Fault diagnosis analysis method and device for wind turbine generator

Abstract

The invention discloses a fault diagnosis analysis method and device for a wind turbine generator, and the method comprises the following steps: building a unit fault diagnosis library after a fault occurs in the unit, and determining a plurality of associated fault systems associated with the fault through the unit fault diagnosis library; wherein the unit fault diagnosis database is composed of a preset historical fault knowledge base and a current fault database; respectively collecting system operation data corresponding to a plurality of time sampling points of the plurality of associated fault systems before and after a fault occurs, preprocessing all the system operation data, then performing fault analysis and positioning on the preprocessed data according to a fault timeline, and determining an abnormal variable; and determining a fault component and a fault reason according to the abnormal variable. The accuracy and efficiency of wind turbine generator fault analysis can be improved.

Description

Technical Field

[0001] This invention relates to the field of mechanical fault analysis, and in particular to a fault diagnosis and analysis method and apparatus for wind turbine generators. Background Technology

[0002] Currently, wind power technology is becoming increasingly mature, and various turbine manufacturers can now output a range of mature wind turbine products. For later operation of these turbines, there are also relatively mature troubleshooting manuals, process guidelines, and video tutorials. However, with the rapid development of the wind power industry, the entire sector has experienced explosive growth, expanding from nearshore to offshore, from suburban areas to remote mountainous regions, and from multi-person to minimally staffed or even unmanned operations. From initial grid connection and commissioning to later turbine maintenance, troubleshooting has always been a significant challenge. Although there are relatively mature troubleshooting manuals for the turbines themselves, the equipment involves multiple professional fields, such as electrical, mechanical, communication, and programming. Therefore, it is virtually impossible for an individual to fully master all the maintenance techniques related to the entire turbine. Furthermore, the design philosophies and methods of different turbine manufacturers vary greatly, making troubleshooting a significant challenge for installation, commissioning, and later operation and maintenance of wind farms.

[0003] Existing fault handling methods for wind turbine systems generally involve the following steps: Sensor feedback signals are processed by the DI / AI input module, converting them into information readable by the control module. This information is then further processed into corresponding status words, such as the fault status words corresponding to the sensor and DI / AI input module feedback information, which are then reported as faults. Maintenance personnel need to consult the fault manual based on the fault status words and then confirm the fault based on their on-site judgment.

[0004] However, this troubleshooting approach is easily limited by the experience and capabilities of maintenance personnel, and requires retrieving data from each subsystem for individual analysis, resulting in a one-sided view. Furthermore, maintenance personnel need to understand the operation of each system and perform real-time data analysis. However, in on-site maintenance, relying solely on individual maintenance personnel's fault data analysis makes it difficult to achieve a unified analysis, leading to significant limitations in the troubleshooting approach. Moreover, it is difficult to correlate related tests of various devices under fault conditions, further resulting in cumbersome, time-consuming, and difficult data analysis. Summary of the Invention

[0005] To overcome the shortcomings of the prior art, this invention provides a fault diagnosis and analysis method and apparatus for wind turbine units, which can improve the accuracy and efficiency of wind turbine unit fault analysis.

[0006] An embodiment of the present invention provides a fault diagnosis and analysis method for wind turbine generators, comprising the following steps: When a unit malfunctions, a unit malfunction diagnosis database is constructed, and several associated malfunction systems related to the malfunction are identified through the unit malfunction diagnosis database; wherein, the unit malfunction diagnosis database consists of a preset historical malfunction knowledge base and a current malfunction database; System operation data of the several related fault systems are collected at several time sampling points before and after the fault occurs, and all the system operation data are preprocessed. Then, the preprocessed data is analyzed and located according to the fault timeline to determine abnormal variables. Based on the abnormal variables, determine the faulty component and the cause of the fault.

[0007] Furthermore, when a unit malfunctions, a unit fault diagnosis database is constructed, and several associated fault systems related to the fault are identified through the unit fault diagnosis database, specifically including: When a unit malfunctions, the unit malfunction message and the current status data of each system of the unit are collected to form the current malfunction database. The unit malfunction diagnosis database is then constructed based on the current malfunction database and the preset historical malfunction knowledge base. The preset historical malfunction knowledge base includes a preset unit malfunction tree, historical unit operating data, and malfunction handling manual. Based on the preset unit fault tree and fault handling manual, several associated systems corresponding to the unit fault message are determined as the associated fault systems.

