Knowledge base construction and application method for intelligent early warning of fan faults

Abstract

The invention discloses a knowledge base construction and application method for intelligent early warning of fan faults, and the knowledge base construction method comprises the steps: collecting fan operation data at different sampling frequencies; performing time alignment, abnormal value detection, data type conversion and data completion processing on the acquired operation data; constructing an early warning knowledge base based on the complemented operation data; and performing fault early warning on the fan by using the constructed early warning knowledge base. Through different sampling frequencies, comprehensive acquisition of fan operation parameters is realized, and meanwhile, time alignment, abnormal value detection, data type conversion and data complementation processing are performed on the acquired operation data, so that the fusion problem of multi-source heterogeneous data is solved; the early warning knowledge base is utilized, and the accuracy of early warning result evaluation is improved in the process of early warning knowledge accumulation and optimization.

Description

Technical Field

[0001] The present invention belongs to the technical field of fan fault early warning, and in particular relates to a knowledge base construction and application method for fan fault intelligent early warning. Background Art

[0002] With the rapid development of the wind power industry, the scale of wind farms continues to expand, and the safe and stable operation of wind turbines has an important impact on the safety and economy of the power grid. Intelligent early warning technology for wind turbine failures collects wind turbine operation data and uses data analysis and artificial intelligence technology to predict and warn potential wind turbine failures in advance, thereby achieving preventive maintenance of wind turbines. This technology is of great significance for improving the reliability of wind turbine operation, reducing maintenance costs, and reducing unplanned downtime.

[0003] At present, the wind turbine fault warning mainly adopts the method based on SCADA system data analysis and CMS vibration monitoring; among them, the method based on SCADA system mainly predicts faults by analyzing the changing trend of wind turbine operating parameters, but due to the low sampling frequency, it is difficult to detect the early faults of components in time; the method based on CMS analyzes the vibration characteristics of key components for fault diagnosis, but due to the limitation of acquisition channels, it is difficult to achieve comprehensive monitoring of the status of the wind turbine. At the same time, most of the existing early warning systems use a single data source and a fixed early warning model, lack the ability to integrate and analyze multi-source heterogeneous data, and the early warning knowledge base is difficult to dynamically update and optimize according to the actual operation situation.

[0004] The existing technology has the following technical problems: first, the standardized processing and time alignment of multi-source data are difficult to solve, and data from different sources and different sampling frequencies are difficult to effectively integrate; second, the generalization ability of the early warning model is insufficient, and it is difficult to adapt to the complex and changeable wind turbine operating conditions by relying solely on a single machine learning model or rule model; third, the early warning results lack a credibility assessment mechanism, which is prone to false alarms and missed alarms; fourth, the knowledge base update mechanism is imperfect, making it difficult to achieve continuous accumulation and optimization of knowledge; finally, the computational efficiency of the early warning analysis is low, and it is difficult to meet the real-time monitoring needs of large-scale wind farms. These problems seriously restrict the actual application effect of the intelligent early warning system for wind turbine failures. Summary of the invention

[0005] The present invention provides a knowledge base construction and application method for intelligent early warning of wind turbine faults, which is used to solve the technical defects of the existing wind turbine fault early warning technology, that is, it is impossible to effectively integrate data from different sources and with different sampling frequencies, so that the generalization ability of the early warning model is insufficient, and it is difficult to adapt to the complex and changeable wind turbine operating conditions, resulting in a lack of credibility evaluation mechanism for the early warning results, prone to false alarms and missed alarms, resulting in low computational efficiency of early warning analysis, and difficulty in meeting the real-time monitoring needs of large-scale wind farms.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: In a first aspect, a knowledge base construction method for intelligent early warning of wind turbine faults is provided, comprising: Collect fan operation data at different sampling frequencies; Performing time alignment, outlier detection, data type conversion and data completion processing on the collected operating data; Build an early warning knowledge base based on the completed operating data; The constructed early warning knowledge base is used to provide fault early warning for the wind turbine.

[0007] Furthermore, collecting the fan operation data at different sampling frequencies specifically includes: The data acquisition unit is connected to the fan control system to collect the fan operating parameters at a collection frequency of 1 Hz; Connect the vibration data acquisition unit to the vibration sensors installed on the key components of the wind turbine to collect vibration signals on the key components at a collection frequency of 25.6 kHz; The environmental data acquisition unit is connected to the weather station to collect meteorological parameters around the fan at a collection frequency of 0.1 Hz.

