Fan control method and system based on laser radar

Abstract

The invention relates to the technical field of fan control, and discloses a fan control method and system based on a laser radar, and the method comprises the steps: carrying out the hierarchical scanning of a wind field in front of a fan through the laser radar, obtaining multi-dimensional wind condition data, carrying out the preprocessing of the collected data, and extracting the wind condition characteristic parameters, the characteristic parameters are compared with a historical database to generate a yaw correction path, a pre-control instruction is generated based on future wind regime prediction, a safety instruction is triggered through wind regime risk judgment to cover the pre-instruction, and after execution is completed, the database is updated to form closed-loop optimization. According to the invention, the fusion of prospective accurate wind facing and intelligent safety protection is realized through the laser radar, the yaw precision and the power generation efficiency are improved, and meanwhile, the operation reliability of the unit under the complex wind condition is enhanced.

Description

Technical Field

[0001] This invention relates to the field of wind turbine control technology, and more specifically, to a wind turbine control method and system based on lidar. Background Technology

[0002] Currently, wind turbines generally use mechanical or ultrasonic anemometers for wind measurement and control. These sensors are passive measures with low temporal resolution and can only acquire instantaneous wind conditions at the turbine's current location. They cannot detect or predict wind speed and direction changes within a range of hundreds of meters in front of the turbine. In complex terrain, such as mountainous wind farms, where turbulence is high and wind direction changes frequently, the lag in this measurement method leads to a response delay in the turbine's yaw system, resulting in insufficient yaw accuracy. The average yaw angle error of most wind farm turbines is above 6 degrees, indicating insufficient yaw accuracy. This not only reduces wind energy capture efficiency but also exacerbates mechanical fatigue in the drivetrain and other components due to the mismatch between yaw adjustments and actual wind conditions, as well as frequent start-stop cycles, potentially leading to serious safety hazards such as blade breakage or impeller overspeed. Currently, the industry lacks effective technical means to continuously and accurately measure and correct this yaw error, and its impact on power generation loss and the lifespan of major components is difficult to quantify.

[0003] Therefore, there is an urgent need to provide a wind turbine control method and system based on lidar to solve this problem. Summary of the Invention

[0004] In view of this, the present invention proposes a wind turbine control method and system based on lidar, aiming to solve the problems of insufficient yaw accuracy, low power generation efficiency, and mechanical load impact and safety risks faced by wind turbines under complex wind conditions caused by the lag in response of traditional wind measuring devices and their inability to sense forward wind field information.

[0005] In one aspect, the present invention proposes a wind turbine control method based on lidar, comprising:

[0006] Wind speed and direction data were collected at several altitude levels and several forward distances using a cabin laser wind radar. Blank or abrupt values ​​were filtered out to obtain a layered raw wind condition dataset.

[0007] The original layered wind condition dataset is denoised, smoothed, and interpolated to form a continuous wind speed time series curve and wind direction change path. Combined with power output, vibration signal, and gearbox data, wind condition characteristic parameters are extracted.

[0008] By comparing wind condition characteristic parameters with historical wind condition characteristic parameters in the historical database, yaw deviations under different wind speed and direction conditions are identified, a yaw correction path is generated, and yaw adjustment is triggered when a continuous change in wind conditions or a rapid change in direction is detected. The yaw adjustment includes the determination of the start time, rotation amplitude, and stopping conditions.

[0009] Based on the yaw correction path, the wind trend of the future time window is collected in advance by the laser wind radar to predict the wind speed and direction that will affect the impeller, generate the prediction result, and match it with the current yaw state. When the prediction result does not match the existing direction, a yaw pre-command is generated and the response priority of the control system is adjusted.

[0010] When receiving the forecast results, the risk assessment is performed using wind condition characteristic parameters to identify gusts, wind shear and extreme wind conditions. When the risk conditions are triggered, safety control instructions are generated, including limiting the yaw adjustment frequency, reducing the speed of operation or triggering emergency braking, to cover or limit the yaw pre-instructions.

[0011] After the safety control command is executed, record the wind speed difference, power change and correction angle before and after yaw adjustment, update the historical database, and use it when the yaw correction module is called.

[0012] Furthermore, when collecting wind speed and direction data, the following are included:

[0013] Wind speed and wind direction information are acquired sequentially at preset time intervals at several altitude levels.

[0014] Establish a continuous sequence of sampling points across several forward distance segments and record the corresponding distance segment markers;

[0015] The collected wind speed and wind direction information were checked item by item for sampling time and numerical range, and data points that did not conform to the continuity of time or the pattern of numerical change were removed.

[0016] After removing the data points, subsequences are created within each height layer according to the distance segments ahead. Within each subsequence, the data points are arranged according to the sampling time. A height layer label, a distance segment label, and a time label are written for each data point. The time label is then associated with the height layer label and the distance segment label, so that the three labels correspond one-to-one in the same record, thus obtaining the original dataset of stratified wind conditions.

[0017] Furthermore, when extracting wind condition characteristic parameters, the following are included:

[0018] The original stratified wind condition dataset was organized into a record sequence arranged according to sampling time, based on both height layer and forward distance segment.

[0019] Time gaps in the record sequence are filled in as follows: when there is a recent valid record on only one side of the gap, the value of the recent valid record on that side is filled in according to the sampling time order until the end of the gap; when there are recent valid records on both sides of the gap, the gap is divided into several positions according to the sampling time interval, and the transition record between the record values ​​on both sides is filled in sequentially, and a time gap supplement mark is added to the supplemented record.

[0020] Spatial gaps in the recorded sequence are filled in: when the same sampling time is missing in a certain forward distance segment, the nearest valid record in the adjacent forward distance segment and at the same height layer is used to fill in the gaps; when the same sampling time is missing in a certain height layer, the nearest valid record in the adjacent height layer and at the same forward distance segment is used to fill in the gaps, and a spatial gap filling mark is added to the supplemented record.

[0021] After completing the temporal and spatial supplementation, the sequences are time-series aligned and spatially aligned to obtain the wind speed time-series curves and wind direction change paths.

[0022] Furthermore, after the original stratified wind condition dataset is supplemented and wind speed time-series curves and wind direction change paths are formed, the directly collected data is labeled as sampled data when stored in the historical database, and the data generated by interpolation or by referencing adjacent height layers and forward distance segments is labeled as supplementary data when stored in the historical database. This distinguishes the sampled data from the supplementary data in the historical database. At the same time, the power output, vibration signal, and gearbox data are organized into unit operation sequences, and a sampling time mark is added to each unit operation sequence when stored in the historical database. This establishes the correspondence between the unit operation sequence and the wind speed time-series curves and wind direction change paths under the same sampling time mark.

[0023] Furthermore, when generating the yaw correction path and implementing the yaw adjustment, the following steps are included:

[0024] Based on the sampling time and spatial markers, historical wind condition characteristic parameter records that are similar to the current wind condition characteristic parameters are retrieved from the historical database to form a comparison set; the current wind condition characteristic parameters are compared with the comparison set item by item to determine the direction of the yaw deviation and the range that needs to be corrected, and a draft yaw correction path is formed.

[0025] The draft yaw correction path is organized into path entries according to the sampling time sequence. The path entries include the start time, rotation direction, segment target and stopping condition, and an entry tag is written for each path entry.

[0026] When the yaw deviation remains in the same direction within a continuous sampling period, or when there is a rapid change in wind direction, yaw adjustment is triggered and executed in the order of path entries;

[0027] After each entry is completed, the current wind condition characteristic parameters are checked, and if necessary, unexecuted path entries are revised and then executed.

[0028] The yaw adjustment ends when the stopping condition is met or the current wind direction is consistent with the cabin orientation; otherwise, the existing yaw state is maintained.

