Intelligent operation and maintenance operation management system based on artificial intelligence

Abstract

The invention discloses an intelligent operation and maintenance operation management system based on artificial intelligence. The system comprises a multi-source data collection module, a data optimization module, a power plant equipment maintenance prediction model building module and an operation and maintenance intelligent operation management module. The invention relates to the technical field of data processing, in particular to an intelligent operation and maintenance management system based on artificial intelligence, which innovatively proposes that equipment is taken as a network node, equipment process dependence is taken as a directed edge, an equipment maintenance network diagram structure is constructed, a risk propagation chain between different equipment can be effectively identified, and the risk propagation efficiency is improved. Power plant operation and maintenance management is converted from local prediction to global correlation modeling; a direct and indirect double-layer spatial dependency joint modeling mechanism is adopted, direct spatial dependency is used for modeling an explicit process relationship between adjacent equipment through a graph attention network, indirect spatial dependency is used for modeling implicit association of non-adjacent equipment through a spatial attention mechanism, the prediction precision is improved, and the accuracy of intelligent operation and maintenance operation decision is improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of data processing, and particularly discloses an intelligent operation and maintenance management system based on artificial intelligence. BACKGROUND

[0002] The intelligent operation and maintenance management system based on artificial intelligence is a system for intelligently operating and maintaining power plant equipment by using artificial intelligence technology. The system can predict possible faults, performance degradation or abnormalities of the equipment by collecting and analyzing real-time data, and thus intelligently takes early operation and maintenance measures, effectively avoids sudden faults and shutdowns of the equipment, ensures that the equipment is always in the best operating state, reduces the operation and maintenance cost of the power plant, and realizes intelligent operation and maintenance management of the power plant.

[0003] However, the traditional intelligent operation and maintenance management system is limited to microscopic data analysis of a single device, and fails to consider the technical problems of multi-device parallel operation characteristics and cross-device process interaction relationship, so that the system cannot construct a global correlation network of the power plant, and is difficult to capture the global chain fault propagation effect caused by local device abnormalities, thereby causing the power plant operation and maintenance management decision to lag behind. The existing power plant equipment maintenance prediction model generally only considers the direct adjacent relationship between devices and ignores the indirect process coupling between non-adjacent devices, thereby causing insufficient prediction accuracy and weak response capability to the global process state, and affecting the reliability and foresight of the operation and maintenance decision. SUMMARY

[0004] In view of the above, in order to overcome the defects of the prior art, the present application provides an intelligent operation and management system based on artificial intelligence, which is aimed at the technical problem that the traditional intelligent operation and management system is limited to micro data analysis of single equipment, and fails to consider the multi-device parallel operation characteristics and cross-device process interaction relationship, resulting in that the system cannot build a global correlation network of the power plant, it is difficult to capture the global chain failure propagation effect caused by local equipment abnormalities, thereby making the power plant operation and management decision lag, the present scheme innovatively proposes to take the equipment as the network node, the equipment process dependence as the directed edge, and construct the equipment maintenance network graph structure, which can effectively identify the risk propagation chain between different devices, improve the identification ability of cross-device associated risks, enhance the global collaboration and dynamic responsiveness of operation and maintenance, improve the comprehensiveness and real-time of prediction, break through the limitations of traditional single device prediction, change the power plant operation and management from local prediction to global correlation modeling, significantly improve the collaboration and decision efficiency of the whole plant maintenance strategy, and promote the scientific allocation of operation and maintenance resources; in view of the technical problem that the existing power plant equipment maintenance prediction model generally only considers the direct adjacent relationship between devices and ignores the indirect process coupling between non-adjacent devices, thereby causing insufficient prediction accuracy and weak response ability to the global process state, affecting the reliability and foresight of operation and maintenance decision, the present scheme innovatively adopts a direct and indirect double-layer spatial dependence joint modeling mechanism, the direct spatial dependence models the explicit process relationship between adjacent devices through a graph attention network, the indirect spatial dependence models the implicit association of non-adjacent devices through a spatial attention mechanism, and finally the two types of spatial features are fused and aggregated to comprehensively capture the multi-level spatial dependence relationship, which can identify the potential risk propagation path and implicit dependence relationship between devices, significantly improve the global cognitive ability and prediction accuracy of the model, and enhance the explainability and robustness of the system to the cooperative operation state of multiple devices, thereby improving the accuracy and scientificity of intelligent operation and decision.

