A wind farm group yawing cooperative control system based on multi-source data fusion

Abstract

The application discloses a kind of wind farm group yaw cooperative control systems based on multi-source data fusion, it is related to wind farm operation control technical field, including multi-source wind farm data acquisition module, to SCADA operation data, cabin wind measurement data, laser radar wind measurement data and weather tower wind measurement data executes time alignment and space mapping processing, generates unified time window wind farm state candidate data;Wind farm state fusion construction module, to unified time window wind farm state candidate data executes multi-source consistency difference calculation, and introduces multi-source credible contraction coefficient, executes multi-source consistency constraint credible contraction fusion algorithm processing, generates unified wind farm state data.In the application, by wind farm state fusion construction module, the wind measurement data of different sources is executed multi-source consistency constraint fusion processing, and unified wind farm state data is generated, so that more accurate, stable wind farm inflow state information can be obtained in complex wind field environment, and the overall operation state perception accuracy of wind farm is improved.

Description

Technical Field

[0001] This invention relates to the field of wind farm operation control technology, and in particular to a wind farm group yaw collaborative control system based on multi-source data fusion. Background Technology

[0002] As large-scale wind farms continue to expand, the number of wind turbines within them has increased significantly. Wake interference between turbines has gradually become a crucial factor affecting the overall power generation efficiency of wind farms. When an upstream turbine is operating, its rotor extracts energy from the incoming wind speed, creating a wake region behind it characterized by velocity attenuation and enhanced turbulence. If a downstream turbine is located within this wake region, it will be significantly affected by wind speed attenuation and incoming flow disturbances, leading to a decrease in power generation and increased load fluctuations.

[0003] Current wind farm operation control typically employs a single-unit control strategy, with each turbine independently executing yaw control based on its measured wind speed, direction, and power status. This lack of systematic analysis of the overall wake propagation relationship of the wind farm makes it difficult to identify the wake propagation paths and impact levels between turbines, resulting in insufficient coordination among multiple units. While some existing technologies introduce overall wind farm control methods, they often rely on single measurement sources such as nacelle or meteorological tower data. In complex wind farm environments or where measurement errors exist, these methods struggle to accurately reflect the true inflow state, affecting the accuracy of wind farm state modeling and control decisions. Furthermore, existing methods often estimate wake relationships through fixed distances or empirical models, lacking a wake coupling relationship modeling mechanism based on real-time operational data and an organizational method for automatically forming turbine cooperative groups based on wake propagation relationships. Consequently, it is difficult to achieve group-oriented coordinated yaw control for wind farms. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a wind farm group yaw coordination control system based on multi-source data fusion.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a wind farm group yaw collaborative control system based on multi-source data fusion, comprising: a multi-source wind farm data acquisition module, which performs time alignment and spatial mapping processing on SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data to generate candidate wind farm state data for a unified time window; a wind farm state fusion construction module, which performs multi-source consistency difference calculation on the candidate wind farm state data for the unified time window, and introduces a multi-source reliable contraction coefficient to perform a multi-source consistency constraint reliable contraction fusion algorithm to generate unified wind farm state data; a wake propagation relationship generation module, which generates wind turbine wake propagation relationship data based on the unified wind farm state data and unit spatial location data; and a wake coupling relationship construction module, which generates a power deviation feedback coefficient based on the unified wind farm state data and wind turbine power output data, and introduces a wake propagation reconstruction system. The system employs a multi-source reliable wake coupling reconstruction algorithm to process the wake propagation relationship data of the wind turbines, generating wake coupling relationship matrix data. The unit group generation module performs sorting, propagation chain identification, and continuous propagation aggregation processing on the wake coupling relationship matrix data, generating unit collaborative group data through a wake influence propagation chain-driven group generation algorithm. The group flow field state reconstruction module performs local flow field reconstruction and power deviation correction processing on the unit collaborative group data, generating local flow field state data. The yaw collaborative decision module generates group yaw decision data based on the unit collaborative group data, local flow field state data, wind turbine power output data, and yaw state data. The control update module generates group yaw collaborative control commands based on the group yaw decision data and performs rolling update processing on the wake coupling relationship matrix data and the unit collaborative group data, forming the group yaw collaborative control result.

[0006] As a further description of the above technical solution: The multi-source wind field data acquisition module receives SCADA operation data, nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data. Simultaneously, it acquires unit spatial location data, turbine power output data, and yaw status data. It reads the corresponding time stamp information from the SCADA operation data, nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data according to the acquisition time, and performs time sorting processing on each type of data based on the time stamp information to generate multi-source time series data. The multi-source time series data is divided into time windows according to the control cycle, and time alignment processing is performed on the SCADA operation data, nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data within each time window to form a multi-source wind field data set for each time window. It also reads unit spatial location data and determines the position of each wind turbine based on the unit spatial location data. Spatial coordinate information is used to perform spatial mapping processing on the nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data in the multi-source wind field data set within the time window. This establishes a correlation between various types of anemometer data and the spatial coordinates of the corresponding wind turbines, generating unit-related anemometer data. The unit-related anemometer data is then processed by wind turbine number to form a multi-source data set for the same wind turbine, consisting of SCADA operation data, nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data. Finally, a unified data structure is used to organize the multi-source data set, enabling each wind turbine to form a unified time window unit data set containing SCADA operation data, nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data within the same time window. Based on this unified time window unit data set, candidate wind field status data for the unified time window is generated.

[0007] As a further description of the above technical solution: The wind field status fusion construction module receives candidate wind field status data from a unified time window and reads the nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data corresponding to each wind turbine from this data. It performs data integrity checks on the nacelle, lidar, and meteorological tower anemometer data, filtering out data items with missing or abnormal records to generate a valid anemometer data set. The valid anemometer data set is then aggregated according to the wind turbine number, forming a multi-source anemometer data set for the same wind turbine. Finally, multi-source consistency difference calculations are performed on the multi-source anemometer data set, analyzing the wind speed and direction differences between the nacelle, lidar, and meteorological tower anemometer data. The differences are compared and processed to generate consistent anemometer data for each turbine. Based on this data, a multi-source consistency constraint credible contraction fusion algorithm is applied. Multi-source consistency constraints are established, and a multi-source credible contraction coefficient is introduced. Credible contraction processing is then applied to the nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data, ensuring that data with smaller consistency differences maintain a higher participation level while reducing the participation level of data with larger consistency differences, thus generating fused anemometer data for each turbine. This fused anemometer data is then reorganized according to the turbine number, creating a turbine inflow state data set for each turbine that includes fused wind speed and fused wind direction data. Finally, unified wind field state data is generated based on this turbine inflow state data set.

[0008] As a further description of the above technical solution: The wake propagation relationship generation module receives unified wind field status data and reads the fused wind speed and fused wind direction data corresponding to each wind turbine from the unified wind field status data. It performs direction conversion processing on the fused wind direction data and wind speed matching processing on the fused wind speed data to generate the incoming flow direction vector corresponding to each wind turbine. It also reads the unit spatial location data and determines the spatial coordinate information of each wind turbine based on the unit spatial location data. It performs pairwise combination processing on all wind turbines according to their numbers to generate wind turbine pairing data. For each pair of wind turbines in the wind turbine pairing data, it calculates the inter-unit line displacement data based on the corresponding spatial coordinate information, and uses the incoming flow direction vector corresponding to one of the wind turbines as the propagation direction reference. The displacement data of the connection between units is subjected to directional projection processing to generate the corresponding longitudinal projection distance. Lateral decomposition processing is performed based on the displacement data of the connection between units and the longitudinal projection distance to generate the corresponding lateral offset distance. Downstream propagation direction determination processing is performed on the longitudinal projection distance; fan pairs whose longitudinal projection distances meet the downstream propagation direction conditions are identified as valid propagation fan pairs, and the remaining fan pairs are removed, generating a set of valid propagation fan pairs. For each fan pair in the set of valid propagation fan pairs, association organization processing is performed according to the upstream fan number, downstream fan number, corresponding incoming flow direction vector, longitudinal projection distance, and lateral offset distance to generate fan wake propagation relationship data.

[0009] As a further description of the above technical solution: The wake coupling relationship construction module receives wake propagation relationship data from wind turbines, unified wind field status data, and wind turbine power output data. It reads the upstream and downstream turbine numbers, incoming flow direction vectors, longitudinal projection distances, and lateral offset distances corresponding to each effective propagating wind turbine pairing from the wake propagation relationship data. It also reads the wind speed data corresponding to each wind turbine from the unified wind field status data and the real-time power data corresponding to each wind turbine from the wind turbine power output data. Based on the wind speed data, it performs theoretical power reference calculations on each wind turbine and performs association processing between the wind speed data and the corresponding power reference results according to the wind turbine number, generating... The theoretical power reference data for each wind turbine is generated; the theoretical power reference data and real-time power data are matched according to the wind turbine number, and the difference between the theoretical power reference data and real-time power data is compared and calculated to generate power deviation data for each wind turbine; the power deviation data is time-collected according to the control time window, and the power deviation data of the same wind turbine within the control time window is statistically processed to generate power deviation statistics data, and then feedback mapping is performed based on the power deviation statistics data to generate power deviation feedback coefficients; wake coupling reconstruction processing is performed on each pair of effectively propagating wind turbines, based on the longitudinal... Propagation distance organization processing is performed on the projected distance, offset distance organization processing is performed on the lateral offset distance, and the propagation distance organization results are correlated with the offset distance organization results to generate propagation spatial relationship data for each effective propagation fan pair. Wake propagation reconstruction coefficients are introduced based on the propagation spatial relationship data, and the wake propagation reconstruction coefficients are correlated with the incoming flow direction vectors corresponding to each effective propagation fan pair to generate propagation reconstruction results for each effective propagation fan pair. Based on the propagation reconstruction results, basic coupling strength construction processing is performed on each effective propagation fan pair to form corresponding basic wake coupling strength data. The power deviation feedback coefficients corresponding to each downstream fan are written into their corresponding basic wake coupling strength data, and feedback correction processing is performed on the basic wake coupling strength data to generate feedback-corrected wake coupling strength data. The feedback-corrected wake coupling strength data is mapped to the upstream and downstream fan numbers using a matrix position mapping process. The feedback-corrected wake coupling strength data for each effective propagation fan pair is written into the corresponding matrix position, and zero-value filling processing is performed on the matrix positions corresponding to fan pairs that do not form effective propagation relationships to generate wake coupling relationship matrix data.

[0010] As a further description of the above technical solution: The unit group generation module receives wake coupling relationship matrix data and reads the row index, column index, and wake coupling intensity data corresponding to each matrix position. It establishes a wind turbine association mapping relationship based on the turbine numbers corresponding to the row and column indices, generating wind turbine association matrix data. It then sorts the wake coupling intensity data in each row of the wind turbine association matrix data according to intensity, generating wake influence intensity ranking data for each upstream wind turbine. Finally, it extracts valid associated wind turbine pairs based on the upstream and downstream wind turbine numbers corresponding to non-zero matrix positions in the wake coupling intensity data, generating wake association pairing data. The wake propagation chain-driven group generation algorithm is applied to the wake association pairing data. Based on the wake influence intensity ranking data, the priority propagation direction corresponding to each upstream turbine is determined. A head-to-tail connection identification process is performed based on the upstream and downstream turbine numbers in the wake association pairing data. Pairs of turbines with matching downstream turbine numbers in the previous pairing are connected to form wake influence propagation chains, generating wake influence propagation chain data. Each wake influence propagation chain in the wake influence propagation chain data is then traversed machine-by-machine according to the propagation order, with the starting point of each wake influence propagation chain... The starting fan is used as the initial aggregation unit, and the next fan adjacent to the initial aggregation unit is used as the current fan to be aggregated. The wake coupling intensity data between the current fan to be aggregated and the initial aggregation unit is read to generate the current wake coupling intensity increment data. The mean coupling value within the current group is calculated based on the wake coupling intensity data between the fans that have entered the same aggregation unit, and the current wake coupling intensity increment data is compared with the mean coupling value within the current group. When the current wake coupling intensity increment data meets the continuous propagation aggregation condition, the current fan to be aggregated is incorporated into the current aggregation unit, and the wake influence propagation chain is read again. The next fan in the chain is designated as the new current fan to be aggregated. When the current wake coupling intensity increment data does not meet the continuous propagation aggregation condition, the current aggregation unit is determined as a unit coordination group, and the next fan in the wake influence propagation chain that has not entered the current aggregation unit is determined as the new initial aggregation unit. The machine-by-machine traversal process continues until all wake influence propagation chains have completed continuous propagation aggregation processing, generating group division result data. The group division result data is then processed according to the group number, the set of fan numbers within the group, and the propagation order of each fan in the wake influence propagation chain to generate unit coordination group data.