[0029] Furthermore, in generating the prediction results and forming the yaw pre-command, the following steps are included:

[0030] Starting from the current sampling time, the order of arrival is determined from farthest to closest distance segments. A future time window index is established, and wind speed and wind direction information corresponding to the index are extracted from each altitude layer to form an arrival order record. The arrival order records are compared between adjacent arrival order records. When the wind speed information is increasing, an increasing mark is written on the next record; when it is decreasing, a decreasing mark is written; when they remain close, a stable mark is written; when the wind direction information changes clockwise, a clockwise mark is written; when it changes counterclockwise, a counterclockwise mark is written.

[0031] The changes in wind speed or wind direction information within adjacent sampling time intervals are marked with fast or slow tags. Based on the arrival order after the tags, the prediction result entries are recorded. The prediction result entries are compared segment by segment with the yaw correction path and cabin orientation. When the wind direction tag in the prediction result entry is consistent with the cabin orientation and the wind speed tag is consistent with the trend of the yaw correction path, it is recorded as a consistent relationship. When the wind direction tag in the prediction result entry deviates from the cabin orientation or the wind speed tag is opposite to the trend of the yaw correction path, it is recorded as an inconsistent relationship. When the inconsistent relationship is recorded for several consecutive sampling time intervals, a draft yaw pre-instruction is generated. The draft yaw pre-instruction is organized into yaw pre-instruction entries according to the arrival order. The yaw pre-instruction entries include the start time, rotation direction, segment target and stop condition, and are written with pre-instruction tags.

[0032] The yaw pre-command entry is set as a pre-execution item of the control system. When the prediction result is consistent with the recovery of the existing yaw state, the yaw pre-command entry that has not yet been executed is withdrawn.

[0033] Furthermore, when conducting risk assessment and generating security instruction entries, the following are included:

[0034] The system receives the prediction results and wind condition characteristic parameters at the same sampling time to form a risk check sequence. In the risk check sequence, when the wind speed information changes abruptly between records of two adjacent sampling times, it is recorded as a gust trigger segment. When the wind speed information changes abruptly between records of adjacent height layers or adjacent forward distance segments, it is recorded as a wind shear trigger segment. When the wind speed information and wind direction information change in different directions at the same sampling time, it is recorded as an extreme wind condition trigger segment.

[0035] When one of the trigger segments is identified, the current yaw pre-command entry is compared with the record corresponding to the trigger segment item by item. When the rotation direction or start timing contained in the yaw pre-command entry is inconsistent with the record of the trigger segment, it is judged as a risk. When it is judged as a risk, a safety command entry is generated. The safety command entry includes emergency braking entry, speed reduction operation entry and yaw adjustment frequency restriction entry, and writes a safety flag according to the type of trigger segment.

[0036] The safety instruction entries are set as the priority execution content of the control system. Unexecuted yaw pre-instruction entries are overridden or restricted. If no trigger segment is identified for several consecutive sampling times, the unexecuted safety instruction entries are withdrawn and the execution of yaw pre-instruction entries is resumed.

[0037] Furthermore, when dealing with situations where both safety instruction entries and yaw pre-instruction entries exist simultaneously, this includes:

[0038] When both safety command entries and yaw pre-command entries are detected at the same sampling time, the safety command entries are executed first, and the yaw pre-command entries that have not yet been executed are marked as frozen.

[0039] In the frozen state, the yaw pre-command entry does not enter the execution process until the safety command entry is completed;

[0040] Once the safety instruction entry has been executed, the freeze state will be lifted if no gust trigger segment, wind shear trigger segment, or extreme wind condition trigger segment is detected in subsequent consecutive sampling periods.

[0041] After the freeze is lifted, the yaw pre-command entries that have not yet been executed are restored and re-entered for execution according to their existing pre-command marking order, so that yaw control can continue after the safety conditions are eliminated.

[0042] Furthermore, when updating the historical database, this includes:

[0043] After each yaw adjustment, wind speed, wind direction, power output, and correction angle were obtained at the sampling time before and after the yaw adjustment, and compiled into a before-and-after comparison record.

[0044] When compiling the before-and-after comparison records, height layer markers were added to each wind speed and wind direction information, forward distance segment markers were added to the collection location of each record, and sampling time markers were added to the sampling time point, so that each record simultaneously has height layer markers, forward distance segment markers, and sampling time markers.

[0045] The before-and-after comparison records with the additional markers are sequentially stored in the historical database. When storing, they are compared item by item with the existing records of the same height layer marker, forward distance segment marker and sampling time marker. The parts with the same content are merged and saved, and the parts with differences are added as new entries.

[0046] The updated historical database is arranged continuously according to the sampling time to obtain a record sequence, which is then used as the historical wind condition characteristic parameter for comparison with the current wind condition characteristic parameter when called again.

[0047] Compared with existing technologies, the advantages of this invention are as follows: By filtering out blank and abrupt values ​​during the acquisition stage, the data entering the stratified wind condition raw dataset remains continuous in the time dimension and maintains a one-to-one correspondence with the height layer and forward distance segment in the spatial dimension, avoiding interference from abnormal records to subsequent processing; the wind speed time series curves and wind direction change paths formed by subsequent denoising, smoothing, and interpolation can effectively compensate for the time and spatial gaps in the sampling process, giving the curves and paths continuous analyzability. Furthermore, after correlating with power output, vibration signals, and gearbox data, wind condition characteristic parameters that truly reflect the relationship between wind conditions and unit operation can be extracted, providing a targeted basis for yaw adjustment; in the comparison stage, the wind condition characteristic parameters correspond one-to-one with records in the historical database, enabling the identification of yaw deviations between the nacelle orientation and the incoming flow, and forming a yaw correction path that includes start-up timing, rotation amplitude, and stopping conditions, thus giving yaw adjustment the characteristics of being itemized, executable, and verifiable; in the prediction stage, the prediction results formed by the wind condition trend in the future time window enable the system to adjust the yaw adjustment based on the incoming flow acting on the wind condition. Yaw pre-commands are generated in advance before the impeller, and rapid execution is achieved by adjusting the response priority of the control system, thereby reducing yaw lag. When the prediction results change frequently, some pre-commands can be merged or postponed to avoid mechanical fatigue caused by excessive nacelle back-and-forth movements. In the risk assessment stage, when gusts, wind shear, or extreme wind conditions are identified, the system can generate safety control commands in a timely manner. By limiting the yaw adjustment frequency, reducing the speed, or triggering emergency braking to cover or limit the yaw pre-commands, the system ensures that the control logic is executed with safety as the priority under abnormal wind conditions, avoiding structural risks caused by excessive movements or sudden load changes. In the update stage, by recording the wind speed difference, power change, and correction angle before and after yaw adjustment and writing them into the historical database, not only are traceable data entries formed, but existing records can also be directly called in subsequent similar conditions. Gradually, a closed-loop operation mechanism of collection, processing, comparison, prediction, control, and updating is built, so that the method can maintain accurate correction under stable wind conditions and ensure operational safety under complex wind conditions. The continuous improvement of the historical database enables long-term adaptation and optimization.

[0048] On the other hand, this application also provides a lidar-based wind turbine control system for implementing the aforementioned lidar-based wind turbine control method, comprising:

[0049] The wind condition sensing module collects wind speed and direction data at several altitude levels and several forward distances using a cabin laser wind radar, filters out blank or abrupt values, and obtains a layered raw wind condition dataset.

[0050] The data processing module performs noise reduction, smoothing, and interpolation on the layered wind condition raw dataset to form a continuous wind speed time series curve and wind direction change path. It also extracts wind condition characteristic parameters by combining power output, vibration signal, and gearbox data.

[0051] The yaw correction module compares wind condition characteristic parameters with historical wind condition characteristic parameters in the historical database, identifies yaw deviations under different wind speeds and directions, generates a yaw correction path, and triggers yaw adjustment when it detects continuous sudden changes in wind conditions or rapid changes in direction. The yaw adjustment includes the determination of the start time, rotation amplitude, and stopping conditions.