[0005] The technical scheme adopted by the present application is as follows: the intelligent operation and management system based on artificial intelligence provided by the present application comprises a multi-source data collection module, a data optimization module, an establishment of a power plant equipment maintenance prediction model module and an intelligent operation and management module;

[0006] The multi-source data collection module specifically obtains power plant operation and management original data through data acquisition operation;

[0007] The data optimization module specifically obtains power plant operation and management optimized data through data cleaning, data standardization and data coding processing;

[0008] The power plant equipment maintenance prediction model establishing module specifically constructs a device maintenance network graph structure by introducing a process topology weighting mechanism, then extracts maintenance time sequence features, and then extracts direct spatial features of explicit process relationships between adjacent devices and global indirect spatial features of implicit relationships between non-adjacent devices based on a graph attention network, and finally globally aggregates multi-layer features to output power plant equipment operation and maintenance demand prediction results, and trains the model based on historical operation and maintenance data to obtain a trained power plant equipment maintenance prediction model.

[0009] The operation and maintenance intelligent operation management module is used for realizing intelligent operation and maintenance operation management of the power plant equipment; specifically, real-time operation and maintenance operation data are input into the trained model to obtain real-time operation and maintenance operation results of the power plant, and based on the operation results, intelligent operation and maintenance management of the power plant equipment is realized.

[0010] Further, the multi-source data collection module specifically obtains power plant operation and maintenance operation raw data through data collection operations; the power plant operation and maintenance operation raw data includes historical operation and maintenance operation data and real-time operation and maintenance operation data; the historical operation and maintenance operation data and the real-time operation and maintenance operation data both include power plant equipment operation data, fault event data and equipment maintenance data; the historical operation and maintenance operation data further includes operation and maintenance operation records.

[0011] Further, the data optimization module specifically includes the following steps:

[0012] Data cleaning, specifically missing value filling, outlier removal and field standardization processing of raw data;

[0013] Data standardization processing, specifically mapping all continuous variables to the [0, 1] interval through the minimum-maximum normalization method;

[0014] Data encoding processing, specifically encoding processing of category fields in raw data by using one-hot encoding method.

[0015] Further, the power plant equipment maintenance prediction model establishing module specifically includes the following steps:

[0016] Constructing a device maintenance network graph structure, specifically defining a graph network node, defining a graph network edge, and embedding a multi-device parallel running relationship by introducing a process topology weighting mechanism to obtain a device maintenance network graph structure, a device process dependency adjacency matrix, and a multi-device graph network node set;

[0017] The graph network node definition specifically defines the power plant operation and maintenance equipment as a graph network node according to the power plant equipment operation data, and each node uniquely associates with the operation and maintenance operation data of the corresponding device;

[0018] The graph network edge definition is specifically defined according to the device process dependency relationship between power plant devices, the graph network edge is defined, and an adjacency matrix is generated based on the device process dependency relationship If there is a device process dependency between device c and device h, the adjacency matrix element Otherwise Where n is the total number of operation and maintenance devices

[0019] The device process dependency relationship is specifically defined as follows: if the operating state of device c directly affects the maintenance requirement of device h, a directed edge is established between the node corresponding to device c and the node corresponding to device h, forming a device direct process dependency relationship; specifically including three types of device process dependencies: energy transmission dependency, material supply dependency and control signal dependency;

[0020] The multi-device parallel operation relationship embedding is specifically defined as follows: a time window is sliced, the multi-device operation and maintenance operation data is sliced by the time window, the operation and maintenance operation data of the multi-device in the same time window is mapped to the initial features of the corresponding nodes, and a multi-device graph network node set is obtained

[0021] Maintenance feature time sequence dependency extraction, specifically from the multi-device graph network node set, extracting the graph network node features of each device node in T consecutive time windows, constructing the time sequence features of the corresponding device node, and then using a time sequence attention mechanism model to calculate the original attention weight between different time windows, obtaining the original attention weight, and then normalizing the original attention weight to obtain the normalized time sequence attention weight, and multiplying it with the time sequence features to obtain the maintenance time sequence features

[0022] Adjacent device direct spatial feature extraction, specifically according to the device process dependency adjacency matrix, locating the adjacent device set of each device node, and inputting the maintenance time sequence features into the graph attention network, mapping the features of the current node and the adjacent node in dimension, calculating the attention weight of the node and the adjacent node, and using a multi-head attention mechanism to aggregate multiple groups of adjacent graph node features and update the current node features to obtain the direct spatial features

[0023] Device global indirect spatial feature extraction is used to capture the global indirect process dependency information between device nodes in the graph network; specifically including the following steps:

[0024] Multi-hop adjacent relationship positioning, specifically based on the device process dependency adjacency matrix, a multi-hop adjacency matrix is generated by matrix power operation Wherein, the element =1 in the multi-hop adjacency matrix indicates that the device and have v-hop indirect process dependency, v represents the number of hops, and all v-hop adjacency matrices are integrated to construct a global indirect adjacency set ;

[0025] The indirect process depends on the indirect association relationship formed by 1 or more intermediate devices between the device nodes without direct process dependence;

[0026] Global indirect attention weight calculation, specifically, input the maintenance time sequence feature into the graph attention network, map the features of the current device and the indirect device through the feature conversion matrix, and calculate the global indirect attention weight of the device and the indirect device; The formula used is as follows:

