A wind turbine fault monitoring and early warning method and system based on big data

Abstract

The present invention provides a wind turbine fault monitoring and early warning method and system based on big data, which relates to the field of data intelligent processing technology. A fault knowledge graph is constructed based on fault record information, and the associated impact data of each fault are analyzed to determine the fault impact factor. Then, the data type of the big data platform is identified, multiple fault early warning models are constructed, and a fault early warning tree is generated. Each fault analysis and early warning is performed on the fault impact factor identification data, and fault early warning information is output. This solves the technical problem that when performing fault detection and early warning for wind turbines in the existing technology, the fault data is not processed intelligently enough, resulting in insufficient completeness of the fault early warning results and inability to perform comprehensive early warning based on the associated impact of various components during the operation of the unit. By constructing a fault knowledge graph and a fault early warning tree, the input fault data is evaluated as a whole based on the associated relationship, and the fault accident and the impact accident are determined, thereby realizing intelligent operation and management of the wind turbine.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent data processing, and in particular to a wind turbine fault monitoring and early warning method and system based on big data. Background Art

[0002] As the main energy supply method in the field of new energy, wind power generation is increasingly widely used. Due to the relatively harsh operating environment of wind power generation and the complex and changeable operating conditions of wind power generation, unit failures are inevitable with the operation of wind turbines. In order to ensure the normal and stable operation of wind turbines, it is necessary to monitor their faults and perform operation and maintenance management on the units based on the monitoring data. Nowadays, traditional operating parameters are mainly determined based on the operating status of the wind turbines, and parameter evaluation is performed to determine the abnormal operating unit structure, and maintenance is performed to avoid abnormal operation of the unit, resulting in low operating efficiency and potential safety hazards.

[0003] In the prior art, when performing fault detection and early warning for wind turbines, the processing method for fault data is not intelligent enough, resulting in incomplete fault warning results and an inability to provide a comprehensive warning based on the associated impact of various components during unit operation. Summary of the Invention

[0004] The present application provides a wind turbine fault monitoring and early warning method and system based on big data, which is used to solve the technical problem in the prior art that when performing fault detection and early warning for wind turbines, the fault data processing method is not intelligent enough, resulting in insufficient completeness of the fault early warning results and inability to provide a comprehensive early warning based on the correlated impact of various components during the operation of the unit.

[0005] In view of the above problems, the present application provides a wind turbine fault monitoring and early warning method and system based on big data.

[0006] In a first aspect, the present application provides a wind turbine fault monitoring and early warning method based on big data, the method comprising:

[0007] Obtaining a wind turbine database from a big data platform, wherein the wind turbine database includes turbine monitoring data and fault record information;

[0008] Building a fault knowledge graph based on the fault record information;

[0009] Analyze the associated impact data of each fault in the fault knowledge graph to determine the fault impact factor;

[0010] According to the fault influencing factors, the data type of the big data platform is identified, and data is automatically acquired based on the data type identification;

[0011] Determine each fault warning model based on the fault knowledge graph and the unit monitoring data, and construct a fault warning tree;

[0012] The fault impact factor identification data obtained from the big data platform is input into the fault warning tree, and each fault analysis and warning is performed to output fault warning information.

[0013] In a second aspect, the present application provides a wind turbine fault monitoring and early warning system based on big data, the system comprising:

[0014] An information acquisition module, configured to acquire a wind turbine database from a big data platform, wherein the wind turbine database includes turbine monitoring data and fault record information;

[0015] A graph construction module, wherein the graph construction module is used to construct a fault knowledge graph based on the fault record information;

[0016] An impact factor determination module, configured to determine a fault impact factor by analyzing associated impact data of each fault in the fault knowledge graph;

[0017] A data identification module, the data identification module is used to identify the data type of the big data platform according to the fault influencing factor, and automatically acquire data based on the data type identification;

[0018] An early warning tree construction module is used to determine each fault early warning model based on the fault knowledge graph and the unit monitoring data, and to construct a fault early warning tree;

[0019] A fault warning module is used to input the fault impact factor identification data obtained from the big data platform into the fault warning tree, perform fault analysis and warning, and output fault warning information.

