure that the choice of direction has a theoretical basis. By comparing the BIC values ​​in the two directions, the optimal direction of the causal relationship between variables can be accurately selected. The causal relationship matrix provides a clear structural representation for the causal network by quantifying the causal relationship between nodes. The introduction of edge weights further enhances the quantifiability of the causal network, which is helpful for system state analysis and decision support.

[0085] S2. Build a digital twin model of the smart operation platform based on the causal network, analyze the operating status of the smart operation platform through the digital twin model, and adjust the causal network based on the operating status;

[0086] Specifically, the digital twin model of the intelligent operation platform is constructed based on the causal network, including:

[0087] According to the edge weights in the causal network G, the key variable set H is screened:

[0088] H={x i ∈G|∑ j ω ij >τ};

[0089] where ωij Output variable x for the node i Its adjacent nodes output variable x j The connection edge weight, τ is the weight threshold;

[0090] Take the key variable set H as the input variable set I, and define the output variable set O based on the input variable set I:

[0091]

[0092] where ω jk Output variable x for the node k Its adjacent nodes output variable x j The connection edge weights of

[0093] For each output variable x in the output variable set O k , extract its adjacent node set P from the causal network a (x k ), through the adjacent node set P a (x k ) for the output variable x k Make adjustments:

[0094]

[0095] where x′ k To adjust the output variable;

[0096] Combined with the time series characteristics, the output variable state update formula is defined as:

[0097] x′ k (t+1) = ReLU(W*x′ k (t)+b);

[0098] where x′ k (t+1) is the updated adjusted output variable at time t+1, W is the updated weight, b is the updated bias, and x′ k (t) is the adjusted output variable at time t, and ReLU is the activation function;

[0099] The mean square error (MSE) is defined as the loss function, and the output variable state update formula is trained using the training set. The weight W and the bias b are iteratively updated based on the gradient descent method. The training is stopped after the loss converges, and the trained output variable state update formula is used as the digital twin model of the smart operation platform.

[0100] Screening the key variable set helps to reduce the input dimension of the model, reduce the computational complexity, and improve the modeling efficiency. The key variables concentrate the core information in the system, which is convenient for the precise analysis and prediction of the subsequent model. Based on the link weight of the causal network, the input variable set and the output variable set are dynamically determined so that the model structure can adapt to the causal characteristics of the system. The definition of the input and output variable sets enhances the interpretability of the model and makes it more in line with the actual operation logic of the system. The adjustment formula of the output variable through the adjacent node set ensures that the state of the output variable can comprehensively consider the dynamic changes of its adjacent nodes, enhances the dynamic response ability of the model, and improves the twin model's modeling ability for the dynamic interaction between complex system variables. The introduction of time series features can capture the dynamic change laws of system variables and improve the accuracy of the model's prediction of future states. Defining the mean square error (MSE) as a loss function can quantify the difference between the predicted value and the true value, providing a clear goal for the optimization of the model. The digital twin model reflects the operating status of the smart operation platform through dynamic modeling and can analyze the dynamic changes of the system in real time.

[0101] Furthermore, the operation status of the smart operation platform is analyzed through the digital twin model, and the causal network is adjusted based on the operation status. The variable x is obtained through the digital twin model of the smart operation platform. k The adjusted output variable at time t+1 The adjacent node set P extracted through the causal network a (x k ) for causal verification:

[0102]

[0103] in To adjust the output variable Check value of

[0104] calculate and The absolute difference between k ), if the anomaly detection index D(x k ) is greater than the preset anomaly detection threshold, then the variable x k is an abnormal variable, and the operation status of the smart operation platform is judged to be abnormal, and the abnormal detection indicator D(x k ) Calculate the abnormal variable x k Strength of influence on adjacent variables:

[0105] Im(x j )=ω jk *D(x k );

[0106] Where Im(x j ) is the abnormal variable xk For adjacent variables x j The influence intensity is based on Im(x j ) for adjacent variables x j Make adjustments:

[0107]

[0108] in is the adjusted variable, γ is the feedback propagation limitation coefficient, (the value range is 0≤γ≤1);

[0109] Substitute the adjusted variables into the causal network to form an adjusted causal network.

[0110] By adjusting the output variables in the prediction time of the digital twin model, it is possible to capture the dynamic changes of the system and reflect the operating status of the smart operation platform in real time. The output variables are adjusted in combination with time series characteristics to enhance the prediction ability of the dynamic behavior of complex systems. The check value of the adjusted output variables is calculated using the adjacent node set to ensure that the relationship between variables conforms to the logic of the causal network, effectively identify abnormal points in the operating status, avoid misjudgment or missed judgment, and quickly judge the abnormality of the variables by setting the abnormal detection threshold, providing a quantitative basis for operating status analysis, quantifying the impact of abnormal variables on adjacent variables, and providing precise direction and amplitude for feedback adjustment, improving the pertinence and effectiveness of abnormal processing, and preventing the spread of abnormalities to the system. The feedback propagation limit coefficient is introduced to control the propagation range of abnormal feedback, avoid excessive adjustment or triggering a chain effect, achieve local isolation and dynamic correction of abnormal states, and improve the stability of system operation. The updated causal network can reflect the adjustment results of the smart operation platform under abnormal conditions and provide a more accurate model for subsequent analysis.