[0027] ;

[0028] In the formula, denotes the normalized global attention weight of the current device node c and the v-hop indirect device node u, denotes the transpose of the global indirect attention coefficient vector, denotes the feature conversion matrix in the global indirect attention weight calculation, denotes the maintenance time sequence feature of the i-th time window device node u, denotes the feature splicing operation;

[0029] Global feature aggregation, specifically, introduce a multi-hop distance decay factor, and first aggregate the time sequence features and global indirect attention weights of the v-hop indirect device according to the number of hops v respectively, then weight the aggregation results of different hop numbers by the hop number weight, and at the same time, combine the multi-head attention mechanism, splice the aggregated features of multiple groups to update the features of the current device node, and obtain the global indirect spatial feature; The formula used is as follows:

[0030] ;

[0031] In the formula, denotes the global indirect spatial feature of the device node c, denotes the number of global attention heads, and V denotes the maximum hop number, denotes the hop number weight, denotes the multi-hop distance decay factor of the hop number v, the value range is between , denotes the v-hop indirect device set of the device c, denotes the normalized attention weight of the v-hop indirect device u to the current device c in the g-th attention head, denotes the global feature conversion matrix of the g-th attention head, denotes the multi-head feature splicing operation;

[0032] The power plant equipment operation and maintenance demand prediction output is specifically that the direct spatial features and the global indirect spatial features are fused across dimensions to generate spatio-temporal fusion features, the spatio-temporal fusion features are input into a full connection layer, after nonlinear feature mapping and dimension adaptation are completed by the full connection layer, the probability distribution of each device corresponding to different operation and maintenance operation categories is quantitatively output by a Softmax activation function, and the category corresponding to the maximum probability is selected as the power plant equipment operation and maintenance demand prediction result;

[0033] The model is constructed and trained, specifically, the power plant equipment maintenance prediction model is constructed by constructing a device maintenance network graph structure, extracting maintenance feature time series dependence, extracting adjacent device direct spatial features, extracting device global indirect spatial features and power plant equipment operation and maintenance demand prediction output, and the model is trained based on historical operation and maintenance operation data as training data, and finally the trained power plant equipment maintenance prediction model is obtained.

[0034] Further, the operation and maintenance intelligent operation management module is specifically that real-time operation and maintenance operation data is input into the trained power plant equipment maintenance prediction model to obtain a power plant real-time operation and maintenance operation result, and corresponding operation and maintenance decision and execution instruction are generated according to the power plant real-time operation and maintenance operation result, intelligent operation and maintenance operation is performed, early warning is automatically sent, maintenance tasks are generated and device maintenance operation personnel are dispatched, and intelligent operation and maintenance operation management of the power plant equipment is realized.

[0035] The above scheme has the following beneficial effects:

[0036] (1) The technical problem that the traditional intelligent operation and maintenance operation management system is limited to micro data analysis of a single device and fails to consider the multi-device parallel operation characteristics and the cross-device process interaction relationship, which leads to the inability of the system to construct a global correlation network of the power plant, the difficulty in capturing the global chain fault propagation effect caused by local device abnormalities, and the lag of the power plant operation and maintenance operation management decision. The present scheme innovatively proposes to take the device as the network node and the device process dependence as the directed edge to construct a device maintenance network graph structure, which can effectively identify the risk propagation chain between different devices, improve the identification ability of cross-device associated risks, enhance the global collaboration and dynamic responsiveness of the operation and maintenance operation, improve the comprehensiveness and real-time performance of the prediction, break through the limitations of traditional single-device prediction, change the power plant operation and maintenance management from local prediction to global correlation modeling, significantly improve the collaboration and decision efficiency of the whole plant maintenance strategy, and promote the scientific allocation of operation and maintenance resources.

[0037] (2) In view of the technical problem that the existing power plant equipment maintenance prediction model generally only considers the direct adjacent relationship between devices and ignores the indirect process coupling between non-adjacent devices, thereby causing insufficient prediction accuracy, weak response ability to the global process state, and affecting the reliability and forward-looking nature of operation and maintenance decisions, the scheme innovatively adopts a direct and indirect double-layer spatial dependence joint modeling mechanism, the direct spatial dependence models the explicit process relationship between adjacent devices through a graph attention network, the indirect spatial dependence models the implicit association between non-adjacent devices through a spatial attention mechanism, and finally the two types of spatial features are fused and aggregated to comprehensively capture the multi-level spatial dependence relationship, identify the potential risk propagation path and implicit dependence relationship between devices, significantly improve the global cognitive ability and prediction accuracy of the model, and enhance the explainability and robustness of the system to the coordinated operation state of multiple devices, thereby improving the accuracy and scientificity of intelligent operation and maintenance operation decisions. BRIEF DESCRIPTION OF DRAWINGS

[0038] Figure 1 A module schematic diagram of an intelligent operation and maintenance management system based on artificial intelligence is provided in the present application.