[0008] Furthermore, the operating parameters of the fan include fan speed, fan power and fan temperature parameters; The key components include bearings and gear boxes; The meteorological parameters include wind speed, wind direction and ambient temperature parameters.

[0009] Furthermore, the collected operation data is subjected to time alignment, outlier detection, data type conversion and data completion processing, specifically including: The collected operation data is time synchronized and a sliding window algorithm is used to align the data; Using the 3σ criterion, outliers in the aligned operating data are detected, and data beyond the normal range is marked or eliminated; Convert running data in different formats into standard floating point or integer data; Missing values ​​in the running data are processed using linear interpolation or nearest neighbor interpolation methods.

[0010] Furthermore, the construction of an early warning knowledge base based on the completed operating data specifically includes: Based on the completed operation data, a warning knowledge base is constructed, wherein the warning knowledge base includes a fault feature base, a fault case base and a diagnosis rule base; Wherein, the fault feature library is used to record the structural definition and extraction rules of fault features; The fault case library is used to store historical fault cases of wind turbines; The diagnosis rule base contains expert diagnosis rules, and the expert diagnosis rules are used for diagnosing fan faults.

[0011] Furthermore, in the fault feature library, the XML format is used to describe the feature structure and establish feature extraction rules; In the fault case library, a hierarchical classification method is adopted to organize historical fault cases; In the diagnostic rule base, a rule editing and verification mechanism is established to enable the import or export of rules.

[0012] Further, using the constructed early warning knowledge base, a fault early warning is performed on the wind turbine, specifically including: Using the constructed early warning knowledge base, respectively extracting time domain features, frequency domain features, operating condition features and environmental features from the operation data; A deep learning model is used to provide fault warning for wind turbines, while expert rules are applied to diagnose faults. The fault results are output by combining deep learning and rule reasoning.

[0013] Furthermore, the deep learning model is used to provide fault warning for the wind turbine, specifically including: Constructing a deep learning model, wherein the deep learning model includes an input layer, a convolutional layer, a long short-term memory network layer, and an output layer; Wherein, in the convolution layer, a 3×3 convolution kernel is used to extract local features; In the long short-term memory network layer, 128 hidden units are configured; Through the output layer, a failure probability prediction value is generated.

[0014] Furthermore, the outputting of the fault result includes: Generate real-time monitoring information to display the operating status of the fan equipment; Generate warning information, display the current fan fault warning level and fault details; Generate trend analysis information to display the trend of fan operating parameters; Generate health score information to display the health status of the fan equipment.

[0015] In a second aspect, a knowledge base application method for intelligent early warning of wind turbine faults is provided, wherein the method is performed using the knowledge base construction method for intelligent early warning of wind turbine faults as described above, and comprises: Process the collected fan operation data; Inputting the processed wind turbine operation data into an early warning knowledge base; Based on the early warning knowledge base, an early warning is issued to the wind turbine.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. Through different sampling frequencies, comprehensive collection of wind turbine operating parameters is achieved. At the same time, time alignment, outlier detection, data type conversion and data completion processing of the collected operating data solve the problem of multi-source heterogeneous data fusion; using the early warning knowledge base, the accuracy of early warning result evaluation is improved in the process of accumulating and optimizing early warning knowledge.

[0017] 2. By collecting the operating parameters of the fan at a frequency of 1 Hz, the basic operating status of the fan can be monitored in real time, providing basic data for fault warning; collecting vibration signals on key components at a high frequency of 25.6 kHz can capture subtle vibration changes, which is crucial for identifying early faults; collecting meteorological parameters around the fan at a frequency of 0.1 Hz can understand changes in the fan operating environment and provide a basis for analyzing the relationship between faults and environmental factors.

[0018] 3. As the most vulnerable key components in the wind turbine, the vibration signals of bearings and gearboxes contain rich fault information. High-frequency collection of these vibration signals can capture tiny vibration changes, such as impact, friction or imbalance, so as to detect bearing wear, gear cracks and other faults in advance; wind speed, wind direction and ambient temperature, these meteorological parameters affect the performance and safety of the wind turbine. By monitoring these parameters, the operating environment of the wind turbine can be evaluated, the operating strategy can be optimized, and faults caused by environmental factors can be reduced.