[0111] S3: Display the operation data and operation status of the smart operation platform in real time and store them in the database;

[0112] Specifically, real-time display of the operation data and operation status of the smart operation platform refers to collecting the operation data of the smart operation platform and obtaining abnormal variables and operation status of the smart operation platform through the operation data, and then visually displaying them, and highlighting the obtained abnormal variables in the operation data for display.

[0113] The operation data collected from the smart operation platform covers multi-dimensional information such as equipment status, user operations and environmental parameters, and can comprehensively reflect the operation status of the platform. The real-time nature of the data collection process ensures the dynamic nature of the analysis results, providing a highly timely data basis for anomaly detection and status analysis. By analyzing the operation data, the abnormal variables of the smart operation platform can be identified, and potential problems or risks in operation can be quickly located. The operation status analysis combined with the anomaly detection algorithm can dynamically evaluate the health of the platform operation and provide efficient anomaly warnings. The highlight mark can present abnormal variables in a visually significant way, helping users to quickly locate problems and improve the efficiency of exception handling. The real-time update mechanism ensures that the displayed operation data and operation status are synchronized with the actual situation of the platform, avoiding misjudgments or decision delays due to information lags.

[0114] Furthermore, storing in a database means that the smart operation platform forms operation records of the collected operation data and operation status according to timestamps and stores them in the database. The database regularly performs integrity and security checks on the stored data and uploads the stored data to the cloud for backup.

[0115] Example 2

[0116] Reference Figure 2 , which is the second embodiment of the present invention, and which is different from the previous embodiment, provides a dynamic data analysis system based on a smart operation platform, which includes:

[0117] Data collection module, used to collect and pre-process the operation data of the smart operation platform;

[0118] The model building module is used to build a causal network based on the operation data and build a digital twin model of the smart operation platform based on the causal network;

[0119] The analysis and adjustment module is used to analyze abnormal variables in the operation data according to the digital twin model, determine the operation status of the smart operation platform based on the abnormal variables, and adjust the causal network through the abnormal variables;

[0120] The display storage module is used to display the operation data and operation status of the smart operation platform in real time and store them in the database.

[0121] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0122] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0123] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0124] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0125] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A dynamic data analysis method based on a smart operation platform, characterized in that: include, Collect the operation data of the smart operation platform for preprocessing, and build a causal network based on the operation data; Build a digital twin model of the smart operation platform based on the causal network, analyze the operating status of the smart operation platform through the digital twin model, and adjust the causal network based on the operating status; The operation data and status of the smart operation platform are displayed in real time and stored in the database.

2. The dynamic data analysis method based on the intelligent operation platform according to claim 1, characterized in that: The preprocessing of the collected operation data of the smart operation platform refers to obtaining the operation data based on the smart operation platform, using wavelet transform and Kalman filtering to jointly perform noise reduction on the operation data, and performing data cleaning on the noise-reduced operation data, and converting the operation data into JSON format after time alignment.

3. The dynamic data analysis method based on the smart operation platform according to claim 2, characterized in that: The construction of a causal network based on operation data refers to defining each type of data in the operation data as a variable, extracting variables from the pre-processed operation data of the smart operation platform and forming a variable set X = {x1, x2, ..., x n }, x n For the nth variable, take each variable as a node to form an undirected association graph, and initialize all nodes so that there are connecting edges between nodes; Get adjacent nodes x i and x j All common neighbor nodes of are exhaustively enumerated to form a combination Z. For each combination Z, check x separately. i and x j Conditional independence of: H0:Q(x i ⊥x j |Z); H0 is a hypothesis, Q is a joint probability distribution, and a statistical test method is used to determine whether H0 is true. If there is a combination Z that makes H0 true, then delete x. i and x j The connecting edges between them, otherwise the connecting edges are retained; Through the conditional independence test, the connection edges between all nodes in the undirected association graph are optimized and the optimized node x is extracted. i The adjacent nodes form the adjacent node set P a (x i ); Calculate the node x through Gaussian distribution i The conditional probability distribution P(x i |P a (x i )): Where N is a normal distribution, β0 is a bias term, and β j is the adjacent node x j For node x i The influence coefficient of 2 is the conditional variance; Based on the conditional probability distribution P(x i |P a (x i ))Define the joint probability distribution P: Where n is the number of nodes; Define the objective function L as: L = -logP + λ‖β‖1; Where λ is the regularization strength, ‖β‖1 is the absolute value and influence coefficient; The BIC value of each connection edge direction is calculated based on the objective function using the Bayesian Information Criterion: For every connecting edge there exists x i →x j and x j →x i In two directions, the BIC values ​​of the two directions are calculated and compared respectively. The direction with the larger BIC value is selected as the optimal direction. The directions of the connecting edges between all nodes are counted. The causal relationship matrix M is constructed based on the causal relationship between nodes. If the node x i →x j If it holds, M(i, j) is 1, otherwise it is 0. The undirected association graph is optimized based on the causal relationship matrix to form a causal network G, and the average of the BIC values ​​of the two connection edge directions is used as the weight of each connection edge in the causal network.