[0039] Figure 2 A flowchart of the power plant equipment maintenance prediction model module is provided.

[0040] The accompanying drawings are used to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation on the present application. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0042] In the description of the present application, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inner", "outer" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the systems or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application.

[0043] Embodiment one, refer to Figure 1The application provides an intelligent operation and management system based on artificial intelligence, which comprises a multi-source data collection module, a data optimization module, an electric power plant equipment maintenance prediction model establishment module and an intelligent operation and management module.

[0044] The multi-source data collection module collects original data of the operation and maintenance of the electric power plant through data acquisition and sends the data to the data optimization module.

[0045] The data optimization module receives the data sent by the multi-source data collection module, improves the quality of the original data and the model adaptability, and obtains optimized data of the operation and maintenance of the electric power plant through data cleaning, data standardization and data coding processing, and sends the data to the electric power plant equipment maintenance prediction model establishment module and the intelligent operation and management module.

[0046] The electric power plant equipment maintenance prediction model establishment module receives the data sent by the data optimization module, constructs a device maintenance network graph structure by introducing a process topology weighting mechanism, extracts maintenance time sequence features, extracts direct spatial features of explicit process relationships between adjacent devices and global indirect spatial features of implicit correlations between non-adjacent devices based on a graph attention network, globally aggregates multi-layer features, outputs a prediction result of the operation and maintenance requirements of the electric power plant equipment, trains the model based on historical operation and maintenance data, obtains a trained electric power plant equipment maintenance prediction model, and sends the data to the intelligent operation and management module.

[0047] The intelligent operation and management module receives the data sent by the data optimization module and the electric power plant equipment maintenance prediction model establishment module, realizes intelligent operation and management of the electric power plant equipment, inputs real-time operation and maintenance data into the trained model, obtains real-time operation and maintenance results of the electric power plant, and realizes intelligent operation and management of the electric power plant equipment based on the operation and maintenance results.

[0048] Embodiment two, refer to Figure 1 Based on the above-mentioned embodiment, the multi-source data collection module is used for collecting multi-dimensional data of the electric power plant equipment in the operation and maintenance management process, specifically, a plurality of types of sensors, data acquisition terminals and industrial internet of things nodes deployed on key devices of the electric power plant are used to collect equipment operation parameter data, and original data of the operation and maintenance of the electric power plant are obtained; the original data of the operation and maintenance of the electric power plant comprise historical operation and maintenance data and real-time operation and maintenance data; the historical operation and maintenance data and the real-time operation and maintenance data both comprise electric power plant equipment operation data, fault event data and equipment maintenance data; the historical operation and maintenance data further comprise operation and maintenance records.

[0049] The power plant equipment operation data includes equipment name, equipment power output, load rate, temperature, pressure, rotating speed, current, voltage, vibration frequency; the fault event data includes historical fault type, fault occurrence time, downtime length, fault site and fault cause analysis; and the equipment maintenance data includes equipment overhaul date, maintenance period, maintenance content, maintenance mode, maintenance time and replacement part information.

[0050] Embodiment three, refer to Figure 1 The embodiment is based on the above embodiment, and the data optimization module is used to improve the integrity, consistency and effectiveness of the original data, specifically for data cleaning, data standardization and data coding processing of the original data to obtain power plant operation and maintenance operation optimization data; including the following steps:

[0051] Data cleaning is used to improve the quality and consistency of the original data, specifically for missing value filling, outlier removal and field standardization processing of the original data;

[0052] The missing value filling is specifically to complete the missing values by the mean filling method; the outlier removal is specifically to detect and remove the extreme values and logical outliers in the original data by the Z-Score algorithm; and the field standardization processing is specifically to standardize the data of different formats by a unified unit conversion rule to ensure the consistency of the model input;

[0053] Data standardization processing is used to unify the dimension of structured numerical data, specifically to map all continuous variables to the [0, 1] interval by the min-max normalization method;

[0054] Data coding processing is used to convert non-numerical data into structured data for model calculation, specifically to encode the category fields in the original data by the one-hot encoding method to convert discrete text and label variables into sparse numerical vectors.

[0055] Embodiment four, refer to Figure 1 And Figure 2 The embodiment is based on the above embodiment, and the power plant equipment maintenance prediction model module is established, specifically including the following steps:

[0056] The equipment maintenance network graph structure is constructed, which is used to convert the scattered equipment operation and maintenance operation data into a structured graph network based on the operation correlation and multi-device parallel characteristics of the power plant equipment; specifically, the graph network node definition, graph network edge definition and multi-device parallel operation relationship embedding are performed by introducing the process topology weighting mechanism to obtain the equipment maintenance network graph structure, equipment process dependency adjacency matrix and multi-device graph network node set.