[0020] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0021] An embodiment of the present application provides a wind turbine fault monitoring and early warning method based on big data. The method obtains a wind turbine database from a big data platform, including unit monitoring data and fault record information, constructs a fault knowledge graph based on the fault record information, performs correlation impact data analysis on each fault to determine the fault impact factor, and then identifies the data type of the big data platform. Data is automatically acquired based on the data type identification. According to the fault knowledge graph and the unit monitoring data, each fault early warning model is determined, and a fault early warning tree is constructed. The fault impact factor identification data obtained from the big data platform is input into the fault early warning tree, and each fault analysis and early warning is performed, and fault early warning information is output. The method solves the technical problem in the prior art that the fault data processing method is not intelligent enough when performing fault detection and early warning for wind turbines, resulting in insufficient completeness of the fault early warning result and inability to perform comprehensive early warning based on the correlation impact of each component during unit operation. By constructing a fault knowledge graph and a fault early warning tree, the input fault data is evaluated as a whole based on the correlation relationship, and the fault accident and the impact accident are determined, thereby realizing intelligent operation and management of the wind turbine. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This application provides a flow chart of a wind turbine fault monitoring and early warning method based on big data;

[0023] Figure 2 This application provides a schematic diagram of the fault knowledge graph construction process in a wind turbine fault monitoring and early warning method based on big data;

[0024] Figure 3 This application provides a schematic diagram of the process of obtaining fault warning information in a wind turbine fault monitoring and early warning method based on big data;

[0025] Figure 4 A structural diagram of a wind turbine fault monitoring and early warning system based on big data is provided for this application.

[0026] Explanation of the accompanying symbols: information acquisition module 11, graph construction module 12, impact factor determination module 13, data identification module 14, warning tree construction module 15, fault warning module 16. DETAILED DESCRIPTION

[0027] The present application provides a wind turbine fault monitoring and early warning method and system based on big data, constructs a fault knowledge graph based on fault record information, analyzes the associated impact data of each fault to determine the fault impact factor, and then identifies the data type of the big data platform, constructs multiple fault early warning models, generates a fault early warning tree, performs each fault analysis and early warning on the fault impact factor identification data, and outputs fault early warning information. The application is used to solve the technical problem in the prior art that when performing fault detection and early warning for wind turbines, the processing method of fault data is not intelligent enough, resulting in insufficient completeness of the fault early warning results and inability to perform comprehensive early warning based on the associated impact of various components during the operation of the unit.

[0028] Example 1

[0029] like Figure 1 As shown, the present application provides a wind turbine fault monitoring and early warning method based on big data, the method comprising:

[0030] Step S100: Acquire a wind turbine database from a big data platform, wherein the wind turbine database includes turbine monitoring data and fault record information;

[0031] Specifically, wind power generation, as the main energy supply method in the field of new energy, is increasingly widely used. Due to the relatively harsh operating environment of wind power generation and the relatively complex and changeable operating conditions of wind power generation, unit failures are inevitable with the operation of wind turbines. In order to ensure the normal and stable operation of wind turbines, the present application provides a wind turbine fault monitoring and early warning method based on big data. By constructing a fault knowledge graph, each node is matched with a fault early warning model to perform multi-level fault monitoring and early warning. First, a predetermined time interval is set, that is, a time period for data collection. Based on the predetermined time interval, the wind turbine operation data is retrieved according to the big data platform to determine the unit monitoring data and the fault record data, wherein the unit monitoring data and the fault record data correspond one-to-one. The unit monitoring data and the fault record data are correspondingly identified based on the time series to generate the wind turbine database. The wind turbine database covers a variety of possible fault data. The acquisition of the wind turbine database provides a basic basis for operating fault analysis.

