Wind field data analysis system and method oriented to runway running direction optimization

Abstract

The invention relates to the field of big data analysis, and discloses a wind field data analysis system and method oriented to runway running direction optimization, and the method comprises the steps: carrying out the real-time block disassembly processing of an obtained high-speed meteorological record, and forming a wind field basic record set with scene specific time sequence deviation information; cross-position mutual verification operation is applied to the wind field basic record set, and multi-source wind field expression reflecting local airflow fine offset is generated; according to the multi-source wind field expression, key airflow conduction veins in the runway extension direction are identified by utilizing an interlayer screening and judging strategy, and a small-scale wind field disturbance form only appearing in a runway scene is locked; the small-scale wind field disturbance form is split again, and a local wind field composite unit reflecting the combined action of the space reconstruction error and the wind trend deflection in the runway direction is generated; and deducing a wind field instant judgment chain suitable for updating the running direction of the runway through a local wind field composite unit. The method has the advantage that the real-time performance of runway running direction optimization is improved.

Description

Technical Field

[0001] This invention relates to the field of big data analytics, specifically to a wind field data analysis system and method optimized for runway operation direction. Background Technology

[0002] As airport flight density continues to increase, operations management departments are increasingly relying on large-scale wind field data to dynamically optimize runway direction. However, the problem of fragmented storage of local wind field data between the runway abutment and mid-section, preventing continuous correlation analysis, remains unresolved. Taking a wind field monitoring system composed of multi-point lidar and ground-based anemometers as an example, the data formats, timestamp accuracy, and sampling periods generated by each device differ, resulting in cross-source heterogeneity, temporal misalignment, and missing spatial labels for key parameters such as wind speed, wind direction, and gust frequency in the data lake. In this specific scenario, when a slight increase in headwind occurs at the runway abutment while crosswind disturbances occur simultaneously in the runway mid-section, this slight local wind direction shift cannot be identified if temporal alignment and spatial reconstruction cannot be performed in a big data environment. Existing solutions generally use single-dimensional statistics or simple threshold judgments, which struggle to establish stable cross-location data correlations within massive wind field records, leading to risks of delays and misjudgments in runway direction recommendations. Therefore, it is essential to design a wind field data analysis system and method for runway direction optimization that improves the real-time performance of runway direction optimization. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a wind field data analysis system and method for runway direction optimization, which has the advantage of improving the real-time performance of runway direction optimization and solves the problems mentioned in the background technology.

[0004] To achieve the aforementioned goal of improving the real-time performance of runway direction optimization, this invention provides the following technical solution: a wind field data analysis method for runway direction optimization, comprising the following steps: The acquired high-speed meteorological records are processed in real time by splitting them into blocks, and short-term anomaly markers are identified during the splitting process to form a basic wind field record set with scene-specific time-series deviation information. Cross-location cross-verification operations are applied to the basic wind field record set, and historical wind direction abrupt changes, crosswind projection deflection, and low-frequency disturbance fragments are included in the comparison to generate a multi-source wind field expression that reflects subtle shifts in local airflow. Based on the multi-source wind field expression, the key airflow transmission network along the runway extension direction is identified by the inter-layer screening strategy, and the directional shift in different time slices is cross-examined to lock the small-scale wind field disturbance pattern that only appears in the runway scene. Small-scale wind field disturbance patterns are re-segmented according to their triggering range, duration, and impact on near-ground airflow distribution. Combined with wind field regions exhibiting stable, transitional, and turbulent characteristics within the same cycle, these are integrated into a local wind field composite unit that reflects the wind direction projection relationship along the runway. By using the local wind field composite unit, the key contact point for directional decision delay within the current operating window is located, and a wind field real-time judgment chain applicable to updating the runway operating direction is derived.

[0005] Preferably, the process of forming a basic wind field record set with scene-specific time-series deviation information is as follows: The continuously collected wind speed, wind direction, temperature, humidity and pressure data are divided into data blocks according to timestamps; For each data block, calculate the short-term standard deviation, slope change, and outlier density, and use a threshold to determine short-term outlier markers to generate a preliminary outlier index; Based on the preliminary anomaly index, local deviations within the data block are weighted and adjusted to generate a basic wind field record set with scene-specific time-series deviation information.

[0006] Preferably, the process of applying cross-location cross-verification operations to the basic wind field record set is as follows: Using the basic wind field record set as input, time alignment is performed on the records corresponding to different observation points, radar stations, and ground sensors. On the aligned records, spatial interpolation and distance weighting methods are used to verify the data at adjacent locations, calculate the consistency index of wind speed and wind direction between each location, and identify deviation characteristics. Based on the verification results, a cross-location consistency matrix is ​​generated, and combined with historical meteorological models, the contribution weight of each location to runway wind direction decision is calculated.

