CN-122020011-A - Knowledge-graph-fused intelligent diagnosis method and system for power generation equipment

CN122020011ACN 122020011 ACN122020011 ACN 122020011ACN-122020011-A

Abstract

The invention discloses an intelligent diagnosis method and system for power generation equipment integrating knowledge maps, and relates to the technical field of equipment diagnosis. The method comprises the steps of collecting multisource operation data of the power generation equipment and historical fault records of the power generation equipment, constructing a power generation equipment fault knowledge graph, monitoring the power generation equipment real-time operation data in real time, carrying out feature extraction and semantic association on the power generation equipment real-time operation data based on the power generation equipment fault knowledge graph to generate a power generation equipment operation state vector, carrying out reasoning analysis on the power generation equipment operation state vector, mining equipment potential fault modes and equipment fault causal chains, carrying out fault positioning diagnosis based on the equipment potential fault modes and the equipment fault causal chains, and determining a power generation equipment diagnosis result. The technical problem that the diagnosis accuracy is low due to the fact that operation data of the power generation equipment are scattered and a fault mechanism is difficult to express explicitly in the prior art is solved, and the technical effect that potential fault modes are intelligently mined by utilizing a graph neural network is achieved, so that the fault diagnosis accuracy is improved.

Inventors

LI YING

Assignees

辽宁大唐国际新能源有限公司锦州热电分公司

Dates

Publication Date: 20260512
Application Date: 20251204

Claims (10)