[0057] The graph network node is defined as a power plant operation and maintenance equipment defined as a graph network node according to power plant equipment operation data, and each node is uniquely associated with the operation and maintenance data of the corresponding equipment.

[0058] The graph network edge is defined according to the equipment process dependency relationship between the power plant equipment, and the adjacency matrix is generated based on the equipment process dependency relationship. The adjacency matrix If there is an equipment process dependency between equipment c and equipment h, the adjacency matrix element Otherwise Where n is the total number of operation and maintenance equipment.

[0059] The equipment process dependency relationship is specifically defined as follows: if the operation state of equipment c directly affects the maintenance requirement of equipment h, a directed edge is established between the node corresponding to equipment c and the node corresponding to equipment h, forming a direct equipment process dependency relationship. Specifically, it includes three types of equipment process dependencies: energy transmission dependency, material supply dependency and control signal dependency.

[0060] The energy transmission dependency is specifically defined as follows: equipment c provides core energy for equipment h, and the energy output state of equipment c directly affects the operation stability and maintenance requirement of equipment h.

[0061] The material supply dependency is specifically defined as follows: equipment c provides key production materials for equipment h, and the material supply state of equipment c directly affects the operation stability and maintenance requirement of equipment h.

[0062] The control signal dependency is specifically defined as follows: equipment c provides control instruction signals for equipment h, and the signal output state of equipment c directly affects the operation stability and maintenance requirement of equipment h.

[0063] The multi-device parallel operation relationship embedding is specifically defined as follows: a time window is set to 1 hour, and the multi-device operation and maintenance data is sliced by the time window. The operation and maintenance data of multiple devices in the same time window is mapped to the initial features of the corresponding nodes to obtain a set of multi-device graph network nodes.

[0064] The multi-device parallel operation is specifically defined as follows: in the same time window, different power plant equipment synchronously execute operation and maintenance related events around the same production target or process link.

[0065] Maintenance feature temporal dependency extraction is used to capture the temporal dynamics of power plant equipment maintenance features. Specifically, it extracts graph network node features for each equipment node over T consecutive time windows from a multi-equipment graph network node set, constructing the temporal features of the corresponding equipment nodes. Then, a temporal attention mechanism model is used to calculate the original attention weights between different time windows, obtaining the original attention weights. Subsequently, the original attention weights are normalized using the Softmax activation function to obtain normalized temporal attention weights. These normalized temporal attention weights are then multiplied by the temporal features to obtain the maintenance temporal features. The formula used is as follows:

[0066] ;

[0067] ;

[0068] In the formula, The elements in the original attention weight matrix represent the influence strength of the temporal features of the device node in the j-th time window on the temporal features of the device node in the i-th time window. This represents a learnable weight vector. , and This represents a learnable parameter matrix used for dimensional mapping and interaction of maintenance features. Indicates the first The temporal characteristics of each time window Indicates the first The temporal characteristics of each time window Learnable bias terms are used to adjust the baseline level of the original attention weights. Indicates the first Maintenance timing characteristics of each time window, Represents the normalized temporal attention weights;

[0069] Direct spatial feature extraction of adjacent devices is used to aggregate explicit process dependency information of adjacent devices in the graph network, capturing the impact of direct process associations between devices on maintenance features. Specifically, based on the device process dependency adjacency matrix, the set of adjacent devices for each device node is located, and the maintenance time-series features are input into the graph attention network. Dimensional mapping is performed on the features of the current node and its adjacent nodes, and the attention weights between the node and its adjacent nodes are calculated. A multi-head attention mechanism is used to aggregate multiple sets of adjacency graph node features and update the current node's features to obtain the direct spatial features. The formula used is as follows:

[0070] ;

[0071] ;

[0072] In the formula, denotes the normalized attention weight of the current device node c and the adjacent device node h, which is used to quantify the direct process dependency influence strength of h on c, denotes the transpose of the attention coefficient vector, which is used to calculate the interaction score of the spliced features, denotes the feature transformation matrix, denotes the maintenance time sequence feature of the i-th time window device node c, denotes the maintenance time sequence feature of the i-th time window device node h, denotes the feature splicing operation, denotes the direct spatial feature of the device node c, denotes the multi-head feature splicing operation, which splices the output features of K attention heads along the dimension, and K represents the number of attention heads, denotes the feature transformation matrix of the k-th attention head, denotes the set of adjacent devices of the current device node c, denotes the normalized attention weight of the current node c and the adjacent node h in the k-th attention head;

[0073] Device global indirect spatial feature extraction is used to capture the global indirect process dependency information between device nodes in the graph network; specifically including the following steps:

[0074] Multi-hop adjacent relationship positioning, specifically based on the device process dependency adjacency matrix, a multi-hop adjacency matrix is generated by matrix power operation , wherein the element in the multi-hop adjacency matrix = 1, indicating that the device has a v-hop indirect process dependency with , v represents the number of hops, and all v-hop adjacency matrices are integrated to construct a global indirect adjacency set ;