[0007] Preferably, the process of generating a multi-source wind field representation that reflects subtle shifts in local airflow is as follows: Retrieve wind direction change events and low-frequency disturbance segments from historical meteorological databases, and extract key waveform features and disturbance amplitude indicators; Historical disturbance features are matched with cross-location consistency matrices for similarity, and local airflow micro-offset vectors are calculated. By weighted superposition of local airflow micro-offset vectors, weighted mapping of contribution weights, and tensor quantization, a multi-source wind field expression that comprehensively reflects the subtle offset of local airflow is formed.

[0008] Preferably, the process of identifying key airflow transmission patterns along the runway extension direction using an interlayer screening strategy is as follows: The multi-source wind field representation is divided into hierarchical data sets according to height level, distance segment, and time slice; Local gradient calculation and time evolution analysis are performed on each layer of data set to identify abrupt airflow change nodes and offset extension directions, and a weighted evaluation is performed by combining the contribution weights of each location. Based on the accumulation of hierarchical gradients and the consistency of offset, key airflow transmission patterns extending along the runway are screened.

[0009] Preferably, the process of identifying small-scale wind field disturbance patterns that only appear in the runway scenario is as follows: The offset vectors and wind speed gradient features in multiple consecutive time slices are organized according to the time series, and combined with the key airflow transmission patterns, the offset direction and amplitude in adjacent time slices are cross-compared and trend analyzed. Within the key airflow transmission network, data from different observation points and historical disturbance paths are compared synchronously to calculate the local disturbance consistency index and abnormal increase coefficient, and to assess the continuity and abruptness of the disturbance in the time dimension. Thresholding is applied to offset vectors and gradient features within the time series and key context, retaining only small-scale wind field disturbances that exhibit instantaneous changes and are consistent in the runway scenario.

[0010] Preferably, the process of re-decomposing small-scale wind field disturbance patterns according to their triggering range, duration, and impact on near-surface airflow distribution is as follows: Small-scale wind field disturbance patterns are divided into local disturbance units, semi-local disturbance units, and cross-segment disturbance units according to their triggering range. Based on the duration and evolution characteristics of the disturbance, each disturbance unit is further divided into the initial stage, the evolution stage, and the dissipation stage. Based on the way disturbances affect the distribution of near-ground airflow, disturbance units at different stages are labeled as stable, transitional, and turbulent.

[0011] Preferably, the process of integrating local wind field composite units to reflect the relationship of wind direction projection along the runway direction is as follows: Spatiotemporal matching and comparison of wind field regions at different disturbance stages within the same cycle are performed to identify their continuity and changing trends. The disturbance patterns, spatial locations, and influence on near-ground airflow direction at different stages are encoded into multi-dimensional feature vectors. Through quantitative fusion and local weighted accumulation, a local wind field composite unit is formed that comprehensively reflects the relationship between wind direction projection along the runway direction.

[0012] Preferably, the process of deriving the wind field instantaneous decision chain suitable for updating the runway running direction is as follows: Multi-dimensional feature scanning is performed on local wind field composite units to identify key points of directional offset hysteresis. Based on the contact point location, offset amplitude, and disturbance stage characteristics, a local wind direction determination chain is constructed; By combining historical wind direction change patterns and multi-source risk contribution indicators, the decision chain is dynamically optimized and reconstructed with weights to derive a real-time wind field decision chain for updating runway operating direction.

[0013] A wind field data analysis system optimized for runway running direction includes: Block processing module: The high-speed meteorological records are broken down into blocks, short-term anomaly markers are extracted, and a basic wind field record set with time-series deviation information is formed; Data comparison module: It incorporates historical wind direction changes, crosswind projection deflection, and low-frequency disturbance segments into the comparison to generate a multi-source wind field expression that reflects subtle shifts in local airflow. The network identification module uses inter-layer screening strategies to analyze the expression of multi-source wind fields, identify key airflow transmission networks along the runway extension direction, and lock in the small-scale wind field disturbance patterns. Disturbance integration module: It breaks down small-scale disturbance patterns and combines them with stable, transitional and disordered regions to generate local wind field composite units that reflect spatial reconstruction errors and wind deflection. Real-time decision module: locates the key points of directional decision delay and derives the real-time decision chain for adjusting the runway's running direction.