[0011] As a further description of the above technical solution: The group flow field state reconstruction module receives unit collaborative group data, unified wind field state data, and turbine power output data. It reads the group number, turbine number set within the group, and propagation order of each turbine in the wake influence propagation chain from the unit collaborative group data. It also reads the unit incoming flow state data corresponding to each turbine within the group from the unified wind field state data and the real-time power data corresponding to each turbine within the group from the turbine power output data. The module performs grouping and aggregation processing on the unit incoming flow state data corresponding to each turbine within each group according to the group number, and performs sequential arrangement processing according to the propagation order of each turbine in the wake influence propagation chain to generate group incoming flow state sequence data. It performs intra-group correlation analysis processing on the fused wind speed and fused wind direction data corresponding to each turbine in the group incoming flow state sequence data to determine the incoming flow transmission relationship between upstream and downstream turbines within the same group, generating intra-group incoming flow correlation data. Based on the intra-group incoming flow correlation data... The group's local flow field is reconstructed by performing local flow field reconstruction processing on the incoming flow state sequence data of the group. The incoming flow state data of the upstream and downstream fans are continuously transmitted and organized according to the propagation order to generate the group's local flow field state. The real-time power data of each fan in the group is aggregated within the group and matched according to the propagation order of each fan in the wake influence propagation chain to generate group power distribution data. Power deviation correction processing is performed on the group's local flow field state based on the group power distribution data. Correction and update processing is performed on the local incoming flow distribution parts in the group's local flow field state that are inconsistent with the changes in real-time power data to generate group corrected flow field state data. The group corrected flow field state data is associated and organized according to the group number, the set of fan numbers in the group, the propagation order of each fan in the wake influence propagation chain, and the local incoming flow distribution results after power deviation correction to generate group local flow field state data.

[0012] As a further description of the above technical solution: The yaw collaborative decision-making module receives unit collaborative group data, group local flow field state data, turbine power output data, and yaw state data. It reads the group number, turbine number set within the group, and propagation order of each turbine in the wake influence propagation chain from the unit collaborative group data. It also reads the local inflow distribution results for each group from the group local flow field state data, the real-time power data for each turbine within the group from the turbine power output data, and the current yaw state information for each turbine within the group from the yaw state data. The module then performs grouping and aggregation processing on the local inflow distribution results, real-time power data, and current yaw state information for each turbine within the group according to the group number, and performs sequential sorting processing based on the propagation order of each turbine in the wake influence propagation chain to generate group decision input data. Finally, it performs group yaw benefit calculation processing based on the local inflow distribution results and real-time power data in the group decision input data to determine the yaw benefit of each turbine within each group. The system analyzes the power changes of turbines under different yaw adjustment states to generate group yaw benefit data. Based on the current yaw state information in the group decision input data, it performs yaw change analysis to determine the yaw adjustment magnitude and order for each turbine within each group, generating yaw execution cost data. Based on the local inflow distribution results in the group decision input data, it performs group inflow fluctuation analysis to determine the continuity and propagation stability of inflow changes for each turbine within each group, generating yaw stability data. Based on the group local flow field state data, it performs group power response calculations for each turbine within each group to determine the power response changes of each group under the current local inflow distribution conditions, generating group power response results. Based on the group power response results, it performs joint scoring calculations on the group yaw benefit data, yaw execution cost data, and yaw stability data to determine the yaw adjustment priority and group yaw coordination direction for each turbine within each group, generating group yaw decision data.

[0013] As a further description of the above technical solution: The control update module receives group yaw decision data, unified wind farm status data, turbine power output data, wake coupling matrix data, and unit coordination group data. It reads the group number, turbine number set within the group, yaw adjustment priority for each turbine, and group yaw coordination direction from the group yaw decision data. From the unified wind farm status data, it reads the turbine inflow status data for each turbine. From the turbine power output data, it reads the real-time power data for each turbine. From the wake coupling matrix data, it reads the wake coupling strength data between each turbine. From the unit coordination group data, it reads the corresponding data for each group. The data is divided into groups; the yaw adjustment priority, group yaw coordination direction, turbine inflow status data, and real-time power data for each group are grouped and aggregated according to group number, and corresponding association processing is performed based on the turbine number set within the group to generate group control input data; based on the group control input data, the yaw control sequence of turbines within each group is organized and processed to determine the control execution order of each turbine, and the yaw adjustment direction and yaw adjustment amount for each turbine are determined in conjunction with the group yaw coordination direction to generate group yaw coordination control commands; the group yaw coordination control commands are mapped according to group number and turbine number. The process involves assigning a unique yaw adjustment direction, yaw adjustment amount, and control execution sequence to each wind turbine, and outputting the mapped group yaw coordinated control commands to the corresponding wind turbine's yaw control execution terminal, forming the group yaw coordinated control command set for the current control cycle. At the end of the current control cycle, updated unified wind farm status data and wind turbine power output data are read and compared with the unified wind farm status data and wind turbine power output data corresponding to the current control cycle, generating post-control state change data. Based on the post-control state change data, the wake coupling relationship moment is... The matrix data undergoes rolling update processing, updating the matrix positions where wake coupling strength changes to generate updated wake coupling matrix data. Based on post-control state change data, the unit coordination group data undergoes rolling update processing, updating the group numbers, turbine number sets within the group, and the propagation order of each turbine in the wake influence propagation chain where group division results change to generate updated unit coordination group data. The updated wake coupling matrix data, unit coordination group data, unified wind farm state data, and turbine power output data undergo periodic correlation organization processing to form the group yaw coordination control result.

[0014] As a further description of the above technical solution: The multi-source wind field data acquisition module receives SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data. It also collects wind turbine power output data and yaw status data. After time alignment and spatial mapping processing, it generates unified time window wind field status candidate data and outputs the unified time window wind field status candidate data to the wind field status fusion construction module. The wind turbine power output data is transmitted to the wake coupling relationship construction module, the group flow field status reconstruction module, the yaw collaborative decision module, and the control update module. The yaw status data is transmitted to the yaw collaborative decision module. The wind field state fusion construction module receives candidate wind field state data with a unified time window, processes it through multi-source consistency difference calculation and multi-source consistency constraint reliable contraction fusion algorithm to generate unified wind field state data, and outputs the unified wind field state data to the wake propagation relationship generation module, the wake coupling relationship construction module, the group flow field state reconstruction module and the control update module respectively. The wake propagation relationship generation module receives unified wind field state data, and generates wind turbine wake propagation relationship data through incoming flow direction vector extraction, longitudinal projection distance calculation, lateral offset distance calculation, and downstream propagation direction determination. The wind turbine wake propagation relationship data is then output to the wake coupling relationship construction module. The wake coupling relationship construction module receives wind turbine wake propagation relationship data, unified wind field status data, and wind turbine power output data. After processing through theoretical power reference calculation, power deviation calculation, power deviation statistical processing, multi-source reliable wake coupling reconstruction algorithm processing, and matrix position mapping, it generates wake coupling relationship matrix data and outputs the wake coupling relationship matrix data to the unit group generation module and the control update module respectively. The unit group generation module receives the wake coupling relationship matrix data, and generates unit cooperative group data through wake influence intensity sorting, wake influence propagation chain identification, continuous propagation aggregation, and wake propagation chain driven group generation algorithm. The unit cooperative group data is then output to the group flow field state reconstruction module, yaw cooperative decision module, and control update module, respectively. The group flow field state reconstruction module receives the unit collaborative group data, unified wind field state data, and wind turbine power output data. After group local flow field reconstruction processing and power deviation correction processing, it generates group local flow field state data and outputs the group local flow field state data to the yaw collaborative decision module. The yaw collaborative decision module receives group data of unit collaboration, group local flow field state data, wind turbine power output data and yaw state data. After processing the group yaw benefit calculation, yaw execution cost calculation, yaw stability data calculation, group power response calculation and joint scoring, it generates group yaw decision data and outputs the group yaw decision data to the control update module. The control update module receives group yaw decision data, unified wind farm status data, wind turbine power output data, wake coupling matrix data, and unit coordination group data. After yaw control sequence organization and processing, it generates group yaw coordinated control commands. After wake coupling matrix data rolling update processing and unit coordination group data rolling update processing, it forms continuously updated group yaw coordinated control results.

[0015] The present invention has the following beneficial effects: 1. In this invention, firstly, a wind farm state fusion construction module performs multi-source consistency constraint fusion processing on wind measurement data from different sources to generate unified wind farm state data. This enables more accurate and stable wind farm inflow state information to be obtained in complex wind farm environments, improving the overall operational status perception accuracy of the wind farm. Secondly, a wake propagation relationship generation module calculates the spatial relationship and inflow direction between wind turbines. A wake coupling relationship construction module, combined with wind turbine power output data, establishes a wake coupling relationship matrix, allowing the wake propagation relationship and wake influence intensity between wind turbines in the wind farm to be expressed in a structured manner. This improves the ability to identify wake propagation relationships and the accuracy of wake coupling relationship modeling. Thirdly, a unit group generation module identifies wake influence propagation chains based on the wake coupling relationship matrix and uses the wake propagation chain to drive a group generation algorithm to form unit collaborative groups. This enables wind turbines with obvious wake correlation relationships in the wind farm to automatically form collaborative operating units, thereby improving the wind turbine group organization capability and enabling subsequent control decisions to be optimized collaboratively on a group basis.

[0016] 2. In this invention, the flow field state reconstruction module reconstructs the local flow field state within the group and performs power deviation correction processing in conjunction with the real-time power data of the wind turbine, so that the propagation relationship of the incoming flow within the group can more accurately reflect the actual operating state, thereby improving the accuracy of the group flow field state modeling. The yaw collaborative decision module performs joint scoring calculation processing on the group yaw benefit data, yaw execution cost data and yaw stability data to generate group yaw decision data, and generates group yaw collaborative control commands through the control update module, so that the wind farm can achieve group-level yaw collaborative control while ensuring operational stability, thereby improving the overall power generation efficiency of the wind farm and reducing the degree of wake influence between wind turbines. Attached Figure Description

[0017] Figure 1 This is a system architecture diagram of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Reference Figure 1 This invention provides an embodiment of a wind farm group yaw collaborative control system based on multi-source data fusion, comprising: a multi-source wind farm data acquisition module, which performs time alignment and spatial mapping processing on SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data to generate unified time window wind farm state candidate data; a wind farm state fusion construction module, which performs multi-source consistency difference calculation on the unified time window wind farm state candidate data, and introduces a multi-source reliable contraction coefficient to perform a multi-source consistency constraint reliable contraction fusion algorithm to generate unified wind farm state data; a wake propagation relationship generation module, which generates wind turbine wake propagation relationship data based on the unified wind farm state data and unit spatial location data; and a wake coupling relationship construction module, which generates a power deviation feedback coefficient based on the unified wind farm state data and wind turbine power output data, and introduces a wake propagation reconstruction coefficient to perform wind... The wake propagation relationship data is processed by a multi-source reliable wake coupling reconstruction algorithm to generate wake coupling relationship matrix data. The unit group generation module performs sorting, propagation chain identification, and continuous propagation aggregation processing on the wake coupling relationship matrix data, and generates unit collaborative group data through a wake influence propagation chain driven group generation algorithm. The group flow field state reconstruction module performs group local flow field reconstruction and power deviation correction processing on the unit collaborative group data to generate group local flow field state data. The yaw collaborative decision module generates group yaw decision data based on the unit collaborative group data, group local flow field state data, turbine power output data, and yaw state data. The control update module generates group yaw collaborative control commands based on the group yaw decision data, and performs rolling update processing on the wake coupling relationship matrix data and the unit collaborative group data to form the group yaw collaborative control result.

