Search

CN-122014546-A - Wind turbine generator set oil temperature regulation and control method and system based on sensor space-time fusion

CN122014546ACN 122014546 ACN122014546 ACN 122014546ACN-122014546-A

Abstract

The invention discloses a wind turbine generator set oil temperature regulation and control method and system based on sensor space-time fusion, and belongs to the technical field of wind power equipment control. The existing regulation and control method does not consider the cooperative evolution relation among multiple variables of the sensor, and influences the regulation and control precision. The invention discloses a wind turbine generator set oil temperature regulation method based on sensor space-time fusion, which comprises the following steps of obtaining sensor input data; the method comprises the steps of obtaining sensor characteristic data, generating relevance data among sensors, generating sensor space-time characteristic data comprising spatial characteristics and time characteristics, predicting oil temperature by taking the sensor space-time characteristic data as input to obtain an oil temperature predicted value, and generating a control strategy and outputting a control signal according to the control strategy. The method effectively fuses the data of multiple types of sensors in the bearing system of the wind turbine, can cope with complex changes of slow drift and sudden transition coexistence, organically unifies prediction and control, and is more accurate in control.

Inventors

  • Lang Chaohao
  • ZHANG SUJUN
  • WU JIANMING

Assignees

  • 浙江远算科技有限公司

Dates

Publication Date
20260512
Application Date
20260415

Claims (10)