1. The intelligent diagnosis method of the power generation equipment integrating the knowledge graph is characterized by comprising the following steps of: Collecting multi-source operation data of the power generation equipment and historical fault records of the power generation equipment, modeling the map relation between the multi-source operation data of the power generation equipment and the historical fault records of the power generation equipment, and constructing a power generation equipment fault knowledge map; real-time monitoring real-time operation data of the power generation equipment, and performing feature extraction and semantic association on the real-time operation data of the power generation equipment based on the fault knowledge graph of the power generation equipment to generate an operation state vector of the power generation equipment; carrying out reasoning analysis on the operation state vector of the power generation equipment by adopting a graph neural network, and mining a potential failure mode and a failure causal chain of the equipment; and carrying out fault positioning diagnosis based on the equipment potential fault mode and the equipment fault causal chain, and determining a power generation equipment diagnosis result, wherein the power generation equipment diagnosis result comprises equipment health state, fault mode occurrence probability and fault influence root cause.
2. The knowledge-graph-fused intelligent diagnosis method for power generation equipment of claim 1, wherein constructing a power generation equipment fault knowledge graph comprises: Performing data cleaning and standardization processing on the multi-source operation data of the power generation equipment and the historical fault records of the power generation equipment to obtain available power generation equipment operation data and available power generation equipment fault data; defining entity types, entity attributes and relationship types according to the fault diagnosis requirements of the power generation equipment; Carrying out knowledge extraction on the operation data of the available power generation equipment and the fault data of the available power generation equipment according to the entity type, the entity attribute and the relationship type to obtain equipment knowledge entity data and knowledge entity relationship data; and carrying out map relation modeling based on the equipment knowledge entity data and the knowledge entity relation data, and constructing a power generation equipment fault knowledge map.
3. The knowledge-graph-integrated intelligent diagnosis method for power generation equipment according to claim 2, wherein the modeling of the graph relationship based on the equipment knowledge-entity data and the knowledge-entity relationship data, and constructing a power generation equipment fault knowledge graph, comprises: Carrying out map driven modeling on the equipment knowledge entity data according to the knowledge entity relation data to generate an initial equipment fault knowledge map; performing fault rule mining based on the available power generation equipment operation data and the available power generation equipment fault data to obtain an equipment fault association rule set; Constructing an equipment fault propagation logic chain according to the equipment fault association rule set; And carrying out map fusion correction on the initial equipment fault knowledge map based on the equipment fault propagation logic chain, and constructing a power generation equipment fault knowledge map.
4. The knowledge-graph-fused intelligent diagnosis method for power generation equipment of claim 1, wherein generating the power generation equipment operation state vector comprises: extracting time-frequency, time domain and depth characteristics of the real-time operation data of the power generation equipment to obtain an equipment operation multidimensional characteristic vector; Dynamically weighting and fusing the equipment operation multidimensional feature vectors according to the operation working conditions of the power equipment to obtain equipment operation fused feature vectors; And mapping the equipment operation fusion feature vector to the power generation equipment fault knowledge graph to perform semantic association matching, and generating a power generation equipment operation state vector.
5. The knowledge-graph-fused intelligent diagnosis method for power generation equipment according to claim 1, wherein the mining of the equipment latent fault mode and the equipment fault causal link comprises: extracting and obtaining the operation state data of the power generation equipment in the power generation equipment fault knowledge graph; Constructing an equipment fault diagram structure according to the power generation equipment fault knowledge map, and selecting a diagram neural network based on the equipment fault diagram structure; performing task training on the power generation equipment operation state data and the power generation equipment fault knowledge graph by adopting the graph neural network to generate a power generation equipment fault reasoning model; and carrying out reasoning analysis on the operation state vector of the power generation equipment based on the power generation equipment fault reasoning model, and outputting an equipment potential fault mode and an equipment fault causal chain.
6. The knowledge-graph-fused intelligent diagnosis method for power generation equipment of claim 5, wherein generating a power generation equipment fault inference model comprises: taking the power generation equipment fault knowledge graph as input, and taking the power generation equipment running state data as node characteristics to obtain power generation equipment fault sample data; Acquiring a power generation equipment fault reasoning task, wherein the power generation equipment fault reasoning task comprises an equipment fault classification task and a fault cause and effect analysis task; and performing task training on the power generation equipment fault sample data based on the power generation equipment fault reasoning task by adopting the graph neural network to generate a power generation equipment fault reasoning model.
7. The knowledge-graph-fused intelligent diagnosis method for power generation equipment according to claim 1, wherein determining the diagnosis result of the power generation equipment comprises: Performing cause positioning based on the equipment fault cause and effect chain to obtain a fault influence cause; performing conditional probability calculation on the potential failure mode of the equipment according to the equipment failure causal chain to obtain failure mode occurrence probability; And determining the health state of the equipment according to the occurrence probability of the fault mode and the root cause of the fault influence, and forming a diagnosis result of the power generation equipment based on the occurrence probability of the fault mode, the root cause of the fault influence and the health state of the equipment.
8. The knowledge-graph-fused intelligent diagnostic method for a power generation device of claim 7, wherein determining the device health status comprises: evaluating the influence degree of the fault influence root cause to obtain a fault root cause influence coefficient; And taking the product of the occurrence probability of the fault mode and the influence coefficient of the fault root as a device health index, defining a health grade threshold value, analyzing the health state of the device health index, and determining the health state of the device.
9. The knowledge-graph-fused intelligent diagnosis method for power generation equipment according to claim 1, further comprising: And collecting power equipment fault updating data in real time, and introducing an online learning mechanism to dynamically learn, update and optimize the power equipment fault knowledge graph based on the power equipment fault updating data.
10. The knowledge-graph-fused intelligent diagnosis system for power generation equipment, which is characterized by being used for implementing the knowledge-graph-fused intelligent diagnosis method for power generation equipment according to any one of claims 1-9, and comprising the following steps: The map construction module is used for collecting multi-source operation data of the power generation equipment and historical fault records of the power generation equipment, carrying out map relation modeling on the multi-source operation data of the power generation equipment and the historical fault records of the power generation equipment, and constructing a power generation equipment fault knowledge map; The feature extraction module is used for monitoring the real-time operation data of the power generation equipment in real time, carrying out feature extraction and semantic association on the real-time operation data of the power generation equipment based on the fault knowledge graph of the power generation equipment, and generating an operation state vector of the power generation equipment; The reasoning analysis module is used for carrying out reasoning analysis on the operation state vector of the power generation equipment by adopting a graph neural network, and excavating an equipment potential fault mode and an equipment fault causal chain; And the fault diagnosis module is used for carrying out fault positioning diagnosis based on the equipment potential fault mode and the equipment fault causal link and determining a power generation equipment diagnosis result, wherein the power generation equipment diagnosis result comprises equipment health state, fault mode occurrence probability and fault influence root cause.