[0052] The feedforward control module, based on the yaw correction path, uses the wind trend of the future time window collected in advance by the laser wind radar to predict the wind speed and wind direction that will affect the impeller, generates the prediction result, and matches it with the current yaw state. When the prediction result does not match the existing direction, it generates a yaw pre-command and adjusts the response priority of the control system.

[0053] The safety control module, upon receiving the forecast results, uses wind condition characteristic parameters to determine the risk, identify gusts, wind shear and extreme wind conditions, and generates safety control commands when risk conditions are triggered. These commands include limiting the yaw adjustment frequency, reducing engine speed, or triggering emergency braking, which are used to cover or limit the yaw pre-commands.

[0054] The status update module records the wind speed difference, power change and correction angle before and after yaw adjustment after the safety control command is executed, updates the historical database, and uses it when the yaw correction module is called.

[0055] It is understandable that the aforementioned wind turbine control method and system based on lidar have the same beneficial effects, and will not be elaborated further here. Attached Figure Description

[0056] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0057] Figure 1 A flowchart illustrating a wind turbine control method based on lidar provided in an embodiment of the present invention;

[0058] Figure 2 This is a structural framework diagram of a wind turbine control system based on lidar provided in an embodiment of the present invention. Detailed Implementation

[0059] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0060] In some embodiments of this application, see Figure 1 As shown, a wind turbine control method based on lidar includes:

[0061] S100: By using the cabin laser wind radar to collect wind speed and direction data at several altitude levels and several forward distances, blank or abrupt values ​​are filtered out to obtain a layered wind condition raw dataset.

[0062] S200: Denoise, smooth and interpolate the layered wind condition raw dataset to form a continuous wind speed time series curve and wind direction change path, and extract wind condition characteristic parameters by combining power output, vibration signal and gearbox data.

[0063] S300: Compares wind condition characteristic parameters with historical wind condition characteristic parameters in the historical database, identifies yaw deviations under different wind speed and direction conditions, generates yaw correction paths, and triggers yaw adjustment when continuous sudden changes in wind conditions or rapid changes in direction are detected. Yaw adjustment includes the determination of start timing, rotation amplitude and stop conditions.

[0064] S400: Based on the yaw correction path, it uses the wind trend of the future time window collected in advance by the laser wind radar to predict the wind speed and wind direction that will affect the impeller, generates the prediction result, and matches it with the current yaw state. When the prediction result does not match the existing direction, it generates a yaw pre-command and adjusts the response priority of the control system.

[0065] S500: When receiving the forecast results, it uses wind condition characteristic parameters to determine the risk, identify gusts, wind shear and extreme wind conditions, and generates safety control commands when risk conditions are triggered. These commands include limiting the yaw adjustment frequency, reducing the speed of operation or triggering emergency braking, which are used to cover or limit the yaw pre-commands.

[0066] S600: After the safety control command is executed, record the wind speed difference, power change and correction angle before and after yaw adjustment, update the historical database, and use it when the yaw correction module is called.

[0067] Specifically, this invention begins by collecting wind speed and direction data from a naval laser wind-measuring radar at several altitude levels and several forward distance segments. During the data acquisition phase, blank or abrupt values ​​are filtered out, and the filtered data is organized into a layered raw wind condition dataset, ensuring a continuous temporal sequence and clear spatial correspondence for subsequent processing. Subsequently, the layered raw wind condition dataset is denoised, smoothed, and interpolated to obtain continuously readable wind speed time-series curves and continuously trackable wind direction change paths. Power output, vibration signals, and gearbox data are then correlated with the wind speed time-series curves and wind direction change paths to extract wind condition characteristic parameters, reflecting the correspondence between wind speed distribution, wind direction shifts, and unit operation. Next, the wind condition characteristic parameters are compared with historical wind condition characteristic parameters in a historical database to identify yaw deviations under different wind speed and direction conditions, generating a yaw correction path that includes start-up timing, rotation amplitude, and stopping conditions. When a continuous abrupt change in wind conditions or a rapid change in direction is detected, yaw adjustment is triggered according to the yaw correction path. Based on this, the wind trend of the future time window is collected in advance by laser wind radar to obtain the prediction result, which is then matched with the current yaw state. When the prediction result does not match the current direction, a yaw pre-command is generated, and the response priority of the control system is adjusted so that the corresponding control can be executed sequentially. Simultaneously, when receiving the prediction result, wind characteristic parameters are used to determine the risk, identifying gusts, wind shear, and extreme wind conditions. When a risk condition is triggered, a safety control command is generated, including limiting the yaw adjustment frequency, reducing the engine speed, or triggering emergency braking, to cover or limit the yaw pre-command. After completing the above process, the wind speed difference, power change, and correction angle before and after yaw adjustment are recorded, the historical database is updated, and this information is used when the yaw correction module is invoked, thus forming a closed loop of acquisition, processing, comparison, prediction, control, and updating. During a certain operating period, the detection of the forward distance segment indicates that the wind trend of the future time window will cause the incoming flow direction to shift, while the current yaw state does not match this direction. First, a yaw correction path is generated by comparing wind condition characteristics with historical databases. Then, a yaw pre-command is generated based on the forecast results, and the response priority of the control system is adjusted to ensure the nacelle completes the turn before the incoming airflow acts on the rotor. If gusts or wind shear are detected simultaneously within this time window, a safety control command is generated to override or limit the yaw pre-command, and to execute reduced engine speed operation or trigger emergency braking. After the event, the wind speed difference, power change, and correction angle before and after the yaw adjustment are recorded, and the historical database is updated for direct recall under similar conditions. In this way, yaw adjustments can be completed systematically based on comparison and forecasting under normal wind conditions, while safety control commands ensure execution sequence and operational safety under abnormal wind conditions, and the historical database is updated to achieve closed-loop operation.

[0068] It is understood that in this invention, blank values ​​and abrupt changes are filtered out to ensure the temporal continuity and spatial correspondence of the original layered wind condition dataset. After denoising, smoothing, and interpolation, wind speed time-series curves and wind direction change paths are formed and correlated with power output, vibration signals, and gearbox data to extract wind condition characteristic parameters reflecting the relationship between wind conditions and unit response. These characteristic parameters are compared with the historical database to identify yaw deviations and generate yaw correction paths that include start-up timing, rotation amplitude, and stopping conditions. Yaw pre-instructions are generated by combining future time window wind condition trends and are executed in advance by adjusting response priorities, thereby reducing yaw lag. When gusts, wind shear, or extreme wind conditions are identified, safety control instructions are generated to override or limit the yaw pre-instructions to maintain safety as the priority. After yaw adjustment is completed, wind speed differences, power changes, and correction angles are recorded and written into the historical database, forming a closed-loop mechanism of acquisition, processing, comparison, prediction, control, and updating. This ensures accurate correction under stable wind conditions, ensures operational safety under complex wind conditions, and achieves long-term adaptation through database updates.

[0069] In some embodiments of this application, the collection of wind speed and direction data includes:

[0070] Wind speed and wind direction information are acquired sequentially at preset time intervals at several altitude levels.

[0071] Establish a continuous sequence of sampling points across several forward distance segments and record the corresponding distance segment markers;

[0072] The collected wind speed and wind direction information were checked item by item for sampling time and numerical range, and data points that did not conform to the continuity of time or the pattern of numerical change were removed.

[0073] After removing the data points, subsequences are created within each height layer according to the distance segments ahead. Within each subsequence, the data points are arranged according to the sampling time. A height layer label, a distance segment label, and a time label are written for each data point. The time label is then associated with the height layer label and the distance segment label, so that the three labels correspond one-to-one in the same record, thus obtaining the original dataset of stratified wind conditions.

