CN-121996952-A - Intelligent wind field wake loss suppression method based on artificial intelligence

CN121996952ACN 121996952 ACN121996952 ACN 121996952ACN-121996952-A

Abstract

The invention relates to the technical field of wind power generation, in particular to an intelligent wind field wake loss suppression method based on artificial intelligence, which comprises the following steps of collecting the running state and environment information of each fan in a wind power field, and carrying out standardized processing on collected data to form a unified data set; and constructing an artificial intelligent prediction model integrating the space-time characteristics and the fan coupling relation based on the data characteristics, and outputting a prediction result comprising the wind speed, turbulence and power state of each fan downstream position in the wind power plant by using the artificial intelligent prediction model. According to the invention, the running state and the environmental information of each fan in the wind power plant are collected, so that the problems of low running efficiency and power generation loss of the fans caused by the fact that the conventional wind power plant wake loss regulation and control mostly adopts an empirical control method are solved.

Inventors

WANG SHIBIAO
HUANG XIAOQIANG
LIU QI

Assignees

江西大唐国际新能源有限公司

Dates

Publication Date: 20260508
Application Date: 20251231

Claims (10)

1. The intelligent wind field wake loss suppression method based on artificial intelligence is characterized by comprising the following steps of: Collecting the running state and environmental information of each fan in the wind power plant, and carrying out standardized processing on the collected data to form a unified data set; extracting the spatial relationship, the upstream and downstream position relationship and the data characteristics of wind field environment from the data set, and encoding the coupling relationship between the fans; Constructing an artificial intelligent prediction model fusing the space-time characteristics and the fan coupling relation based on the data characteristics, and outputting a prediction result comprising wind speed, turbulence and power states of all fans at downstream positions in the wind power plant by utilizing the artificial intelligent prediction model; Generating an initial fan cooperative control strategy by using a reinforcement learning algorithm, and introducing fan mechanical constraint conditions in the strategy generation process; according to the iteratively updated fan cooperative control strategy, dynamically adjusting the rotating speed and the blade pitch of each fan in the wind power plant; Constraint optimization is carried out on the positions of the fans in the wind power plant based on the prediction result and the control strategy, and the arrangement sequence of the fans is adjusted; And issuing the control strategy and the layout optimization scheme to a fan control system, monitoring the fan state, and correcting the control strategy according to the monitoring data.
2. The intelligent wind farm wake loss suppression method based on artificial intelligence of claim 1, wherein the normalizing the collected data comprises: interpolation or supplementation is carried out on the missing value; detecting and eliminating abnormal values; Normalizing or standardizing the characteristic data; carrying out alignment and resampling on the time series data; a unified dataset is constructed for subsequent feature extraction and prediction.
3. The intelligent wind farm wake loss suppression method based on artificial intelligence according to claim 1, wherein the extracting the spatial relationship between fans, the upstream and downstream positional relationship and the data features of wind farm environment comprises: Acquiring geographic position data of a fan; Calculating horizontal distance, vertical distance and relative azimuth angle among fans; Judging the upstream and downstream position relation of the fan according to the wind direction information; extracting wind field environmental characteristics including wind speed, wind direction, turbulence intensity and other environmental parameters.
4. The intelligent wind farm wake loss suppression method based on artificial intelligence of claim 1, wherein the encoding of the coupling relationship between fans comprises: Constructing a fan coupling matrix to represent the space between fans and the upstream-downstream relationship; Combining the state information of each fan with the coupling matrix to form a feature vector; Integrating all fan feature vectors to form a unified data representation; the data representation is input as an artificial intelligence predictive model.
5. The intelligent wind farm wake loss suppression method based on artificial intelligence of claim 1, wherein the constructing an artificial intelligence prediction model integrating space-time features and fan coupling relations comprises: Inputting the characteristics and the coupling relation of the fan into an artificial intelligent model; Selecting an artificial intelligent model framework integrating space-time characteristics and fan coupling relations, wherein the framework is a fusion framework of a convolutional neural network for extracting space characteristics, capturing time sequence rules by a long-term and short-term memory network and encoding the coupling relations among fans by a graph neural network; training the model, and outputting the wind speed, turbulence and power state of the downstream position of the fan; and verifying the consistency of the model output and the historical data through the root mean square error or the average absolute error, and storing training parameters.
6. The artificial intelligence based intelligent wind farm wake loss suppression method of claim 1, wherein the generating an initial fan cooperative control strategy comprises: Generating an initial fan cooperative control strategy by using a reinforcement learning algorithm; introducing a fan mechanical constraint condition in the generation process of the initial fan cooperative control strategy; The initial fan cooperative control strategy is updated by adopting multi-working condition training, field randomization and strategy regularization iteration, wherein the multi-working condition training comprises data enhancement training based on combination of simulation data generated by computational fluid dynamics or large vortex simulation and actual wind field data; And outputting a control strategy which can be issued to the fan for execution.
7. The intelligent wind farm wake loss suppression method based on artificial intelligence of claim 1, wherein dynamically adjusting the rotational speed and blade pitch of each fan in the wind farm comprises: issuing the generated control strategy to a fan control system; adjusting the rotating speed of the fan according to the strategy instruction; Adjusting the blade pitch according to the strategy instruction; And recording the adjustment result to form fan operation state data.
8. The intelligent wind farm wake loss suppression method based on artificial intelligence of claim 1, wherein the constrained optimization of the position of the wind turbine in the wind farm comprises: Generating constraint conditions according to the prediction result and the control strategy; adopting a genetic algorithm to calculate and optimize the arrangement sequence and the relative position of the fans; Generating a fan layout scheme meeting constraint conditions; and outputting the fan position adjustment scheme to a control system.
9. The intelligent wind farm wake loss suppression method based on artificial intelligence of claim 1, wherein the modifying the control strategy according to the monitoring data comprises: Inputting the monitoring data into a strategy updating module; Adjusting reinforcement learning control strategy parameters according to fan power output deviation data; And outputting the updated control strategy and issuing the updated control strategy to the fan for execution.
10. An artificial intelligence based intelligent wind park wake loss suppression system for use in an artificial intelligence based intelligent wind park wake loss suppression method according to any one of claims 1-9, the system comprising: the data acquisition module is used for acquiring the running state and the environment information of each fan in the wind power plant and carrying out standardized processing on the acquired data to form a unified data set; The feature extraction module is used for extracting the data features of the space relationship, the upstream and downstream position relationship and the wind field environment among the fans from the data set and encoding the coupling relationship among the fans; The wake flow prediction module is used for constructing an artificial intelligent prediction model integrating the space-time characteristics and the fan coupling relation based on the data characteristics, and outputting a prediction result comprising wind speed, turbulence and power states of the downstream positions of all fans in the wind power plant by utilizing the artificial intelligent prediction model; The intelligent control module is used for generating an initial fan cooperative control strategy by utilizing a reinforcement learning algorithm and introducing fan mechanical constraint conditions in the strategy generation process; the strategy updating module is used for dynamically adjusting the rotating speed and blade pitch of each fan in the wind power plant according to the iteratively updated fan cooperative control strategy; The layout optimization module is used for carrying out constraint optimization on the positions of the fans in the wind power plant based on the prediction result and the control strategy, and adjusting the arrangement sequence of the fans; And the execution and monitoring module is used for issuing the control strategy and the layout optimization scheme to a fan control system, monitoring the fan state and correcting the control strategy according to the monitoring data.