4. The dynamic data analysis method based on the intelligent operation platform according to claim 3, characterized in that: The digital twin model of the intelligent operation platform constructed according to the causal network includes: According to the edge weights in the causal network G, the key variable set H is screened: H={x i ∈G|∑ j oh ij >t}; where ω ij Output variable x for the node i Its adjacent nodes output variable x j The connection edge weight, τ is the weight threshold; Take the key variable set H as the input variable set I, and define the output variable set O based on the input variable set I: where ω jk Output variable x for the node k Its adjacent nodes output variable x j The connection edge weights of For each output variable x in the output variable set O k , extract its adjacent node set P from the causal network a (x k ), through the adjacent node set P a (x k ) for the output variable x k Make adjustments: where x′ k To adjust the output variable; Combined with the time series characteristics, the output variable state update formula is defined as: x′ k (t+1)=ReLU(W*x′ k (t)+b)? where x′ k (t+1) is the updated adjusted output variable at time t+1, W is the updated weight, b is the updated bias, and x′ k (t) is the adjusted output variable at time t, and ReLU is the activation function; The mean square error (MSE) is defined as the loss function, and the output variable state update formula is trained using the training set. The weight W and the bias b are iteratively updated based on the gradient descent method. The training is stopped after the loss converges, and the trained output variable state update formula is used as the digital twin model of the smart operation platform.

5. The dynamic data analysis method based on the smart operation platform according to claim 4, characterized in that: The analysis of the operation status of the smart operation platform through the digital twin model and the adjustment of the causal network based on the operation status refer to obtaining the variable x through the digital twin model of the smart operation platform. k The adjusted output variable at time t+1 The adjacent node set P extracted through the causal network a (x k ) for causal verification: in To adjust the output variable Check value of calculate and The absolute difference between k ), if the anomaly detection index D(x k ) is greater than the preset anomaly detection threshold, then the variable x k is an abnormal variable, and the operation status of the smart operation platform is judged to be abnormal, and the abnormal detection indicator D(x k ) Calculate the abnormal variable x k Strength of influence on adjacent variables: Im(x j )=ω jk *D(x k ); Where Im(x j ) is the abnormal variable x k For adjacent variables x j The influence intensity is based on Im(x j ) for adjacent variables x j Make adjustments: in is the adjusted variable, γ is the feedback propagation limitation coefficient; Substitute the adjusted variables into the causal network to form an adjusted causal network.

6. The dynamic data analysis method based on the intelligent operation platform according to claim 5, characterized in that: The real-time display of the operation data and operation status of the smart operation platform refers to collecting the operation data of the smart operation platform and visually displaying the abnormal variables and operation status of the smart operation platform obtained through the operation data, and highlighting the obtained abnormal variables in the operation data for display.

7. The dynamic data analysis method based on the smart operation platform according to claim 6, characterized in that: The storage in the database refers to the intelligent operation platform storing the collected operation data and operation status into operation records according to the timestamp in the database. The database regularly performs integrity and security checks on the stored data, and uploads the stored data to the cloud for backup.

8. A dynamic data analysis system based on a smart operation platform and a dynamic data analysis method based on a smart operation platform according to any one of claims 1 to 7, characterized in that: include, Data collection module, used to collect and pre-process the operation data of the smart operation platform; The model building module is used to build a causal network based on the operation data and build a digital twin model of the smart operation platform based on the causal network; The analysis and adjustment module is used to analyze abnormal variables in the operation data according to the digital twin model, determine the operation status of the smart operation platform based on the abnormal variables, and adjust the causal network through the abnormal variables; The display storage module is used to display the operation data and operation status of the smart operation platform in real time and store them in the database.

9. A computer device comprising: Memory and processor; The memory stores a computer program, characterized in that when the processor executes the computer program, the steps of the dynamic data analysis method based on the smart operation platform described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the dynamic data analysis method based on the smart operation platform described in any one of claims 1 to 7 are implemented.

Images