[0074] Understandably, the data acquisition process, while sequentially acquiring wind speed and direction information at preset time intervals, automatically writes the sampling time to each data point to ensure a continuous timeline during subsequent processing. When establishing a continuous sampling point sequence across several forward distance segments, each distance segment is mapped to its corresponding altitude layer, ensuring that records from the same sampling time are aligned both vertically and horizontally. During the process of verifying the sampling time and numerical range item by item, any data points that do not match adjacent records or whose variation exceeds a reasonable range are discarded, ensuring that the remaining data accurately reflects the gradual change in wind speed and direction. After discarding, data points within each altitude layer are organized into subsequences based on the distance segments, and then arranged within each subsequence according to the sampling time, ensuring that the records at each altitude layer maintain consistent order across different distance segments. Simultaneously, each record is written with an altitude layer marker, a distance segment marker, and a time marker, and a one-to-one correspondence between these three types of markers is established in the storage structure, allowing any record to be located simultaneously in the timeline, spatial position, and altitude layer dimensions. The resulting layered wind condition dataset not only covers sampling results at different heights and distances ahead, but also possesses complete temporal continuity and spatial distribution characteristics, laying a reliable data foundation for subsequent denoising, smoothing, and interpolation processing.

[0075] In some embodiments of this application, the extraction of wind condition characteristic parameters includes:

[0076] The original stratified wind condition dataset was organized into a record sequence arranged according to sampling time, based on both height layer and forward distance segment.

[0077] Time gaps in the record sequence are filled in as follows: when there is a recent valid record on only one side of the gap, the value of the recent valid record on that side is filled in according to the sampling time order until the end of the gap; when there are recent valid records on both sides of the gap, the gap is divided into several positions according to the sampling time interval, and the transition record between the record values ​​on both sides is filled in sequentially, and a time gap supplement mark is added to the supplemented record.

[0078] Spatial gaps in the recorded sequence are filled in: when the same sampling time is missing in a certain forward distance segment, the nearest valid record in the adjacent forward distance segment and at the same height layer is used to fill in the gaps; when the same sampling time is missing in a certain height layer, the nearest valid record in the adjacent height layer and at the same forward distance segment is used to fill in the gaps, and a spatial gap filling mark is added to the supplemented record.

[0079] After completing the temporal and spatial supplementation, the sequences are time-series aligned and spatially aligned to obtain the wind speed time-series curves and wind direction change paths.

[0080] Understandably, after the stratified wind condition dataset is organized into a record sequence, gaps are checked sequentially according to the sampling time. When a time period is found to be missing, if only one side has the most recent valid record, the wind speed and direction values ​​recorded on that side are extended point by point to the end of the gap according to the sampling time interval, ensuring continuity of that time period with the surrounding area. If valid records exist on both sides of the gap, the gap is divided into several positions according to the sampling time interval, and transitional records between the records on both sides are sequentially filled into each position, so that the values ​​of the transitional records change gradually with the sampling time without abrupt jumps. These records are uniformly marked with a time gap supplementation tag to distinguish them from the original records. During spatial supplementation, if a record is missing in a certain forward distance segment at the same sampling time, the most recent valid record at the same height layer in the adjacent forward distance segment is called to fill it in, ensuring the spatial distribution corresponding to that sampling time remains complete. If a record is missing in a certain height layer at the same sampling time, the most recent valid record at the same forward distance segment in the adjacent height layer is called to fill it in, maintaining the continuity of the vertical stratification, and the supplemented records are marked with a spatial gap supplementation tag. Through the above processing, the recording sequence at each sampling time can simultaneously cover the complete altitude layer and forward distance segment. After completing temporal and spatial supplementation, all recording sequences are temporally and spatially aligned, ensuring that each altitude layer and each forward distance segment at the same sampling time can form a corresponding relationship. This yields a wind speed time-series curve that continuously reflects changes over time, and a wind direction change path that completely describes spatial transfer. The data processed in this way not only possesses continuity and completeness but also provides a directly callable foundation for subsequent comparison, prediction, and control steps. For example, in a certain sampling, the third segment of the forward distance segment is missing a record at a sampling time, while the second and fourth segments of the same altitude layer have valid data. The gap is filled by referencing adjacent valid records, and a spatial gap supplementation mark is written for the supplemented data. Similarly, if the record of the highest altitude layer is missing at a certain time point, the record of the next-lower altitude layer in the same forward distance segment is used to supplement it, ensuring that the vertical layering remains complete at that time point. After this supplementation and alignment, the resulting wind speed time series curve and wind direction change path can continuously reflect the evolution of wind conditions and will not cause an overall interruption due to the absence of a single point.

[0081] In some embodiments of this application, after the original stratified wind condition dataset is supplemented and wind speed time-series curves and wind direction change paths are formed, the directly collected data is marked as sampled data when stored in the historical database, and the data generated by interpolation or by referencing adjacent height layers and forward distance segments is marked as supplementary data when stored in the historical database, so that sampled data and supplementary data can be distinguished in the historical database. At the same time, the power output, vibration signal and gearbox data are organized into unit operation sequences, and a sampling time mark is added to each unit operation sequence when stored in the historical database, establishing the correspondence between the unit operation sequence and the wind speed time-series curves and wind direction change paths under the same sampling time mark.

[0082] Specifically, after the stratified wind condition raw dataset is supplemented and wind speed time-series curves and wind direction change paths are generated, the data sources are classified during storage. Wind speed and direction data directly acquired by the nacelle laser anemometer are labeled as sampled data when written to the historical database to indicate that they originate from actual observations. Data generated through interpolation or referencing valid records of adjacent altitude layers and adjacent forward distance segments are labeled as supplementary data to distinguish them from sampled data. In this way, each record in the historical database has a clear source attribute, allowing subsequent retrieval to select only sampled data or, when continuity is required, to simultaneously retrieve supplementary data. Simultaneously, power output, vibration signals, and gearbox data are organized into unit operation sequences, and each record is appended with a sampling time marker when stored in the historical database. A one-to-one correspondence is established between the unit operation sequences and the wind speed time-series curves and wind direction change paths under the same sampling time marker, enabling simultaneous retrieval of wind condition data and unit operation data at any given time point. Therefore, when performing comparisons or predictions, the wind speed distribution, wind direction shift, and power and vibration status of the generator unit at the same sampling time can be directly retrieved, providing a consistent data foundation for the generation of yaw correction paths, the formation of yaw pre-commands, and the triggering of safety control commands. For example, during a certain operating period, when calling the historical database for yaw correction comparison, the sampled data can be selected first to ensure that the comparison results are entirely from actual observations; if there are gaps in the required time period, supplementary data can be called simultaneously to keep the wind speed time series curve and wind direction change path continuous. At this sampling time, the generator unit operating sequence can also be retrieved synchronously to obtain power output and vibration status, and correlate them with wind condition changes, so that the comparison process reflects not only the incoming flow conditions but also the generator unit response status. In this way, the database can ensure clear data sources and provide a continuous and complete comparison foundation when needed.

[0083] In some embodiments of this application, generating a yaw correction path and implementing yaw adjustments includes:

[0084] Based on the sampling time and spatial markers, historical wind condition characteristic parameter records that are similar to the current wind condition characteristic parameters are retrieved from the historical database to form a comparison set; the current wind condition characteristic parameters are compared with the comparison set item by item to determine the direction of the yaw deviation and the range that needs to be corrected, and a draft yaw correction path is formed.

[0085] The draft yaw correction path is organized into path entries according to the sampling time sequence. The path entries include the start time, rotation direction, segment target and stopping condition, and an entry tag is written for each path entry.