[0008] Furthermore, the step of collecting system operation data for the several related fault systems at several time points before and after the fault occurred, and preprocessing all the system operation data, specifically includes: Starting from the time of the fault occurrence, and with a preset time threshold as the interval, several time sampling points are determined before and after the fault occurrence. System operation data for all the associated fault systems were collected at the time sampling points; the system operation data included variable data corresponding to several different types of variables. The data types of the collected system operation data are uniformly converted into computer-readable data types to complete the preprocessing.

[0009] Preferably, the data type of the collected system operation data is uniformly converted into a computer-readable data type, specifically including: The decimal data in the system operation data is converted into binary data using a base conversion algorithm; The pulse quantity document-type data packets in the variable data are read and parsed into tabular data using a preset framework.

[0010] Furthermore, the step of performing fault analysis and location on the preprocessed data based on the fault timeline to determine abnormal variables specifically includes: The fault timeline is formed by the time sampling points, and a variable data coordinate system is created with the fault timeline as the horizontal axis. The variable data of each type of variable is input into the variable data coordinate system respectively to obtain the fault data point set of the variable data changing over time; The fault data point set is compared with the unit's historical operating data. When the fault data point set shows an anomaly compared with the unit's historical operating data, the variable corresponding to the fault data point set is determined to be an abnormal variable.

[0011] Preferably, the abnormal phenomena include, but are not limited to: the fault data point set showing abrupt changes or missing data points compared to the unit's historical operating data, or the existence of variable data exceeding the preset normal range.

[0012] Furthermore, determining the faulty component and cause of the fault based on the abnormal variables specifically includes: Extract the operational data of each component associated with the abnormal variable on the fault timeline, and create a fault data point set for the associated component; The faulty component and the cause of the fault are determined based on the fault data point set of the associated components.

[0013] Preferably, when no abnormality is found in the associated components of the abnormal variable, the method further includes: Send a secondary fault message to the human operator and obtain feedback from the human operator regarding the components to be investigated; Extract the operational data of the component to be investigated from the fault timeline and create a fault data point set for the component to be investigated. The faulty component and the cause of the fault are determined based on the fault data point set of the component to be investigated.

[0014] Another embodiment of the present invention provides a fault diagnosis and analysis device for wind turbines, comprising: an association module, a data acquisition module, and an analysis module; The association module is used to construct a unit fault diagnosis database after a unit failure occurs, and to identify several associated fault systems related to the fault through the unit fault diagnosis database; wherein, the unit fault diagnosis database consists of a preset historical fault knowledge base and a current fault database; The acquisition module is used to collect system operation data corresponding to several time sampling points before and after the occurrence of the fault in the several related fault systems, preprocess all the system operation data, and then perform fault analysis and location on the preprocessed data according to the fault timeline to determine abnormal variables. The analysis module is used to determine the faulty component and the cause of the fault based on the abnormal variables.

[0015] Furthermore, the association module is used to construct a unit fault diagnosis library after a unit fault occurs, and to identify several associated fault systems related to the fault through the unit fault diagnosis library, specifically including: When a unit malfunctions, the unit malfunction message and the current status data of each system of the unit are collected to form the current malfunction database. The unit malfunction diagnosis database is then constructed based on the current malfunction database and the preset historical malfunction knowledge base. The preset historical malfunction knowledge base includes a preset unit malfunction tree, historical unit operating data, and malfunction handling manual. Based on the preset unit fault tree and fault handling manual, several associated systems corresponding to the unit fault message are determined as the associated fault systems.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: By uniformly collecting, processing, and analyzing data from all fault-related systems at several time points before and after the fault occurs, problems such as incomplete system analysis, inaccurate fault location, and parameter changes during actual handling that prevent the original fault handling approach from becoming unusable, lack of proficiency in handling non-logical faults and replacing components, and difficulty in handling faults in harsh environments are avoided during routine maintenance. This improves the efficiency and accuracy of fault location for wind turbine units. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a fault diagnosis and analysis method for wind turbine units, provided as an embodiment of the present invention.