[0032] Step S200: constructing a fault knowledge graph based on the fault record information;

[0033] Step S300: Analyze the associated impact data of each fault in the fault knowledge graph to determine the fault impact factor;

[0034] Step S400: Identify the data type of the big data platform according to the fault impact factor, and automatically acquire data based on the data type identification;

[0035] Specifically, the corresponding fault events are determined based on the multiple fault records contained in the fault record information, the fault events are divided based on the fault components, and multiple groups of fault events are determined. The fault risks of the multiple groups of fault events are further determined, and fault risk levels are set, corresponding to different risk intervals. The multiple groups of fault times are layered based on the fault risks, and multi-level nodes are determined as the main architecture of the graph, wherein the higher the fault risk, the higher the corresponding graph level. Fault impact analysis is performed on the fault events corresponding to the multi-level nodes, and the fault impact relationship between the sub-nodes corresponding to the nodes and the association relationship with the remaining nodes are determined. They are connected to generate a fault event association network as the fault knowledge graph, and fault analysis can be performed in the fault knowledge graph through logical reasoning.

[0036] Furthermore, the associated impact data of each fault in the fault knowledge graph is analyzed. For example, power may affect the motor speed, transmission structure operation, control cabinet temperature, etc., and power is used as the fault influencing factor of the above fault events. The same fault influencing factor may cause multiple faults, and the same fault may correspond to multiple fault influencing factors. The fault influencing factors are used as identification data to identify the data type of the big data platform. When data is retrieved based on the big data platform, data identification and extraction can be directly performed through the identification data to determine multiple groups of associated data that may have faults, while avoiding data omissions.

[0037] Furthermore, if Figure 2 As shown, based on the fault record information, a fault knowledge graph is constructed. Step S200 of this application also includes:

[0038] Step S210: Analyze the fault component according to the fault record information to determine the type of the fault component;

[0039] Step S220: performing a fault risk analysis based on the fault component type information and the fault record information to determine the fault risk level;

[0040] Step S230: classifying the fault component type information according to the fault risk to determine fault risk classification information;

[0041] Step S240: obtaining a map location level of each fault based on the fault risk classification information;

[0042] Step S250: Analyze and extract the impact data of each fault based on the map location level of each fault to determine the impact information of each fault;

[0043] Step S260: Connecting the associated faults in the graph position level of each fault based on the fault impact information to construct the fault knowledge graph.

[0044] Specifically, the fault record information is extracted based on the fault database, the fault components of the fault record information are identified, the fault component types corresponding to multiple fault information in the fault record information are determined, and the fault component type information is associated and matched to determine the fault component type information. Based on the fault component type information, the fault risk level is analyzed according to the fault record information corresponding to each fault component type. For example, the wind turbine operation fault risk level can be set, the fault component risk level matching is performed based on the fault record information, and the fault risk level is determined based on the matching result.

[0045] The fault risk is used as a classification standard to classify the fault component type information, and multi-level fault component type information is determined as the fault risk classification information, that is, multi-level fault events corresponding to different fault component types. Further, based on the fault risk classification information, the fault map position is determined. The higher the fault risk level, the higher the corresponding map level, and the higher the corresponding maintenance priority.

[0046] Further, fault impact data analysis is performed based on the map position level of each fault. For example, a generator fault may affect the transmission device, causing insufficient tension of the toothed belt, etc., and there is a multi-level association relationship. The fault impact information of each fault is determined, and the fault events in the map position level are connected based on the association influence relationship to construct the fault knowledge spectrum, wherein the fault knowledge spectrum contains multiple levels, corresponding to different fault levels. There may be multiple fault type events at the same level, and any fault type event corresponds to multiple branches, i.e., fault impact information. For example, a generator fault may correspond to generator bearing damage, abnormal winding heat dissipation, abnormal speed, etc., among which there may be an influence relationship. For example, bearing damage may affect winding heat dissipation and may further affect the transmission device. Based on the possible association influence relationship, they are connected to generate a fault network as the fault knowledge graph to ensure the accuracy of the construction of the fault knowledge graph. When there is a wind turbine operation fault, a prediction analysis is performed in the fault knowledge graph based on the association connection relationship.