[0075] The indirect process dependency refers to an indirect association relationship formed by one or more intermediate devices between device nodes without direct process dependency;

[0076] Global indirect attention weight calculation, specifically inputting the maintenance time sequence feature into the graph attention network, and performing high-dimensional mapping on the features of the current device and the indirect device through the feature transformation matrix to calculate the global indirect attention weight of the device and the indirect device; the formula used is as follows:

[0077] ;

[0078] In the formula, denotes the normalized global attention weight of the current device node c and the v-hop indirect device node u, denotes the transpose of the global indirect attention coefficient vector, denotes a feature conversion matrix in global indirect attention weight calculation, denotes a maintenance time sequence feature of an i-th time window device node u;

[0079] Global feature aggregation, specifically introducing a multi-hop distance attenuation factor, and first aggregating the time sequence features of indirect devices and global indirect attention weights of v-hop respectively according to the number of hops v, then weighting and summing the aggregation results of different hop numbers through hop number weights, and combining a multi-head attention mechanism to update the features of the current device node by concatenating multiple groups of aggregated features to obtain global indirect spatial features; the formula used is as follows:

[0080] ;

[0081] In the formula, denotes a global indirect spatial feature of a device node c, denotes a global attention head number, and V denotes a maximum hop number, denotes a hop number weight, denotes a multi-hop distance attenuation factor of hop number v, the value range is between , denotes a v-hop indirect device set of device c, denotes a normalized attention weight of a v-hop indirect device u to the current device c in the g-th attention head, denotes a global feature conversion matrix of the g-th attention head, denotes a multi-head feature concatenation operation;

[0082] The power plant device operation and maintenance demand prediction output specifically cross-dimensionally fuses the direct spatial features and the global indirect spatial features to generate spatio-temporal fusion features, then inputs the spatio-temporal fusion features into a fully connected layer, and after completing nonlinear feature mapping and dimension adaptation through the fully connected layer, quantitatively outputs the probability distribution of different operation and maintenance operation categories corresponding to each device through a Softmax activation function, and selects the category corresponding to the maximum probability as the power plant device operation and maintenance demand prediction result.

[0083] The device operation and maintenance operation category includes emergency shutdown maintenance, planned shutdown maintenance, online component replacement, enhanced inspection, and routine monitoring.

[0084] The model is constructed and trained, specifically by constructing a device maintenance network graph structure, extracting maintenance feature time sequence dependencies, extracting adjacent device direct spatial features, extracting device global indirect spatial features, and outputting power plant device operation and maintenance demand prediction, constructing a power plant device maintenance prediction model, and based on historical operation and maintenance operation data as training data, the model is trained, and finally a trained power plant device maintenance prediction model is obtained.

[0085] By performing the above operations, this solution addresses the technical problem of traditional intelligent operation and maintenance management systems, which are limited to micro-data analysis of single devices and fail to consider the parallel operation characteristics of multiple devices and cross-device process interactions. This results in the system's inability to construct a global interconnected network for the power plant, making it difficult to capture the global cascading failure propagation effect caused by local device anomalies, thus leading to lagging power plant operation and maintenance management decisions. This solution innovatively proposes constructing a device maintenance network graph structure with devices as network nodes and device process dependencies as directed edges. This effectively identifies risk propagation chains between different devices, improves the ability to identify cross-device associated risks, enhances the global coordination and dynamic responsiveness of operation and maintenance operations, and improves the comprehensiveness and real-time nature of predictions. It breaks through the limitations of traditional single-device prediction, transforming power plant operation and maintenance management from local prediction to global interconnected modeling, significantly improving the coordination and decision-making efficiency of plant-wide maintenance strategies, and promoting the scientific management of operation and maintenance resources. To address the technical problem that existing power plant equipment maintenance prediction models generally only consider direct adjacency relationships between equipment and ignore indirect process coupling between non-adjacent equipment, resulting in insufficient prediction accuracy and weak responsiveness to global process states, thus affecting the reliability and foresight of operation and maintenance decisions, this solution innovatively adopts a joint modeling mechanism of direct and indirect spatial dependencies. Direct spatial dependencies model explicit process relationships between adjacent equipment through graph attention networks, while indirect spatial dependencies model implicit associations between non-adjacent equipment through spatial attention mechanisms. Finally, the two types of spatial features are fused and aggregated to achieve a comprehensive capture of multi-level spatial dependencies. This enables the identification of potential risk propagation paths and implicit dependencies between equipment, significantly improving the model's global cognitive ability and prediction accuracy, enhancing the system's interpretability and robustness to the collaborative operation status of multiple equipment, and thus improving the accuracy and scientific nature of intelligent operation and maintenance decision-making.