[0086] When the yaw deviation remains in the same direction within a continuous sampling period, or when there is a rapid change in wind direction, yaw adjustment is triggered and executed in the order of path entries;

[0087] After each entry is completed, the current wind condition characteristic parameters are checked, and if necessary, unexecuted path entries are revised and then executed.

[0088] The yaw adjustment ends when the stopping condition is met or the current wind direction is consistent with the cabin orientation; otherwise, the existing yaw state is maintained.

[0089] Understandably, when generating a yaw correction path and implementing yaw adjustments, the system first retrieves historical wind condition characteristic parameter records from the historical database that are similar to the current wind condition characteristic parameters according to sampling time and spatial markers. These records are then compiled into a reference set, providing multiple benchmarks for comparison of the current parameters. The current wind condition characteristic parameters are then compared item by item with the reference set to determine the direction of the yaw deviation and the range requiring correction at different sampling times, thus forming a draft yaw correction path. This draft is then organized into path entries in chronological order of sampling time. Each path entry includes the initiation time, rotation direction, segment target, and stopping condition, and is marked with an entry tag to distinguish the order and execution status of the entries. Yaw adjustment is triggered when the yaw deviation remains in the same direction within consecutive sampling times, or when a rapid change in wind direction occurs between adjacent sampling times, and is executed sequentially according to the path entries. After each path entry is executed, the latest wind condition parameters are compared with the records from before execution. If the comparison shows that the deviation direction or correction range has changed, the path entries that have not yet been executed are revised to ensure that subsequent actions are consistent with the latest wind conditions. The entire process ends when the stopping condition is met or the current wind direction is consistent with the cabin orientation. If the trigger condition is not met, the existing yaw state is maintained without action. In this way, yaw adjustments always maintain a correspondence with real-time wind conditions during execution, avoiding inappropriate adjustments due to changes in conditions. For example, the wind direction gradually changes from south-southwest to due west. The draft yaw correction path generated after comparing with the historical database is organized into multiple path entries. The first entry includes the start time being the most recent sampling time, the rotation direction being clockwise, the segment target being southwest-southwest, and the stopping condition being alignment with wind conditions; the second entry includes the rotation direction continuing to be clockwise, the segment target being due west, and the stopping condition being consistency with wind conditions. After triggering a yaw adjustment, the first step is executed. Once completed, the current wind characteristics are checked again. If it is confirmed that the wind direction has further stabilized in a westward direction, the second step is revised and executed again. Finally, the yaw adjustment ends when the cabin orientation is aligned with the wind direction. In this way, the adjustment action can closely follow the actual changes in wind direction, avoiding excessive actions and ensuring the final alignment effect.

[0090] In some embodiments of this application, the process of generating prediction results and forming yaw pre-commands includes:

[0091] Starting from the current sampling time, the order of arrival is determined from farthest to closest distance segments. A future time window index is established, and wind speed and wind direction information corresponding to the index are extracted from each altitude layer to form an arrival order record. The arrival order records are compared between adjacent arrival order records. When the wind speed information is increasing, an increasing mark is written on the next record; when it is decreasing, a decreasing mark is written; when they remain close, a stable mark is written; when the wind direction information changes clockwise, a clockwise mark is written; when it changes counterclockwise, a counterclockwise mark is written.

[0092] The changes in wind speed or wind direction information within adjacent sampling time intervals are marked with fast or slow tags. Based on the arrival order after the tags, the prediction result entries are recorded. The prediction result entries are compared segment by segment with the yaw correction path and cabin orientation. When the wind direction tag in the prediction result entry is consistent with the cabin orientation and the wind speed tag is consistent with the trend of the yaw correction path, it is recorded as a consistent relationship. When the wind direction tag in the prediction result entry deviates from the cabin orientation or the wind speed tag is opposite to the trend of the yaw correction path, it is recorded as an inconsistent relationship. When the inconsistent relationship is recorded for several consecutive sampling time intervals, a draft yaw pre-instruction is generated. The draft yaw pre-instruction is organized into yaw pre-instruction entries according to the arrival order. The yaw pre-instruction entries include the start time, rotation direction, segment target and stop condition, and are written with pre-instruction tags.

[0093] Set the yaw pre-command entry as a prerequisite for the control system. When the prediction result is consistent with the current yaw state, withdraw the yaw pre-command entry that has not yet been executed.

[0094] Specifically, starting from the current sampling time, a future time window index is established according to the order of distance segments from farthest to nearest. Wind speed and direction information corresponding to this index are extracted at each altitude level to form an arrival sequence record. When comparing two adjacent arrival sequence records, the trend of change in the latter record is marked: an increasing mark is written when wind speed increases compared to the previous record, a decreasing mark is written when it decreases, and a stable mark is written when it remains approximately the same; a clockwise mark is written when wind direction changes clockwise compared to the previous record, and a counterclockwise mark is written when it changes counterclockwise. Simultaneously, the magnitude of change in wind speed or direction information within adjacent sampling time intervals is determined; large changes are marked as fast, and small changes as slow. After the above processing, all arrival sequence records have marked information and are organized into prediction result entries. The predicted results are compared segment by segment with the yaw correction path and nacelle orientation. A consistent relationship is recorded when the wind direction marker in the predicted result is aligned with the nacelle orientation and the wind speed marker follows the trend of the yaw correction path. An inconsistent relationship is recorded when there is a deviation between the wind direction marker and the nacelle orientation, or when the wind speed marker follows the opposite trend of the yaw correction path. When inconsistent relationships occur in several consecutive sampling periods, a draft yaw pre-command is generated and organized into yaw pre-command entries in arrival order. Each entry includes the start time, rotation direction, segment target, and stop condition, and a pre-command marker is added during writing to distinguish its source and status. Yaw pre-command entries are set as pre-execution items for the control system to complete adjustments before the predicted incoming flow acts on the impeller. If the predicted results return to consistency with the existing yaw state in subsequent sampling periods, unexecuted yaw pre-command entries are withdrawn to avoid unnecessary adjustments. For example, within a certain period, the future time window index shows a continuous increase in wind speed in the distant segment, accompanied by a gradual shift of wind direction from southeast to due east. In the generated arrival sequence record, ascending and clockwise markers are written one by one, and rapid markers are added in adjacent sampling time intervals. When the compiled prediction results show multiple inconsistencies with the current cabin orientation, a draft yaw pre-instruction is generated, forming yaw pre-instruction entries. The initiation time is set to the next sampling cycle, the rotation direction to be clockwise, the segment target to be due east, and the stopping condition to be consistent with the wind direction. When the prediction results in subsequent sampling time indicate that the wind direction has stabilized and the cabin orientation is aligned with it, the pre-instructions that have not yet been executed are withdrawn to ensure that the adjustment actions are consistent with the actual changes in wind conditions.

[0095] In some embodiments of this application, the process of risk assessment and generating security instruction entries includes:

[0096] The system receives the prediction results and wind condition characteristic parameters at the same sampling time to form a risk check sequence. In the risk check sequence, when the wind speed information changes abruptly between records of two adjacent sampling times, it is recorded as a gust trigger segment. When the wind speed information changes abruptly between records of adjacent height layers or adjacent forward distance segments, it is recorded as a wind shear trigger segment. When the wind speed information and wind direction information change in different directions at the same sampling time, it is recorded as an extreme wind condition trigger segment.

[0097] When one of the trigger segments is identified, the current yaw pre-command entry is compared with the record corresponding to the trigger segment item by item. When the rotation direction or start timing contained in the yaw pre-command entry is inconsistent with the record of the trigger segment, it is judged as a risk. When it is judged as a risk, a safety command entry is generated. The safety command entry includes emergency braking entry, speed reduction operation entry and yaw adjustment frequency restriction entry, and writes a safety flag according to the type of trigger segment.