[0020] In this embodiment, the multi-source wind field data acquisition module continuously receives SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data, and simultaneously acquires unit spatial location data, wind turbine power output data, and yaw status data; it reads the corresponding time stamp information for each of the SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data; and performs time sorting processing on the SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data according to the time stamp information, so that all types of data are arranged in chronological order on a unified time axis, generating multi-source time series data.

[0021] The multi-source time series data is divided into time windows according to the control cycle; within each time window, the corresponding time period's SCADA operation data, cabin wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data are read; based on the time stamp information, the SCADA operation data, cabin wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data within the same time window are time-aligned; the time-aligned SCADA operation data, cabin wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data are organized into a time window multi-source wind field data set.

[0022] Read the spatial location data of the wind turbine units and determine the spatial coordinate information corresponding to each wind turbine based on the spatial location data of the wind turbine units; read the nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data from the multi-source wind field data set of the time window; establish a correspondence between the nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data and the corresponding wind turbine spatial coordinate information; perform spatial mapping processing on the nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data based on the wind turbine spatial coordinate information to generate wind measurement data associated with the wind turbine units.

[0023] Data aggregation processing is performed on the wind measurement data associated with the wind turbines according to their turbine numbers. During the aggregation process, the SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data corresponding to the same wind turbine are organized and processed according to the wind turbine numbers. This ensures that the same wind turbine forms a unified data set structure within the same time window. A multi-source data set for the wind turbines is generated based on the aggregation results.

[0024] Perform unified data structure organization processing on the multi-source data set of the unit; perform structural arrangement processing on the SCADA operation data, nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data of each wind turbine in the same time window according to the unified data structure; so that each wind turbine forms a unified data organization structure in the same time window; generate a unified time window unit data set based on the structural organization result.

[0025] Read the unified time window unit data set; perform overall organization and processing on SCADA operation data, nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data based on the unified time window unit data set; generate candidate data for wind field status of the unified time window based on the organization and processing results.

[0026] In this embodiment, after receiving the candidate wind field status data for a unified time window, the wind field status fusion construction module first performs data reading processing on the candidate wind field status data for the unified time window. The wind field status fusion construction module sequentially reads the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data corresponding to each wind turbine from the candidate wind field status data for the unified time window, so that the wind measurement data from different sources can form a corresponding data reading structure within the unified time window.

[0027] After data reading is complete, the wind field status fusion construction module performs data integrity checks on the acquired nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data. During the check, each data record corresponding to each wind turbine undergoes integrity verification; data items with missing records are removed, and data items with abnormal records are filtered out, ensuring that the remaining data are all eligible for subsequent fusion calculations, thus generating a valid anemometer data set. The valid anemometer data set includes the nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data that have passed the integrity check.

[0028] After the effective wind measurement data set is formed, the wind field status fusion construction module performs data aggregation processing on the effective wind measurement data set according to the wind turbine number. During the aggregation process, the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data corresponding to the same wind turbine are uniformly organized according to the wind turbine number, so that the same wind turbine forms a corresponding data set structure within a unified time window, thereby generating a multi-source wind measurement data set for the unit. The multi-source wind measurement data set for the unit reflects the data combination relationship of different wind measurement sources for the same wind turbine within the same time window.

[0029] After the multi-source wind measurement data set of the unit is formed, the wind field state fusion construction module performs multi-source consistency difference calculation processing on the multi-source wind measurement data set. In this process, the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data in the multi-source wind measurement data set are compared separately. By calculating the differences in wind speed and wind direction between different wind measurement sources, the degree of difference between the wind measurement results from different sources is quantified, thereby generating unit wind measurement consistency difference data. Unit wind measurement consistency difference data is used to reflect the degree of consistency of wind measurements from different sources at the same wind turbine. The formula for calculating unit wind measurement consistency difference is as follows: ; ; ; The results of the difference between cabin wind measurement data and lidar wind measurement data. The results of the discrepancy between cabin wind measurement data and meteorological tower wind measurement data. The results of the differences between lidar wind measurement data and meteorological tower wind measurement data. Wind speed data from the cabin wind measurement data. Wind speed data from lidar wind measurement data. Wind speed data from meteorological tower anemometers. Wind direction data in the cabin wind measurement data. Wind direction data in lidar wind measurement data, Wind direction data from meteorological tower wind measurement data.

[0030] The wind field status fusion construction module reads the wind measurement consistency difference data of each wind turbine according to its turbine number, and extracts the difference results between the nacelle wind measurement data and the lidar wind measurement data, the difference results between the nacelle wind measurement data and the meteorological tower wind measurement data, and the difference results between the lidar wind measurement data and the meteorological tower wind measurement data from the unit's wind measurement consistency difference data. By performing corresponding aggregation processing on each difference result, all difference results corresponding to the same wind turbine are formed into a consistent difference set structure. Each difference result is categorized according to the wind measurement data source, so that the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data each correspond to the difference result set they are comparing, and the difference degree statistics corresponding to each wind measurement data source are calculated based on each difference result set. After obtaining the difference degree statistics, the wind field status fusion construction module establishes multi-source consistency constraints based on the difference degree statistics. In the multi-source consistency constraints, wind measurement data sources with smaller difference degree statistics are identified as data sources with higher consistency, and wind measurement data sources with larger difference degree statistics are identified as data sources with lower consistency. Based on the statistical values ​​of the degree of difference corresponding to each wind measurement data source, a multi-source reliable contraction coefficient is determined. This ensures that wind measurement data sources with smaller statistical values ​​of degree of difference correspond to larger multi-source reliable contraction coefficients, and wind measurement data sources with larger statistical values ​​of degree of difference correspond to smaller multi-source reliable contraction coefficients. This generates multi-source reliable contraction coefficients for engine room wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data. The formula for calculating the statistical value of degree of difference is as follows: ; ; ; Statistical values ​​of the degree of difference corresponding to the wind measurement data in the engine room. Statistical values ​​of the degree of difference corresponding to lidar wind measurement data. Statistical values ​​of the degree of difference corresponding to the wind measurement data from the weather tower. The results of the difference between cabin wind measurement data and lidar wind measurement data. The results of the discrepancy between cabin wind measurement data and meteorological tower wind measurement data. The results show the discrepancy between lidar wind measurement data and meteorological tower wind measurement data. The formula for calculating the multi-source reliable contraction coefficient is as follows: ; ; ; : Multi-source reliable contraction coefficient corresponding to nacelle wind measurement data : Multi-source reliable contraction coefficient corresponding to lidar wind measurement data The multi-source reliable contraction coefficient corresponding to the wind measurement data from the weather tower. Statistical values ​​of the degree of difference corresponding to the wind measurement data in the engine room. Statistical values ​​of the degree of difference corresponding to lidar wind measurement data. : Statistical values ​​of the degree of difference corresponding to the wind measurement data from the meteorological tower.

[0031] After generating the unit wind measurement consistency difference data, the wind field state fusion construction module executes a multi-source consistency constraint reliable contraction fusion algorithm based on the unit wind measurement consistency difference data. In this process, multi-source consistency constraints are established based on the unit wind measurement consistency difference data, and a multi-source reliable contraction coefficient is introduced based on these constraints. This ensures that the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data participate in the fusion calculation according to their consistency level. Data with smaller consistency differences maintain a higher participation level during the fusion process, while data with larger consistency differences have a lower participation level. By performing reliable contraction processing on various types of wind measurement data, wind measurement data from different sources are unified under a unified fusion rule to form a unified result, thereby generating the unit fused wind measurement data. The unit fused wind measurement data reflects the fused wind measurement result obtained by integrating multi-source wind measurement information at the same wind turbine. The fused wind speed calculation formula is as follows: ; Integrating wind speed data, : Multi-source reliable contraction coefficient corresponding to nacelle wind measurement data : Multi-source reliable contraction coefficient corresponding to lidar wind measurement data The multi-source reliable contraction coefficient corresponding to the wind measurement data from the weather tower. Wind speed data from the cabin wind measurement data. Wind speed data from lidar wind measurement data. Wind speed data from meteorological tower anemometers. Combined with the wind direction calculation formula: ; Integrating wind direction data, : Multi-source reliable contraction coefficient corresponding to nacelle wind measurement data : Multi-source reliable contraction coefficient corresponding to lidar wind measurement data The multi-source reliable contraction coefficient corresponding to the wind measurement data from the weather tower. Wind direction data in the cabin wind measurement data. Wind direction data in lidar wind measurement data, Wind direction data from meteorological tower wind measurement data.

[0032] After the fused wind measurement data is generated, the wind farm status fusion construction module performs data recombination processing on the fused wind measurement data according to the wind turbine number. During the data recombination process, the fused wind measurement results corresponding to the same wind turbine are uniformly organized, so that each wind turbine forms a data structure containing fused wind speed data and fused wind direction data, and the data corresponding to each wind turbine forms a set of inflow status data for the unit. The set of inflow status data for the unit reflects the inflow status information of each wind turbine within a unified time window.

[0033] After the incoming flow status data set of the turbines is formed, the wind farm status fusion construction module performs unified organization processing on the data of each turbine based on the incoming flow status data set. This ensures that the incoming flow status data of all turbines forms a data set with a unified structure within a unified time window, thereby generating unified wind farm status data. Unified wind farm status data reflects the unified incoming flow status information of the entire wind farm within the same time window and serves as the data input for the subsequent wake propagation relationship generation module and wake coupling relationship construction module. Unified wind farm status data organization expression: ; Unified wind field status data Fan serial number : No. The fused wind speed data for each wind turbine : No. The integrated wind direction data corresponding to each wind turbine The number of wind turbines in a wind farm.

[0034] In this embodiment, the wake propagation relationship generation module receives unified wind field status data during system operation and reads the fused wind speed data and fused wind direction data corresponding to each wind turbine from the unified wind field status data. During the reading process, the corresponding fused wind speed data and fused wind direction data are read one by one according to the wind turbine number order, so that each wind turbine forms a corresponding data reading structure within the same time window.

[0035] After data reading is complete, the fused wind direction data undergoes direction conversion processing, transforming it into a directional representation of the incoming flow propagation direction, enabling it to be used to represent the wind flow propagation direction. Subsequently, the fused wind speed data undergoes wind speed matching processing. During this process, the fused wind speed data is matched with its corresponding directional representation according to the wind turbine number, creating a unified representation structure between the fused wind speed data and the propagation direction, thereby generating the incoming flow direction vector for each wind turbine.

[0036] After the incoming flow direction vector is generated, the spatial location data of the turbine units is read, and the spatial coordinate information of each turbine is determined based on the spatial location data. Then, all turbines are paired according to their numbers, so that any two turbines form a pair, thereby generating turbine pairing data.

[0037] After the wind turbine pairing data is generated, spatial relationship calculations are performed on each pair of wind turbines in the pairing data. Based on the spatial coordinate information corresponding to the wind turbine pairing, the inter-unit line displacement data between the two wind turbines is calculated. Then, using the incoming flow direction vector corresponding to one of the wind turbines as the propagation direction reference, directional projection processing is performed on the inter-unit line displacement data to calculate the component of the inter-unit line displacement data in the propagation direction, thereby generating the corresponding longitudinal projection distance.

[0038] After the longitudinal projection distance is generated, a lateral decomposition process is performed based on the inter-unit line displacement data and the longitudinal projection distance. The part of the inter-unit line displacement data that does not belong to the propagation direction component is decomposed and calculated, so that the distance relationship of the inter-unit line displacement data in the direction perpendicular to the propagation direction can be expressed, thereby generating the corresponding lateral offset distance.

[0039] After the longitudinal projection distance and lateral offset distance are generated, the downstream propagation direction is determined for the longitudinal projection distance. When the longitudinal projection distance meets the downstream propagation direction condition, the corresponding wind turbine pair is determined as a valid propagation wind turbine pair; when the longitudinal projection distance does not meet the downstream propagation direction condition, the corresponding wind turbine pair is removed from the wind turbine pairing data, thereby generating a set of valid propagation wind turbine pairs.