[0014] Compared with existing technologies, the present invention provides a wind field data analysis system and method for runway operation direction optimization, which has the following beneficial effects: This invention achieves high timeliness and refined capture of wind field data by real-time segmentation of high-speed meteorological records and identification of short-term anomaly markers. The resulting wind field baseline record set possesses scene-specific temporal deviation information, accurately reflecting local wind field changes along the runway. Through cross-location verification and comparison of historical wind direction abrupt changes, crosswind projection deflection, and low-frequency disturbance segments, a multi-source wind field representation is generated. This achieves high-resolution representation of minute shifts, local disturbances, and low-altitude wind direction changes, overcoming the shortcomings of traditional wind field analysis methods in identifying short-term, localized wind speed abrupt changes. By employing an inter-layer screening strategy to identify key airflow transmission patterns along the runway extension direction and combining time-series cross-examination to pinpoint small-scale disturbances, the method can accurately identify transient disturbance events occurring only in the runway scene, thus providing sensitive node references for aircraft takeoff and landing safety. This method breaks down disturbance patterns by trigger range, duration, and impact on near-ground airflow, and integrates these patterns with stable, transitional, and turbulent characteristics within the same period. This results in a local wind field composite unit reflecting the wind direction projection along the runway, enabling multi-dimensional quantitative expression and dynamic modeling of the wind field's spatiotemporal structure. Based on the local composite unit, key points of directional decision delay are located, and an instantaneous wind field judgment chain is derived. This allows runway direction updates to respond in real-time to current wind field changes, balancing short-term abrupt changes and continuous trends, significantly improving the accuracy and safety of runway scheduling decisions. With high precision, high timeliness, and quantifiability, this method effectively reduces flight delays and safety risks caused by local wind field disturbances. It is suitable for real-time airport runway operation optimization and low-altitude wind field monitoring, providing reliable and operable technical support for aviation operations management. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the method of the present invention; Figure 2 This is a schematic diagram of the structure of the present invention. Detailed Implementation

[0016] 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.

[0017] Example 1: Please refer to Figure 1 As shown, an embodiment of the present invention provides a wind field data analysis method for runway direction optimization, comprising the following steps: S1: Perform real-time block splitting processing on the acquired high-speed meteorological records, and identify short-term anomaly markers during the block splitting process to form a basic wind field record set with scene-specific time-series deviation information.

[0018] The process of forming a basic wind field record set with scene-specific time-series deviation information in S1 is as follows: The continuously collected wind speed, wind direction, temperature, humidity, and pressure data are segmented into data blocks according to timestamps. Parameters such as wind speed, wind direction, temperature, humidity, and atmospheric pressure are continuously sampled by multiple meteorological sensors installed at the runway and surrounding areas. The sampling frequency can be set to 1 to 5 times per second to ensure the capture of short-term micro-disturbances. The collected data are arranged sequentially according to timestamps and divided into separate data blocks according to preset time windows to facilitate the analysis of short-term fluctuations and anomaly detection. The data blocks can contain all meteorological parameters of each observation point to ensure the completeness of the record for each time window. For each data block, calculate the short-term standard deviation, slope change, and outlier density, and use a threshold to mark short-term anomalies, generating a preliminary anomaly index. Within each time window of the data block, calculate the short-term standard deviation for parameters such as wind speed and wind direction to reflect the instantaneous fluctuation amplitude of the wind field. Calculate the slope of changes at continuous time points to capture short-term trends such as sharp wind direction deviations or sudden increases in wind speed. Assess the intensity of local disturbances by statistically analyzing outlier density (e.g., the proportion of data points deviating from the average by more than 2 standard deviations). Compare the standard deviation, slope change, and outlier density with preset thresholds. If the threshold is exceeded, it is marked as a short-term anomaly. Based on the anomaly distribution of each data block, generate a preliminary anomaly index. Based on the preliminary anomaly index, local deviations within data blocks are weighted and adjusted to generate a basic wind field record set with scene-specific time-series deviation information. Within each data block, a weighted smoothing method is used to adjust the time points or parameters marked as anomalies. The weighting is allocated according to the severity of the anomaly and the stability of neighboring time points. For example, the larger the anomaly amplitude, the lower the weight, and the higher the weight of neighboring normal data. This corrects local deviations without weakening the real wind field changes. Combined with moving average or exponential weighted filtering algorithms, the wind field data of each data block is smoothed while retaining instantaneous disturbance characteristics. By processing all data blocks, a basic wind field record set with scene-specific time-series deviation information is formed, reflecting the short-term wind field dynamics of the runway and surrounding areas, while eliminating acquisition noise and anomaly interference.

[0019] S2: Apply cross-location cross-verification operations to the basic wind field record set, and include historical wind direction abrupt changes, crosswind projection deflection, and low-frequency disturbance fragments in the comparison to generate a multi-source wind field expression that reflects subtle shifts in local airflow.