[0047] Furthermore, step S260 of the present application also includes:

[0048] Step S261: performing a trend cycle analysis on each fault based on the fault record information to determine trend cycle information for each fault type;

[0049] Step S262: determining the risk level of each cycle based on the trend cycle information of each fault type;

[0050] Step S263: establishing a periodic risk map position based on the risk level of each period;

[0051] Step S264: embedding the periodic risk map position into the map position level of each fault.

[0052] Specifically, any fault record is extracted based on the fault record data to determine the complete derivative cycle of the fault. For example, it is divided into multiple cycle nodes, namely, the early stage of the fault, the middle stage of the fault, and the late stage of the fault. Different fault cycles correspond to different fault degrees, and at the same time, the degree of impact on other components of the wind turbine is different. Based on multiple fault record information, the trend cycle information of each fault type is determined, and risk assessment is further performed based on the trend cycle information of each fault type. Different cycle nodes correspond to different risks, including the risk of the fault component and the risk of the fault correlation impact. The cycle risk map position is established based on the cycle risk. Under different cycle nodes, the corresponding fault correlation impact components can be regarded as different fault events, which facilitates targeted analysis based on the cycle node of the fault and improves the accuracy of fault prediction. The cycle risk map position is embedded in the map position level of each fault, and the map position level of each fault is associated with the fault connection. The associated faults corresponding to different cycle nodes are connected separately, and different connection methods can be set for differentiation, so as to match the corresponding associated network for different fault cycle nodes, complete fault correlation identification, and ensure the targetedness and accuracy of fault identification.

[0053] Step S500: determining each fault warning model based on the fault knowledge graph and the unit monitoring data, and constructing a fault warning tree;

[0054] Step S600: inputting the fault impact factor identification data obtained from the big data platform into the fault warning tree, performing fault analysis and warning, and outputting fault warning information.

[0055] Specifically, multiple warning nodes are determined based on the fault knowledge graph, and the fault warning model is constructed based on the correlation between the graph nodes and the unit monitoring data. The multiple warning nodes correspond to a fault warning model respectively. The multiple fault warning models are connected based on the connection relationship between the nodes to form the fault warning tree. The fault warning tree corresponds to the fault knowledge graph and is used to perform fault accident warning alerts for each node in the fault knowledge graph. Furthermore, the fault influencing factor identification data is extracted based on the big data platform and input into the fault warning tree. The fault influencing factor identification data is used to identify the influencing nodes based on the correlation between the fault factor and the top event. The corresponding fault warning model is activated based on the identification result. By performing single node fault analysis and multi-related node comprehensive analysis, the fault warning information is generated for model output, and the operation and maintenance management of the wind turbine is performed based on the fault and mechanism information.

[0056] Furthermore, each fault warning model is determined based on the fault knowledge graph and the unit monitoring data. Step S500 of the present application further includes:

[0057] Step S510: extracting fault information of a preset risk level according to the fault knowledge graph, wherein the fault information of the preset risk level includes multiple wind turbine faults;

[0058] Step S520: sequentially treating multiple wind turbine failures as top events and analyzing failure factors;

[0059] Step S530: Based on the fault factor and the unit monitoring data, perform minimum segmentation of the fault corresponding type to determine a minimum segmentation set;

[0060] Step S540: determining a fault warning result according to the association relationship between the minimum partition set and the wind turbine fault.

[0061] Specifically, the preset risk level, i.e., the risk level critical value for risk accident extraction, is obtained. The fault knowledge graph includes a variety of possible fault events in the operation of the wind turbine. Based on the fault knowledge graph, fault information is extracted according to the preset risk level to obtain multiple fault events, including multiple wind turbine faults. The multiple wind turbine faults are sequentially used as the top events, which can be divided into three types of top events: serious fault accidents, major fault accidents, and general fault accidents. All cause events related to wind turbine faults are analyzed from various factors, such as generator failure and pitch system failure. Pitch system failure may be caused by problems with the pitch motor, control cabinet, toothed belt, etc., which are used as accident factors of the top events.