Description

Knowledge-graph-fused intelligent diagnosis method and system for power generation equipment Technical Field The invention relates to the technical field of equipment diagnosis, in particular to an intelligent diagnosis method and system for power generation equipment fused with a knowledge graph. Background In the long-term operation process of the power generation equipment, the operation state of the power generation equipment is influenced by load fluctuation, environmental condition change, equipment aging and other factors, and the fault type is characterized by diversification and coupling. The existing power generation equipment diagnosis method mainly depends on single-source monitoring data or an empirical rule-based fault judgment model, and is difficult to fully utilize multi-source heterogeneous data in the equipment operation process, so that fault characterization information is incomplete. Meanwhile, due to the fact that the internal mechanism of the equipment is complex, implicit association relations and causal links often exist among various fault phenomena, and the traditional model is difficult to effectively model and mine the implicit relations, so that fault diagnosis depends on manual experience, and the diagnosis accuracy and the interpretation are low. Disclosure of Invention The application provides an intelligent diagnosis method and system for power generation equipment, which are integrated with a knowledge graph, and solve the technical problem that in the prior art, the operation data of the power generation equipment are scattered, and the failure mechanism is difficult to be explicitly expressed, so that the diagnosis accuracy is low. In a first aspect of the present application, there is provided a knowledge-graph-fused intelligent diagnosis method for power generation equipment, the method comprising: The method comprises the steps of collecting multi-source operation data of power generation equipment and historical fault records of the power generation equipment, modeling the map relation between the multi-source operation data of the power generation equipment and the historical fault records of the power generation equipment, constructing a power generation equipment fault knowledge map, monitoring the real-time operation data of the power generation equipment in real time, carrying out feature extraction and semantic association on the real-time operation data of the power generation equipment based on the power generation equipment fault knowledge map, generating a power generation equipment operation state vector, carrying out reasoning analysis on the power generation equipment operation state vector by adopting a graph neural network, excavating equipment potential fault modes and equipment fault causal chains, carrying out fault positioning diagnosis based on the equipment potential fault modes and the equipment fault causal chains, and determining a power generation equipment diagnosis result, wherein the power generation equipment diagnosis result comprises equipment health state, fault mode occurrence probability and fault influence cause. In a second aspect of the present application, there is provided a knowledge-graph-fused intelligent diagnostic system for power generation equipment, the system comprising: the system comprises a map construction module, a characteristic extraction module, an inference analysis module and a fault diagnosis module, wherein the map construction module is used for collecting multi-source operation data of power generation equipment and historical fault records of the power generation equipment, modeling the multi-source operation data of the power generation equipment and the historical fault records of the power generation equipment to construct a power generation equipment fault knowledge map, the characteristic extraction module is used for monitoring the real-time operation data of the power generation equipment in real time, carrying out characteristic extraction and semantic association on the real-time operation data of the power generation equipment based on the power generation equipment fault knowledge map to generate a power generation equipment operation state vector, the inference analysis module is used for carrying out inference analysis on the power generation equipment operation state vector by adopting a graph neural network, excavating equipment potential fault modes and equipment fault causal chains, and the fault diagnosis module is used for carrying out fault location diagnosis based on the equipment potential fault modes and the equipment fault causal chains to determine a power generation equipment diagnosis result, wherein the power generation equipment diagnosis result comprises equipment health state, fault mode occurrence probability and fault influence root cause. One or more technical schemes provided by the application have at least the following technical effects or advantages: Firstly, collec