[0020] The process of applying cross-location cross-verification operations to the basic wind field record set in S2 is as follows: Using the basic wind field record set as input, the records corresponding to different observation points, radar stations, and ground sensors are time-aligned. Data such as wind speed, wind direction, temperature, and humidity collected from multi-source meteorological observation equipment (including ground sensors, wind towers, and radar stations) in the runway and surrounding areas are uniformly summarized. Since there may be slight differences in the sampling time of different devices, the records are precisely aligned according to the timestamp to achieve a synchronous data benchmark across devices and locations. The alignment method can adopt linear time interpolation or high-precision clock synchronization strategy to ensure that the data of each observation point within the same analysis time window can be directly compared and to avoid wind field misjudgment caused by time offset. On the aligned records, spatial interpolation and distance weighting methods are used to verify the data of adjacent locations, calculate the consistency index of wind speed and wind direction between each location, and identify deviation characteristics. The distance between adjacent points is calculated based on the geographical coordinates of each observation point, and continuous spatial estimates are generated using inverse distance weighting or Kriging interpolation methods. At the same time, the interpolation weights are adjusted in combination with the influence of terrain and obstacles. By comparing the actual observed data with the interpolated predicted values, the wind speed difference and wind direction angle deviation are calculated. For areas with large deviations, their local disturbance characteristics are further analyzed, such as anomalous micro-disturbances such as short-term wind direction changes or sudden increases in wind speed. Based on the verification results, a cross-location consistency matrix is ​​generated, and combined with historical meteorological models, the contribution weight of each location to runway wind direction decision is calculated. The consistency indices between the observation points are organized into a cross-location consistency matrix, where each matrix element represents the degree of matching and relative deviation of wind speed and direction between two points. This matrix is ​​compared and analyzed with a historical meteorological model database, including historical crosswind events, local wind direction abrupt changes, and low-frequency disturbance segments. By combining the matrix consistency value with the matching degree of historical events, the contribution weight of each observation point to runway wind direction decision can be calculated. These weights reflect the reference value of each location's data within the current operating window. The weights are applied to dynamically adjust the sampling frequency, data priority, or calculation priority of wind field data, thereby achieving real-time and reliable wind direction decision support within a big data analysis framework.

[0021] The process of generating a multi-source wind field representation reflecting subtle shifts in local airflow in S2 is as follows: Retrieve wind direction change events and low-frequency disturbance segments from historical meteorological databases, and extract key waveform features and disturbance amplitude indicators. Utilize historical meteorological databases accumulated by airport or regional meteorological monitoring systems to screen out key meteorological events that occurred in the runway and surrounding areas in the past, such as sudden wind direction deviation events, crosswind anomalies, and low-frequency wind field disturbance segments. For each historical event, use signal processing methods to extract key waveform features, such as the instantaneous slope of wind direction change, peak wind speed, disturbance duration, and frequency components. At the same time, calculate disturbance amplitude indicators, including local wind speed gradient, deviation amplitude, and disturbance energy magnitude. Historical disturbance features are matched with cross-location consistency matrices for similarity, and local airflow micro-offset vectors are calculated. Using the weighted correlation coefficient method, each historical disturbance feature is compared with real-time cross-location data point by point to identify time slices and spatial locations with high matching degree. After matching, local airflow micro-offset vectors are calculated through vector operations, including wind direction deviation, local increase or decrease in wind speed, and local disturbance propagation direction. The micro-offset vectors not only reflect the real-time airflow offset amplitude, but also reflect the transmission path of disturbances between different observation points. By weighted superposition of local airflow micro-offset vectors, weighted mapping of contribution weights, and tensor representation, a multi-source wind field expression that comprehensively reflects subtle local airflow offsets is formed. Each micro-offset vector is multiplied by its corresponding contribution weight to form a weighted vector. The weighted vectors are accumulated and superimposed in the spatial and temporal dimensions to obtain a comprehensive index of the impact of local disturbances. The accumulated result is mapped into a tensor structure, including the temporal dimension, spatial location dimension, and disturbance amplitude dimension, for multi-level analysis and big data processing. The resulting tensor is the multi-source wind field expression, which can comprehensively reflect the subtle local airflow offsets, disturbance intensity, and spatial distribution, providing a high-precision, real-time available wind field data foundation for runway operation direction optimization.

[0022] S3: Based on the multi-source wind field expression, the key airflow transmission network along the runway extension direction is identified using the inter-layer screening strategy, and the directional shift in different time slices is cross-examined to lock the small-scale wind field disturbance pattern that only appears in the runway scene.

