A method for early push notification of operation and maintenance faults based on big data analysis

Abstract

The present invention discloses a method for early push of operation and maintenance faults based on big data analysis, which relates to the technical field of operation and maintenance fault prediction, and includes the following steps: collecting various types of data in the operation and maintenance system and preprocessing the data; constructing a topological analysis strategy of a graph neural network based on the collected data, analyzing the topological map, and updating the hidden state of the nodes; constructing a real-time adaptive analysis model, dynamically adjusting the parameters and structure of the model with real-time data, and performing fault prediction; using the topological analysis results based on the graph neural network and the prediction results of the adaptive analysis model to identify potential fault points. The present invention dynamically adjusts the prediction parameters and structure through a real-time adaptive analysis model to adapt to changes in the operation and maintenance environment, thereby improving the accuracy and timeliness of fault prediction. The topological analysis based on the graph neural network makes full use of the nodes and their connection relationships in the system, identifies complex interaction patterns and potential faults, and realizes comprehensive fault management.

Description

Technical Field

[0001] The present invention relates to the technical field of operation and maintenance fault prediction, and in particular to a method for early push of operation and maintenance faults based on big data analysis. Background Art

[0002] At present, with the rapid development and application of information technology, the importance of operation and maintenance systems in various enterprises and institutions has become increasingly prominent. The reliability and efficiency of the operation and maintenance system are directly related to the stability and economic benefits of enterprise operations. However, traditional operation and maintenance systems still have many shortcomings in fault prediction and management, and it is difficult to meet the needs of modern complex operation and maintenance environments.

[0003] Existing technologies mainly rely on static rules and simple statistical methods for fault prediction and management. These methods are usually based on predefined thresholds and historical data for analysis, lack real-time adaptive capabilities, and have difficulty coping with dynamic changes in the operation and maintenance environment. With the increase in system complexity and the rapid expansion of data volume, the prediction accuracy and timeliness of these traditional methods have significantly decreased, resulting in the inability of operation and maintenance personnel to detect and handle potential faults in a timely manner, which leads to a series of problems, including system downtime, production losses, and increased maintenance costs.

[0004] In addition, existing fault prediction methods usually ignore the complex associations and interactions between nodes in the system. The nodes and connections in the operation and maintenance system form a complex network structure. Relying solely on the status information of individual nodes for fault prediction often fails to capture the overall operating status and potential risks of the system. This limitation leads to the lack of effective early warning and management methods when the operation and maintenance system faces complex multi-node faults, further increasing the risk and uncertainty of system operation. Summary of the invention

[0005] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0006] In view of the problems existing in the above-mentioned existing method for early push of operation and maintenance faults based on big data analysis, the present invention is proposed.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] S1: Collect various data in the operation and maintenance system and pre-process the data;

[0009] S2: Construct a topological analysis strategy for the graph neural network based on the collected data, analyze the topological graph, and update the hidden state of the nodes;

[0010] S3: Build a real-time adaptive analysis model to dynamically adjust the model parameters and structure based on real-time data to predict faults;

[0011] S4: Utilize the topological analysis results based on graph neural networks and the prediction results of the adaptive analysis model to identify potential failure points and push warning information to operation and maintenance personnel in a timely manner through multiple channels.

[0012] As a preferred solution of the method for early push of operation and maintenance faults based on big data analysis described in the present invention, in the S2 process, the topological analysis strategy for constructing the graph neural network includes the following steps:

[0013] S21: The components and devices in the operation and maintenance system are regarded as nodes of the graph, and the connections between nodes are regarded as edges, that is, the network topology graph of the operation and maintenance system G = (V, E), where V is the node set and E is the edge set;

[0014] S22: Set x according to the network topology diagram i The eigenvector of i For node v i The abnormal state represents that the following formula is defined to update and propagate the state of the node;

[0015]

[0016] Where N(i) represents the node v i The set of neighbor nodes, d i ,d j The nodes v i and v j The degree, w (k) and It represents the weight matrix of the kth layer, and σ represents the activation function.

