CN-121976926-A - Wind turbine generator system fault diagnosis method and system based on meteorological factors

CN121976926ACN 121976926 ACN121976926 ACN 121976926ACN-121976926-A

Abstract

The invention relates to the technical field of wind power failure intelligent diagnosis, in particular to a wind turbine generator failure diagnosis method and system based on meteorological factors, comprising the following steps: and collecting the environmental meteorological parameters and the body operation parameters of the unit, performing time alignment and data fusion on the two types of parameters, generating a meteorological-operation coupling data set representing the association relation between the meteorological and the unit operation, and extracting three weather-sensitive fault characteristics of yaw system overload caused by vibration impact triggered by sudden change of wind speed, lubricating oil viscosity abnormality caused by low temperature and high humidity and rapid change of wind direction. The method comprises the steps of constructing a meteorological layering fault diagnosis model comprising a wind speed layering, temperature-humidity combination and wind direction-yaw linkage diagnosis module, identifying weather strong related potential faults by a characteristic input model, and generating a diagnosis report containing fault types, positions and weather causes. The invention can clearly determine the corresponding relation between the meteorological cause and the fault, and realize the accurate identification and clear representation of the cause of the meteorological cause fault.

Inventors

MA NING
HUANG BO
LI XIAOLONG
LONG FEI
YANG YIHUI
ZHAO QI
Yue kaixuan
Hui Xuefeng

Assignees

陕西省水电开发集团股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260408

Claims (10)