[0062] Further, based on the fault factors and the unit monitoring data, the corresponding types of the faults are subjected to minimum accident segmentation. Multi-level segmentation can be performed until the minimum fault accident is determined, which is used as the most severe segmentation set. The minimum segmentation set is then associated with the wind turbine fault to determine the wind turbine fault that may be caused by the minimum segmentation accident, including the impact faults generated as the minimum segmentation accident develops, as the fault warning result. By performing the minimum accident segmentation, reverse fault association analysis can be performed layer by layer based on the minimum segmentation accident and the fault knowledge graph, and possible faults with associated relationships can be used as analysis results for warning, thereby ensuring the completeness of the fault query.

[0063] Furthermore, the step S540 of constructing the fault warning tree in this application further includes:

[0064] Step S541: determining the graph locations of multiple wind turbine faults based on the fault knowledge graph;

[0065] Step S542: Based on the map position, the fault warning models of the wind turbine faults are connected to construct the fault warning tree.

[0066] Specifically, the fault knowledge graph has multiple levels of nodes, each corresponding to a wind turbine fault. The graph positions corresponding to the multiple wind turbine faults are determined, and a fault warning model is constructed based on a machine learning algorithm. The fault warning model is a multi-level network layer, including an association identification layer, a level analysis layer, and a warning output layer. The fault warning model can perform multi-level warnings. A fault warning model is linked to each wind turbine fault corresponding to the graph position. For the same type of fault warning models, the multiple fault warning models are connected based on the connection relationship of the fault knowledge graph to generate the fault warning tree. When there is a fault node, the corresponding model performs fault analysis and warning. At the same time, connection analysis is performed based on the association identification layer. When there is a connection relationship, the warning model is enabled, and fault analysis is performed based on the fault cycle and correlation. When there is a fault impact, the corresponding fault level is determined, and then warning information is generated based on the warning output layer for warning and alerting, thereby realizing automated analysis and warning of fault accidents.

[0067] Furthermore, if Figure 3 As shown, the fault impact factor identification data obtained from the big data platform is input into the fault warning tree, each fault analysis and warning is performed, and fault warning information is output. Step S600 of this application also includes:

[0068] Step S610: inputting the fault impact factor identification data into the fault warning tree, issuing a warning for each fault, and obtaining warning information for each graph node;

[0069] Step S620: Determine whether the warning information of each graph node meets the preset requirements;

[0070] Step S630: When the conditions are met, a fault warning message is sent;

[0071] Step S640: If not satisfied, determine the fault risk weight according to the risk of each node fault in the fault knowledge graph;

[0072] Step S650: Calculate based on the fault association relationship and fault risk weight in the fault knowledge graph to obtain comprehensive warning information, and determine the fault warning information according to the comprehensive warning information.

[0073] Specifically, based on the big data platform, the fault impact factor identification data is retrieved and input into the fault warning tree, the position associated with the fault impact factor is determined, and the corresponding fault warning model is activated. The warning information of each graph node is obtained by performing model analysis to perform early warning and alert. It is further determined whether the warning information of each graph node meets the preset requirements. The preset requirements are critical values ​​for the occurrence of node failures. When the requirements are met, it indicates that the fault impact factor has caused the failure to occur, and the fault warning information is sent for early warning and alert so that the failure can be corrected in time. When the requirements are not met, it indicates that the fault accident corresponding to the current node has not occurred, but there are certain potential fault hazards. The risk weight of each node is determined based on the risk degree of each node failure in the fault knowledge graph, wherein the fault risk weight is proportional to the fault risk degree. Based on the fault association relationship in the fault knowledge graph, the associated node corresponding to the warning node to be analyzed is determined, and the risk level of the node is obtained by performing risk weighted calculation on the associated node. Based on this, the comprehensive warning information is generated and output as the fault warning information. Operation and maintenance management is performed based on the fault warning information.