[0017] As a preferred solution of the method for early push of operation and maintenance faults based on big data analysis described in the present invention, the specific method for constructing a real-time adaptive analysis model in the S3 process is:

[0018] S31: Topological analysis based on neural network to calculate node failure score S i ;

[0019]

[0020] Among them, t 0 and t 1 represents the time interval, K represents the number of layers of GNN, represents the weight coefficient at the kth layer, and tanh represents the hyperbolic tangent function;

[0021] S32: Calculate the comprehensive fault push value P according to the obtained node fault score i ;

[0022]

[0023] Where, |V| is the total number of nodes; It represents the rate of change of the node turntable over time;

[0024] Complex Fault Score S i : The integral and exponential functions make S i The value range of is in the positive range, and the specific value depends on the system running time and the characteristics of the neighboring nodes;

[0025] Comprehensive fault push value P i : Through averaging and normalization, P i The value range of is usually within the positive range, indicating the comprehensive failure risk of the system. The larger the value, the higher the failure risk.

[0026] As a preferred solution of the method for early push of operation and maintenance faults based on big data analysis described in the present invention, for each node v i Comprehensive fault push value P i It can be used as an indicator to measure its failure risk. A threshold T is set and the comprehensive failure push value P is i Nodes exceeding the threshold are identified as potential failure points:

[0027] Fault point set = {v i |P i >T};

[0028] According to the nodes where the fault points are concentrated, early warning information is pushed to the operation and maintenance personnel in advance.

[0029] As a preferred solution of the method for early push of operation and maintenance faults based on big data analysis described in the present invention, wherein: in the S1 process, the collected data includes: topology data, node feature data, historical data and current status data of each node;

[0030] The data preprocessing comprises the following steps:

[0031] Data cleaning; removing noise data and processing missing values;

[0032] Data normalization: normalize data of different dimensions for subsequent analysis;

[0033] Feature extraction: extracting features useful for analysis from raw data;

[0034] Time series processing: Perform time series analysis and processing on real-time data and historical data.

[0035] As a preferred solution of the method for early push of operation and maintenance faults based on big data analysis described in the present invention, the fault records of all nodes in the system in the past year are collected to determine the failure rate and false alarm rate of the system;

[0036] By calculating the failure rate and false alarm rate under different thresholds, the receiver operating characteristic (ROC) curve is drawn;

[0037] According to business needs, select a balance point on the ROC curve as the threshold θ;

[0038] And S i Compare with the threshold θ to determine whether to push the subsequent comprehensive fault value P i calculate.

[0039] As a preferred solution of the method for early push of operation and maintenance faults based on big data analysis described in the present invention, it is assumed that the fault score of node A is calculated as S according to the calculation node fault score formula. A ;

[0040] If S A ≥ threshold θ, it is directly determined that there is no fault;

[0041] If S A <threshold θ, then continue to calculate the comprehensive push value P A .

[0042] The present invention also discloses a system for pushing operation and maintenance faults in advance based on big data analysis. Based on the above-mentioned method for pushing operation and maintenance faults in advance based on big data analysis, the system includes the following functional modules:

[0043] Data collection module, which is used to collect various types of data in the operation and maintenance system and pre-process the data;

[0044] The neural network analysis model building module builds the topological analysis strategy of the graph neural network based on the collected data, analyzes the topological graph, and updates the hidden state of the nodes;

[0045] The real-time adaptive analysis module builds a real-time adaptive analysis model based on the updated node hidden status, dynamically adjusts the model parameters and structure based on real-time data, and performs fault prediction;

[0046] The early warning information push module uses the topological analysis results based on the graph neural network and the prediction results of the adaptive analysis model to identify potential fault points and push the early warning information to the operation and maintenance personnel in a timely manner through email, text message or application.

[0047] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned method for early push of operation and maintenance faults based on big data analysis are implemented.

[0048] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned method for early push of operation and maintenance faults based on big data analysis.