1. A wind turbine generator system fault diagnosis method based on meteorological factors is characterized by comprising the following steps: collecting a meteorological parameter set of an environment where a wind turbine is located and an operation parameter set of a wind turbine body, Performing time alignment and data fusion processing on the meteorological parameter set and the operation parameter set to generate a meteorological-operation coupling data set, wherein the meteorological-operation coupling data set is used for representing the association relation between meteorological conditions and unit operation states; Extracting weather-sensitive fault characteristics from the weather-operation coupling data set, wherein the weather-sensitive fault characteristics comprise vibration impact characteristics triggered by sudden change of wind speed, lubricating oil viscosity abnormal characteristics caused by low temperature and high humidity and yaw system overload characteristics caused by rapid change of wind direction; constructing a fault diagnosis model based on meteorological layering, wherein the fault diagnosis model comprises a wind speed layering diagnosis module, a temperature-humidity combined diagnosis module and a wind direction-yaw linkage diagnosis module; and inputting the weather-sensitive fault characteristics into the fault diagnosis model, identifying a potential fault mode which is strongly related to the current weather conditions, and generating a preliminary diagnosis report containing the fault type, the fault position and the weather cause.
2. The wind turbine generator system fault diagnosis method based on meteorological factors according to claim 1, wherein performing time alignment and data fusion processing on the meteorological parameter set and the operation parameter set to generate a meteorological-operation coupling data set comprises: the meteorological parameter set comprises a wind speed change time sequence, a wind direction angle time sequence, atmospheric humidity distribution data, environment temperature change data and atmospheric pressure gradient data; The operation parameter set comprises a main shaft bearing vibration spectrum, gearbox oil temperature data, generator winding temperature, yaw system current value and blade pitch angle time sequence data; Uniformly processing the wind speed change time sequence, the wind direction angle time sequence, the atmospheric humidity distribution data, the environmental temperature change data and the atmospheric pressure gradient data into a synchronous timestamp format to form synchronous meteorological parameters; Uniformly processing the vibration spectrum of the main shaft bearing, the oil temperature data of the gear box, the temperature of a generator winding, the current value of a yaw system and the time sequence data of the pitch angle of the blades into a synchronous time stamp format to form synchronous operation parameters; Establishing a meteorological influence transmission chain, and defining an influence factor matrix of each meteorological parameter on a specific operation parameter, wherein the influence factor matrix comprises experience weight coefficients between meteorological parameter change and operation parameter response; carrying out weighted fusion calculation on the synchronized meteorological parameters and the corresponding synchronized operation parameters by utilizing the influence factor matrix to generate a meteorological-operation coupling feature vector; And arranging all the weather-operation coupling characteristic vectors obtained by calculation at different time points in time sequence to form the weather-operation coupling data set.
3. The wind turbine generator system fault diagnosis method based on meteorological factors according to claim 2, wherein extracting meteorological-sensitive fault features from the meteorological-operational coupling data set comprises: In the meteorological-operational coupling data set, monitoring a moment point when the absolute value of the gradient of the wind speed change time sequence exceeds a preset mutation threshold value, and extracting energy mutation characteristics of a main shaft bearing vibration spectrum in a time period before and after the moment point to form the vibration impact characteristics triggered by wind speed mutation; In the meteorological-operational coupling data set, identifying a working condition period which simultaneously satisfies that the environmental temperature change data is lower than a low temperature threshold value and the atmospheric humidity distribution data is higher than a high humidity threshold value, extracting deviation characteristics of gear box oil temperature data and a nominal viscosity calculated value of lubricating oil in the working condition period, and forming the lubricating oil viscosity abnormal characteristics caused by low temperature and high humidity; And capturing a time interval when the change rate of the wind direction angle time sequence exceeds a preset rapid change threshold value in the meteorological-operation coupling data set, and synchronously extracting abnormal fluctuation characteristics of a yaw system current value in the time interval to form the yaw system overload characteristics caused by rapid change of the wind direction.
4. The wind turbine generator system fault diagnosis method based on meteorological factors according to claim 3, wherein the constructing a fault diagnosis model based on meteorological layering comprises: Dividing a plurality of wind speed working condition layers according to the amplitude range of the wind speed change time sequence, and training a sub-model specially used for identifying typical faults in the wind speed range aiming at each wind speed working condition layer to form the wind speed layering diagnosis module; establishing a two-dimensional combined working condition space of environmental temperature change data and atmospheric humidity distribution data, dividing the two-dimensional combined working condition space into a plurality of discrete temperature-humidity combined intervals, and training a submodel specially used for diagnosing component degradation or performance deviation under temperature-humidity conditions aiming at each temperature-humidity combined interval to form the temperature-humidity combined diagnosis module; establishing a correlation map of a wind direction change mode and a yaw system load, and respectively training a corresponding yaw system health state evaluation sub-model aiming at each wind direction change mode of wind direction mutation, wind direction gradual change and wind direction continuous abnormality to form a wind direction-yaw linkage diagnosis module; The wind speed layering diagnosis module, the temperature and humidity combined diagnosis module and the wind direction and yaw linkage diagnosis module are integrated under a unified diagnosis frame, and the diagnosis frame automatically selects and activates the corresponding diagnosis module according to the input meteorological parameters.