[0040] After the effective propagation turbine pairing set is formed, an association organization process is performed on each pairing within the set. This process associates the upstream turbine number, downstream turbine number, corresponding incoming flow direction vector, longitudinal projection distance, and lateral offset distance according to a unified data structure, creating a complete wake propagation relationship representation structure for each pairing. Subsequently, a unified organization process is performed on all effective propagation turbine pairings, arranging the wake propagation relationships between each turbine according to a unified data structure, thereby generating turbine wake propagation relationship data.

[0041] The generated wind turbine wake propagation relationship data serves as the data input for the subsequent wake coupling relationship construction module to perform wake coupling relationship calculation and processing.

[0042] In this embodiment, the wake coupling relationship construction module receives wind turbine wake propagation relationship data, unified wind field status data, and wind turbine power output data, and performs data reading according to a unified processing order. From the wind turbine wake propagation relationship data, it reads the upstream wind turbine number, downstream wind turbine number, incoming flow direction vector, longitudinal projection distance, and lateral offset distance corresponding to each effective propagation wind turbine pair in turn, ensuring that each effective propagation wind turbine pair forms complete wake propagation relationship reading content. Simultaneously, it reads the wind speed data corresponding to each wind turbine according to its number from the unified wind field status data, and reads the real-time power data corresponding to each wind turbine according to its number from the wind turbine power output data, enabling the wind speed data and real-time power data to establish a correspondence according to the wind turbine number, providing an input basis for subsequent theoretical power reference calculation and power deviation calculation.

[0043] Wind speed data is read sequentially from each wind turbine according to its number, and a power reference calculation result is generated based on the wind speed data corresponding to the same turbine. Then, the wind speed data and corresponding power reference results are associated and organized according to the turbine number, so that each wind turbine has a unified expression of wind speed data and corresponding power reference results, thereby generating theoretical power reference data for each wind turbine. The theoretical power reference data is used to characterize the power reference level of each wind turbine under the current wind speed conditions and serves as direct input for subsequent power deviation calculations. Theoretical power reference calculation formula: ; : No. The theoretical power reference results for each wind turbine : No. Wind speed data for each fan. : Reference calculation factor for wind turbine power.

[0044] The turbine numbers in the theoretical power reference data are matched one-to-one with the turbine numbers in the real-time power data. Then, the difference between the theoretical power reference data and the real-time power data for the same turbine is compared and calculated, resulting in a corresponding power difference result for each turbine. These power difference results are then organized into power deviation data for each turbine. The power deviation data reflects the difference between the theoretical power reference result and the actual power output result, and serves as input for subsequent statistical processing and feedback mapping. Power deviation calculation formula: ; : No. Power deviation data for each wind turbine : No. Real-time power data for each wind turbine : No. Theoretical power reference data for each wind turbine.

[0045] The control time window, using the control cycle as its time scale, is a continuous time interval used by the system during operation to perform data statistics, status assessment, and control decisions. Based on the control time window, power deviation data for the same fan are time-merged, resulting in continuous power deviation changes for the same fan within the same control time window. Statistical processing is then performed on the power deviation data for the same fan within the control time window, generating statistical results of power deviation changes and producing power deviation statistics. Feedback mapping is then performed on the power deviation statistics to establish a mapping relationship between the statistical data and the corresponding feedback results, generating a power deviation feedback coefficient. The power deviation feedback coefficient reflects the degree of power deviation feedback of downstream fans affected by wake within the control time window and participates in the feedback correction processing of subsequent foundation wake coupling strength data. Power deviation statistical calculation formula: ; : No. Statistical data on the power deviation of each fan within the control time window. : No. Each fan in time The corresponding power deviation data, The number of sampling times included within the control time window. Power deviation feedback coefficient calculation formula: ; : No. The power deviation feedback coefficient corresponding to each wind turbine. : No. Statistical data on the power deviation of each fan within the control time window. The maximum absolute value among all the statistical data on power deviation of all wind turbines.

[0046] First, the longitudinal projection distances corresponding to each effective propagation fan pair are read, and propagation distance calculations are performed on the longitudinal projection distances according to the fan pairing order, so that the longitudinal projection distances corresponding to each effective propagation fan pair form stable propagation distance calculation results under a unified processing order. Next, the lateral offset distances corresponding to each effective propagation fan pair are read, and offset distance calculations are performed on the lateral offset distances according to the fan pairing order, so that the lateral offset distances corresponding to each effective propagation fan pair form stable offset distance calculation results under a unified processing order. Then, the longitudinal projection distances and lateral offset distances are correlated, so that the propagation distance relationship and lateral offset relationship of the same effective propagation fan pair are combined on the same correlation basis, generating propagation spatial relationship data. The propagation spatial relationship data reflects the distance relationship and lateral position relationship of each effective propagation fan pair in the wake propagation direction, and serves as the basic input for the wake propagation reconstruction coefficient calculation. Propagation spatial relationship calculation formula: ; Upstream wind turbines Downstream wind turbines The distance of the propagation space relationship between them Upstream wind turbines Downstream wind turbines The longitudinal projection distance between them Upstream wind turbines Downstream wind turbines The lateral offset distance between them.

[0047] The incoming flow direction vector and propagation space relationship data corresponding to the same effective propagation fan pairing are matched group by group. This allows the incoming flow direction information and propagation space relationship to form a correlation calculation result based on the unified fan pairing, thereby generating a wake propagation reconstruction coefficient for each effective propagation fan pairing. The wake propagation reconstruction coefficient is used to characterize the degree of wake propagation reconstruction under the combined effect of the incoming flow direction vector and the propagation space relationship, and serves as the direct basis for generating the propagation reconstruction result. The formula for calculating the wake propagation reconstruction coefficient is as follows: ; Upstream wind turbines Downstream wind turbines The corresponding wake propagation reconstruction coefficient, Upstream wind turbines The corresponding incoming flow direction vector, Upstream wind turbines Pointing to downstream wind turbines The propagation direction vector, : Distance of spatial relationships in propagation.

[0048] The wake propagation reconstruction coefficients formed by each pair of effective propagating wind turbines are written into the propagation spatial relationship data corresponding to that pair. Then, the written propagation spatial relationship data is uniformly organized with the corresponding incoming flow direction vector, so that each pair of effective propagating wind turbines forms a complete organizational result containing the incoming flow direction vector, propagation spatial relationship data, and wake propagation reconstruction coefficients. This generates the propagation reconstruction result. The propagation reconstruction result is used to characterize the wake propagation relationship of each effective propagating wind turbine pair after processing by the multi-source reliable wake coupling reconstruction algorithm.

[0049] The propagation reconstruction results are read group by group according to the effective propagation turbine pairing order. The reconstruction content corresponding to each effective propagation turbine pairing in the propagation reconstruction results is converted into a basic coupling strength expression result, so that each group of effective propagation turbine pairings generates corresponding basic wake coupling strength data. The basic wake coupling strength data is used to characterize the wake coupling relationship strength before the introduction of power deviation feedback correction, and serves as the input for subsequent feedback correction processing. The basic wake coupling strength calculation formula is as follows: ; Upstream wind turbines Downstream wind turbines Basic wake coupling strength data between them Wake propagation reconstruction coefficient : Distance of spatial relationships in propagation.

[0050] Based on the downstream wind turbine number, the power deviation feedback coefficient is mapped one-to-one with the foundation wake coupling strength data. Then, the power deviation feedback coefficient for each downstream wind turbine is written into its corresponding foundation wake coupling strength data, ensuring that the foundation wake coupling strength data simultaneously includes propagation reconstruction results and power deviation feedback results. Next, feedback correction processing is performed on the written foundation wake coupling strength data, allowing the power deviation feedback coefficient to correct the foundation coupling strength, thereby generating feedback-corrected wake coupling strength data. The feedback-corrected wake coupling strength data is used to characterize the wake coupling strength after incorporating the influence of power deviation feedback. The formula for calculating the feedback-corrected wake coupling strength is as follows: ; Feedback corrected wake coupling strength data, Basic wake coupling strength data, Downstream wind turbines The corresponding power deviation feedback coefficient.

[0051] Based on the upstream and downstream turbine numbers, the corresponding positions of each effective propagating turbine pair in the wake coupling relationship matrix are determined. Then, the feedback-corrected wake coupling strength data for each effective propagating turbine pair is written into the corresponding matrix position, ensuring that each effective propagating turbine pair occupies a unique position in the matrix. Subsequently, zero-value filling is performed on the matrix positions corresponding to turbine pairs that do not form an effective propagation relationship, ensuring that all turbine pairs maintain a uniform arrangement in the matrix structure, thereby generating the wake coupling relationship matrix data. The wake coupling relationship matrix data is used to characterize the wake coupling strength distribution relationship among all turbines in the wind farm and serves as input data for the subsequent unit group generation module. The wake coupling relationship matrix expression is as follows: ; Wake coupling matrix data Upstream wind turbines For downstream wind turbines data on wake coupling strength. The number of wind turbines in a wind farm.

[0052] In this embodiment, the unit group generation module receives the wake coupling relationship matrix data and reads the row index, column index, and wake coupling intensity data corresponding to each matrix position in the wake coupling relationship matrix data row by row according to the matrix structure order. Simultaneously, it reads the corresponding wind turbine number information based on the matrix position corresponding to the row and column indices, so that each matrix position in the matrix forms a correspondence between upstream and downstream wind turbine numbers. The wake coupling intensity data and the corresponding wind turbine numbers are then associated and organized, so that all wake coupling intensity data in the matrix structure are paired with corresponding wind turbines to form a unified expression structure, thereby generating the wind turbine association matrix data. Wind turbine association matrix data expression: ; Wake coupling matrix data Upstream wind turbines For downstream wind turbines data on wake coupling strength. : Row index, corresponding to the upstream wind turbine number, : Column index, corresponding to the downstream wind turbine number, Number of wind turbines.

[0053] The wake coupling intensity data in the wind turbine correlation matrix is ​​read row by row in row index order. All wake coupling intensity data in the same row are sorted according to their numerical values, maintaining the correspondence between matrix positions and corresponding downstream wind turbine numbers. Matrix positions with larger wake coupling intensity values ​​are placed first, resulting in a structure where all downstream wind turbines corresponding to the same upstream wind turbine are arranged in descending order of wake coupling intensity. This generates the wake influence intensity ranking data for each upstream wind turbine. The formula for ranking the wake influence intensity for the same upstream wind turbine is as follows: ; : No. After sorting the upstream wind turbines, the first one... The location of the downstream wind turbine number. : No. After sorting the upstream wind turbines, the first one... Data on the wake coupling strength of a downstream wind turbine. Upstream wind turbine number Number of wind turbines.

[0054] The wake coupling relationship matrix data is scanned row by row, and the wake coupling intensity data corresponding to each matrix position is processed for non-zero identification. When the wake coupling intensity data is not zero, the upstream and downstream fan numbers corresponding to that matrix position are read, and this fan combination is identified as a fan pair with wake influence relationship. Simultaneously, all identified fan pairs are uniformly organized according to the matrix reading order, arranging each fan pair according to a consistent data structure, thereby generating wake association pairing data. Wake association pairing identification formula: ; Wake flow association pairing data, Wind turbine pairings that have wake effect relationships Upstream wind turbines For downstream wind turbines data on wake coupling strength. Number of wind turbines.

[0055] The wake propagation chain-driven group generation algorithm is applied to the wake association pairing data. A head-to-tail connection identification process is then performed based on the upstream and downstream turbine numbers in the wake association pairing data. A step-by-step matching process is then performed between the downstream turbine number in the previous pairing and the upstream turbine number in the subsequent pairing. When two numbers match, the corresponding turbine pairing is connected, creating a continuous propagation relationship between the two turbine pairings in the data structure. Simultaneously, number matching is continuously performed, gradually connecting multiple turbine pairings to form a continuous propagation structure, thus generating wake influence propagation chain data. The formula for the wake influence propagation chain connection condition is as follows: Propagation chain expression: ; Wake waves affect propagation chain data. The first in the transmission chain The fan number corresponding to each position in the propagation sequence. A single wake affects the number of wind turbines in the propagation chain. Wake coupling strength data between two adjacent propagation sequence positions.