[0023] The process of identifying key airflow transmission patterns along the runway extension direction using an inter-layer screening strategy in S3 is as follows: The multi-source wind field is divided into a hierarchical data set according to height layers, distance segments, and time slices. Raw wind field data such as wind speed, wind direction, temperature, and humidity are obtained from various monitoring devices deployed along the runway. Based on the actual runway length and height range, the wind field data is divided into several height layers vertically (e.g., one layer every 10 meters), into fixed distance segments along the runway direction (e.g., one segment every 50 meters), and into time slices according to the sampling time interval (e.g., one second or one minute). This generates a multi-dimensional hierarchical data set that not only preserves the spatial distribution characteristics of the wind field but also reflects the dynamic changes over time. Local gradient calculation and temporal evolution analysis are performed on each layer of data set to identify abrupt airflow change nodes and offset extension directions. A weighted evaluation is then conducted, incorporating the contribution weights of each location. For each height layer and distance segment, spatial gradient operators are used to calculate the local rate of change of wind speed and direction to identify abrupt airflow change nodes—locations of sudden wind speed increases, abrupt wind direction changes, or intensified turbulence. In the temporal dimension, time series analysis is performed on continuous time slices of data to calculate the instantaneous rate of change and cumulative trend of wind field parameters, thereby determining the extension direction and evolution trend of airflow. Location contribution weights are introduced, which can be determined based on the node's distance from the runway centerline, local wind field intensity, or historical influence factors. By integrating spatial and temporal information through weighted evaluation, the identified key nodes not only reflect local abrupt changes but also embody the overall impact on runway operational safety. Based on the layered gradient accumulation and offset consistency judgment, key airflow transmission patterns extending along the runway are screened. After completing the local gradient calculation and node weight evaluation, the gradient information at each altitude layer and distance segment is accumulated to quantify the transmission intensity of airflow in space and time. The offset direction of adjacent nodes along the runway is analyzed for consistency, that is, continuous node sequences with high gradient accumulation and consistent wind direction extension trends are screened to form a preliminary airflow transmission path. Combining the continuity requirements from the runway inlet to the terminal and the node contribution weight, the preliminary path is pruned and optimized to remove abnormal nodes with short-term or weak influence, thereby finally determining the key airflow transmission patterns extending along the runway.

[0024] The process of locking the small-scale wind field disturbance pattern that only appears in the runway scene in S3 is as follows: The offset vectors and wind speed gradient features in multiple consecutive time slices are organized into a time series, and combined with the key airflow transmission network, the offset direction and amplitude in adjacent time slices are cross-compared and trend analyzed. The wind speed, wind direction and spatial gradient information of consecutive time slices are obtained from multi-source wind field monitoring equipment, and these data are organized into a time series structure. For the wind field offset vector in each time slice, combined with the key airflow transmission network identified above, the offset direction, amplitude and gradient changes of each node are cross-compared with adjacent time slices. For example, the angle difference and amplitude difference analysis are performed on the node offset vectors of consecutive time slices to identify wind field changes with significant abrupt changes or consistent extension trends in the time series, forming a preliminary disturbance trend assessment. Within the key airflow transmission network, data from different observation points and historical disturbance paths are synchronously compared to calculate local disturbance consistency indices and anomalous amplification coefficients, assessing the continuity and abrupt changes of disturbances over time. For each node within the key airflow transmission network, real-time wind field data from different observation points are extracted and compared with historical similar disturbance paths to form a benchmark. By synchronously comparing the offset vectors and wind speed gradients of each node in space and time, local disturbance consistency indices are calculated, such as the similarity of offset directions and amplitude correlation between different observation points, as well as the anomalous amplification coefficient, i.e., the ratio of instantaneous gradient or wind speed change to the historical average level. Through these indices, the continuity and abrupt changes of disturbances over time are quantified, thereby screening out the small-scale disturbance events most sensitive to runway operation. Thresholding is performed on offset vectors and gradient features within the time series and key context range, retaining only small-scale wind field disturbance patterns that exhibit instantaneous abrupt changes and are consistent in the runway scenario. After completing the consistency and abrupt change assessment, in order to accurately locate small-scale disturbances in the runway scenario, thresholding is performed on offset vectors and gradient features. The threshold is determined based on historical wind field statistical characteristics and runway safety requirements. For example, nodes with wind speed gradients exceeding 1.5 times the historical average and offset directions consistent with the extension direction of the key context are judged as valid disturbances. Only disturbance patterns that meet the instantaneous abrupt change conditions, have strong temporal continuity, and spatial distribution consistent with the runway direction are retained, while noise and non-critical changes are eliminated. This yields a set of highly correlated small-scale wind field disturbances in the runway scenario, which can be used for real-time monitoring, flight safety warnings, or runway usage scheduling decisions.

[0025] S4: The small-scale wind field disturbance pattern is re-segmented according to the triggering range, duration stage and the way it affects the distribution of near-ground airflow, and combined with the wind field regions that exhibit stable, transitional and turbulent characteristics within the same period, and integrated into a local wind field composite unit to reflect the relationship of wind direction projection along the runway direction.