  1. 1. A wind turbine generator system oil temperature regulation and control method based on sensor space-time fusion is characterized by comprising the following steps: firstly, collecting original data of multiple types of sensors of a bearing system of a wind turbine generator in real time, and cleaning the data to obtain sensor input data, wherein the original data of the sensors comprise bearing lubricating oil temperature data; setting a time window based on the sensor input data obtained in the first step, and extracting data features in the time window to obtain sensor feature data, wherein the data features are used for reflecting real-time data, short-term dynamic change trend and time sequence statistics rules of the sensor; Thirdly, based on the sensor characteristic data obtained in the second step, obtaining the dynamic data correlation among the sensors, and synthesizing the physical coupling relation among the sensors on site to generate the correlation data among the sensors; Generating sensor space-time characteristic data comprising spatial characteristics and time characteristics based on the sensor characteristic data obtained in the second step, and predicting the oil temperature at the future moment by taking the sensor space-time characteristic data as input to obtain an oil temperature predicted value; And fifthly, generating a control strategy based on the oil temperature predicted value, and outputting a control signal for controlling the oil temperature regulating device according to the control strategy.
  2. 2. The method for regulating and controlling the oil temperature of the wind turbine based on sensor space-time fusion, as set forth in claim 1, is characterized in that: Firstly, collecting original data of multiple types of sensors of a bearing system of a wind turbine generator in real time, and cleaning the data to obtain input data of the sensors, wherein the method comprises the following steps of: Step 11, sampling multi-type sensor raw data of a bearing system of a wind turbine generator system at a fixed frequency, and marking time marks, wherein the sensor raw data comprises bearing lubricating oil temperature data, environment data and related equipment operation data; step 12, eliminating abnormal values of the data acquired in the step 11; and 13, carrying out normalization processing on the data obtained in the step 12, and filling a default value to obtain sensor input data.
  3. 3. The method for regulating and controlling the oil temperature of the wind turbine based on the sensor space-time fusion according to claim 1, wherein in the second step, the method for extracting the data characteristics in the time window and obtaining the sensor characteristic data comprises the following steps: step 21, setting a sliding window for sensor input data in a time window and sliding the sliding window according to a fixed step length; Step 22, calculating multi-dimensional feature vectors of sensor input data in each sliding window, wherein the multi-dimensional feature vectors comprise normalized values of sensor real-time data, time difference between adjacent moments of the sensors and mean, variance and skewness of the sensors in the sliding windows; and step 23, outputting a sensor characteristic matrix, namely the sensor characteristic data, according to all the multi-dimensional characteristic vectors obtained in the step 22.
  4. 4. The method for regulating and controlling the oil temperature of the wind turbine based on sensor space-time fusion, as set forth in claim 1, is characterized in that: the third step, based on the sensor characteristic data obtained in the second step, obtaining the dynamic data correlation among the sensors, and integrating the physical coupling relation among the sensors on site, and generating the correlation data among the sensors, wherein the method comprises the following steps: step 31, constructing a graph structure with each sensor as a vertex and the adjacent relation among the sensors as a side; Step 32, distributing static weights to edges of the graph structure, wherein the static weights are determined according to physical coupling relations among sensors in the scene; Step 33, setting a dynamic calculation window based on the sensor characteristic data, and calculating correlation indexes among sensors in the dynamic calculation window to obtain dynamic weights; Step 34, fusing the static weight and the dynamic weight to obtain a new edge weight, and updating the edge weight of the graph structure; And step 35, outputting a sensor adjacency relation matrix corresponding to the graph structure, namely the relevance data.
  5. 5. The method for regulating and controlling the oil temperature of the wind turbine based on sensor space-time fusion, as set forth in claim 1, is characterized in that: And a fourth step, based on the sensor characteristic data obtained in the second step, generating sensor space-time characteristic data comprising spatial characteristics and time characteristics according to the relevance data, wherein the method comprises the following steps: Step 41, calculating attention scores between the sensor and each neighbor sensor according to the relevance data for one sensor based on the sensor characteristic data; step 42, processing the attention scores of all the neighbor sensors of the sensor to obtain attention weights; Step 43, replacing different attentiveness to pay attention to different coupling modes among the sensors, and repeating the steps 41 and 42 for a plurality of times to obtain a plurality of attentiveness weights; Step 44, according to the plurality of attention weights obtained in step 43, weighting and aggregating the data features of the neighbor sensors to the sensors to obtain the spatial aggregation feature values of the sensors; Step 45, calculating the space aggregation characteristic values of all the sensors, and outputting a sensor space characteristic matrix according to the space aggregation characteristic values; step 46, setting a plurality of layers of capturing time with different granularities based on the sensor characteristic data; Step 47, capturing sensor time characteristic values in each layer of time, and outputting a sensor time characteristic matrix according to the sensor time characteristic values; and 48, fusing the sensor space feature matrix obtained in the step 45 and the sensor time feature matrix obtained in the step 47 to obtain a sensor space-time feature matrix, namely the sensor space-time feature data.
  6. 6. The method for regulating and controlling the oil temperature of the wind turbine based on sensor space-time fusion, as set forth in claim 1, is characterized in that: In the fourth step, the method further comprises the step of predicting the abnormal probability at the future moment by taking the sensor space-time characteristic data as input to obtain an abnormal probability prediction value.
  7. 7. The method for regulating and controlling the oil temperature of the wind turbine based on sensor space-time fusion as set forth in claim 6, wherein the method is characterized in that: and a sixth step of collecting feedback data of the oil temperature adjusting device and optimizing the control strategy by referring to the feedback data.
  8. 8. The method for regulating and controlling the oil temperature of the wind turbine based on sensor space-time fusion as claimed in claim 7, wherein the method is characterized by comprising the following steps: the sixth step is that the feedback data of the oil temperature adjusting device is collected and the control strategy is optimized by referring to the feedback data, and the method is as follows: Step 61, collecting feedback data of the oil temperature adjusting device in real time, and calculating an energy consumption index of the oil temperature adjusting device; step 62, calculating a safety index according to the abnormal probability predicted value obtained in the fourth step; And 63, constructing a reward function taking energy consumption and safety as control targets, and optimizing a control strategy according to the energy consumption index obtained in the step 61 and the safety index obtained in the step 62.
  9. 9. The method for regulating and controlling the oil temperature of the wind turbine based on the sensor space-time fusion according to any one of claims 1 to 8, which is characterized by comprising the following steps: In the fourth step, the predicting the oil temperature at the future time refers to predicting the oil temperature at the future time through a neural network, the neural network is deployed at the edge end of the wind turbine generator set site, the training process of the neural network is deployed at the cloud, the cloud performs centralized modeling and training on the data uploaded by the edge end and transmits training result parameters to the edge end, and the training of the cloud comprises offline training, incremental training and cross-scene migration.
  10. 10. A wind turbine generator system oil temperature regulation and control system based on sensor space-time fusion is characterized in that: A wind turbine generator set oil temperature regulation method based on sensor space-time fusion as claimed in any one of claims 1-6; The wind turbine generator system oil temperature regulation and control system comprises the following modules: The data acquisition and cleaning module is used for acquiring the original data of the multiple types of sensors of the bearing system of the wind turbine generator in real time, and carrying out data cleaning to obtain the input data of the sensors; the characteristic data extraction module is used for setting a time window based on sensor input data, extracting data characteristics in the time window and obtaining sensor characteristic data; The relevance data generation module is used for obtaining dynamic data relevance among the sensors based on the sensor characteristic data, and synthesizing physical coupling relations among the sensors on site to generate relevance data among the sensors; the time-space characteristic data generation and prediction module is used for generating sensor time-space characteristic data comprising space characteristics and time characteristics according to the relevance data based on the sensor characteristic data, and predicting the oil temperature at the future moment by taking the sensor time-space characteristic data as input to obtain an oil temperature predicted value; and the control module is used for generating a control strategy based on the oil temperature predicted value and outputting a control signal for controlling the oil temperature regulating device according to the control strategy.