[0074] Furthermore, the present application further includes step S700, comprising:

[0075] Step S710: Obtain minimum segmentation fault information and fault area location information;

[0076] Step S720: performing regional work order matching based on the minimum segmentation fault information and the fault area location information to determine matching work order information;

[0077] Step S730: Obtaining work order processing plan information based on the matching work order information;

[0078] Step S740: Determine whether the work order processing plan information meets the processing requirements, wherein the processing requirements include impact information and timeliness information;

[0079] Step S750: When both the impact information and the timeliness information are satisfied, a processing work order information is generated.

[0080] Specifically, the fault information of the minimum split accident is determined, and then the fault is located. The wind turbine is positioned, the fault area positioning information is obtained, and the unit operation and maintenance management area is used as the area range. Work orders are matched based on the minimum split fault information and the fault area positioning information, the best operation and maintenance group is determined, and the matching work order information is generated. Then, fault correction information is extracted based on the matching work order information, and the work order processing plan information is generated. The impact information and the timeliness information are set. The timeliness information is the fault processing event interval. Exceeding the timeliness information may cause the fault level to be upgraded. The impact information is the associated impact range of the current fault. Whether it causes a joint fault is used as the processing requirement to determine whether the work order processing plan information meets the processing requirement. When the impact information and the timeliness information are both met, it indicates that the work order processing plan information can complete the fault correction, the processing work order information is generated and transferred to the corresponding operation and maintenance group, and the fault repair is carried out in time to ensure the normal operation of the unit.

[0081] Example 2

[0082] Based on the same inventive concept as the wind turbine fault monitoring and early warning method based on big data in the above embodiment, Figure 4 As shown, the present application provides a wind turbine fault monitoring and early warning system based on big data, the system comprising:

[0083] An information acquisition module 11 is used to obtain a wind turbine database from a big data platform, wherein the wind turbine database includes turbine monitoring data and fault record information;

[0084] A graph construction module 12, the graph construction module 12 is used to construct a fault knowledge graph based on the fault record information;

[0085] An impact factor determination module 13 is configured to determine a fault impact factor by analyzing associated impact data of each fault in the fault knowledge graph;

[0086] A data identification module 14 is configured to identify data types on the big data platform according to the fault influencing factors, and automatically acquire data based on the data type identification;

[0087] An early warning tree construction module 15 is used to determine each fault early warning model based on the fault knowledge graph and the unit monitoring data, and to construct a fault early warning tree;

[0088] The fault warning module 16 is used to input the fault impact factor identification data obtained from the big data platform into the fault warning tree, perform fault analysis and warning, and output fault warning information.

[0089] Furthermore, the system further comprises:

[0090] An information determination module, configured to analyze a faulty component based on the fault record information and determine the type of the faulty component;

[0091] a risk analysis module, configured to perform a fault risk analysis based on the fault component type information and the fault record information to determine the fault risk level;

[0092] An information classification module, the information classification module being configured to classify the fault component type information according to the fault risk and determine fault risk classification information;

[0093] A level acquisition module, the level acquisition module is used to obtain the map location level of each fault according to the fault risk classification information;

[0094] An impact information determination module, configured to analyze and extract impact data of each fault based on the map location level of each fault, and determine the impact information of each fault;

[0095] A knowledge graph construction module is used to connect the associated faults in the graph position level of each fault based on the fault impact information to construct the fault knowledge graph.

[0096] Furthermore, the system further comprises:

[0097] A cycle analysis module, the cycle analysis module is used to perform trend cycle analysis on each fault based on the fault record information, and determine trend cycle information of each fault type;

[0098] A risk determination module, configured to determine the risk of each period based on the trend period information of each fault type;

[0099] A map position establishment module, the map position establishment module is used to establish a period risk map position based on the risk degree of each period;

[0100] A position embedding module is used to embed the periodic risk map position into the map position level of each fault.