[0026] The process of re-decomposing small-scale wind field disturbance patterns in S4 according to their triggering range, duration, and impact on near-surface airflow distribution is as follows: Small-scale wind field disturbance patterns are divided into local disturbance units, semi-local disturbance units, and cross-segment disturbance units according to their triggering range. Based on the locked small-scale wind field disturbance patterns, the spatial impact range of each disturbance event is quantified. For example, by calculating the spatial diffusion radius of the disturbance node and its surrounding wind field offset vector, the influence range is divided into local disturbance units if it is less than 1 / 3 of the runway length, semi-local disturbance units if it is between 1 / 3 and 2 / 3 of the runway length, and cross-segment disturbance units if it exceeds 2 / 3 of the runway length. Combining the disturbance intensity weight and key airflow transmission patterns, the division results are ensured to reflect both the spatial diffusion range and the potential impact on the runway wind field. Based on the duration and evolution characteristics of the disturbance, each disturbance unit is further divided into the initial stage, the evolution stage, and the dissipation stage. After determining the disturbance unit, the evolution of each disturbance event in the time dimension is analyzed. The change curve of the disturbance intensity is calculated based on the time series data. The period when the disturbance first appears and the intensity rises rapidly is marked as the initial stage. The stage when the intensity remains high or continues to increase and shows a spatial diffusion trend is marked as the evolution stage. The stage when the disturbance intensity gradually weakens, the spatial diffusion shrinks, and until it disappears is marked as the dissipation stage. Combined with threshold judgment (such as the duration period when the wind speed gradient or offset vector exceeds 1.2 times the historical average), the division of each stage is ensured to conform to the actual disturbance evolution characteristics. Based on the way disturbances affect near-ground airflow distribution, disturbance units at different stages are labeled as stable, transitional, and turbulent. The impact pattern of each disturbance unit on the near-ground wind field at each time stage is analyzed. If the disturbance mainly causes slight shifts or slight increases in local wind speed, and the wind direction shift is basically consistent with the key airflow transmission path, it can be labeled as stable. If the disturbance causes significant changes in local wind speed and short-term inconsistencies in the shift direction, but still extends along the transmission path overall, it is labeled as transitional. If the disturbance causes violent fluctuations in wind speed, variable shift direction, or the formation of turbulent masses, causing significant turbulence in near-ground airflow distribution, it is labeled as turbulent. In the labeling process, the local disturbance consistency index and the abnormal increase coefficient are combined to achieve a quantitative description of the actual impact of disturbances on the runway wind field, thereby forming a complete disturbance unit classification system.

[0027] The process of integrating S4 into a local wind field composite unit to reflect the wind direction projection relationship along the runway direction is as follows: Spatiotemporal matching and comparison of wind field regions at different disturbance stages within the same period are performed to identify their continuity and trends. Information on each disturbance unit in time and space is extracted from the disturbance decomposition and classification results, including time slices, spatial locations, and impact ranges of each stage of initiation, evolution, and dissipation. Spatiotemporal matching of each disturbance stage within the same period (e.g., runway operation cycle or observation cycle) is performed. By comparing the characteristics of adjacent stages in terms of location, intensity, wind speed shift direction, and gradient changes, the continuity of disturbance events in the time dimension and their spatial extension trends are identified. For example, by calculating the trajectory shift and intensity change curve of the disturbance center point, it can be determined whether the disturbance extends along the runway direction or deflects, thereby establishing a correlation matrix between disturbance stages. The disturbance morphology, spatial location, and impact on near-ground airflow direction at different stages are encoded into multi-dimensional feature vectors. After completing the matching and trend analysis of disturbance stages, the key information of each disturbance stage is encoded into a multi-dimensional feature vector. The feature vector may include disturbance type (stable, transitional, turbulent), spatial coordinates (x, y, z), disturbance intensity (wind speed gradient, offset amplitude), directional impact (angle of offset direction relative to runway direction), duration, and stage identifier (initial, evolution, dissipation). Through this encoding method, the spatiotemporal characteristics of each disturbance stage and its impact on airflow near the runway can be expressed quantitatively and structurally, providing standardized input for cross-stage and cross-disturbance unit data fusion. By employing tensor fusion and local weighted accumulation, a local wind field composite unit is formed that comprehensively reflects the wind direction projection relationship along the runway direction. The eigenvectors are organized into a three-dimensional or multi-dimensional tensor structure according to the disturbance stage and spatial location to reflect the composite relationship of disturbances in time, space, and type. The influence of each disturbance stage in the tensor is locally weighted and accumulated. The weights are determined based on the disturbance intensity, contribution to key airflow transmission patterns, and stage duration. For example, higher weights are given when the evolution stage has a significant impact on airflow direction, while lower weights are given when the initial stage has a weaker impact. Through tensor fusion and weighted accumulation, a local wind field composite unit is formed, which can comprehensively reflect the wind direction projection relationship along the runway direction, providing quantitative references for wind field prediction, runway operation optimization, and safety management.

[0028] S5: Through the local wind field composite unit, locate the key contact point of directional decision delay within the current operating window, and derive the wind field real-time judgment chain applicable to updating the runway operating direction.