Description

Intelligent wind field wake loss suppression method based on artificial intelligence Technical Field The invention relates to the technical field of wind power generation, in particular to an intelligent wind field wake loss suppression method based on artificial intelligence. Background The wind power plant is a renewable energy power generation facility for converting wind energy into electric energy by arranging a plurality of wind power generator sets in a certain area. With the development of wind power generation technology, the scale of wind farms is gradually expanding, and a single wind farm typically contains tens to hundreds of fans. The operation of the wind power plant involves a plurality of links such as acquisition of wind energy resources, optimization of fan layout, control of a generator set, management of wind power output and the like. The traditional wind farm wake loss regulation and control mostly adopts an empirical control method, and the problems of low running efficiency and power generation loss of the fans are caused because complex coupling relations among the fans cannot be predicted. Disclosure of Invention In order to make up for the defects, the invention provides an intelligent wind farm wake loss suppression method based on artificial intelligence, which aims to solve the problems of low running efficiency and power generation loss of fans caused by incapability of predicting complex coupling relations among fans because the traditional wind farm wake loss regulation and control mostly adopts an empirical control method. In a first aspect, the present invention provides a method for suppressing wake loss in an intelligent wind farm based on artificial intelligence, comprising the steps of: Collecting the running state and environmental information of each fan in the wind power plant, and carrying out standardized processing on the collected data to form a unified data set; extracting the spatial relationship, the upstream and downstream position relationship and the data characteristics of wind field environment from the data set, and encoding the coupling relationship between the fans; Constructing an artificial intelligent prediction model fusing the space-time characteristics and the fan coupling relation based on the data characteristics, and outputting a prediction result comprising wind speed, turbulence and power states of all fans at downstream positions in the wind power plant by utilizing the artificial intelligent prediction model; Generating an initial fan cooperative control strategy by using a reinforcement learning algorithm, and introducing fan mechanical constraint conditions in the strategy generation process; according to the iteratively updated fan cooperative control strategy, dynamically adjusting the rotating speed and the blade pitch of each fan in the wind power plant; Constraint optimization is carried out on the positions of the fans in the wind power plant based on the prediction result and the control strategy, and the arrangement sequence of the fans is adjusted; And issuing the control strategy and the layout optimization scheme to a fan control system, monitoring the fan state, and correcting the control strategy according to the monitoring data. By adopting the technical scheme, the operation state and the environmental information of each fan in the wind power plant are collected, the collected data are subjected to standardized processing, and then an artificial intelligent prediction model is constructed based on the extracted spatial relationship, the upstream and downstream position relationship and the wind power plant environmental characteristics, so that the wind speed, turbulence and power state of the downstream position of each fan are predicted, and the problems that the operation efficiency of the fans is low and the generated energy is lost due to the fact that complicated coupling relationship between the fans cannot be predicted due to the fact that the conventional wind power plant wake loss regulation and control mostly adopts an empirical control method are solved. Further, the normalizing the collected data includes: interpolation or supplementation is carried out on the missing value; detecting and eliminating abnormal values; Normalizing or standardizing the characteristic data; carrying out alignment and resampling on the time series data; a unified dataset is constructed for subsequent feature extraction and prediction. Further, the extracting the data features of the spatial relationship, the upstream and downstream position relationship and the wind field environment between the fans includes: Acquiring geographic position data of a fan; Calculating horizontal distance, vertical distance and relative azimuth angle among fans; Judging the upstream and downstream position relation of the fan according to the wind direction information; extracting wind field environmental characteristics including wind speed, wind