[0056] Following the propagation order in the wake influence propagation chain data, each wake influence propagation chain is read sequentially. Within each chain, the first fan number in the propagation order is read and designated as the first fan in the initial aggregation unit. Simultaneously, the next fan number in the chain is read according to the propagation order and designated as the current fan to be aggregated. Furthermore, the wake coupling strength data between the current fan to be aggregated and the fans in the initial aggregation unit is read, forming the current wake coupling strength increment data. The formula for calculating the current wake coupling strength increment data is as follows: ; Current fans awaiting aggregation Relative to the current aggregation unit Current wake coupling intensity increment data, : The set of fan numbers in the current aggregation unit The number of fans in the current aggregation unit. : The fan number in the current aggregation unit : The number of the blower to be aggregated. : Fan in the current aggregation unit For the current blower to be aggregated Data on wake coupling strength.

[0057] The algorithm reads the wake coupling intensity data between all wind turbines within the current aggregation unit and performs statistical calculations on this data to form a statistical set. Simultaneously, it performs an average calculation on all wake coupling intensity data within the statistical set to obtain the mean coupling value within the current group. Finally, it compares the incremental wake coupling intensity data with the mean coupling value within the current group. The formula for calculating the mean coupling value within the current group is as follows: ; : The mean of coupling within the current group : The set of fan numbers in the current aggregation unit The set of fan pairs within the current aggregation unit where all fans have wake coupling relationships. The number of wind turbine pairs with wake coupling relationships within the current aggregation unit. : Current fan in the aggregation unit With wind turbine Data on the wake coupling strength between them.

[0058] When the current wake coupling intensity increment data meets the continuous propagation aggregation condition, the fan to be aggregated is added to the current aggregation unit, and its number is written into the fan number set of the current aggregation unit. Simultaneously, the next fan is read according to the propagation order in the wake influence propagation chain as the new current fan to be aggregated. The wake coupling intensity data between this fan and all fans in the current aggregation unit is read again, causing this wake coupling intensity data to regenerate new current wake coupling intensity increment data. At the same time, the calculation and comparison of the coupling mean within the group are repeated. Continuous propagation aggregation condition determination formula: ; Current wake coupling strength increment data, : The mean of coupling within the current group : The number of the blower to be aggregated. : The set of fan numbers in the current aggregation unit.

[0059] When the current wake coupling intensity increment data does not meet the continuous propagation aggregation condition, the set of all turbine numbers in the current aggregation unit is determined as a unit coordination group, and a corresponding group number is assigned to this unit coordination group. Simultaneously, the next turbine in the wake influence propagation chain that has not yet entered the current aggregation unit is determined as the new initial aggregation unit, and the turbine-by-turbine traversal process continues according to the wake influence propagation chain order, ensuring that the turbines in the chain complete the aggregation judgment process sequentially, thereby generating the group partitioning result data. The expression for the group partitioning result data is as follows: ; : No. Group division results data corresponding to each unit's collaborative group. Group ID : No. The propagation order position in the unit coordination group is the [number]. The fan number, : No. The number of wind turbines in a unit coordination group.

[0060] The system reads all group information from the group partitioning results and performs unified association organization processing on the group number, the set of turbine numbers within each group, and the propagation order of each turbine in the wake influence propagation chain. This ensures that all turbine numbers within the same group form a unified arrangement structure according to the propagation order. The system then performs unified structure arrangement processing on all groups, thereby generating unit collaborative group data. Unit collaborative group data expression: ;in, ; Unit collaboration group data, : No. Complete group data corresponding to each group Number of groups Group ID : No. The first group The fan number corresponding to each position in the propagation sequence. : No. The number of wind turbines in each group : No. The propagation order of each wind turbine within a group in the wake influence propagation chain.

[0061] In this embodiment, the group flow field state reconstruction module receives unit collaborative group data, unified wind field state data, and wind turbine power output data. It then reads the group number, the set of wind turbine numbers within the group, and the propagation order of each wind turbine in the wake influence propagation chain from the unit collaborative group data, group by group according to the group number order. It also reads the unit incoming flow state data from the unified wind field state data, turbine by turbine, according to the wind turbine number set within the group. Finally, it reads the real-time power data from the wind turbine power output data, turbine by turbine, according to the same wind turbine number order. This ensures a one-to-one correspondence between the unit incoming flow state data, real-time power data, and propagation order for the same wind turbine number within the same group, and maintains that the propagation order remains unchanged during the reading process.

[0062] The incoming flow status data of the units are grouped and collected according to the group number, and then arranged according to the propagation order of each wind turbine in the wake influence propagation chain. The incoming flow status data of the units with smaller propagation order is written to the beginning of the arrangement sequence, and the incoming flow status data of the units with larger propagation order is written to the end of the arrangement sequence. The correspondence between the wind turbine number and the corresponding unit incoming flow status data is kept unchanged, so that the incoming flow status data of all units in the same group form a sequential arrangement structure along the wake influence propagation chain, generating the group incoming flow status sequence data.

[0063] The system reads the wind turbine records from the incoming flow status sequence data of each group according to the group number, and reads the fused wind speed data and fused wind direction data of adjacent wind turbines according to the propagation order. The fused wind speed data and fused wind direction data of the earlier positions in the propagation order are written to the upstream position, and the fused wind speed data and fused wind direction data of the later positions in the propagation order are written to the downstream position. The data between adjacent positions are processed sequentially to form an incoming flow transmission relationship between adjacent positions in the propagation order. The association processing is continuously performed along the propagation order to form a continuous incoming flow transmission structure among all wind turbines in the same group, generating incoming flow association data within the group.

[0064] According to the group number, the upstream and downstream fan numbers in the incoming flow association data within the group are read one by one, and the unit incoming flow status data at the corresponding position in the group incoming flow status sequence data is read. The unit incoming flow status data corresponding to the upstream fan is written to the downstream fan position along the propagation order, and the propagation order is kept unchanged during the transmission process, so that the unit incoming flow status data corresponding to the upstream fan and the unit incoming flow status data corresponding to the downstream fan form a continuous transmission relationship. The same transmission organization processing is continuously executed along the propagation order, so that all propagation positions in the same group form a continuous incoming flow transmission structure, generating the local flow field state of the group.

[0065] Real-time power data belonging to the same group is read from the wind turbine power output data according to the group number, and aggregation processing is performed according to the order of wind turbine numbers in the set of wind turbine numbers within the group. Position matching processing is performed according to the propagation order of each wind turbine in the wake influence propagation chain, so that each position in the propagation order corresponds to a unique real-time power data, and the one-to-one correspondence between wind turbine number, propagation order and real-time power data is maintained, so that all real-time power data in the same group form an arrangement structure along the propagation order, generating group power distribution data.

[0066] According to the group number, the local incoming flow distribution results in the local flow field state of the group are read one by one, and the real-time power data corresponding to the same propagation sequence position in the group power distribution data is read. The local incoming flow distribution results corresponding to the same fan are matched with the real-time power data. For the positions where the local incoming flow distribution results and the real-time power data change in the same way, the original local incoming flow distribution results are kept unchanged. For the positions where the local incoming flow distribution results and the real-time power data change in different ways, the correction and update processing is performed. The local incoming flow distribution results are corrected position by position along the propagation sequence to make the local incoming flow distribution results and the corresponding real-time power data keep the change relationship consistent, and the group corrected flow field state data is generated.

[0067] The local inflow distribution results in the corrected flow field state data of each group are read one by one according to the group number. The group number is written into the corresponding data structure position, the set of fan numbers in the group is written into the position corresponding to the group number, the propagation order of each fan in the wake influence propagation chain is written into the position consistent with the corresponding fan number, and the local inflow distribution results after power deviation correction are written into the position consistent with the propagation order. This makes the group number, the set of fan numbers in the group, the propagation order, and the local inflow distribution results in the same group form a fixed correspondence. All groups are then uniformly arranged to generate local flow field state data for the group.

[0068] After the local flow field state data of each group is generated, the group number corresponding to each group, the set of fan numbers within the group, the propagation order of each fan in the wake influence propagation chain, and the local incoming flow distribution results after power deviation correction are output in the same data structure. The yaw collaborative decision module can read the local incoming flow distribution results according to the group number, and read the positional relationship of each fan within the group according to the set of fan numbers and propagation order within the group. The data correspondence between the local flow field state data of the group, the unit collaborative group data, the unified wind field state data, and the fan power output data is maintained continuously.

[0069] In this embodiment, the yaw cooperative decision-making module receives unit cooperative group data, group local flow field state data, turbine power output data, and yaw state data. It reads the group number, the set of turbine numbers within the group, and the propagation order of each turbine in the wake influence propagation chain from the unit cooperative group data in the order of group numbers. It also reads the local incoming flow distribution results of the corresponding turbine in the group's local flow field state data according to the turbine number set within the group. Simultaneously, it reads the real-time power data of the corresponding turbine in the turbine power output data according to the same turbine number order, and the current yaw state information of the corresponding turbine in the yaw state data according to the same turbine number order. This ensures a one-to-one correspondence between the local incoming flow distribution results, real-time power data, and current yaw state information for the same turbine number within the same group, while maintaining the propagation order of each turbine in the wake influence propagation chain unchanged during the reading process.

[0070] The local incoming flow distribution results, real-time power data, and current yaw status information are grouped and aggregated according to group number. All local incoming flow distribution results corresponding to the same group number are written into the same group data set, all real-time power data corresponding to the same group number are written into the same group data set, and all current yaw status information corresponding to the same group number are written into the same group data set. At the same time, the above data are arranged according to the propagation order recorded in the unit coordination group data, so that the data corresponding to the wind turbines with earlier propagation order are arranged at the beginning of the sequence, and the data corresponding to the wind turbines with later propagation order are arranged at the end of the sequence. This maintains a fixed correspondence between the wind turbine number and the corresponding local incoming flow distribution results, real-time power data, and current yaw status information, thereby generating group decision input data.

[0071] The local flow distribution results and real-time power data in the group decision input data are read one by one according to the group number. The local flow distribution results and real-time power data corresponding to the same propagation position are read one by one according to the propagation order. The real-time power data of the same wind turbine under the local flow distribution result is used as the current power expression result. Based on the change relationship of the local flow distribution results in the propagation order, the power changes that may be formed under different yaw adjustment states are calculated and processed one by one, so that the power changes of each wind turbine under different yaw adjustment states form a continuous change result. All calculation results are uniformly organized and processed according to the group number to generate group yaw benefit data.

[0072] The current yaw status information is read from the group decision input data group by group number, and the current yaw status information of each wind turbine is read from the propagation sequence. The yaw change is calculated for the current yaw status information of adjacent wind turbines in the propagation sequence, so that the degree of yaw change of each wind turbine in the propagation chain when performing yaw adjustment is expressed. According to the degree of yaw change of each wind turbine, all wind turbines in the same group are arranged in order, so that wind turbines with larger yaw change are placed at the beginning of the sequence and wind turbines with smaller yaw change are placed at the end of the sequence, while keeping the correspondence between wind turbine number and corresponding yaw change unchanged. This determines the yaw adjustment range and yaw adjustment order of each wind turbine. The unified organization processing is performed according to the group number to generate yaw execution cost data.

[0073] The local flow distribution results in the group decision input data are read one by one according to the group number, and the local flow distribution results corresponding to each wind turbine are read one by one according to the propagation order. Continuous change calculation processing is performed on the local flow distribution results between adjacent wind turbines in the propagation order to express the flow change relationship between adjacent wind turbines in the propagation chain. Change continuity identification processing is performed on all wind turbines in the propagation chain to form a corresponding expression result of the continuity of flow change of each wind turbine in the propagation chain. At the same time, stability arrangement processing is performed on the continuity of flow change according to the propagation order to keep the position with high continuity of flow change in a stable arrangement order and form change records for the position with low continuity of flow change, thereby generating yaw stability data.