[0049] Beneficial effects of the present invention: The present invention combines a real-time adaptive analysis model with a topological analysis based on a graph neural network to achieve accurate fault prediction and advance push. This method dynamically adjusts the prediction parameters and structure through a real-time adaptive analysis model to adapt to changes in the operation and maintenance environment, thereby improving the accuracy and timeliness of fault prediction. At the same time, the topological analysis based on the graph neural network makes full use of the nodes and their connection relationships in the system to identify complex interaction patterns and potential faults, thereby achieving comprehensive fault management. The combination of the two not only improves the operation and maintenance efficiency and system reliability, but also enhances the scalability and maintainability of the system, providing an efficient and reliable solution for modern complex operation and maintenance environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0051] Figure 1 This is a schematic diagram of the overall structure of a method for early push of operation and maintenance faults based on big data analysis proposed by the present invention;

[0052] Figure 2 A schematic diagram of a receiver operating characteristic ROC curve in an embodiment of an operation and maintenance fault advance push method based on big data analysis proposed by the present invention;

[0053] Figure 3 A schematic diagram of a network diagram structure including nodes A, B, C and their initial states and updated states in an embodiment of an operation and maintenance fault advance push method based on big data analysis proposed by the present invention; DETAILED DESCRIPTION

[0054] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0055] 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.

[0056] 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 does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0057] Reference Figure 1 , is an embodiment of the present invention, and provides a method for early push of operation and maintenance faults based on big data analysis, the method comprising the following steps:

[0058] S1: Collect various data from the operation and maintenance system and pre-process the data; clean the collected data to remove noise and outliers. Normalize structured data and unstructured data respectively;

[0059] Specifically, the collected data includes: topological data, node feature data, historical data, and the current status data of each node;

[0060] The data preprocessing comprises the following steps:

[0061] Data cleaning; removing noise data and processing missing values;

[0062] Data normalization: normalize data of different dimensions for subsequent analysis;

[0063] Feature extraction: extracting features useful for analysis from raw data;

[0064] Time series processing: Perform time series analysis and processing on real-time data and historical data.

[0065] S2: Construct a topological analysis strategy for the graph neural network based on the collected data, analyze the topological graph, and update the hidden state of the nodes;

[0066] Specifically, the topological analysis strategy for building a graph neural network includes the following steps:

[0067] S21: The components and devices in the operation and maintenance system are regarded as nodes of the graph, and the connections between nodes are regarded as edges, that is, the network topology graph of the operation and maintenance system G = (V, E), where V is the node set and E is the edge set;

[0068] S22: Set x according to the network topology diagram iThe eigenvector of i For node v i The abnormal state represents that the following formula is defined to update and propagate the state of the node;

[0069]

[0070] Where N(i) represents the node v i The set of neighbor nodes, d i ,d j The nodes v i and v j The degree, w (k) and It represents the weight matrix of the kth layer, and σ represents the activation function;

[0071] In summary, neighborhood aggregation and update formula The output range depends on the activation function σ, such as ReLU outputs non-negative values.

[0072] Here we take a simple operation and maintenance system as an example, which includes three nodes A, B, and C. Their connection relationship is as follows:

[0073] Node A is connected to node B and node C; node B is connected to node A; node C is connected to node A.

[0074] Assume the initial vector feature

[0075] Suppose the weight matrix in the first layer is:

[0076]

[0077] Degree d A =2,d B =1,d C =1;

[0078] Compute the hidden state update of A:

[0079] The neighbor node set N(A) includes nodes B and C, so:

[0080]

[0081] Plus the feature update of node A:

[0082]

[0083] Apply activation function;

[0084] Finally, the hidden state of node A is updated as:

[0085] Here we use ReLU as the activation function;

[0086]

[0087] Through the above steps, the hidden state of node A is updated from the initial state [1,0] to the new hidden state [0.88,1.01]. Similarly, the hidden states of other nodes can also be updated through the corresponding formulas and steps.

[0088] This process demonstrates how to use graph neural networks to update node status. By aggregating neighbor node features and its own features, the hidden state of each node can reflect its overall information in the network, thus providing a basis for subsequent fault prediction and push notification.

[0089] S3: Build a real-time adaptive analysis model, dynamically adjust the model parameters and structure based on real-time data, and predict faults. In a real-time operation and maintenance environment, combine historical data with real-time data to dynamically adjust and optimize the model so that the model can adaptively capture changes in node status, thereby more accurately predicting faults.

[0090] Furthermore, the specific method of constructing a real-time adaptive analysis model is:

[0091] S31: Topological analysis based on neural network to calculate node failure score S i ;

[0092]

[0093] Among them, t 0 and t 1 represents the time interval, K represents the number of layers of GNN, represents the weight coefficient at the kth layer, and tanh represents the hyperbolic tangent function;

[0094] Furthermore, the failure records of all nodes in the system in the past year are collected to determine the failure rate and false alarm rate of the system;

[0095] Reference Figure 2 , each point on the curve corresponds to a different threshold θ, showing the false positive rate (FPR) and true positive rate (TPR) under that threshold. Ideally, the true positive rate is as high as possible, and the false positive rate is as low as possible.