5. The method for diagnosing a wind turbine generator system fault based on meteorological factors of claim 4, wherein inputting the meteorological-sensitive fault characteristics into the fault diagnosis model identifies a potential fault mode that is strongly related to current meteorological conditions, comprising: Inputting the vibration impact characteristics triggered by the sudden change of the wind speed to the wind speed layering diagnosis module, and calling a corresponding sub-model by the wind speed layering diagnosis module according to the current wind speed working condition layer to judge whether the vibration impact characteristics are matched with characteristic fingerprints of structural fatigue, bolt looseness or unbalanced faults of the fan; inputting the abnormal characteristics of the viscosity of the lubricating oil caused by low temperature and high humidity into the temperature and humidity combined diagnosis module, and calling a corresponding sub-model by the temperature and humidity combined diagnosis module according to the current temperature and humidity combined interval to judge whether the abnormal characteristics of the viscosity of the lubricating oil indicate the increase of the abrasion of a gear box, poor lubrication of a bearing or the emulsion risk of the oil; Inputting the yaw system overload characteristic caused by the rapid change of the wind direction into the wind direction-yaw linkage diagnosis module, and calling a corresponding sub-model by the wind direction-yaw linkage diagnosis module according to the current wind direction change mode to judge whether the yaw system overload characteristic caused by the rapid change of the wind direction indicates overheating of a yaw motor, damage of a reduction gear or abrasion of a brake system; And summarizing the judging results of the diagnosis modules to generate a comprehensive diagnosis list, wherein the comprehensive diagnosis list comprises a plurality of candidate potential fault modes and corresponding confidence probabilities thereof.
6. The wind turbine generator system fault diagnosis method based on meteorological factors according to claim 5, further comprising a cause association analysis step based on historical fault cases: Establishing a historical fault case library, wherein each historical fault case records historical weather-operation coupling data, a final fault component and a determined root cause of the fault when the fault occurs; Retrieving a historical fault case set with similarity with the current weather-operating coupling data set exceeding a preset threshold value from the historical fault case library; Counting the occurrence frequency of each fault cause from the retrieved historical fault case set, and calculating the posterior probability of each fault cause under the current meteorological condition; Cross-verifying candidate potential fault modes in a comprehensive diagnosis list generated by the fault diagnosis model and posterior probability of each fault cause obtained by the historical fault case analysis; and for the fault modes which are simultaneously in the comprehensive diagnosis list and have high posterior probability, the confidence level of the fault modes is improved, and typical historical case numbers of the fault modes are associated and marked so as to enhance the reliability of the primary diagnosis report.
7. The wind turbine generator system fault diagnosis method based on meteorological factors according to claim 6, further comprising a fault risk early warning step based on meteorological forecast: receiving refined numerical weather forecast data within a specified future time period, wherein the weather forecast data comprises a predicted wind speed, a predicted wind direction, a predicted temperature, a predicted humidity and a predicted air pressure; Inputting the weather forecast data into the fault diagnosis model, and simulating the evolution trend of weather-sensitive fault characteristics encountered by the wind turbine in a future time period; Based on the simulated evolution trend, predicting potential fault modes output by each fault diagnosis module and predicted occurrence time and severity level of the potential fault modes in a future designated time period; generating a risk early warning schedule comprising fault risk levels at various time points in a future time period, wherein the risk early warning schedule is used for guiding operation and maintenance personnel to conduct preventive inspection and maintenance in advance.
8. The wind turbine generator system fault diagnosis method based on meteorological factors according to claim 7, wherein the weather forecast data is input to the fault diagnosis model to simulate the evolution trend of meteorological-sensitive fault characteristics encountered by the wind turbine generator system in a future time period, and the method comprises the following steps: Extrapolation of a predicted wind speed, a predicted wind direction, a predicted temperature, a predicted humidity and a predicted air pressure within a future specified time period according to the latest state of a currently acquired operation parameter set to generate a simulated weather-operation coupling data set within the future time period; Calculating simulated vibration impact characteristics triggered by predicted wind speed mutation, which occur at each future time point, based on the simulated meteorological-operation coupling data set; Calculating simulated lubricating oil viscosity abnormal characteristics caused by predicted low temperature and predicted high humidity at each future time point based on the simulated weather-operation coupling data set; calculating overload characteristics of the simulated yaw system, which occur at each future time point and are caused by rapid change of the predicted wind direction, based on the simulated meteorological-operation coupling data set; And arranging the calculated simulated vibration impact characteristic, simulated lubricating oil viscosity abnormal characteristic and simulated yaw system overload characteristic according to a time sequence to form an evolution trend graph of the weather-sensitive fault characteristic.
9. The wind turbine generator system fault diagnosis method based on meteorological factors according to claim 8, further comprising a step of dynamically adaptively updating a fault diagnosis model: After the wind turbine generator actually breaks down and completes maintenance, a maintenance report is obtained, wherein the maintenance report comprises confirmed actual fault components, actual fault reasons and complete meteorological parameters and operation parameter records before and after the occurrence of the faults; Recording complete meteorological parameters and operation parameters before and after the occurrence of faults as new meteorological-operation coupling data samples, and taking actual fault components and reasons in a maintenance report as sample labels; Adding the new weather-operating coupling data sample and sample label to a training sample set of the fault diagnosis model; And (3) carrying out periodic fine tuning training on parameters of the fault diagnosis model by using the updated training sample set, so that the diagnosis rules of the fault diagnosis model adapt to new modes caused by fan part aging and local climate change of the wind power plant.
10. A wind turbine fault diagnosis system based on meteorological factors, comprising a central data processing device and a distributed data acquisition network, wherein the distributed data acquisition network is in data communication with the central data processing device and is used for acquiring and transmitting a meteorological parameter set and an operation parameter set, and the central data processing device is used for executing a wind turbine fault diagnosis method based on meteorological factors according to any one of claims 1 to 9.