[0074] The local inflow distribution results in the local flow field state data of each group are read one by one according to the group number, and the local inflow distribution results corresponding to each fan are read one by one according to the propagation order. The local inflow distribution results corresponding to the same propagation position are processed with the local inflow distribution results corresponding to adjacent propagation positions in the group to perform continuous response calculation processing, so that the power response change relationship of the local inflow distribution change in the propagation chain can be expressed. The unified calculation processing is performed on all propagation positions in the same group, so that the group forms a complete power change expression result under the current local inflow distribution conditions, thereby generating the group power response result.

[0075] The power response results of each wind turbine in the group are read sequentially according to the group number. The yaw benefit data, yaw execution cost data, and yaw stability data of the same group are also read. The group power response results are matched with the group yaw benefit data, the group power response results with the yaw execution cost data, and the group power response results with the yaw stability data. A joint scoring calculation is performed on these three types of data to rank the yaw adjustment priorities of each wind turbine within the group. The yaw adjustment execution order of the wind turbines within the group is determined based on the ranking results. Simultaneously, the group yaw coordination direction is determined based on the joint scoring results. The yaw adjustment priorities and group yaw coordination directions of each group are then uniformly organized according to the group number, thereby generating group yaw decision data.

[0076] In this embodiment, the control update module receives group yaw decision data, unified wind farm status data, turbine power output data, wake coupling matrix data, and unit coordination group data during system operation. It reads the group number, turbine number set within the group, yaw adjustment priority for each turbine, and group yaw coordination direction from the group yaw decision data group by group number, following the group number order. During the reading process, the correspondence between the group number and the turbine number set within the group remains unchanged. Simultaneously, it reads the turbine incoming flow status data from the unified wind farm status data machine by machine, following the turbine number order within the turbine number set within the group, and reads the real-time power data from the turbine power output data machine by machine, following the same turbine number order. This ensures that the same turbine number forms a corresponding reading structure of turbine incoming flow status data and real-time power data within the same group. Furthermore, it reads the wake coupling strength data between each turbine item by item from the wake coupling matrix data according to row and column index order, and reads the group division result data corresponding to each group from the unit coordination group data, ensuring that the wake coupling strength data and group division result data form a continuous reading structure within the same reading sequence.

[0077] According to the group number, the read yaw adjustment priority, group yaw coordination direction, unit inflow status data, and real-time power data are grouped and aggregated. During the aggregation process, all yaw adjustment priorities belonging to the same group number are written into the same group data sequence, all group yaw coordination directions belonging to the same group number are written into the same group data sequence, and the unit inflow status data belonging to the same group number are written into the same group data sequence in the order of the turbine numbers. The real-time power data of the corresponding turbine numbers are also written into the same group data sequence in the same order. This ensures that the group number, the set of turbine numbers within the group, the yaw adjustment priority, the group yaw coordination direction, the unit inflow status data, and the real-time power data form a fixed positional correspondence within the same group, and maintains that the correspondence between the turbine numbers and the unit inflow status data and real-time power data remains unchanged, thereby generating group control input data.

[0078] The yaw adjustment priority in the group control input data is read sequentially according to the group number. Then, the corresponding yaw adjustment priority is read sequentially according to the turbine number set in the group. The order of yaw adjustment priorities of all turbines in the same group is compared. The turbine record with higher yaw adjustment priority is moved to the beginning of the sequence, and the turbine record with lower yaw adjustment priority is moved to the end of the sequence. At the same time, the correspondence between the turbine number and the corresponding unit flow status data and real-time power data is kept unchanged, so that all turbines in the same group are arranged in the execution order according to the yaw adjustment priority.

[0079] The system reads the group yaw coordination direction of each wind turbine according to the execution order, and reads the corresponding unit's incoming flow status data and real-time power data according to the turbine number. It performs corresponding matching processing on the changes in the incoming flow status data and the power changes in the real-time power data, so that the power output of the same wind turbine under the current incoming flow status conditions forms a continuous change record. Based on the matched change record, the corresponding yaw adjustment direction is determined for each wind turbine. At the same time, the yaw adjustment amount is calculated for each wind turbine according to the relationship between the changes in the incoming flow status and the power changes, so that each wind turbine has a unique yaw adjustment direction and a unique yaw adjustment amount. The results are written to the corresponding wind turbine number position according to the execution order, thereby generating the group yaw coordination control command.

[0080] The group yaw coordinated control command is mapped according to the group number and the turbine number. During the mapping process, the yaw adjustment direction, yaw adjustment amount and control execution order in the group yaw coordinated control command are written into the position consistent with the corresponding turbine number. The mapped group yaw coordinated control command is then output to the yaw control execution terminal of the corresponding turbine, so that each turbine executes the yaw adjustment operation in sequence according to the control execution order, thereby forming the group yaw coordinated control command set for the current control cycle.

[0081] After the current control cycle ends, the updated unified wind farm status data and updated wind turbine power output data are read. Then, the updated unit inflow status data and updated real-time power data are read one by one according to the wind turbine number. At the same time, the unit inflow status data and real-time power data corresponding to the current control cycle are read. The updated unit inflow status data and the unit inflow status data corresponding to the current control cycle are compared before and after to form a continuous change record of the inflow status change of the same wind turbine between two control cycles. At the same time, the updated real-time power data and the real-time power data corresponding to the current control cycle are compared before and after to form a continuous change record of the power change of the same wind turbine between two control cycles. Based on the above change records, the post-control status change data is generated.

[0082] The matrix positions in the wake coupling relationship matrix data are read row by row in the matrix structure order. The matrix positions where the wake coupling intensity changes are identified based on the post-control state change data. The wake coupling intensity data corresponding to the identified matrix positions is updated so that the updated wake coupling intensity data is rewritten to the corresponding matrix positions. At the same time, the correspondence between the row index, column index and the corresponding wind turbine number is kept unchanged, thereby generating the updated wake coupling relationship matrix data.

[0083] The group division results in the unit collaborative group data are read one by one according to the group number order. Based on the post-control state change data, the group number whose group division structure has changed is identified. The set of wind turbine numbers in the group corresponding to the identified group number is updated. At the same time, the propagation order of each wind turbine in the group is reorganized according to the propagation order in the wake effect propagation chain. The updated group number, the set of wind turbine numbers in the group and the propagation order form a new group division structure, thereby generating the updated unit collaborative group data.

[0084] The updated wake coupling matrix data, updated unit coordination group data, unified wind farm status data, and wind turbine power output data are processed according to a unified data structure. This ensures that all types of data form a unified data arrangement structure within the same control cycle, while maintaining the correspondence between wind turbine numbers, group numbers, and matrix positions. This allows the system to continue performing wake coupling calculations, group divisions, and yaw coordination decision-making in the next control cycle according to the updated data structure, thereby generating continuously updated group yaw coordination control results.

[0085] Specific Implementation Example 1: In this embodiment, the system runs on a wind farm control platform. The system continuously performs data acquisition, wind farm status construction, and wake propagation relationship identification and processing according to the control cycle.

[0086] The system first receives SCADA operation data, nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data from a multi-source wind field data acquisition module. Simultaneously, it acquires unit spatial location data, turbine power output data, and yaw status data. The multi-source wind field data acquisition module then reads the time stamp information from each data type and performs time sorting processing on the SCADA operation data, nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data based on the time stamp information. This ensures that all data forms a continuous arrangement on a unified time axis, thereby generating multi-source time series data.

[0087] After the multi-source time series data is generated, the system performs time window division processing on the multi-source time series data according to the control cycle, forming a time window multi-source wind field data set within each control cycle. Subsequently, the system reads the spatial location data of the wind turbines and determines the spatial coordinate information of each wind turbine. Then, it performs spatial mapping processing on the nacelle anemometer data, lidar anemometer data, and meteorological tower anemometer data in the time window multi-source wind field data set, establishing a correspondence between various anemometer data and the corresponding wind turbine spatial coordinates, thereby generating unit-related anemometer data.

[0088] Subsequently, the system performs data aggregation processing on the wind measurement data associated with the wind turbines according to the turbine number, so that the SCADA operation data, nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data corresponding to the same wind turbine form a multi-source data set of the unit. The system then performs unified data structure organization processing on the multi-source data set of the unit to generate a unified time window unit data set, and generates candidate data of wind field status for a unified time window based on this.

[0089] After receiving candidate wind field status data for a unified time window, the wind field status fusion construction module reads the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data corresponding to each wind turbine from the data, and performs data integrity checks. Data items with missing or abnormal records are removed, thereby generating a valid set of wind measurement data.

[0090] The system then aggregates the valid wind measurement data sets according to the wind turbine number, forming a multi-source wind measurement data set for the unit from the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data corresponding to the same wind turbine. The system then performs multi-source consistency difference calculation on this set, generating unit wind measurement consistency difference data by comparing the differences in wind speed and wind direction between wind measurement data from different sources.

[0091] After generating the wind measurement consistency difference data, a multi-source consistency constraint relationship is established based on the difference results. A multi-source reliable contraction coefficient is introduced to perform reliable contraction processing on the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data, thereby generating the unit fused wind measurement data. Subsequently, data recombination processing is performed on the unit fused wind measurement data according to the wind turbine number, so that each wind turbine forms a set of unit inflow state data containing fused wind speed data and fused wind direction data, and a unified wind field state data is generated.

[0092] After receiving the unified wind field status data, the wake propagation relationship generation module reads the fused wind speed data and fused wind direction data corresponding to each wind turbine from the data, performs direction conversion processing on the fused wind direction data, and performs wind speed matching processing on the fused wind speed data, thereby generating the incoming flow direction vector corresponding to each wind turbine.

[0093] The system then reads the spatial location data of the units and determines the spatial coordinate information of each wind turbine. It then performs pairwise combination processing on all wind turbines according to their numbers to generate turbine pairing data. For each pair of wind turbines, the system calculates the displacement data of the line connecting the units based on the spatial coordinate information. Using the incoming flow direction vector corresponding to one of the wind turbines as the propagation direction reference, the system performs directional projection processing on the displacement data of the line connecting the units to generate the longitudinal projection distance.

[0094] After obtaining the longitudinal projection distance, the displacement data of the inter-unit connection is processed by lateral decomposition to generate lateral offset distance. Then, the longitudinal projection distance is processed by downstream propagation direction determination. The turbine pairings that meet the downstream propagation direction condition are identified as the effective propagation turbine pairing set. Finally, the upstream turbine number, downstream turbine number, incoming flow direction vector, longitudinal projection distance, and lateral offset distance in the effective propagation turbine pairing set are processed by association organization to generate turbine wake propagation relationship data.

[0095] Specific Implementation Example 2: In this embodiment, after the system completes the calculation of the wake coupling relationship and the division of the unit group, it performs coordinated yaw control on the group of wind turbines and continuously updates the wind farm's operating status.

[0096] The system first receives wind turbine wake propagation relationship data, unified wind field status data, and wind turbine power output data through the wake coupling relationship construction module. It also reads the upstream wind turbine number, downstream wind turbine number, incoming flow direction vector, longitudinal projection distance, and lateral offset distance corresponding to each effective propagation wind turbine pairing. At the same time, it reads the wind speed data in the unified wind field status data and the real-time power data in the wind turbine power output data.

[0097] The system performs theoretical power reference calculations based on wind speed data to generate theoretical power reference data for each wind turbine. It then performs matching processing between the theoretical power reference data and real-time power data, calculating the power deviation data for each turbine. Subsequently, it performs time-based aggregation processing on the power deviation data according to the control time window, and performs statistical processing on the power deviation data of the same wind turbine within the control time window to generate power deviation statistics. Based on the statistical results, it generates a power deviation feedback coefficient.

[0098] The system then performs propagation distance organization and offset distance organization processing based on the longitudinal projection distance and lateral offset distance, generating propagation spatial relationship data. Next, it calculates the wake propagation reconstruction coefficients based on the propagation spatial relationship data and the incoming flow direction vector, generating the propagation reconstruction result. The propagation reconstruction result is then converted into basic wake coupling strength data, and the power deviation feedback coefficient is written into the basic wake coupling strength data, thereby generating feedback-corrected wake coupling strength data. Finally, wake coupling relationship matrix data is generated according to matrix position mapping processing.