[0096] The red point in the figure represents the optimal threshold value θ, which is determined by finding the position closest to the ideal point (0,1). In the example in the figure above, the optimal threshold value is about 0.46, at which the model performs best in balancing the fault detection rate and the false alarm rate;

[0097] According to business needs, select a balance point on the ROC curve as the threshold θ;

[0098] And S i Compare with the threshold θ to determine whether to push the subsequent comprehensive fault value P i calculate;

[0099] Assume node A, calculate node fault score S according to the calculation node fault score formula A ;

[0100] If S A ≥ threshold θ, it is directly determined that there is no fault;

[0101] If S A <threshold θ, then continue to calculate the comprehensive push value P A .

[0102] S32: Calculate the comprehensive fault push value P according to the obtained node fault score i ;

[0103]

[0104] Where, |V| is the total number of nodes; It represents the rate of change of the node turntable over time;

[0105] Complex Fault Score S i : The integral and exponential functions make S i The value range of is in the positive range, and the specific value depends on the system running time and the characteristics of the neighboring nodes;

[0106] Comprehensive fault push value P i : Through averaging and normalization, P i The value range of is usually within the positive range, indicating the comprehensive failure risk of the system. The larger the value, the higher the failure risk.

[0107] S4: Utilize the topological analysis results based on graph neural networks and the prediction results of the adaptive analysis model to identify potential failure points and push warning information to operation and maintenance personnel in a timely manner through multiple channels (such as email, SMS, in-app notifications, etc.).

[0108] In addition, for each node v i Comprehensive fault push value P i It can be used as an indicator to measure its failure risk. A threshold T is set and the comprehensive failure push value P is i Nodes exceeding the threshold are identified as potential failure points:

[0109] Fault point set = {v i |P i >T};

[0110] According to the nodes where the fault points are concentrated, early warning information is pushed to the operation and maintenance personnel in advance.

[0111] The present invention also discloses a system for pushing operation and maintenance faults in advance based on big data analysis. Based on the above-mentioned method for pushing operation and maintenance faults in advance based on big data analysis, the system is characterized in that the system includes the following functional modules:

[0112] Data collection module, which is used to collect various types of data in the operation and maintenance system and pre-process the data;

[0113] The neural network analysis model building module builds the topological analysis strategy of the graph neural network based on the collected data, analyzes the topological graph, and updates the hidden state of the nodes;

[0114] The real-time adaptive analysis module builds a real-time adaptive analysis model based on the updated node hidden status, dynamically adjusts the model parameters and structure based on real-time data, and performs fault prediction;

[0115] The early warning information push module uses the topological analysis results based on the graph neural network and the prediction results of the adaptive analysis model to identify potential fault points and push the early warning information to the operation and maintenance personnel in a timely manner through email, text message or application.

[0116] In summary, the present invention combines the real-time adaptive analysis model and the topological analysis based on the graph neural network to achieve accurate fault prediction and advance push. This method dynamically adjusts the prediction parameters and structure through the real-time adaptive analysis model to adapt to changes in the operation and maintenance environment, thereby improving the accuracy and timeliness of fault prediction. At the same time, the topological analysis based on the graph neural network makes full use of the nodes and their connection relationships in the system to identify complex interaction patterns and potential faults, thereby achieving comprehensive fault management. The combination of the two not only improves the operation and maintenance efficiency and system reliability, but also enhances the scalability and maintainability of the system, providing an efficient and reliable solution for modern complex operation and maintenance environments.

[0117] This embodiment also provides a computer device, which is suitable for a method for early push of operation and maintenance faults based on big data analysis, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement a method for early push of operation and maintenance faults based on big data analysis as proposed in the above embodiment.

[0118] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0119] The present embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for early push of operation and maintenance faults based on big data analysis as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, disk or optical disk.