[0098] Set safety command entries as priority execution items for the control system, cover or restrict yaw pre-command entries that have not yet been executed, and withdraw unexecuted safety command entries and resume execution of yaw pre-command entries if no trigger segment is identified for several consecutive sampling times.

[0099] Understandably, at the same sampling time, the prediction results and wind condition characteristic parameters are combined into a risk check sequence, so that each time point includes both the predicted future trend and the current actual state. When checking the risk check sequence sequentially, if a sudden change in wind speed information occurs between two adjacent sampling times, a gust trigger segment is written for that time period; if a sudden change in wind speed information occurs between adjacent height layers or adjacent forward distance segments, a wind shear trigger segment is written for that position; if the changing directions of wind speed and wind direction information contradict each other in the records of the same sampling time, an extreme wind condition trigger segment is written. Through this marking method, the risk check sequence can clearly identify potential sources of risk. When any trigger segment is identified, the current yaw pre-command entry is compared item by item with the record corresponding to the trigger segment. If the rotation direction or initiation timing in the yaw pre-command entry is inconsistent with the conditions reflected by the trigger segment, it is determined to be a risk. After a risk determination is established, a safety command entry is generated. This entry may include emergency braking entries, speed reduction operation entries, and yaw adjustment frequency restriction entries, and is written into a safety flag according to the type of trigger segment for identification in the database. The generated safety instruction entries are set as priority execution items for the control system, used to cover or restrict yaw pre-instruction entries that have not yet been executed. If no trigger segment appears in the risk check sequence within a certain number of consecutive sampling periods, the unexecuted safety instruction entries are withdrawn, and the normal execution sequence of yaw pre-instruction entries is restored. For example, if the forecast shows that the wind speed will gradually increase in the future time window, and the wind condition characteristic parameters indicate a sudden change in the wind speed records between two adjacent sampling times, a gust trigger segment is written into the risk check sequence. At this time, if the start time set in the yaw pre-instruction entry is too late and does not match the sudden wind speed, it is judged as a risk, a safety instruction entry is immediately generated, and an emergency braking entry and safety flag are written, so that the control system prioritizes emergency braking. In the subsequent sampling period, when no sudden change flag appears in the risk check sequence, the unexecuted safety instruction entries are withdrawn, and the execution of yaw pre-instruction entries is restored, so that the unit operation remains in response to wind condition changes under the premise of safety priority.

[0100] In some embodiments of this application, when handling the simultaneous existence of safety instruction entries and yaw pre-instruction entries, the following methods are included:

[0101] When both safety command entries and yaw pre-command entries are detected at the same sampling time, the safety command entries are executed first, and the yaw pre-command entries that have not yet been executed are marked as frozen.

[0102] In the frozen state, the yaw pre-command entry does not enter the execution process until the safety command entry is completed;

[0103] Once the safety instruction entry has been executed, the freeze state will be lifted if no gust trigger segment, wind shear trigger segment, or extreme wind condition trigger segment is detected in subsequent consecutive sampling periods.

[0104] After the freeze is lifted, the yaw pre-command entries that have not yet been executed are restored and re-entered for execution according to their existing pre-command marking order, so that yaw control can continue after the safety conditions are eliminated.

[0105] Understandably, when both safety command items and yaw pre-command items are detected simultaneously during the same sampling time, the control flow will be immediately adjusted, prioritizing the execution of safety command items and marking any unexecuted yaw pre-command items as frozen. Frozen yaw pre-command items do not enter the execution flow and no longer affect the engine rotation direction, start-up timing, or stop conditions until all current safety command items are executed. After the safety command items are executed, subsequent sampling time periods are monitored. If no gust triggering segment, wind shear triggering segment, or extreme wind condition triggering segment is detected within several consecutive sampling time periods, the risk condition is considered resolved, and the frozen state is automatically lifted. After the frozen state is lifted, the unexecuted yaw pre-command items will resume execution according to their existing pre-command marking order, allowing yaw control actions that were suspended due to risk conditions to continue after the safety conditions are eliminated, thus creating a seamless connection between safety priority and yaw correction. For example, in one operational scenario, the prediction results for a future time window generate a yaw pre-command entry, setting the rotation direction to clockwise and initiating it at the next sampling time. However, a wind shear trigger segment is identified in the risk check sequence at the same sampling time, immediately generating a safety command entry containing reduced speed operation. At this point, the yaw pre-command entry is marked as frozen and no longer enters the execution process, while the safety command entry is executed first. As operation continues, no new trigger segments appear in the subsequent sampling time, indicating that the risk has been eliminated. Therefore, the frozen state is lifted, and the previously unexecuted yaw pre-command entries are restored and executed according to the pre-command marking order, allowing the engine room to continue yaw adjustment after the risk has been eliminated.

[0106] In some embodiments of this application, updating the historical database includes:

[0107] After each yaw adjustment, wind speed, wind direction, power output, and correction angle were obtained at the sampling time before and after the yaw adjustment, and compiled into a before-and-after comparison record.

[0108] When compiling the before-and-after comparison records, height layer markers were added to each wind speed and wind direction information, forward distance segment markers were added to the collection location of each record, and sampling time markers were added to the sampling time point, so that each record simultaneously has height layer markers, forward distance segment markers, and sampling time markers.

[0109] The before-and-after comparison records with the additional markers are sequentially stored in the historical database. When storing, they are compared item by item with the existing records of the same height layer marker, forward distance segment marker and sampling time marker. The parts with the same content are merged and saved, and the parts with differences are added as new entries.

[0110] The updated historical database is arranged continuously according to the sampling time to obtain a record sequence, which is then used as the historical wind condition characteristic parameter for comparison with the current wind condition characteristic parameter when called again.

[0111] Specifically, after yaw adjustment, wind speed, wind direction, power output, and correction angle information from the sampling time before and after the adjustment are immediately acquired and compiled into before-and-after comparison records. During compilation, a height layer marker is added to each wind speed and wind direction information entry, a forward distance segment marker is added to the collection location of each record, and a sampling time marker is added to each sampling time point. This ensures that each record, when stored in the historical database, simultaneously possesses a height layer marker, a forward distance segment marker, and a sampling time marker, allowing for location in the longitudinal, lateral, and time axes. The before-and-after comparison records with added markers are sequentially stored in the historical database and compared item by item with existing records in the database that have the same height layer marker, forward distance segment marker, and sampling time marker. If the content is completely identical, it is merged and saved as a single entry to avoid redundancy; if there are differences, it is added as a new entry to ensure the database has a complete record of changes. After the update is completed, the historical database is arranged continuously by sampling time to obtain a record sequence, so that it can be directly compared with the current wind condition characteristic parameters in the next call, thus providing a stable reference basis. For example, during yaw correction, the engine room rotates from one direction to another. The wind direction information collected before and after the correction differs, and the power output and correction angle also change accordingly. These differences are marked in the before-and-after comparison records and stored in the historical database. After comparison with existing records, the differences are added as new entries, allowing for a complete understanding of the changes before and after the correction during subsequent calls, and serving as a crucial basis for generating the yaw correction path.

[0112] In another preferred embodiment based on the above embodiments, see [reference] Figure 2 As shown, this embodiment provides a lidar-based wind turbine control system for applying the aforementioned lidar-based wind turbine control method, including:

[0113] The wind condition sensing module collects wind speed and direction data at several altitude levels and several forward distances using a cabin laser wind radar, filters out blank or abrupt values, and obtains a layered raw wind condition dataset.

[0114] The data processing module performs noise reduction, smoothing, and interpolation on the layered wind condition raw dataset to form a continuous wind speed time series curve and wind direction change path. It also extracts wind condition characteristic parameters by combining power output, vibration signal, and gearbox data.

[0115] The yaw correction module compares wind condition characteristic parameters with historical wind condition characteristic parameters in the historical database, identifies yaw deviations under different wind speeds and directions, generates a yaw correction path, and triggers yaw adjustment when it detects continuous sudden changes in wind conditions or rapid changes in direction. The yaw adjustment includes the determination of the start time, rotation amplitude, and stopping conditions.