Description

Wind turbine generator system fault diagnosis method and system based on meteorological factors Technical Field The invention relates to the technical field of wind power fault intelligent diagnosis, in particular to a wind turbine generator fault diagnosis method and system based on meteorological factors. Background The conventional wind turbine generator system fault diagnosis technology only collects the running parameters of the turbine generator system body, carries out fault recognition work by means of single running data, monitors and analyzes the running states of the turbine generator system body such as machinery, electricity and the like, and does not incorporate the environmental meteorological parameters into a fault diagnosis core data system. The existing diagnosis mode does not build data processing logic aiming at the association relation between meteorological conditions and the running state of the unit, the diagnosis model adopts a generalized structural design, and the specific fault diagnosis adaptation is not developed by combining with meteorological dimensions. The existing fault diagnosis scheme cannot perform time alignment and data fusion processing on meteorological parameters and operation parameters, is difficult to generate coupling data representing the correlation between meteorological and unit operation, and cannot extract exclusive meteorological sensitive fault characteristics corresponding to sudden changes of wind speed, low temperature and high humidity and rapid changes of wind direction. The general diagnosis model is not designed in a layering manner according to meteorological dimensions, does not have a special diagnosis module for wind speed layering, temperature and humidity combination and wind direction-yaw linkage, cannot identify unit faults induced by meteorological conditions, and cannot correlate meteorological cause information with diagnosis results. According to the invention, time alignment and data fusion of weather and operation parameters are required to be completed, three types of weather sensitive fault characteristics are extracted, a weather layered fault diagnosis model is built, a corresponding diagnosis module is configured, and a diagnosis report containing fault types, positions and weather causes is output. Disclosure of Invention The invention aims to solve the defects in the prior art, and provides a wind turbine generator fault diagnosis method and system based on meteorological factors. In order to achieve the purpose, the invention adopts the following technical scheme that the wind turbine generator fault diagnosis method based on meteorological factors comprises the following steps: collecting a meteorological parameter set of an environment where a wind turbine is located and an operation parameter set of a wind turbine body, Performing time alignment and data fusion processing on the meteorological parameter set and the operation parameter set to generate a meteorological-operation coupling data set, wherein the meteorological-operation coupling data set is used for representing the association relation between meteorological conditions and unit operation states; Extracting weather-sensitive fault characteristics from the weather-operation coupling data set, wherein the weather-sensitive fault characteristics comprise vibration impact characteristics triggered by sudden change of wind speed, lubricating oil viscosity abnormal characteristics caused by low temperature and high humidity and yaw system overload characteristics caused by rapid change of wind direction; constructing a fault diagnosis model based on meteorological layering, wherein the fault diagnosis model comprises a wind speed layering diagnosis module, a temperature-humidity combined diagnosis module and a wind direction-yaw linkage diagnosis module; and inputting the weather-sensitive fault characteristics into the fault diagnosis model, identifying a potential fault mode which is strongly related to the current weather conditions, and generating a preliminary diagnosis report containing the fault type, the fault position and the weather cause. As a further aspect of the present invention, performing a time alignment and data fusion process on the weather parameter set and the operation parameter set to generate a weather-operation coupled data set, including: the meteorological parameter set comprises a wind speed change time sequence, a wind direction angle time sequence, atmospheric humidity distribution data, environment temperature change data and atmospheric pressure gradient data; The operation parameter set comprises a main shaft bearing vibration spectrum, gearbox oil temperature data, generator winding temperature, yaw system current value and blade pitch angle time sequence data; Uniformly processing the wind speed change time sequence, the wind direction angle time sequence, the atmospheric humidity distribution data, the environmental temperature