[0019] 4. Time synchronization and sliding window algorithms are used to ensure that data from different sources and different sampling frequencies can be accurately aligned in time sequence, ensuring the time consistency of the data, so that subsequent data analysis can be based on accurate time series, avoiding misjudgment or missed judgment due to time dislocation. At the same time, it also provides a reliable basis for timing analysis when a fault occurs. The 3σ criterion is used to detect outliers in the aligned operating data, and mark or remove data that exceeds the normal range, effectively identifying and processing outliers in the data. These outliers may be caused by sensor failure, data transmission errors, or external interference. After removing these outliers, the accuracy and reliability of the data have been significantly improved, thereby improving the accuracy of fault warning.

[0020] 5. The fault feature library records the structural definition and extraction rules of fault features, and can automatically and accurately extract fault-related features from the operating data. These features are the basis for accurate fault warning and help improve the accuracy and timeliness of warning. At the same time, with continuous operation and continuous accumulation of data, the fault feature library can be continuously updated and optimized to adapt to changes in the operating status of the fan. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 A flow chart of a knowledge base construction method for intelligent early warning of fan faults provided by the present invention; Figure 2 A flow chart of the data preprocessing steps in the knowledge base construction method for intelligent early warning of fan faults provided by the present invention; Figure 3 A flow chart of the knowledge base construction steps in the knowledge base construction method for intelligent early warning of fan faults provided by the present invention; Figure 4 A flow chart of the early warning analysis steps in the knowledge base construction method for intelligent early warning of fan faults provided by the present invention; Figure 5 The overall architecture diagram of the knowledge base construction system for intelligent early warning of fan failure provided by the present invention; Figure 6 A schematic diagram of hardware deployment of a knowledge base construction system for intelligent early warning of wind turbine failures provided by the present invention; Figure 7 A schematic diagram of a visual interface of a knowledge base construction system for intelligent early warning of fan failures provided by the present invention; Figure 8 A schematic diagram of a knowledge base construction system for intelligent early warning of fan faults provided by the present invention; Fig. 9 A flow chart of the knowledge base application method for intelligent early warning of fan failure provided by the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0024] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0026] With the rapid development of the wind power industry, the scale of wind farms continues to expand, and the safe and stable operation of wind turbines has an important impact on the safety and economy of the power grid. Intelligent early warning technology for wind turbine failures collects wind turbine operation data and uses data analysis and artificial intelligence technology to predict and warn potential wind turbine failures in advance, thereby achieving preventive maintenance of wind turbines. This technology is of great significance for improving the reliability of wind turbine operation, reducing maintenance costs, and reducing unplanned downtime.

[0027] At present, the wind turbine fault warning mainly adopts the method based on SCADA system data analysis and CMS vibration monitoring; among them, the method based on SCADA system mainly predicts faults by analyzing the changing trend of wind turbine operating parameters, but due to the low sampling frequency, it is difficult to detect the early faults of components in time; the method based on CMS analyzes the vibration characteristics of key components for fault diagnosis, but due to the limitation of acquisition channels, it is difficult to achieve comprehensive monitoring of the status of the wind turbine. At the same time, most of the existing early warning systems use a single data source and a fixed early warning model, lack the ability to integrate and analyze multi-source heterogeneous data, and the early warning knowledge base is difficult to dynamically update and optimize according to the actual operation situation.

[0028] The existing technology has the following technical problems: first, the standardized processing and time alignment of multi-source data are difficult to solve, and data from different sources and different sampling frequencies are difficult to effectively integrate; second, the generalization ability of the early warning model is insufficient, and it is difficult to adapt to the complex and changeable wind turbine operating conditions by relying solely on a single machine learning model or rule model; third, the early warning results lack a credibility assessment mechanism, which is prone to false alarms and missed alarms; fourth, the knowledge base update mechanism is imperfect, making it difficult to achieve continuous accumulation and optimization of knowledge; finally, the computational efficiency of the early warning analysis is low, and it is difficult to meet the real-time monitoring needs of large-scale wind farms. These problems seriously restrict the actual application effect of the intelligent early warning system for wind turbine failures.

[0029] In order to solve the above technical defects, the inventor provides a knowledge base construction and application method for intelligent early warning of fan failure.