[0018] Figure 2 This is a schematic diagram of a fault diagnosis and analysis device for wind turbines, provided as another embodiment of the present invention. Detailed Implementation

[0019] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent. The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0020] Reference Figure 1 The following is a flowchart illustrating a fault diagnosis and analysis method for wind turbine generators according to an embodiment of the present invention, comprising the following steps: S1: When a unit malfunctions, a unit malfunction diagnosis database is constructed, and several associated malfunction systems related to the malfunction are identified through the unit malfunction diagnosis database; wherein, the unit malfunction diagnosis database consists of a preset historical malfunction knowledge base and a current malfunction database; S2: Collect system operation data corresponding to several time sampling points before and after the occurrence of the fault in the several related fault systems respectively, preprocess all the system operation data, and then perform fault analysis and location on the preprocessed data according to the fault timeline to determine abnormal variables; S3: Based on the abnormal variables, determine the faulty component and the cause of the fault.

[0021] For step S1, specifically, when a unit malfunctions, a unit malfunction diagnosis database is constructed, and several associated malfunction systems related to the malfunction are identified through the unit malfunction diagnosis database, specifically including: When a unit malfunctions, the unit malfunction message and the current status data of each system of the unit are collected to form the current malfunction database. The unit malfunction diagnosis database is then constructed based on the current malfunction database and the preset historical malfunction knowledge base. The preset historical malfunction knowledge base includes a preset unit malfunction tree, historical unit operating data, and malfunction handling manual. Based on the preset unit fault tree and fault handling manual, several associated systems corresponding to the unit fault message are determined as the associated fault systems.

[0022] In a preferred embodiment, after a fault occurs, this application incorporates all relevant fault manuals, fault trees, historical operating data, fault data, and real-time status information of each subsystem and equipment of the unit into a fault diagnosis library, and performs correlation analysis on the fault based on this library. Specifically, when a fault occurs, several systems associated with the fault status word (i.e., the unit fault message) are first determined based on preset information such as the fault manual and fault tree. By constructing the fault diagnosis library in real time for faults, the relevance and comprehensiveness of fault analysis can be significantly improved.

[0023] For example, if the unit reports a fault message of "power generation not matching theoretical value", the system involved in the fault can be determined based on the unit's fault tree and fault manual, including the wind speed prediction system, converter system, pitch system, and gearbox system. These systems are the associated fault systems.

[0024] For step S2, specifically, the process of collecting system operation data for the several associated fault systems at several time points before and after the fault occurred, and preprocessing all the system operation data, specifically includes: Starting from the time of the fault occurrence, and with a preset time threshold as the interval, several time sampling points are determined before and after the fault occurrence. System operation data for all the associated fault systems were collected at the time sampling points; the system operation data included variable data corresponding to several different types of variables. The data types of the collected system operation data are uniformly converted into computer-readable data types to complete the preprocessing.

[0025] Preferably, the data type of the collected system operation data is uniformly converted into a computer-readable data type, specifically including: The decimal data in the system operation data is converted into binary data using a base conversion algorithm; The pulse quantity document-type data packets in the variable data are read and parsed into tabular data using a preset framework.

[0026] In a preferred embodiment, after identifying the associated faulty system, system operation data of all associated systems can be collected at preset time intervals (e.g., once every 1 second) for a period of time before and after the fault occurs (e.g., 60 seconds before the fault and 20 seconds after the fault).

[0027] Subsequently, in this preferred embodiment, the time of occurrence is used as a common reference coordinate system. Different data from these different systems are first converted into the same type of data through the information processing module, and then the variable data that caused the anomaly is determined through subsequent steps.

[0028] In this preferred embodiment, the process of unifying the data into the same type includes: converting the decimal data of the generator set into binary data using a decimal-to-binary conversion formula to facilitate the generation of statistical tables by the computer; or parsing the document-type data packets of pulse quantities (current, voltage, speed, torque, etc.) measured by the generator set into computer-viewable tabular data through a pre-edited framework; after processing the data packets of different forms into intuitive tables, filtering the data by keywords, similar reference values, etc., extracting data columns, and then generating corresponding chart analysis.

[0029] For step S2, specifically, the step of performing fault analysis and location on the preprocessed data based on the fault timeline to determine abnormal variables includes: The fault timeline is formed by the time sampling points, and a variable data coordinate system is created with the fault timeline as the horizontal axis. The variable data of each type of variable is input into the variable data coordinate system respectively to obtain the fault data point set of the variable data changing over time; The fault data point set is compared with the unit's historical operating data. When the fault data point set shows an anomaly compared with the unit's historical operating data, the variable corresponding to the fault data point set is determined to be an abnormal variable.

[0030] Preferably, the abnormal phenomena include, but are not limited to: the fault data point set showing abrupt changes or missing data points compared to the unit's historical operating data, or the existence of variable data exceeding the preset normal range.