[0029] The process of deriving the wind field instantaneous decision chain applicable to updating runway running direction in S5 is as follows: Multidimensional feature scanning is performed on the local wind field composite unit to identify key touchpoints of directional offset delay. The generated local wind field composite unit is scanned in multiple dimensions, including spatial location, wind speed gradient, offset vector, disturbance stage, and the way it affects the near-ground airflow direction. Through node-by-node analysis, the degree of deviation and delay time of wind direction change from the runway direction are calculated. Nodes where the wind direction deviates from the runway direction but does not recover immediately are marked as key touchpoints. Combined with sliding time window analysis, the delayed response after instantaneous change is captured to ensure that the wind field location most sensitive to the runway running direction can be identified. Based on the contact point location, offset amplitude, and disturbance stage characteristics, a local wind direction determination chain is constructed. After identifying key contact points, the contact points are arranged in spatial order along the runway direction. Combining the offset amplitude and disturbance stage (initial, evolution, and dissipation) characteristics of each contact point, a local wind direction determination chain is established. The determination chain is represented as a sequence of nodes along the runway direction. Each node includes the contact point location, instantaneous offset amplitude, disturbance duration, and influence weight. Through the continuity of the chain and the analysis of node contribution, the projection trend of wind direction changes on the local runway segment is preliminarily assessed, thereby providing a quantitative basis for real-time runway operation direction adjustment. By combining historical wind direction change patterns and multi-source risk contribution indicators, the decision chain is dynamically optimized and reconstructed with weights to derive an instant wind field decision chain for updating runway operating direction. After completing the construction of the local decision chain, it is compared with historical wind direction change data, including typical wind direction evolution patterns, seasonal changes, and recorded abnormal wind field events. Multi-source risk contribution indicators, such as sensitive nodes at the runway inlet, low-altitude turbulence influence areas, and aircraft take-off and landing safety thresholds, are introduced to dynamically weight the chain nodes. Through iterative optimization and chain reconstruction, the contribution of key touchpoints to runway operating direction determination is enhanced, while the impact of non-critical or abnormal disturbances is weakened. This results in the derivation of an instant wind field decision chain that comprehensively reflects the current local wind field conditions, enabling real-time updates of runway operating direction and safety decision support.

[0030] Example 2: As Figure 2 As shown, a wind field data analysis system for runway direction optimization is characterized by comprising: Block processing module: The high-speed meteorological records are broken down into blocks, short-term anomaly markers are extracted, and a basic wind field record set with time-series deviation information is formed; Data comparison module: It incorporates historical wind direction changes, crosswind projection deflection, and low-frequency disturbance segments into the comparison to generate a multi-source wind field expression that reflects subtle shifts in local airflow. The network identification module uses inter-layer screening strategies to analyze the expression of multi-source wind fields, identify key airflow transmission networks along the runway extension direction, and lock in the small-scale wind field disturbance patterns. Disturbance integration module: It breaks down small-scale disturbance patterns and combines them with stable, transitional and disordered regions to generate local wind field composite units that reflect spatial reconstruction errors and wind deflection. Real-time decision module: locates the key points of directional decision delay and derives the real-time decision chain for adjusting the runway's running direction.

[0031] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0032] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A wind field data analysis method for runway running direction optimization, characterized in that, Includes the following steps: The acquired high-speed meteorological records are processed in real time by splitting them into blocks, and short-term anomaly markers are identified during the splitting process to form a basic wind field record set with scene-specific time-series deviation information. Cross-location cross-verification operations are applied to the basic wind field record set, and historical wind direction abrupt changes, crosswind projection deflection, and low-frequency disturbance fragments are included in the comparison to generate a multi-source wind field expression that reflects subtle shifts in local airflow. Based on the multi-source wind field expression, the key airflow transmission network along the runway extension direction is identified by the inter-layer screening strategy, and the directional shift in different time slices is cross-examined to lock the small-scale wind field disturbance pattern that only appears in the runway scene. Small-scale wind field disturbance patterns are re-segmented according to their triggering range, duration, and impact on near-ground airflow distribution. Combined with wind field regions exhibiting stable, transitional, and turbulent characteristics within the same cycle, these are integrated into a local wind field composite unit that reflects the wind direction projection relationship along the runway. By using the local wind field composite unit, the key contact point for directional decision delay within the current operating window is located, and a wind field real-time judgment chain applicable to updating the runway operating direction is derived.

2. The wind field data analysis method for runway running direction optimization according to claim 1, characterized in that, The process of creating a basic wind field record set with scene-specific time-series deviation information is as follows: The continuously collected wind speed, wind direction, temperature, humidity and pressure data are divided into data blocks according to timestamps; For each data block, calculate the short-term standard deviation, slope change, and outlier density, and use a threshold to determine short-term outlier markers to generate a preliminary outlier index; Based on the preliminary anomaly index, local deviations within the data block are weighted and adjusted to generate a basic wind field record set with scene-specific time-series deviation information.

3. The wind field data analysis method for runway running direction optimization according to claim 2, characterized in that, The process of applying cross-location cross-verification operations to the basic wind field record set is as follows: Using the basic wind field record set as input, time alignment is performed on the records corresponding to different observation points, radar stations, and ground sensors. On the aligned records, spatial interpolation and distance weighting methods are used to verify the data at adjacent locations, calculate the consistency index of wind speed and wind direction between each location, and identify deviation characteristics. Based on the verification results, a cross-location consistency matrix is ​​generated, and combined with historical meteorological models, the contribution weight of each location to runway wind direction decision is calculated.