[0099] After receiving the wake coupling relationship matrix data, the unit group generation module sorts the wake coupling intensity data in the matrix and identifies the upstream and downstream turbine numbers corresponding to the non-zero matrix positions, thereby generating wake association pairing data. Subsequently, based on the wake association pairing data, the wake propagation chain-driven group generation algorithm is executed. Through the head-to-tail connection identification process, wake influence propagation chain data is formed, and the propagation chain is traversed machine by machine. Based on the comparison result of the wake coupling intensity increment data and the current group's internal coupling average, continuous propagation aggregation processing is performed to generate group division result data, and finally, unit collaborative group data is generated.

[0100] The group flow field state reconstruction module receives unit collaborative group data, unified wind field state data, and wind turbine power output data. It also reads the group number, the set of wind turbine numbers within each group, and the propagation order. Simultaneously, it reads the incoming flow state data and real-time power data for each wind turbine within the group. The system performs incoming flow correlation analysis processing according to the propagation order and continuously transmits the incoming flow state data of the upstream wind turbines along the propagation order, thereby generating the local flow field state of the group. Finally, it performs power deviation correction processing based on the real-time power data to generate the local flow field state data for the group.

[0101] The yaw collaborative decision-making module receives group data, local flow field state data, turbine power output data, and yaw state data. It also reads the local inflow distribution results, real-time power data, and current yaw state information according to the group number. The system performs group yaw benefit calculations based on the local inflow distribution results and real-time power data. It then performs yaw change analysis based on the current yaw state information to generate yaw execution cost data, and performs inflow fluctuation analysis based on the local inflow distribution results to generate yaw stability data. Subsequently, it performs group power response calculations based on the group's local flow field state data, and performs joint scoring calculations on the group yaw benefit data, yaw execution cost data, and yaw stability data based on the group power response results, thereby generating group yaw decision data.

[0102] After receiving the group yaw decision data, the control update module generates group yaw coordinated control instructions based on the group yaw decision data, and performs instruction mapping processing according to the group number and the wind turbine number, so that each wind turbine forms a unique yaw adjustment direction, yaw adjustment amount and control execution sequence, and outputs the group yaw coordinated control instructions to the corresponding wind turbine execution end.

[0103] After the control cycle ends, the system reads the updated unified wind farm status data and turbine power output data, and performs before-and-after comparison processing to generate post-control status change data. Subsequently, based on the post-control status change data, the system performs rolling update processing on the wake coupling matrix data and the unit coordination group data. The updated wake coupling matrix data, unit coordination group data, unified wind farm status data, and turbine power output data are then subjected to periodic correlation organization processing to form a continuously updated group yaw coordinated control result.

[0104] The multi-source wind field data acquisition module receives SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data. It also collects wind turbine power output data and yaw status data. After time alignment and spatial mapping processing, it generates unified time window wind field status candidate data and outputs the unified time window wind field status candidate data to the wind field status fusion construction module. The wind turbine power output data is transmitted to the wake coupling relationship construction module, the group flow field status reconstruction module, the yaw collaborative decision module, and the control update module. The yaw status data is transmitted to the yaw collaborative decision module. The wind field state fusion construction module receives candidate wind field state data with a unified time window, processes it through multi-source consistency difference calculation and multi-source consistency constraint reliable contraction fusion algorithm to generate unified wind field state data, and outputs the unified wind field state data to the wake propagation relationship generation module, the wake coupling relationship construction module, the group flow field state reconstruction module and the control update module respectively. The wake propagation relationship generation module receives unified wind field state data, and generates wind turbine wake propagation relationship data through incoming flow direction vector extraction, longitudinal projection distance calculation, lateral offset distance calculation, and downstream propagation direction determination. The wind turbine wake propagation relationship data is then output to the wake coupling relationship construction module. The wake coupling relationship construction module receives wind turbine wake propagation relationship data, unified wind field status data, and wind turbine power output data. After processing through theoretical power reference calculation, power deviation calculation, power deviation statistical processing, multi-source reliable wake coupling reconstruction algorithm processing, and matrix position mapping, it generates wake coupling relationship matrix data and outputs the wake coupling relationship matrix data to the unit group generation module and the control update module respectively. The unit group generation module receives the wake coupling relationship matrix data, and generates unit cooperative group data through wake influence intensity sorting, wake influence propagation chain identification, continuous propagation aggregation, and wake propagation chain driven group generation algorithm. The unit cooperative group data is then output to the group flow field state reconstruction module, yaw cooperative decision module, and control update module, respectively. The group flow field state reconstruction module receives the unit collaborative group data, unified wind field state data, and wind turbine power output data. After group local flow field reconstruction processing and power deviation correction processing, it generates group local flow field state data and outputs the group local flow field state data to the yaw collaborative decision module. The yaw collaborative decision module receives group data of unit collaboration, group local flow field state data, wind turbine power output data and yaw state data. After processing the group yaw benefit calculation, yaw execution cost calculation, yaw stability data calculation, group power response calculation and joint scoring, it generates group yaw decision data and outputs the group yaw decision data to the control update module. The control update module receives group yaw decision data, unified wind farm status data, wind turbine power output data, wake coupling matrix data, and unit coordination group data. After yaw control sequence organization and processing, it generates group yaw coordinated control commands. After wake coupling matrix data rolling update processing and unit coordination group data rolling update processing, it forms continuously updated group yaw coordinated control results.

[0105] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A wind farm group yaw coordination control system based on multi-source data fusion, characterized in that: include: The multi-source wind field data acquisition module performs time alignment and spatial mapping processing on SCADA operation data, cabin wind measurement data, lidar wind measurement data and meteorological tower wind measurement data to generate candidate wind field status data for a unified time window. The wind field state fusion construction module performs multi-source consistency difference calculation on candidate wind field state data of a unified time window, and introduces multi-source reliable contraction coefficient to perform multi-source consistency constraint reliable contraction fusion algorithm to generate unified wind field state data; The wake propagation relationship generation module generates wind turbine wake propagation relationship data based on unified wind field status data and unit spatial location data. The wake coupling relationship construction module generates a power deviation feedback coefficient based on unified wind field state data and wind turbine power output data, introduces a wake propagation reconstruction coefficient, and performs a multi-source reliable wake coupling reconstruction algorithm on the wind turbine wake propagation relationship data to generate wake coupling relationship matrix data. The unit group generation module performs sorting, propagation chain identification, and continuous propagation aggregation processing based on the wake coupling relationship matrix data, and generates unit collaborative group data through the wake influence propagation chain driven group generation algorithm. The group flow field state reconstruction module performs group local flow field reconstruction and power deviation correction processing on the unit collaborative group data to generate group local flow field state data. The yaw coordination decision module generates group yaw decision data based on unit coordination group data, group local flow field state data, wind turbine power output data, and yaw state data. The control update module generates group yaw coordinated control commands based on group yaw decision data, and performs rolling update processing on the wake coupling relationship matrix data and the crew coordinated group data to form group yaw coordinated control results.

2. The wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The multi-source wind field data acquisition module receives SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data. It also collects unit spatial location data, wind turbine power output data, and yaw status data. The module reads the corresponding time stamp information of SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data according to the acquisition time, and performs time sorting processing on various types of data based on the time stamp information to generate multi-source time series data. The multi-source time series data is divided into time windows according to the control cycle, and time alignment processing is performed on SCADA operation data, cabin wind measurement data, lidar wind measurement data and meteorological tower wind measurement data within each time window to form a multi-source wind field data set for the time window; Read the spatial location data of the wind turbines and determine the spatial coordinate information of each wind turbine based on the spatial location data of the wind turbines. Perform spatial mapping processing on the nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data in the multi-source wind field data set of the time window to establish a correlation between various types of wind measurement data and the corresponding wind turbine spatial coordinates, and generate wind measurement data associated with the wind turbines. The wind measurement data associated with the unit is aggregated and processed according to the wind turbine number, so that the SCADA operation data, nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data corresponding to the same wind turbine are formed into a multi-source data set of the unit. The multi-source data set of the units is processed using a unified data structure, so that each wind turbine forms a unified time window unit data set containing SCADA operation data, nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data within the same time window. Candidate data for wind field status in the unified time window are generated based on the unified time window unit data set.

3. The wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The wind field status fusion construction module receives candidate wind field status data for a unified time window, and reads the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data for each wind turbine from the candidate wind field status data for a unified time window. It performs data integrity checks on the nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data, filters out data items with missing or abnormal records, and generates a valid set of wind measurement data. The valid wind measurement data set is processed by wind turbine number to form a multi-source wind measurement data set for the unit, which includes the nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data corresponding to the same wind turbine. Multi-source consistency difference calculation processing is performed on the multi-source wind measurement data set of the unit. By performing corresponding comparison processing on the differences in wind speed and wind direction between the nacelle wind measurement data, lidar wind measurement data and meteorological tower wind measurement data, the unit wind measurement consistency difference data is generated. Based on the differences in the wind measurement consistency data of the unit, a multi-source consistency constraint credible contraction fusion algorithm is executed. A multi-source consistency constraint relationship is established based on the differences in the wind measurement consistency data of the unit, and a multi-source credible contraction coefficient is introduced based on the multi-source consistency constraint relationship. Credible contraction processing is performed on the wind measurement data of the engine room, the wind measurement data of the lidar, and the wind measurement data of the weather tower respectively, so that the data with small consistency differences maintain a high degree of participation, and the data with large consistency differences reduce the degree of participation, thereby generating the fused wind measurement data of the unit. The wind measurement data of the units are reorganized according to the wind turbine number to form a set of unit inflow status data containing the integrated wind speed data and the integrated wind direction data for each wind turbine, and a unified wind field status data is generated based on the set of unit inflow status data.

4. A wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The wake propagation relationship generation module receives unified wind field status data and reads the fused wind speed data and fused wind direction data corresponding to each wind turbine from the unified wind field status data. It performs direction conversion processing on the fused wind direction data and wind speed matching processing on the fused wind speed data to generate the incoming flow direction vector corresponding to each wind turbine. Read the spatial location data of the units, determine the spatial coordinate information of each wind turbine based on the spatial location data of the units, and perform pairwise combination processing on all wind turbines according to the wind turbine number to generate wind turbine pairing data; For each group of wind turbines in the wind turbine pairing data, the inter-group line displacement data is calculated based on the corresponding spatial coordinate information. The inflow direction vector corresponding to one of the wind turbines is used as the propagation direction reference. The inter-group line displacement data is then processed by directional projection to generate the corresponding longitudinal projection distance. Based on the inter-unit line displacement data and longitudinal projection distance, perform lateral decomposition processing to generate the corresponding lateral offset distance; The downstream propagation direction determination process is performed on the longitudinal projection distance. The fan pairings that meet the downstream propagation direction condition are identified as valid propagation fan pairings, and the remaining fan pairings are removed to generate a set of valid propagation fan pairings. For each pair of wind turbines in the effective propagation wind turbine pairing set, perform association organization processing according to the upstream wind turbine number, downstream wind turbine number, corresponding incoming flow direction vector, longitudinal projection distance, and lateral offset distance to generate wind turbine wake propagation relationship data.