[0086] Example 5, see Figure 1 This embodiment is based on the above embodiment. Specifically, the intelligent operation and maintenance management module inputs real-time operation and maintenance data into the trained power plant equipment maintenance prediction model to obtain the real-time operation and maintenance results of the power plant. Based on the real-time operation and maintenance results of the power plant, it generates corresponding operation and maintenance decisions and execution instructions, executes intelligent operation and maintenance operations, automatically issues early warnings, generates maintenance tasks, and dispatches equipment maintenance personnel to realize intelligent operation and maintenance management of power plant equipment.

[0087] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting; it is not intended to exclude myriad other embodiments of the present application that other inventors can develop based on the same general inventive concepts embodied by the described embodiments. That is, although the present application is described in terms of particular embodiments and implementations, it is to be understood that the terminology used is for the purpose of descriptive clarity and that it is intended to be limited only by the claims.

[0088] While the embodiments of the application have been shown and described, it is to be understood that the embodiments described are only by way of example and that modifications, changes, substitutions and variations can be made by those skilled in the art without departing from the spirit and scope of the application.

[0089] The above description of the application and its embodiments is not intended to limit the application, as described by the appended claims, to the embodiments described above. Rather, it is intended to cover all adaptations, modifications and variations of the specific embodiments of the application chosen by the inventors as coming within the scope of the application.

Claims

1. An intelligent operation and maintenance management system based on artificial intelligence, characterized in that: It includes a multi-source data collection module, a data optimization module, a module for establishing predictive models for power plant equipment maintenance, and an intelligent operation and management module for operation and maintenance; The multi-source data collection module specifically obtains raw data of power plant operation and maintenance through data acquisition operations. The data optimization module specifically obtains optimized power plant operation and maintenance data through data cleaning, data standardization, and data encoding. The module for establishing a power plant equipment maintenance prediction model specifically involves constructing an equipment maintenance network graph structure by introducing a process topology weighting mechanism, then extracting maintenance time-series features, and then extracting the direct spatial features of explicit process relationships between adjacent equipment and the global indirect spatial features of implicit relationships between non-adjacent equipment based on a graph attention network. Finally, the multi-layer features are globally aggregated to output the power plant equipment operation and maintenance demand prediction results. The model is then trained based on historical operation and maintenance data to obtain the trained power plant equipment maintenance prediction model. The intelligent operation and maintenance management module specifically inputs real-time operation and maintenance data into the trained model to obtain real-time operation and maintenance results of the power plant, and realizes intelligent operation and maintenance management of power plant equipment based on these results.

2. The intelligent operation and maintenance management system based on artificial intelligence according to claim 1, characterized in that: The module for establishing a power plant equipment maintenance prediction model specifically includes the following steps: Construct the equipment maintenance network diagram structure; The maintenance feature temporal dependency extraction is specifically carried out by extracting the graph network node features of each device node for T consecutive time windows from the multi-device graph network node set, constructing the temporal features of the corresponding device nodes, and then using a temporal attention mechanism model to calculate the original attention weights between different time windows to obtain the original attention weights. Subsequently, the original attention weights are normalized to obtain normalized temporal attention weights, which are then multiplied by the temporal features to obtain the maintenance temporal features. The direct spatial feature extraction of adjacent devices is specifically carried out by locating the set of adjacent devices for each device node based on the device process dependency adjacency matrix, inputting the maintenance time-series features into the graph attention network, performing dimensional mapping on the features of the current node and adjacent nodes, calculating the attention weights of the node and adjacent nodes, and using a multi-head attention mechanism to aggregate multiple sets of adjacency graph node features and update the current node features to obtain the direct spatial features. Extraction of global indirect spatial features of equipment; The power plant equipment operation and maintenance demand prediction output specifically involves cross-dimensional fusion of direct spatial features and global indirect spatial features to generate spatiotemporal fusion features. These spatiotemporal fusion features are then input into a fully connected layer. After nonlinear feature mapping and dimension adaptation are completed through the fully connected layer, the probability distribution of different operation and maintenance operation categories corresponding to each device is quantized and output through the Softmax activation function. The category with the highest probability is selected as the power plant equipment operation and maintenance demand prediction result. The model is constructed and trained by building a network graph structure for equipment maintenance, extracting the temporal dependency of maintenance features, extracting the direct spatial features of adjacent equipment, extracting the global indirect spatial features of equipment, and predicting the operation and maintenance needs of power plant equipment. The model is then trained using historical operation and maintenance data to obtain the final trained power plant equipment maintenance prediction model.