[0101] Furthermore, the system further comprises:

[0102] A fault information extraction module, configured to extract fault information of a preset risk level based on the fault knowledge graph, wherein the fault information of the preset risk level includes multiple wind turbine faults;

[0103] A fault factor analysis module, wherein the fault factor analysis module is used to sequentially treat multiple wind turbine faults as top events and analyze fault factors;

[0104] An accident segmentation module, the accident segmentation module is used to perform minimum segmentation of the accident corresponding to the fault type based on the fault factor and the unit monitoring data, and determine a minimum segmentation set;

[0105] The early warning result determination module is used to determine the fault early warning result according to the association relationship between the minimum segmentation set and the wind turbine fault.

[0106] Furthermore, the system further comprises:

[0107] A location determination module, configured to determine the map locations of multiple wind turbine faults based on the fault knowledge map;

[0108] A fault warning tree construction module is used to connect the fault warning models of each wind turbine fault based on the map position to construct the fault warning tree.

[0109] Furthermore, the system further comprises:

[0110] A node warning information acquisition module, which is used to input fault impact factor identification data into the fault warning tree, issue a warning for each fault, and obtain warning information for each graph node;

[0111] An information judgment module, which is used to judge whether the warning information of each graph node meets the preset requirements;

[0112] A warning information sending module, configured to send fault warning information when conditions are met;

[0113] A weight determination module, wherein the weight determination module is used to determine the fault risk weight according to the risk degree of each node fault in the fault knowledge graph when the conditions are not met;

[0114] A fault warning information determination module is used to calculate based on the fault association relationship and fault risk weight in the fault knowledge graph to obtain comprehensive warning information, and determine the fault warning information according to the comprehensive warning information.

[0115] Furthermore, the system further comprises:

[0116] A fault information acquisition module, wherein the fault information acquisition module is used to obtain minimum segmentation fault information and fault area location information;

[0117] A work order matching module, configured to perform regional work order matching based on the minimum segmentation fault information and the fault area location information, and determine matching work order information;

[0118] A plan acquisition module, configured to obtain work order processing plan information based on the matching work order information;

[0119] A plan determination module, configured to determine whether the work order processing plan information meets processing requirements, wherein the processing requirements include impact information and timeliness information;

[0120] The work order information generation module is used to generate processing work order information when both the impact information and the timeliness information are met.

[0121] Through the above detailed description of a wind turbine fault monitoring and early warning method based on big data in this specification, those skilled in the art can clearly understand a wind turbine fault monitoring and early warning method and system based on big data in this embodiment. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0122] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A wind turbine fault monitoring and early warning method based on big data, characterized in that: The method comprises: Obtaining a wind turbine database from a big data platform, wherein the wind turbine database includes turbine monitoring data and fault record information; Building a fault knowledge graph based on the fault record information; Analyze the associated impact data of each fault in the fault knowledge graph to determine the fault impact factor; According to the fault influencing factors, the data type of the big data platform is identified, and data is automatically acquired based on the data type identification; Determine each fault warning model based on the fault knowledge graph and the unit monitoring data, and construct a fault warning tree; Input the fault impact factor identification data obtained from the big data platform into the fault warning tree, perform fault analysis and warning, and output fault warning information; According to the fault knowledge graph and the unit monitoring data, each fault warning model is determined and a fault warning tree is constructed, including: Extracting fault information of a preset risk level according to the fault knowledge graph, wherein the fault information of the preset risk level includes multiple wind turbine faults; Multiple wind turbine failures are sequentially treated as top events to analyze failure factors; Based on the fault factors and the unit monitoring data, performing minimum segmentation of the fault corresponding types to determine a minimum segmentation set; Determining a fault warning result according to the correlation between the minimum segmentation set and the wind turbine fault; Determining the map locations of multiple wind turbine faults based on the fault knowledge map; Based on the position of the graph, the fault warning models of the wind turbine faults are connected to construct the fault warning tree; The method further comprises: Obtain minimum segmentation fault information and fault area location information; Performing regional work order matching based on the minimum segmentation fault information and the fault area location information to determine matching work order information; Obtaining work order processing plan information based on the matching work order information; Determine whether the work order processing plan information meets the processing requirements, wherein the processing requirements include impact information and timeliness information; When both the impact information and the timeliness information are met, the processing work order information is generated.