5. A wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The wake coupling relationship construction module receives wind turbine wake propagation relationship data, unified wind field status data, and wind turbine power output data. It reads the upstream wind turbine number, downstream wind turbine number, incoming flow direction vector, longitudinal projection distance, and lateral offset distance corresponding to each effective propagating wind turbine pair from the wind turbine wake propagation relationship data, reads the wind speed data corresponding to each wind turbine from the unified wind field status data, and reads the real-time power data corresponding to each wind turbine from the wind turbine power output data. Theoretical power reference calculations are performed on each wind turbine based on the wind speed data, and the wind speed data and corresponding power reference results are associated and organized according to the wind turbine number to generate theoretical power reference data for each wind turbine. The theoretical power reference data and real-time power data are matched according to the wind turbine number, and the difference between the theoretical power reference data and real-time power data is compared and calculated to generate power deviation data for each wind turbine. The power deviation data is processed by time aggregation according to the control time window, and the power deviation data of the same fan within the control time window is processed by statistical processing to generate power deviation statistics. Then, feedback mapping processing is performed based on the power deviation statistics to generate power deviation feedback coefficients. For each pair of effective propagating wind turbines, wake coupling reconstruction processing is performed. Propagation distance organization processing is performed based on the longitudinal projection distance and offset distance organization processing is performed based on the lateral offset distance. The propagation distance organization results and offset distance organization results are then correlated to generate propagation spatial relationship data for each pair of effective propagating wind turbines. Based on the propagation spatial relationship data, a wake propagation reconstruction coefficient is introduced, and the wake propagation reconstruction coefficient is associated with the incoming flow direction vector corresponding to each effective propagation fan pairing to generate the propagation reconstruction result corresponding to each effective propagation fan pairing. Based on the propagation reconstruction results, the basic coupling strength construction process is performed on each effective propagation fan pairing to form corresponding basic wake coupling strength data for each effective propagation fan pairing. Write the power deviation feedback coefficient corresponding to each downstream wind turbine into its corresponding foundation wake coupling strength data, perform feedback correction processing on the foundation wake coupling strength data, and generate feedback corrected wake coupling strength data. The feedback-corrected wake coupling intensity data is processed by matrix position mapping according to the upstream and downstream fan numbers. The feedback-corrected wake coupling intensity data corresponding to each pair of effectively propagating fans is written into the corresponding matrix position. Zero-value filling is performed on the matrix positions corresponding to fan pairs that have not formed an effective propagation relationship, generating wake coupling relationship matrix data.

6. A wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The unit group generation module receives the wake coupling relationship matrix data and reads the row index, column index, and wake coupling intensity data corresponding to each matrix position in the wake coupling relationship matrix data. Based on the turbine number corresponding to the row index and column index, it establishes the turbine association mapping relationship and generates turbine association matrix data. The wake coupling intensity data of each row in the wind turbine correlation matrix data is sorted according to the intensity, generating wake influence intensity sorting data for each upstream wind turbine. Then, based on the upstream and downstream wind turbine numbers corresponding to the non-zero matrix positions in the wake coupling intensity data, valid associated wind turbine pairings are extracted to generate wake association pairing data. The wake association pairing data is processed by the wake propagation chain driven group generation algorithm. The priority propagation direction of each upstream wind turbine is determined according to the wake influence intensity sorting data. The head-to-tail connection identification processing is performed according to the upstream wind turbine number and downstream wind turbine number in the wake association pairing data. The associated wind turbine pairings that match the downstream wind turbine number in the previous wind turbine pairing and the upstream wind turbine number in the next wind turbine pairing are connected to form the wake influence propagation chain, and the wake influence propagation chain data is generated. Each wake influence propagation chain in the wake influence propagation chain data is traversed machine by machine in the propagation order. The starting fan in each wake influence propagation chain is taken as the initial aggregation unit, and the next fan adjacent to the initial aggregation unit is taken as the current fan to be aggregated. The wake coupling strength data between the current fan to be aggregated and the initial aggregation unit is read, and the current wake coupling strength increment data is generated. The mean coupling strength of the current group is calculated based on the wake coupling strength data of each fan that has entered the same aggregation unit. The incremental data of the current wake coupling strength is compared with the mean coupling strength of the current group. When the incremental data of the current wake coupling strength meets the continuous propagation aggregation condition, the fan to be aggregated is merged into the current aggregation unit, and the next fan in the wake influence propagation chain is read as the new fan to be aggregated. When the current wake coupling intensity increment data does not meet the continuous propagation aggregation condition, the current aggregation unit is determined as a unit coordination group, and the next wind turbine in the wake influence propagation chain that has not entered the current aggregation unit is determined as the new initial aggregation unit. The turbine-by-turbine traversal process continues until all wake influence propagation chains have completed the continuous propagation aggregation process, and the group division result data is generated. The group division results data are processed according to the group number, the set of fan numbers within the group, and the propagation order of each fan in the wake influence propagation chain to generate unit collaborative group data.

7. A wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The group flow field state reconstruction module receives unit collaborative group data, unified wind field state data, and wind turbine power output data. It reads the group number corresponding to each group, the set of wind turbine numbers in the group, and the propagation order of each wind turbine in the wake influence propagation chain from the unit collaborative group data. It reads the unit incoming flow state data corresponding to each wind turbine in the group from the unified wind field state data, and reads the real-time power data corresponding to each wind turbine in the group from the wind turbine power output data. The incoming flow status data of the wind turbines in each group are grouped and collected according to the group number, and then sorted according to the propagation order of each wind turbine in the wake influence propagation chain to generate the group incoming flow status sequence data. Perform intra-group correlation analysis on the fused wind speed and fused wind direction data corresponding to each fan in the group inflow state sequence data to determine the inflow transmission relationship between upstream and downstream fans in the same group and generate inflow correlation data within the group. Based on the inflow correlation data within the group, the inflow state sequence data of the group is processed to perform local flow field reconstruction. The inflow state data of the upstream wind turbine and the inflow state data of the downstream wind turbine are processed to perform continuous transmission organization in the order of propagation to generate local flow field state of the group. The real-time power data of each wind turbine in the group is aggregated within the group, and the corresponding matching process is performed according to the propagation order of each wind turbine in the wake influence propagation chain to generate group power distribution data. Based on the group power distribution data, power deviation correction processing is performed on the local flow field state of the group, and correction and update processing is performed on the local inflow distribution part in the local flow field state of the group that is inconsistent with the real-time power data, generating group corrected flow field state data. The group-corrected flow field state data is processed by association according to the group number, the set of fan numbers within the group, the propagation order of each fan in the wake influence propagation chain, and the local inflow distribution results after power deviation correction, to generate local flow field state data for the group.

8. A wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The yaw cooperative decision module receives unit cooperative group data, group local flow field state data, turbine power output data, and yaw state data. It reads the group number, turbine number set within the group, and propagation order of each turbine in the wake influence propagation chain from the unit cooperative group data. It reads the local inflow distribution results corresponding to each group from the group local flow field state data. It reads the real-time power data corresponding to each turbine within the group from the turbine power output data. It reads the current yaw state information corresponding to each turbine within the group from the yaw state data. The local inflow distribution results, real-time power data and current yaw status information of the wind turbines in each group are grouped and collected according to the group number, and then sorted according to the propagation order of each wind turbine in the wake influence propagation chain to generate group decision input data. Based on the local inflow distribution results and real-time power data in the group decision input data, the group yaw benefit calculation is performed to determine the power change of each wind turbine in each group under different yaw adjustment states, and generate group yaw benefit data. Based on the current yaw status information in the group decision input data, perform yaw change analysis and processing to determine the yaw adjustment magnitude and yaw adjustment order for each wind turbine in each group, and generate yaw execution cost data. Based on the local inflow distribution results in the group decision input data, perform group inflow fluctuation analysis to determine the continuity and propagation stability of inflow changes for each wind turbine in each group, and generate yaw stability data; Based on the local flow field state data of the group, the group power response calculation is performed on the fans in each group to determine the power response changes of each group under the current local inflow distribution conditions and generate the group power response results; Based on the group power response results, the group yaw benefit data, yaw execution cost data, and yaw stability data are jointly scored and processed to determine the yaw adjustment priority and group yaw coordination direction of each wind turbine in the group, and generate group yaw decision data.

9. A wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The control update module receives group yaw decision data, unified wind farm status data, turbine power output data, wake coupling matrix data, and unit coordination group data. It reads the group number, turbine number set within the group, yaw adjustment priority, and group yaw coordination direction corresponding to each group from the group yaw decision data. It reads the turbine inflow status data corresponding to each turbine from the unified wind farm status data. It reads the real-time power data corresponding to each turbine from the turbine power output data. It reads the wake coupling strength data between each turbine from the wake coupling matrix data. It reads the group division result data corresponding to each group from the unit coordination group data. The yaw adjustment priority, group yaw coordination direction, unit inflow status data and real-time power data corresponding to each group are grouped and collected according to the group number, and corresponding association processing is performed according to the set of wind turbine numbers in the group to generate group control input data. Based on the group control input data, the yaw control sequence of each wind turbine in each group is organized and processed to determine the control execution order of each wind turbine. Combined with the group yaw coordination direction, the yaw adjustment direction and yaw adjustment amount of each wind turbine are determined, and the group yaw coordination control command is generated. The group yaw coordinated control command is processed by mapping the command to the group number and the wind turbine number, so that each wind turbine corresponds to a unique yaw adjustment direction, yaw adjustment amount and control execution sequence. The mapped group yaw coordinated control command is then output to the yaw control execution terminal of the corresponding wind turbine, forming the group yaw coordinated control command set for the current control cycle. After the current control cycle ends, read the updated unified wind farm status data and wind turbine power output data, and perform before-and-after comparison processing with the unified wind farm status data and wind turbine power output data corresponding to the current control cycle to generate post-control status change data. Based on the post-control state change data, the wake coupling matrix data is rolled up and updated accordingly. The matrix positions where the wake coupling strength changes are updated to generate the updated wake coupling matrix data. Based on the post-control state change data, the unit coordination group data is rolled up and updated accordingly. The group number, the set of fan numbers in the group, and the propagation order of each fan in the wake influence propagation chain are updated to generate the updated unit coordination group data. The updated wake coupling matrix data, unit coordination group data, unified wind farm status data, and wind turbine power output data are subjected to periodic correlation organization processing to form the group yaw coordination control results.

10. A wind farm group yaw coordination control system based on multi-source data fusion according to claim 1, characterized in that: The multi-source wind field data acquisition module receives SCADA operation data, nacelle wind measurement data, lidar wind measurement data, and meteorological tower wind measurement data. It also collects wind turbine power output data and yaw status data. After time alignment and spatial mapping processing, it generates unified time window wind field status candidate data and outputs the unified time window wind field status candidate data to the wind field status fusion construction module. The wind turbine power output data is transmitted to the wake coupling relationship construction module, the group flow field status reconstruction module, the yaw collaborative decision module, and the control update module. The yaw status data is transmitted to the yaw collaborative decision module. The wind field state fusion construction module receives candidate wind field state data with a unified time window, processes it through multi-source consistency difference calculation and multi-source consistency constraint reliable contraction fusion algorithm to generate unified wind field state data, and outputs the unified wind field state data to the wake propagation relationship generation module, the wake coupling relationship construction module, the group flow field state reconstruction module and the control update module respectively. The wake propagation relationship generation module receives unified wind field state data, and generates wind turbine wake propagation relationship data through incoming flow direction vector extraction, longitudinal projection distance calculation, lateral offset distance calculation, and downstream propagation direction determination. The wind turbine wake propagation relationship data is then output to the wake coupling relationship construction module. The wake coupling relationship construction module receives wind turbine wake propagation relationship data, unified wind field status data, and wind turbine power output data. After processing through theoretical power reference calculation, power deviation calculation, power deviation statistical processing, multi-source reliable wake coupling reconstruction algorithm processing, and matrix position mapping, it generates wake coupling relationship matrix data and outputs the wake coupling relationship matrix data to the unit group generation module and the control update module respectively. The unit group generation module receives the wake coupling relationship matrix data, and generates unit cooperative group data through wake influence intensity sorting, wake influence propagation chain identification, continuous propagation aggregation, and wake propagation chain driven group generation algorithm. The unit cooperative group data is then output to the group flow field state reconstruction module, yaw cooperative decision module, and control update module, respectively. The group flow field state reconstruction module receives the unit collaborative group data, unified wind field state data, and wind turbine power output data. After group local flow field reconstruction processing and power deviation correction processing, it generates group local flow field state data and outputs the group local flow field state data to the yaw collaborative decision module. The yaw collaborative decision module receives group data of unit collaboration, group local flow field state data, wind turbine power output data and yaw state data. After processing the group yaw benefit calculation, yaw execution cost calculation, yaw stability data calculation, group power response calculation and joint scoring, it generates group yaw decision data and outputs the group yaw decision data to the control update module. The control update module receives group yaw decision data, unified wind farm status data, wind turbine power output data, wake coupling matrix data, and unit coordination group data. After yaw control sequence organization and processing, it generates group yaw coordinated control commands. After wake coupling matrix data rolling update processing and unit coordination group data rolling update processing, it forms continuously updated group yaw coordinated control results.

Images