3. The intelligent operation and maintenance management system based on artificial intelligence according to claim 1, characterized in that: The construction of the equipment maintenance network graph structure specifically involves defining graph network nodes, graph network edges, and embedding multi-device parallel operation relationships by introducing a process topology weighting mechanism, thereby obtaining the equipment maintenance network graph structure, the equipment process dependency adjacency matrix, and the set of multi-device graph network nodes. The graph network node definition is specifically as follows: power plant operation and maintenance equipment is defined as graph network nodes based on power plant equipment operation data, and each node is uniquely associated with the operation and maintenance data of the corresponding equipment. The graph network edge definition specifically involves defining graph network edges based on the equipment process dependencies between power plant equipment, and generating an adjacency matrix based on these dependencies. If device c has a process dependency on device h, then the adjacency matrix elements ,otherwise , where n is the total number of maintenance equipment; The equipment process dependency relationship is specifically defined as follows: if the operating status of equipment c directly affects the maintenance requirements of equipment h, then a directed edge is established between the node corresponding to equipment c and the node corresponding to equipment h, forming a direct equipment process dependency relationship; specifically, it includes three types of equipment process dependencies: energy transfer dependency, material supply dependency, and control signal dependency. The embedding of the multi-device parallel operation relationship specifically adopts time window slicing. The operation and maintenance data of multiple devices are sliced ​​through time windows, and the operation and maintenance data of multiple devices within the same time window are mapped to the initial features of the corresponding nodes to obtain a set of nodes in the multi-device graph network.

4. The intelligent operation and maintenance management system based on artificial intelligence according to claim 2, characterized in that: The extraction of the device's global indirect spatial features specifically includes the following steps: Multi-hop adjacency location; The calculation of global indirect attention weights involves inputting the maintained temporal features into the graph attention network, performing a high-dimensional mapping of the features of the current device and indirect devices using a feature transformation matrix, and then calculating the global indirect attention weights between the device and the indirect devices. The formula used is as follows: ； In the formula, This represents the normalized global attention weight between the current device node c and the device node u hop from v. This represents the transpose of the global indirect attention coefficient vector. This represents the feature transformation matrix used in the calculation of global indirect attention weights. This represents the maintenance timing characteristics of device node u in the i-th time window. Indicates feature concatenation operation; Global feature aggregation specifically involves introducing a multi-hop distance decay factor, and first aggregating the temporal features of v-hop indirect devices and global indirect attention weights according to the number of hops v. Then, the aggregation results of different hop counts are weighted and summed. Simultaneously, combined with a multi-head attention mechanism, multiple sets of aggregated features are concatenated to update the features of the current device node, resulting in global indirect spatial features. The formula used is as follows: ； In the formula, This represents the global indirect space characteristics of device node c. This represents the number of global attention heads, and V represents the maximum number of hops. Indicates the weight of the number of hops. This represents the multi-hop distance decay factor, which indicates the number of hops v. The range of values ​​is between, This represents the set of devices in the v-hop indirection of device c. Let represent the normalized attention weight of the v-hop indirect device u in the g-th attention head with respect to the current device c. This represents the global feature transformation matrix of the g-th attention head. This indicates a multi-head feature splicing operation.

5. The intelligent operation and maintenance management system based on artificial intelligence according to claim 4, characterized in that: The multi-hop adjacency relationship localization is specifically based on the equipment process-dependent adjacency matrix, and generates the multi-hop adjacency matrix through matrix exponentiation. Among them, the elements in the multi-hop adjacency matrix =1 indicates the device and There exists a v-hop indirect process dependency, where v represents the hop count. By integrating all v-hop adjacency matrices, a global indirect adjacency set is constructed. The indirect process dependency refers to an indirect relationship formed through one or more intermediate devices, where there is no direct process dependency between equipment nodes.

6. The intelligent operation and maintenance management system based on artificial intelligence according to claim 1, characterized in that: The multi-source data collection module specifically obtains raw power plant operation and maintenance data through data acquisition operations. The raw power plant operation and maintenance data includes historical operation and maintenance data and real-time operation and maintenance data. Both the historical operation and maintenance data and the real-time operation and maintenance data include power plant equipment operation data, fault event data, and equipment maintenance data. The historical operation and maintenance data also includes operation and maintenance records.

7. The intelligent operation and maintenance management system based on artificial intelligence according to claim 1, characterized in that: The data optimization module specifically includes the following steps: Data cleaning specifically involves imputing missing values, removing outliers, and normalizing fields in the original data. Data standardization is specifically performed by mapping all continuous variables to the [0,1] interval using the min-max normalization method. The data encoding process specifically involves using one-hot encoding to encode the category fields in the original data.

8. The intelligent operation and maintenance management system based on artificial intelligence according to claim 1, characterized in that: The intelligent operation and maintenance management module specifically inputs real-time operation and maintenance data into the trained power plant equipment maintenance prediction model to obtain real-time operation and maintenance results of the power plant. Based on the real-time operation and maintenance results of the power plant, it generates corresponding operation and maintenance decisions and execution instructions, executes intelligent operation and maintenance operations, automatically issues early warnings, generates maintenance tasks, and dispatches equipment maintenance personnel to realize intelligent operation and maintenance management of power plant equipment.

Images