[0031] In a preferred embodiment, in the variable data coordinate system constructed with the timeline as the horizontal axis, by capturing the same variable (such as torque, speed, wind speed, power, etc.) and comparing it vertically with the normal variable data in the historical operation data, it is possible to clearly determine which variable is the abnormal variable by its variable change trend or data point mutation, missing or exceeding the preset normal range.

[0032] For step S3, specifically, determining the faulty component and the cause of the fault based on the abnormal variables includes: Extract the operational data of each component associated with the abnormal variable on the fault timeline, and create a fault data point set for the associated component; The faulty component and the cause of the fault are determined based on the fault data point set of the associated components.

[0033] In a preferred embodiment, after obtaining the abnormal variable, the components associated with the abnormal variable can be determined again based on the unit fault tree and other data. Then, corresponding fault data point sets are made from the operating data of these associated components for fault analysis. This will determine which associated component is the faulty component and the reason for the abnormality of the faulty component. For example, it can be determined that the transmission frequency of a certain signal transmitter is abnormal, and there is a problem of signal transmission interruption or lag.

[0034] Preferably, when no abnormality is found in the associated components of the abnormal variable, the method further includes: Send a secondary fault message to the human operator and obtain feedback from the human operator regarding the components to be investigated; Extract the operational data of the component to be investigated from the fault timeline and create a fault data point set for the component to be investigated. The faulty component and the cause of the fault are determined based on the fault data point set of the component to be investigated.

[0035] In a preferred embodiment, if no fault is found after checking all associated components related to the abnormal variable automatically confirmed by the system, a secondary fault message can be sent to the manual terminal for inspection and intervention by maintenance personnel. Maintenance personnel can provide the system with several other components to be checked, excluding the associated components, based on the actual situation and their own experience, and the system will then conduct a secondary check on these components.

[0036] In this preferred embodiment, during the actual fault handling process, the fault information can be updated based on real-time detection information, and the manual intervention module can be used to perform relevant tests for fault judgment to assist in fault location.

[0037] In summary, the fault diagnosis and analysis method for wind turbines provided by this invention, by uniformly collecting, processing and analyzing data from all fault-related systems at several time points before and after the fault occurs, avoids the problems of incomplete system analysis, inaccurate fault location, and parameter changes during actual handling that lead to the unusability of the original fault handling approach, unfamiliarity with handling non-logical faults and component replacement, and difficulty in handling faults in harsh environments. This improves the efficiency and accuracy of fault location for wind turbines.

[0038] Reference Figure 2 The diagram below shows a fault diagnosis and analysis device for wind turbines according to another embodiment of the present invention, which includes: an association module 101, an acquisition module 102, and an analysis module 103. The association module is used to construct a unit fault diagnosis database after a unit failure occurs, and to identify several associated fault systems related to the fault through the unit fault diagnosis database; wherein, the unit fault diagnosis database consists of a preset historical fault knowledge base and a current fault database; The acquisition module is used to collect system operation data corresponding to several time sampling points before and after the occurrence of the fault in the several related fault systems, preprocess all the system operation data, and then perform fault analysis and location on the preprocessed data according to the fault timeline to determine abnormal variables. The analysis module is used to determine the faulty component and the cause of the fault based on the abnormal variables.

[0039] Furthermore, the association module 101 is used to construct a unit fault diagnosis library after a unit fault occurs, and to determine several associated fault systems related to the fault through the unit fault diagnosis library, specifically including: When a unit malfunctions, the unit malfunction message and the current status data of each system of the unit are collected to form the current malfunction database. The unit malfunction diagnosis database is then constructed based on the current malfunction database and the preset historical malfunction knowledge base. The preset historical malfunction knowledge base includes a preset unit malfunction tree, historical unit operating data, and malfunction handling manual. Based on the preset unit fault tree and fault handling manual, several associated systems corresponding to the unit fault message are determined as the associated fault systems.

[0040] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A method for failure diagnostic analysis for a wind turbine, characterized in that, The method comprises the following steps: After the unit fails, a unit fault diagnosis library is constructed, and a plurality of associated fault systems associated with the fault are determined through the unit fault diagnosis library; wherein the unit fault diagnosis library is composed of a preset historical fault knowledge base and a current fault database; Respectively collect system operation data corresponding to a plurality of time sampling points before and after the failure of the plurality of associated fault systems, and pre-process all the system operation data, and then perform fault analysis and positioning on the pre-processed data according to a fault timeline to determine abnormal variables; According to the abnormal variables, determine the fault components and the fault causes.