Description

Wind turbine generator set oil temperature regulation and control method and system based on sensor space-time fusion Technical Field The invention relates to a wind turbine generator set oil temperature regulation and control method and system based on sensor space-time fusion, and belongs to the technical field of wind power equipment control. Background In wind power systems, generator bearing lubricating oil temperature is an important parameter for measuring equipment operation safety and energy efficiency level. An oil temperature anomaly is often indicative of potential failure such as mechanical wear, lubrication degradation, or cooling system failure, the change of which has the complex feature of slow drift coexisting with abrupt transitions. However, the existing oil temperature monitoring and control technology has the following main disadvantages: (1) The monitoring means is delayed and false alarm frequently occurs, namely, the current industrial field mostly adopts a fixed threshold alarm or single-point temperature trend-dependent judging method, and once the oil temperature exceeds a set upper limit, a shutdown or load-reducing instruction is triggered. The method ignores the co-evolution relation among the multiple variables, cannot capture early abnormal symptoms in time, is sensitive to environmental disturbance, and has high false alarm rate. (2) The modeling method ignores the spatial coupling between the sensors, namely, the prior research attempts to introduce a time sequence model such as a long-short-term memory network and the like to predict the oil temperature, but usually only processes a temperature sequence per se, and the spatial structure and the dynamic coupling relation between the sensors such as the oil temperature, the rotating speed and the like are not fused effectively, so that the response to composite faults (such as slow change and abrupt change) is slow, and the prediction capability is limited. (3) The prediction and control strategy is split, the energy efficiency is not optimized, the existing prediction system only provides an alarm signal, and the cooling pump and the lubricating oil pump still operate according to rated power or traditional PID logic, so that not only is the energy consumption waste caused by excessive cooling possibly caused, but also a flow regulation mechanism is lacked in the abnormal evolution process, and the fault risk and the energy loss cannot be reduced cooperatively. Because of the above-described technical problems, improvements to existing control techniques are needed. The information disclosed in this background is only for the understanding of the background of the inventive concept and therefore it may comprise information that does not form the prior art. Disclosure of Invention In view of the above technical problems, one of the objects of the present invention is: the method is characterized in that the correlation data among the sensors is obtained by extracting the sensor characteristic data and integrating the dynamic data correlation and physical coupling relation, the sensor space-time characteristic data is obtained according to the sensor characteristic data and the correlation data, the sensor space-time characteristic data is used as a prediction basis, the prediction accuracy is effectively improved, and the prediction result and the control strategy are organically unified, so that the control is more accurate and effective. In order to solve the technical problems, the invention aims to provide a wind turbine generator system oil temperature regulation system based on sensor space-time fusion, which can effectively fuse multiple types of sensor data in a wind turbine generator system bearing system, can cope with complex oil temperature changes coexisting with slow drift and sudden transition, organically unifies prediction and control, obtains relevance data among sensors by extracting sensor characteristic data and integrating dynamic data relevance and physical coupling relation, obtains sensor space-time characteristic data according to the sensor characteristic data and the relevance data, effectively improves prediction accuracy by using the sensor space-time characteristic data as a basis of prediction, organically unifies a prediction result and a control strategy, and enables control to be more accurate and effective. In order to achieve one of the above objects, a first technical solution of the present invention is: a wind turbine generator system oil temperature regulation and control method based on sensor space-time fusion comprises the following steps: firstly, collecting original data of multiple types of sensors of a bearing system of a wind turbine generator in real time, and cleaning the data to obtain sensor input data, wherein the original data of the sensors comprise bearing lubricating oil temperature data; setting a time window based on the sensor input data obtained in the first step