2. The method according to claim 1, wherein Based on the fault record information, a fault knowledge graph is constructed, including: Analyze the fault component according to the fault record information to determine the type of the fault component; Based on the fault component type information and the fault record information, a fault risk degree analysis is performed to determine the fault risk degree; classifying the fault component type information according to the fault risk to determine fault risk classification information; Obtaining a map location level of each fault based on the fault risk classification information; Based on the map location level of each fault, the impact data of each fault is analyzed and extracted to determine the impact information of each fault; Based on the fault impact information, the associated faults in the graph position level of each fault are connected to construct the fault knowledge graph.

3. The method according to claim 2, wherein The method further comprises: Performing a trend cycle analysis on each fault based on the fault record information to determine trend cycle information for each fault type; Determining the risk level of each period based on the trend period information of each fault type; Based on the risk level of each period, a period risk map position is established; The periodic risk map position is embedded into the map position level of each fault.

4. The method according to claim 1, wherein Input the fault impact factor identification data obtained from the big data platform into the fault warning tree, perform fault analysis and warning, and output fault warning information, including: Input the fault impact factor identification data into the fault warning tree, issue a warning for each fault, and obtain warning information for each graph node; Determine whether the warning information of each graph node meets the preset requirements; When the conditions are met, a fault warning message is sent; If it is not satisfied, the fault risk weight is determined based on the risk of each node fault in the fault knowledge graph; Calculation is performed based on the fault association relationship and the fault risk weight in the fault knowledge graph to obtain comprehensive warning information, and the fault warning information is determined based on the comprehensive warning information.

5. A wind turbine fault monitoring and early warning system based on big data, characterized in that: The system comprises: An information acquisition module, configured to acquire a wind turbine database from a big data platform, wherein the wind turbine database includes turbine monitoring data and fault record information; A graph construction module, wherein the graph construction module is used to construct a fault knowledge graph based on the fault record information; An impact factor determination module, configured to determine a fault impact factor by analyzing associated impact data of each fault in the fault knowledge graph; A data identification module, the data identification module is used to identify the data type of the big data platform according to the fault influencing factor, and automatically acquire data based on the data type identification; An early warning tree construction module is used to determine each fault early warning model based on the fault knowledge graph and the unit monitoring data, and to construct a fault early warning tree; A fault warning module, which is used to input the fault impact factor identification data obtained from the big data platform into a fault warning tree, perform fault analysis and warning, and output fault warning information; The system further comprises: A fault information extraction module, configured to extract fault information of a preset risk level based on the fault knowledge graph, wherein the fault information of the preset risk level includes multiple wind turbine faults; A fault factor analysis module, wherein the fault factor analysis module is used to sequentially treat multiple wind turbine faults as top events and analyze fault factors; An accident segmentation module, the accident segmentation module is used to perform minimum segmentation of the accident corresponding to the fault type based on the fault factor and the unit monitoring data, and determine a minimum segmentation set; An early warning result determination module, configured to determine a fault early warning result based on an association between the minimum segmentation set and the wind turbine fault; A location determination module, configured to determine the map locations of multiple wind turbine faults based on the fault knowledge map; A fault warning tree construction module is used to connect the fault warning models of each wind turbine fault based on the map position to construct the fault warning tree; A fault information acquisition module, wherein the fault information acquisition module is used to obtain minimum segmentation fault information and fault area location information; A work order matching module, configured to perform regional work order matching based on the minimum segmentation fault information and the fault area location information, and determine matching work order information; A plan acquisition module, configured to obtain work order processing plan information based on the matching work order information; A plan determination module, configured to determine whether the work order processing plan information meets processing requirements, wherein the processing requirements include impact information and timeliness information; The work order information generation module is used to generate processing work order information when both the impact information and the timeliness information are met.

Images