[0030] In a first aspect, an embodiment of the present invention provides a method for constructing a knowledge base for intelligent early warning of wind turbine faults, such as Figure 1-Figure 4 As shown, including: S101, collecting fan operation data at different sampling frequencies; for example, Figure 1 As shown in the figure, the SCADA data acquisition unit is connected to the fan control system through the OPC UA protocol, and the fan operation parameters are collected at a collection frequency of 1Hz; the CMS vibration data acquisition unit is connected to the vibration sensors installed on each key component of the fan using the RS485 bus, and the vibration signals on the key components are collected at a collection frequency of 25.6kHz; the environmental data acquisition unit is connected to the weather station through the standard Modbus protocol, and the meteorological parameters around the fan are collected at a collection frequency of 0.1Hz. Among them, the fan operation parameters include fan speed, fan power and fan temperature parameters; key components include bearings and gearboxes and other components that can generate vibration; meteorological parameters include wind speed, wind direction and ambient temperature parameters. In this process, the comprehensive collection of fan operation parameters is achieved through different sampling frequencies.

[0031] S102, performing time alignment, outlier detection, data type conversion and data completion processing on the collected operation data; illustratively, as Figure 2As shown in the figure, the collected fan operation data is transmitted to the data preprocessing server for processing. The data preprocessing server adopts a dual-machine hot standby configuration. The time alignment unit is used to synchronize the operation data from different sources, and a sliding window algorithm is used to achieve data alignment. The sliding window size can be configured, and the default setting is 10 seconds. Then, the data cleaning unit uses the 3σ criterion to detect abnormal values ​​in the operation data, and the data that exceeds the normal range is marked or eliminated. Subsequently, the data type conversion unit is used to uniformly convert the operation data in different formats into standard floating point or integer data, and the missing values ​​in the data are processed by the data completion unit using linear interpolation or nearest neighbor interpolation methods.

[0032] S103, based on the completed operation data, construct an early warning knowledge base; illustratively, Figure 3 As shown in the figure, based on the completed operation data, an early warning knowledge base is constructed, which includes a fault feature base, a fault case base and a diagnostic rule base; wherein the fault feature base is used to record the structural definition and extraction rules of the fault feature; the fault case base is used to store the historical fault cases of the fan; the diagnostic rule base has expert diagnostic rules, which are used to diagnose the fan fault. Specifically, in the fault feature base, the feature definition unit uses the XML format to describe the feature structure and establish the feature extraction rules; then, the feature association analysis unit uses the correlation coefficient to calculate the correlation relationship between the features; finally, the feature verification unit uses the cross-validation method to verify the validity of the feature.

[0033] In the fault case library, a hierarchical classification method is adopted to organize historical fault cases, a standardized format is used to record historical fault cases, a hierarchical classification system is established, multi-dimensional retrieval is supported, and regular updates and maintenance are performed; in the diagnostic rule library, a rule editing and verification mechanism is established to enable the import or export of rules; at the same time, a rule editing interface is provided in the diagnostic rule library to support rule editing, verification and version management, and also supports the import and export functions of rules.

[0034] S104: Using the constructed warning knowledge base, a fault warning is performed on the wind turbine. like Figure 4As shown in the figure, the collected fan operation data is input into the early warning knowledge base, and the constructed early warning knowledge base is used to extract the time domain characteristics, frequency domain characteristics, operating condition characteristics and environmental characteristics of the operation data respectively; then, a deep learning model is used to perform fault warning for the fan, and expert rules are applied to perform fault diagnosis, and the fault results are output by combining deep learning and rule reasoning. This process also includes feature extraction module, fault warning model, rule reasoning unit, model fusion unit and warning result evaluation module; among them, the feature extraction module includes time domain feature extraction unit, frequency domain feature extraction unit, working condition feature extraction unit and environmental feature extraction unit, which are responsible for calculating statistical features, spectrum features, performance indicators and environmental parameters respectively; the fault warning model includes deep learning prediction unit and rule reasoning unit, among which the deep learning model includes input layer, convolution layer, long short-term memory network layer and output layer, the deep learning prediction unit adopts the combination structure of convolution neural network and long short-term memory network, the convolution layer adopts 3×3 convolution kernel to extract local features, and the long short-term memory network layer contains 128 hidden layer units to capture time series features; the rule reasoning unit adopts forward reasoning strategy to perform rule matching; the model fusion unit adopts weighted voting method to fuse the prediction results of multiple models, and the weights are dynamically adjusted according to the historical accuracy; the warning result evaluation module calculates the warning credibility by comprehensively considering the model confidence and data quality, verifies the warning results by comparing with historical cases, and initiates the expert review process for high-risk warnings. In terms of early warning output of fault results, it includes generating real-time monitoring information to display the operating status of the fan equipment; generating early warning information to display the current fan fault warning level and fault details; generating trend analysis information to display the changing trend of the fan operating parameters; generating health score information to display the health status of the fan equipment.