[0120] 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 method for early push of operation and maintenance faults based on big data analysis, characterized in that: The following steps are involved: S1: Collect various data in the operation and maintenance system and pre-process the data; S2: Construct a topological analysis strategy for the graph neural network based on the collected data, analyze the topological graph, and update the hidden state of the nodes; The topological analysis strategy for building a graph neural network includes the following steps: S21: The components and devices in the operation and maintenance system are regarded as nodes of the graph, and the connections between nodes are regarded as edges, that is, the network topology graph of the operation and maintenance system G=(V,E), where V is the node set and E is the edge set; S22: According to the network topology diagram The characteristic vector of For Node The abnormal state represents that the following formula is defined to update and propagate the state of the node; ； in, Representation Node The set of neighbor nodes of Neighbor node The abnormal state indicates that Node and The degree of and It represents the weight matrix of the kth layer, represents the activation function; S3: Build a real-time adaptive analysis model to dynamically adjust the model parameters and structure based on real-time data to predict faults; In the S3 process, the specific method of building a real-time adaptive analysis model is as follows: S31: Topological analysis based on neural network to calculate node failure scores ; dt; in, and represents the time interval, K represents the number of layers of GNN, represents the weight coefficient at the kth layer, represents the hyperbolic tangent function; S32: Calculate the comprehensive fault push value based on the obtained node fault score ; ； Where, |V| is the total number of nodes; Indicates the rate of change of the node turntable over time; Complex Fault Scoring : Integrals and exponential functions make The value range of is in the positive range, and the specific value depends on the system running time and the characteristics of the neighboring nodes; Comprehensive fault push value : Through averaging and normalization, The value range of is usually in the positive range, indicating the comprehensive failure risk of the system. The larger the value, the higher the failure risk. For each node Comprehensive fault push value It can be used as an indicator to measure its failure risk. A threshold T is set to push the comprehensive failure value Nodes exceeding the threshold are identified as potential failure points: Fault point set = ; According to the nodes where the fault points are concentrated, early warning information is pushed to the operation and maintenance personnel in advance; S4: Utilize the topological analysis results based on graph neural networks and the prediction results of the adaptive analysis model to identify potential failure points and push warning information to operation and maintenance personnel in a timely manner through multiple channels.

2. According to the method of claim 1, the method is characterized by: In the S1 process, the collected data include: topological data, node feature data, historical data and current status data of each node; The pre-processing comprises the following steps: Data cleaning; removing noise data and processing missing values; Data normalization: normalize data of different dimensions for subsequent analysis; Feature extraction: extracting features useful for analysis from raw data; Time series processing: Perform time series analysis and processing on real-time data and historical data.

3. According to claim 2, a method for early push of operation and maintenance faults based on big data analysis is characterized in that: Collect the failure records of all nodes in the system in the past year to determine the system failure rate and false alarm rate; By calculating the failure rate and false alarm rate under different thresholds, the receiver operating characteristic (ROC) curve is drawn; According to business needs, select a balance point on the ROC curve as the threshold ; and will Compare with the threshold θ to determine whether to push the subsequent comprehensive fault value calculate.

4. According to claim 3, a method for early push of operation and maintenance faults based on big data analysis is characterized in that: Assume that node A calculates the fault score according to the node fault score calculation formula: ; like , it is directly determined that there is no fault; like , then continue to calculate the comprehensive push value .

5. A system for pushing operation and maintenance faults in advance based on big data analysis, based on a method for pushing operation and maintenance faults in advance based on big data analysis according to any one of claims 1 to 4, characterized in that: The system includes the following functional modules: Data collection module, which is used to collect various types of data in the operation and maintenance system and pre-process the data; The neural network analysis model building module builds the topological analysis strategy of the graph neural network based on the collected data, analyzes the topological graph, and updates the hidden state of the nodes; The real-time adaptive analysis module builds a real-time adaptive analysis model based on the updated node hidden status, dynamically adjusts the model parameters and structure based on real-time data, and performs fault prediction; The early warning information push module uses the topological analysis results based on the graph neural network and the prediction results of the adaptive analysis model to identify potential fault points and push the early warning information to the operation and maintenance personnel in a timely manner through email, text message or application.

6. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of a method for early push of operation and maintenance faults based on big data analysis as described in any one of claims 1 to 4 are implemented.

7. 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 a method for early push of operation and maintenance faults based on big data analysis as described in any one of claims 1 to 4 are implemented.

Images