[0116] The feedforward control module, based on the yaw correction path, uses the wind trend of the future time window collected in advance by the laser wind radar to predict the wind speed and wind direction that will affect the impeller, generates the prediction result, and matches it with the current yaw state. When the prediction result does not match the existing direction, it generates a yaw pre-command and adjusts the response priority of the control system.

[0117] The safety control module, upon receiving the forecast results, uses wind condition characteristic parameters to determine the risk, identify gusts, wind shear and extreme wind conditions, and generates safety control commands when risk conditions are triggered. These commands include limiting the yaw adjustment frequency, reducing engine speed, or triggering emergency braking, which are used to cover or limit the yaw pre-commands.

[0118] The status update module records the wind speed difference, power change and correction angle before and after yaw adjustment after the safety control command is executed, updates the historical database, and uses it when the yaw correction module is called.

[0119] Understandably, the wind turbine control system uses a wind condition sensing module as its front end. It continuously collects wind speed and direction data at different altitudes and distances ahead using a nacelle laser anemometer. During the data collection phase, blank or abrupt values ​​are directly filtered out, forming a layered raw wind condition dataset with temporal continuity and spatial distribution. The data processing module further denoises, smooths, and interpolates the layered raw wind condition dataset to obtain a wind speed time-series curve reflecting wind speed evolution and a wind direction change path describing direction shifts. Combined with power output, vibration signals, and gearbox data collected during unit operation, wind condition characteristic parameters are extracted to form inputs characterizing the relationship between wind conditions and unit response. When the yaw correction module accesses the historical database, it compares the current wind condition characteristic parameters with historical wind condition characteristic parameters item by item to identify deviations between the nacelle orientation and the incoming wind conditions, generating a yaw correction path that includes start-up timing, rotation amplitude, and stopping conditions. When the system detects a continuous abrupt change in wind conditions or a rapid change in direction, this module triggers yaw adjustment to ensure that the nacelle's yaw direction corresponds to the wind condition changes. Based on the yaw correction path, the feedforward control module utilizes wind trend data collected in advance by laser wind radar over a future time window to generate a prediction result, which is then compared with the current yaw state. When the prediction result differs from the current yaw state, the system generates a yaw pre-command and adjusts the response priority of the control system, ensuring that yaw adjustment is completed before the incoming wind conditions actually affect the rotor, thereby reducing the risk of yaw lag. The safety control module, while receiving the prediction result, uses wind characteristic parameters to assess risk, identifying the presence of gusts, wind shear, or extreme wind conditions. When a risk condition is triggered, the system immediately generates a safety control command, overriding or limiting the current yaw pre-command. Safety control commands include limiting the yaw adjustment frequency, reducing engine speed, or triggering emergency braking to ensure safe unit operation under special wind conditions. After the safety control command is executed, the status update module compares the wind speed difference, power change, and correction angle before and after the yaw adjustment, compiles a before-and-after comparison record, and updates it to the historical database. After the database is updated, the system can use the new record to compare with the current wind conditions upon subsequent calls, achieving a closed-loop cycle of data accumulation and path correction. For example, when the wind perception module detects that the incoming wind direction is gradually changing from southeast to due east, the wind direction change path output by the data processing module shows a continued trend. The yaw correction module generates a corresponding yaw correction path and triggers yaw adjustment. Simultaneously, the feedforward control module confirms in the forecast results of the future time window that the wind speed still has an upward trend and generates a clockwise yaw pre-instruction. However, within the same sampling time, the safety control module identifies a sudden change in wind speed at an adjacent altitude layer. The system immediately generates a safety control instruction to limit the frequency of yaw adjustment, temporarily freezing the pre-instruction. After the risk is eliminated, the status update module writes the data records before and after this yaw adjustment into the historical database, providing a more complete reference for the next yaw correction.

[0120] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A wind turbine control method based on lidar, characterized in that, include: Wind speed and direction data were collected at several altitude levels and several forward distances using a cabin laser wind radar. Blank or abrupt values ​​were filtered out to obtain a layered raw wind condition dataset. The original layered wind condition dataset is denoised, smoothed, and interpolated to form a continuous wind speed time series curve and wind direction change path. Combined with power output, vibration signal, and gearbox data, wind condition characteristic parameters are extracted. By comparing wind condition characteristic parameters with historical wind condition characteristic parameters in the historical database, yaw deviations under different wind speed and direction conditions are identified, a yaw correction path is generated, and yaw adjustment is triggered when a continuous change in wind conditions or a rapid change in direction is detected. The yaw adjustment includes the determination of the start time, rotation amplitude, and stopping conditions. Based on the yaw correction path, the wind trend of the future time window is collected in advance by the laser wind radar to predict the wind speed and direction that will affect the impeller, generate the prediction result, and match it with the current yaw state. When the prediction result does not match the existing direction, a yaw pre-command is generated and the response priority of the control system is adjusted. When receiving the forecast results, the risk assessment is performed using wind condition characteristic parameters to identify gusts, wind shear and extreme wind conditions. When the risk conditions are triggered, safety control instructions are generated, including limiting the yaw adjustment frequency, reducing the speed of operation or triggering emergency braking, to cover or limit the yaw pre-instructions. After the safety control command is executed, record the wind speed difference, power change and correction angle before and after yaw adjustment, update the historical database, and use it when the yaw correction module is called.

2. The wind turbine control method based on lidar according to claim 1, characterized in that, When collecting wind speed and direction data, the following should be included: Wind speed and wind direction information are acquired sequentially at preset time intervals at several altitude levels. Establish a continuous sequence of sampling points across several forward distance segments and record the corresponding distance segment markers; The collected wind speed and wind direction information were checked item by item for sampling time and numerical range, and data points that did not conform to the continuity of time or the pattern of numerical change were removed. After removing the data points, subsequences are created within each height layer according to the distance segments ahead. Within each subsequence, the data points are arranged according to the sampling time. A height layer label, a distance segment label, and a time label are written for each data point. The time label is then associated with the height layer label and the distance segment label, so that the three labels correspond one-to-one in the same record, thus obtaining the original dataset of stratified wind conditions.

3. The wind turbine control method based on lidar according to claim 2, characterized in that, When extracting wind condition characteristic parameters, the following are included: The original stratified wind condition dataset was organized into a record sequence arranged according to sampling time, based on both height layer and forward distance segment. Time gaps in the record sequence are filled in as follows: when there is a recent valid record on only one side of the gap, the value of the recent valid record on that side is filled in according to the sampling time order until the end of the gap; when there are recent valid records on both sides of the gap, the gap is divided into several positions according to the sampling time interval, and the transition record between the record values ​​on both sides is filled in sequentially, and a time gap supplement mark is added to the supplemented record. Spatial gaps in the recorded sequence are filled in: when the same sampling time is missing in a certain forward distance segment, the nearest valid record in the adjacent forward distance segment and at the same height layer is used to fill in the gaps; when the same sampling time is missing in a certain height layer, the nearest valid record in the adjacent height layer and at the same forward distance segment is used to fill in the gaps, and a spatial gap filling mark is added to the supplemented record. After completing the temporal and spatial supplementation, the sequences are time-series aligned and spatially aligned to obtain the wind speed time-series curves and wind direction change paths.

4. The wind turbine control method based on lidar according to claim 3, characterized in that, After the original stratified wind condition dataset is supplemented and wind speed time-series curves and wind direction change paths are formed, the directly collected data is labeled as sampled data when stored in the historical database, and the data generated by interpolation or by referencing adjacent height layers and forward distance segments is labeled as supplementary data when stored in the historical database. This distinguishes the sampled data from the supplementary data in the historical database. At the same time, the power output, vibration signal, and gearbox data are organized into unit operation sequences, and a sampling time mark is added to each unit operation sequence when stored in the historical database. This establishes the correspondence between the unit operation sequence and the wind speed time-series curves and wind direction change paths under the same sampling time mark.