[0035] In a second aspect, a knowledge base application method for intelligent early warning of wind turbine faults is provided, wherein the method is performed by using the knowledge base construction method for intelligent early warning of wind turbine faults as described above. Fig. 9 As shown, including: S1. Processing the collected fan operation data; S2, inputting the processed wind turbine operation data into an early warning knowledge base; S3. Based on the warning knowledge base, issue a warning to the wind turbine.

[0036] In a third aspect, an embodiment of the present invention provides a knowledge base construction system for intelligent early warning of wind turbine faults, including: Data acquisition module, used to collect fan operation data; The data processing module is used to perform time alignment, outlier detection, data type conversion, and data completion on the running data; An early warning knowledge base construction module is used to construct an early warning knowledge base; The fault warning module is used to warn of fan faults.

[0037] When applied, such as Figure 5-Figure 8 As shown, the system includes five layers: data acquisition layer, data processing layer, knowledge base layer, analysis layer and application layer. Each layer is connected through industrial Ethernet and uses TCP / IP protocol for communication. In the data acquisition layer, a data acquisition module is set up, which includes a SCADA data acquisition unit, a CMS vibration data acquisition unit and an environmental data acquisition unit. The SCADA data acquisition unit uses an industrial-grade programmable controller as a data collector and communicates with the fan control system through an OPC UA server. The CMS vibration data acquisition unit uses an industrial-grade data acquisition card to support the simultaneous acquisition of 4-channel vibration signals. The environmental data acquisition unit uses a standard meteorological station and is connected to the system network through an RS485 to Ethernet module. Figure 6 As shown, the hardware deployment of the system adopts a distributed architecture.

[0038] The data preprocessing layer uses dual hot standby servers, configured with server-level processors, 256GB memory, and a commercial Linux operating system. The data preprocessing software runs on the server. The software is developed in C++ and includes a data receiving module, a preprocessing module, and a data distribution module. Data exchange between modules is achieved through shared memory. The distributed storage layer uses an open source distributed computing framework to build a storage cluster. The master node server is configured with dual hot standby servers. Each server is configured with a server-level processor and 512GB memory.

[0039] There are 6 data node servers in total, each equipped with a server-level processor, 128GB memory, and 12TB storage space. There are 2 index node servers with the same configuration as the data nodes, which are used to maintain data indexes. The knowledge base layer uses a relational database to store the fault feature library, fault case library, and diagnostic rule library. The database server adopts a high-availability architecture. The table structure of the feature library, case library, and rule library adopts a standardized design, and appropriate indexes are established to improve query efficiency. The knowledge base management software is developed in Java and provides a web version management interface. The early warning analysis layer uses a GPU server cluster, each server is equipped with a dual-channel server-level processor, 4 high-performance GPU acceleration cards, and 512GB memory. The deep learning framework uses an open source deep learning framework, and the rule reasoning engine uses an open source rule engine. The early warning analysis software is developed in Python and uses GPU to accelerate the calculation process. The system supports horizontal expansion and can dynamically increase server nodes according to the computing load. Figure 7As shown, the application layer adopts the Web architecture, the application server adopts the mainstream Web server and application server architecture, and is deployed through containerization; the Web server adopts containerization deployment, including the application server, which adopts the mainstream Web server architecture; the data display module adopts a responsive design; the authority management module supports multi-level authority configuration; the log recording module records the system operation log. The server is equipped with a server-level processor and 256GB of memory. The display end is developed using the mainstream front-end framework, adopts a responsive design, and supports PC and mobile terminal access. The system provides functional modules such as real-time monitoring, early warning display, and maintenance management, and supports multiple data visualization methods. The visualization interface includes components such as the system navigation bar, real-time monitoring panel, early warning information panel, trend analysis panel, and health score panel, and each panel supports custom configuration. The present invention realizes the effective fusion of multi-source heterogeneous data by establishing a sound data acquisition and preprocessing mechanism; by building a complete knowledge base system, it ensures the continuous accumulation and optimization of fault characteristics, cases, and rules; through the hybrid early warning model combining deep learning with rule reasoning, the accuracy and reliability of fault early warning are significantly improved; through the deployment of distributed computing architecture, the real-time monitoring needs of large-scale wind farms are met. The present invention has strong practicality and promotion value, can effectively improve the intelligent operation and maintenance level of wind farms, and is of great significance to the healthy development of the wind power industry.