2. The wind turbine generator oriented fault diagnostic analysis method of claim 1, wherein, After the unit fails, a unit fault diagnosis library is constructed, and a plurality of associated fault systems associated with the fault are determined through the unit fault diagnosis library; wherein the unit fault diagnosis library is composed of a preset historical fault knowledge base and a current fault database; After the unit fails, collect unit fault messages and current state data of each system of the unit to form the current fault database, and construct the unit fault diagnosis library according to the current fault database and the preset historical fault knowledge base; wherein the preset historical fault knowledge base includes a preset unit fault tree, unit historical operation data and a fault handling manual; According to the preset unit fault tree and the fault handling manual, a plurality of associated systems corresponding to the unit fault message are determined as the associated fault systems.

3. The wind turbine generator oriented fault diagnostic analysis method of claim 2, wherein, The method comprises the following steps: Determine a plurality of time sampling points before and after the failure, taking the failure time as the starting point and a preset time threshold as the interval; Respectively collect system operation data of all the associated fault systems at the time sampling points; the system operation data includes variable data corresponding to a plurality of different types of variables; Convert the data types of the collected system operation data into computer-readable data types to complete the pre-processing.

4. The wind turbine generator oriented fault diagnostic analysis method of claim 3, wherein, The method comprises the following steps: Convert decimal data in the system operation data into binary data through a base conversion algorithm; Read the pulse amount document type data packet through a preset framework to parse it into table data.

5. The wind turbine generator oriented fault diagnostic analysis method of claim 3, wherein, The method comprises the following steps: The fault timeline is composed of the time sampling points, and a variable data coordinate system is created with the fault timeline as the horizontal axis; Respectively input the variable data of each type of variable into the variable data coordinate system to obtain a fault data point set of the variable data changing with time; Compare the fault data point set with the unit historical operation data, and when the fault data point set is abnormal compared with the unit historical operation data, determine the variable corresponding to the fault data point set as an abnormal variable.

6. The wind turbine generator oriented fault diagnostic analysis method of claim 5 wherein, The abnormal phenomenon includes but is not limited to: the fault data point set has data point mutation, missing or variable data out of the preset normal data range compared with the unit historical operation data.

7. The wind turbine generator oriented fault diagnostic analysis method of claim 5 wherein, The abnormal variable is used to determine the fault component and the fault cause, and specifically includes: Respectively extracting each operation data of the associated component on the fault timeline, and making a fault data point set of the associated component; The fault component and the fault cause are determined according to the fault data point set of the associated component.

8. The wind turbine generator oriented fault diagnostic analysis method of claim 7 wherein, When no abnormal associated component of the abnormal variable is found, it further includes: Send a secondary fault message to the manual end, and obtain the feedback of the component to be checked from the manual end; Respectively extracting each operation data of the associated component on the fault timeline, and making a fault data point set of the associated component; The fault component and the fault cause are determined according to the fault data point set of the associated component.

9. A wind turbine generator oriented fault diagnostic analysis apparatus characterized by, It includes: The association module, the collection module and the analysis module; The association module is used to build a unit fault diagnosis library after the unit fails, and determine a plurality of associated fault systems related to the fault through the unit fault diagnosis library; wherein the unit fault diagnosis library is composed of a preset historical fault knowledge base and a current fault database; The collection module is used to collect system operation data corresponding to a plurality of time sampling points before and after the fault of the plurality of associated fault systems, and pre-process all the system operation data, and then perform fault analysis and positioning on the pre-processed data according to the fault timeline to determine the abnormal variable; The analysis module is used to determine the fault component and the fault cause according to the abnormal variable.

10. The wind turbine generator oriented fault diagnostic analysis apparatus of claim 9 wherein, The association module is used to build a unit fault diagnosis library after the unit fails, and determine a plurality of associated fault systems related to the fault through the unit fault diagnosis library, and specifically includes: After the unit fails, collect the unit fault message and the current state data of each system of the unit to form the current fault database, and build the unit fault diagnosis library according to the current fault database and the preset historical fault knowledge base; wherein the preset historical fault knowledge base includes a preset unit fault tree, unit historical operation data and a fault handling manual; According to the preset unit fault tree and the fault handling manual, a plurality of associated systems corresponding to the unit fault message are determined as the associated fault systems.

Images