4. The wind field data analysis method for runway running direction optimization according to claim 3, characterized in that, The process of generating a multi-source wind field representation that reflects subtle shifts in local airflow is as follows: Retrieve wind direction change events and low-frequency disturbance segments from historical meteorological databases, and extract key waveform features and disturbance amplitude indicators; Historical disturbance features are matched with cross-location consistency matrices for similarity, and local airflow micro-offset vectors are calculated. By weighted superposition of local airflow micro-offset vectors, weighted mapping of contribution weights, and tensor quantization, a multi-source wind field expression that comprehensively reflects the subtle offset of local airflow is formed.

5. The wind field data analysis method for runway running direction optimization according to claim 4, characterized in that, The process of identifying key airflow transmission patterns along the runway extension direction using an inter-layer screening strategy is as follows: The multi-source wind field representation is divided into hierarchical data sets according to height level, distance segment, and time slice; Local gradient calculation and time evolution analysis are performed on each layer of data set to identify abrupt airflow change nodes and offset extension directions, and a weighted evaluation is performed by combining the contribution weights of each location. Based on the accumulation of hierarchical gradients and the consistency of offset, key airflow transmission patterns extending along the runway are screened.

6. The wind field data analysis method for runway running direction optimization according to claim 5, characterized in that, The process of identifying small-scale wind field disturbance patterns that only appear in the runway scenario is as follows: The offset vectors and wind speed gradient features in multiple consecutive time slices are organized according to the time series, and combined with the key airflow transmission patterns, the offset direction and amplitude in adjacent time slices are cross-compared and trend analyzed. Within the key airflow transmission network, data from different observation points and historical disturbance paths are compared synchronously to calculate the local disturbance consistency index and abnormal increase coefficient, and to assess the continuity and abruptness of the disturbance in the time dimension. Thresholding is applied to offset vectors and gradient features within the time series and key context, retaining only small-scale wind field disturbances that exhibit instantaneous changes and are consistent in the runway scenario.

7. The wind field data analysis method for runway running direction optimization according to claim 6, characterized in that, The process of re-decomposing small-scale wind field disturbance patterns according to their triggering range, duration, and impact on near-surface airflow distribution is as follows: Small-scale wind field disturbance patterns are divided into local disturbance units, semi-local disturbance units, and cross-segment disturbance units according to their triggering range. Based on the duration and evolution characteristics of the disturbance, each disturbance unit is further divided into the initial stage, the evolution stage, and the dissipation stage. Based on the way disturbances affect the distribution of near-ground airflow, disturbance units at different stages are labeled as stable, transitional, and turbulent.

8. The wind field data analysis method for runway running direction optimization according to claim 7, characterized in that, The process of integrating local wind field composite units to reflect the projection relationship of wind direction along the runway direction is as follows: Spatiotemporal matching and comparison of wind field regions at different disturbance stages within the same cycle are performed to identify their continuity and changing trends. The disturbance patterns, spatial locations, and influence on near-ground airflow direction at different stages are encoded into multi-dimensional feature vectors. Through quantitative fusion and local weighted accumulation, a local wind field composite unit is formed that comprehensively reflects the relationship between wind direction projection along the runway direction.

9. A wind field data analysis method for runway running direction optimization according to claim 8, characterized in that, The process of deriving the wind field instant decision chain applicable to updating runway running direction is as follows: Multi-dimensional feature scanning is performed on local wind field composite units to identify key points of directional offset hysteresis. Based on the contact point location, offset amplitude, and disturbance stage characteristics, a local wind direction determination chain is constructed; By combining historical wind direction change patterns and multi-source risk contribution indicators, the decision chain is dynamically optimized and reconstructed with weights to derive a real-time wind field decision chain for updating runway operating direction.

10. A wind field data analysis system for runway direction optimization, applied to the method described in any one of claims 1-9, characterized in that, include: Block processing module: The high-speed meteorological records are broken down into blocks, short-term anomaly markers are extracted, and a basic wind field record set with time-series deviation information is formed; Data comparison module: It incorporates historical wind direction changes, crosswind projection deflection, and low-frequency disturbance segments into the comparison to generate a multi-source wind field expression that reflects subtle shifts in local airflow. The network identification module uses inter-layer screening strategies to analyze the expression of multi-source wind fields, identify key airflow transmission networks along the runway extension direction, and lock in the small-scale wind field disturbance patterns. Disturbance integration module: It breaks down small-scale disturbance patterns and combines them with stable, transitional and disordered regions to generate local wind field composite units that reflect spatial reconstruction errors and wind deflection. Real-time decision module: locates the key points of directional decision delay and derives the real-time decision chain for adjusting the runway's running direction.

Images