[0040] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

[0041] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0042] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0043] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0044] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0045] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A knowledge base construction method for intelligent early warning of fan failure, characterized in that: include: Collect fan operation data at different sampling frequencies; Performing time alignment, outlier detection, data type conversion and data completion processing on the collected operating data; Build an early warning knowledge base based on the completed operating data; The constructed early warning knowledge base is used to provide fault early warning for the wind turbine.

2. The knowledge base construction method for intelligent early warning of wind turbine failure according to claim 1 is characterized in that: The fan operation data is collected at different sampling frequencies, specifically including: The data acquisition unit is connected to the fan control system to collect the fan operating parameters at a collection frequency of 1 Hz; Connect the vibration data acquisition unit to the vibration sensors installed on the key components of the wind turbine to collect vibration signals on the key components at a collection frequency of 25.6 kHz; The environmental data acquisition unit is connected to the weather station to collect meteorological parameters around the fan at a collection frequency of 0.1 Hz.

3. The knowledge base construction method for intelligent early warning of wind turbine failure according to claim 2 is characterized in that: The operating parameters of the fan include fan speed, fan power and fan temperature parameters; The key components include bearings and gear boxes; The meteorological parameters include wind speed, wind direction and ambient temperature parameters.

4. The knowledge base construction method for intelligent early warning of wind turbine failure according to claim 1 is characterized in that: The collected operation data is subjected to time alignment, outlier detection, data type conversion and data completion processing, specifically including: The collected operation data is time synchronized and a sliding window algorithm is used to align the data; Using the 3σ criterion, outliers in the aligned operating data are detected, and data beyond the normal range is marked or eliminated; Convert running data in different formats into standard floating point or integer data; Missing values ​​in the running data are processed using linear interpolation or nearest neighbor interpolation methods.

5. The knowledge base construction method for intelligent early warning of wind turbine failure according to claim 1 is characterized in that: The construction of the early warning knowledge base based on the completed operation data specifically includes: Based on the completed operation data, a warning knowledge base is constructed, wherein the warning knowledge base includes a fault feature base, a fault case base and a diagnosis rule base; Wherein, the fault feature library is used to record the structural definition and extraction rules of fault features; The fault case library is used to store historical fault cases of wind turbines; The diagnosis rule base contains expert diagnosis rules, and the expert diagnosis rules are used for diagnosing fan faults.

6. The knowledge base construction method for intelligent early warning of wind turbine failure according to claim 5 is characterized in that: In the fault feature library, XML format is used to describe the feature structure and establish feature extraction rules; In the fault case library, a hierarchical classification method is adopted to organize historical fault cases; In the diagnostic rule base, a rule editing and verification mechanism is established to enable the import or export of rules.

7. The knowledge base construction method for intelligent early warning of wind turbine failure according to claim 1 is characterized in that: Using the constructed early warning knowledge base to provide fault early warning for the wind turbine specifically includes: Using the constructed early warning knowledge base, respectively extracting time domain features, frequency domain features, operating condition features and environmental features from the operation data; A deep learning model is used to provide fault warning for wind turbines, while expert rules are applied to diagnose faults. The fault results are output by combining deep learning and rule reasoning.

8. The knowledge base construction method for intelligent early warning of wind turbine failure according to claim 7 is characterized in that: The deep learning model is used to provide fault warning for the fan, specifically including: Constructing a deep learning model, wherein the deep learning model includes an input layer, a convolutional layer, a long short-term memory network layer, and an output layer; Wherein, in the convolution layer, a 3×3 convolution kernel is used to extract local features; In the long short-term memory network layer, 128 hidden units are configured; Through the output layer, a failure probability prediction value is generated.

9. The knowledge base construction method for intelligent early warning of wind turbine failure according to claim 8 is characterized in that: The output failure result includes: Generate real-time monitoring information to display the operating status of the fan equipment; Generate warning information, display the current fan fault warning level and fault details; Generate trend analysis information to display the trend of fan operating parameters; Generate health score information to display the health status of the fan equipment.

10. A knowledge base application method for intelligent early warning of fan failure, characterized in that: The method is carried out by using the knowledge base construction method of wind turbine fault intelligent early warning according to any one of claims 1 to 9, comprising: Process the collected fan operation data; Inputting the processed wind turbine operation data into an early warning knowledge base; Based on the early warning knowledge base, an early warning is issued to the wind turbine.

Images