5. The wind turbine control method based on lidar according to claim 4, characterized in that, When generating the yaw correction path and implementing the yaw adjustment, the following are included: Based on the sampling time and spatial markers, historical wind condition characteristic parameter records that are similar to the current wind condition characteristic parameters are retrieved from the historical database to form a comparison set; the current wind condition characteristic parameters are compared with the comparison set item by item to determine the direction of the yaw deviation and the range that needs to be corrected, and a draft yaw correction path is formed. The draft yaw correction path is organized into path entries according to the sampling time sequence. The path entries include the start time, rotation direction, segment target and stopping condition, and an entry tag is written for each path entry. When the yaw deviation remains in the same direction within a continuous sampling period, or when there is a rapid change in wind direction, yaw adjustment is triggered and executed in the order of path entries; After each entry is completed, the current wind condition characteristic parameters are checked, and if necessary, unexecuted path entries are revised and then executed. The yaw adjustment ends when the stopping condition is met or the current wind direction is consistent with the cabin orientation; otherwise, the existing yaw state is maintained.

6. The wind turbine control method based on lidar according to claim 5, characterized in that, When generating prediction results and forming yaw pre-commands, the following are included: Starting from the current sampling time, the order of arrival is determined from farthest to closest distance segments. A future time window index is established, and wind speed and wind direction information corresponding to the index are extracted from each altitude layer to form an arrival order record. The arrival order records are compared between adjacent arrival order records. When the wind speed information is increasing, an increasing mark is written on the next record; when it is decreasing, a decreasing mark is written; when they remain close, a stable mark is written; when the wind direction information changes clockwise, a clockwise mark is written; when it changes counterclockwise, a counterclockwise mark is written. The changes in wind speed or wind direction information within adjacent sampling time intervals are marked with fast or slow tags. Based on the arrival order after the tags, the prediction result entries are recorded. The prediction result entries are compared segment by segment with the yaw correction path and cabin orientation. When the wind direction tag in the prediction result entry is consistent with the cabin orientation and the wind speed tag is consistent with the trend of the yaw correction path, it is recorded as a consistent relationship. When the wind direction tag in the prediction result entry deviates from the cabin orientation or the wind speed tag is opposite to the trend of the yaw correction path, it is recorded as an inconsistent relationship. When the inconsistent relationship is recorded for several consecutive sampling time intervals, a draft yaw pre-instruction is generated. The draft yaw pre-instruction is organized into yaw pre-instruction entries according to the arrival order. The yaw pre-instruction entries include the start time, rotation direction, segment target and stop condition, and are written with pre-instruction tags. The yaw pre-command entry is set as a pre-execution item of the control system. When the prediction result is consistent with the recovery of the existing yaw state, the yaw pre-command entry that has not yet been executed is withdrawn.

7. The wind turbine control method based on lidar according to claim 6, characterized in that, When conducting risk assessment and generating security instruction entries, the following are included: The system receives the prediction results and wind condition characteristic parameters at the same sampling time to form a risk check sequence. In the risk check sequence, when the wind speed information changes abruptly between records of two adjacent sampling times, it is recorded as a gust trigger segment. When the wind speed information changes abruptly between records of adjacent height layers or adjacent forward distance segments, it is recorded as a wind shear trigger segment. When the wind speed information and wind direction information change in different directions at the same sampling time, it is recorded as an extreme wind condition trigger segment. When one of the trigger segments is identified, the current yaw pre-command entry is compared with the record corresponding to the trigger segment item by item. When the rotation direction or start timing contained in the yaw pre-command entry is inconsistent with the record of the trigger segment, it is judged as a risk. When it is judged as a risk, a safety command entry is generated. The safety command entry includes emergency braking entry, speed reduction operation entry and yaw adjustment frequency restriction entry, and writes a safety flag according to the type of trigger segment. The safety instruction entries are set as the priority execution content of the control system. Unexecuted yaw pre-instruction entries are overridden or restricted. If no trigger segment is identified for several consecutive sampling times, the unexecuted safety instruction entries are withdrawn and the execution of yaw pre-instruction entries is resumed.

8. The wind turbine control method based on lidar according to claim 7, characterized in that, When dealing with situations where both safety instruction entries and yaw pre-instruction entries exist simultaneously, the following applies: When both safety command entries and yaw pre-command entries are detected at the same sampling time, the safety command entries are executed first, and the yaw pre-command entries that have not yet been executed are marked as frozen. In the frozen state, the yaw pre-command entry does not enter the execution process until the safety command entry is completed; Once the safety instruction entry has been executed, the freeze state will be lifted if no gust trigger segment, wind shear trigger segment, or extreme wind condition trigger segment is detected in subsequent consecutive sampling periods. After the freeze is lifted, the yaw pre-command entries that have not yet been executed are restored and re-entered for execution according to their existing pre-command marking order, so that yaw control can continue after the safety conditions are eliminated.

9. The wind turbine control method based on lidar according to claim 8, characterized in that, When updating the historical database, the following is included: After each yaw adjustment, wind speed, wind direction, power output, and correction angle were obtained at the sampling time before and after the yaw adjustment, and compiled into a before-and-after comparison record. When compiling the before-and-after comparison records, height layer markers were added to each wind speed and wind direction information, forward distance segment markers were added to the collection location of each record, and sampling time markers were added to the sampling time point, so that each record simultaneously has height layer markers, forward distance segment markers, and sampling time markers. The before-and-after comparison records with the additional markers are sequentially stored in the historical database. When storing, they are compared item by item with the existing records of the same height layer marker, forward distance segment marker and sampling time marker. The parts with the same content are merged and saved, and the parts with differences are added as new entries. The updated historical database is arranged continuously according to the sampling time to obtain a record sequence, which is then used as the historical wind condition characteristic parameter for comparison with the current wind condition characteristic parameter when called again.

10. A lidar-based wind turbine control system, used to implement the lidar-based wind turbine control method as described in any one of claims 1-9, characterized in that, include: The wind condition sensing module collects wind speed and direction data at several altitude levels and several forward distances using a cabin laser wind radar, filters out blank or abrupt values, and obtains a layered raw wind condition dataset. The data processing module performs noise reduction, smoothing, and interpolation on the layered wind condition raw dataset to form a continuous wind speed time series curve and wind direction change path. It also extracts wind condition characteristic parameters by combining power output, vibration signal, and gearbox data. The yaw correction module compares wind condition characteristic parameters with historical wind condition characteristic parameters in the historical database, identifies yaw deviations under different wind speeds and directions, generates a yaw correction path, and triggers yaw adjustment when it detects continuous sudden changes in wind conditions or rapid changes in direction. The yaw adjustment includes the determination of the start time, rotation amplitude, and stopping conditions. The feedforward control module, based on the yaw correction path, uses the wind trend of the future time window collected in advance by the laser wind radar to predict the wind speed and wind direction that will affect the impeller, generates the prediction result, and matches it with the current yaw state. When the prediction result does not match the existing direction, it generates a yaw pre-command and adjusts the response priority of the control system. The safety control module, upon receiving the forecast results, uses wind condition characteristic parameters to determine the risk, identify gusts, wind shear and extreme wind conditions, and generates safety control commands when risk conditions are triggered. These commands include limiting the yaw adjustment frequency, reducing engine speed, or triggering emergency braking, which are used to cover or limit the yaw pre-commands. The status update module records the wind speed difference, power change and correction angle before and after yaw adjustment after the safety control command is executed, updates the historical database, and uses it when the yaw correction module is called.

Images