Wind speed uniformity measuring method based on wind speed sensor

Abstract

The invention discloses a wind speed uniformity measurement method based on a wind speed sensor, and relates to the technical field of environmental test equipment detection, and the method comprises the steps: deploying a sensor array in a to-be-detected region of an environmental test box, synchronously collecting original data, and carrying out the detection and elimination of an abnormal purified flow; aligning the data to the same time slice by a timestamp to form a time-space synchronization data set; constructing a wind speed matrix, and calculating a wind speed mean value and a sensor absolute deviation of each time slice; summarizing the deviations to generate a deviation sequence of each sensor, and performing statistical analysis to extract statistical characteristic quantity; and generating a wind speed uniformity characteristic field covering a measurement surface through spatial interpolation, and quantifying to obtain a grade distribution diagram. According to the method, time-space dislocation is eliminated, data consistency is improved, time fluctuation and spatial interpolation are integrated, evaluation comprehensiveness and spatial resolution are enhanced, and accurate measurement of wind speed uniformity is achieved.

Description

Technical Field

[0001] This invention belongs to the field of environmental testing equipment technology, specifically a method for measuring wind speed uniformity based on a wind speed sensor. Background Technology

[0002] Traditional wind speed uniformity measurement involves deploying an array of wind speed sensors in the test area of ​​an environmental test chamber to collect data and assess regional uniformity based on differences in wind speed distribution. While existing technologies attempt to characterize the spatial distribution of wind speed, they have inherent limitations in practical applications, especially in the scenario of wind speed uniformity testing in environmental test chambers. Asynchronous sensor sampling clocks or data transmission delays result in data from different times within the same batch, creating a spatiotemporal misalignment. Evaluation often relies on single-time-point data to calculate biases, failing to consider the cumulative effect of temporal fluctuations, leading to results that only partially reflect the instantaneous state. Spatial characterization is limited to discrete sensor locations, making it impossible to generate continuous uniformity characteristics covering the entire measurement surface and failing to capture subtle regional differences.

[0003] Existing methods need to address two key issues: First, the purification wind speed data stream suffers from spatiotemporal misalignment due to differences in acquisition time. Precise synchronization and alignment of the data in the time dimension are required to ensure that the data within each time slice represents synchronous observations from all sensors at the same moment, eliminating the interference of time differences on spatial distribution. Second, the deviation values ​​of discrete sensor positions cannot fully characterize the uniformity of the entire measurement surface. The deviation information of each sensor over long time series needs to be transformed into a continuous expression of regional uniformity characteristics, overcoming the locality limitations of discrete point evaluation and achieving a holistic characterization of the uniformity of the measurement surface. Summary of the Invention

[0004] This invention aims to solve at least one of the technical problems existing in the prior art; Therefore, this invention proposes a method for measuring wind speed uniformity based on a wind speed sensor, comprising: An array of wind speed sensors is deployed on the measurement surface of the test area in the environmental test chamber to synchronously collect raw wind speed data. After inspection and anomaly removal, the purified air velocity data stream is obtained; The timestamps are used to align each piece of purified wind speed data in the purified wind speed data stream to the same time slice, forming a spatiotemporally synchronized wind speed dataset. The spatiotemporally synchronized wind speed dataset is structured in the spatial domain to construct a wind speed matrix corresponding to the layout of the wind speed sensor array. The elements of the wind speed matrix are the wind speed values ​​of each sensor in the corresponding time slice. Based on the wind speed matrix, the mean wind speed of the measurement surface of the test area of ​​the environmental test chamber under each time slice and the absolute deviation of the wind speed at each sensor location from the mean wind speed are calculated. Summarize the absolute deviations across all time slices, and generate a wind speed deviation sequence for each sensor location. Statistical analysis is performed on the wind speed deviation sequence at each sensor location to extract the statistical features of the wind speed deviation sequence; Spatial interpolation of the statistical features is performed to generate a wind speed uniformity feature field covering the measurement surface; Based on a preset uniformity evaluation threshold, the wind speed uniformity feature field is quantified into a wind speed uniformity level distribution map.

[0005] Furthermore, the purified wind speed data stream obtained after inspection and anomaly removal includes: Multiple wind speed sensors are deployed on the measurement surface of the area to be tested in the environmental test chamber to form a wind speed sensor array; A measurement period is set, and within the measurement period, real-time raw wind speed data from all sensors in the wind speed sensor array are collected simultaneously to generate a raw wind speed data stream. Anomalies are removed from each real-time raw wind speed data in the raw wind speed data stream to obtain a purified wind speed data stream, including: A dynamic threshold window is established for the real-time raw wind speed data of each sensor, and the boundary of the dynamic threshold window is dynamically determined by the wind speed statistics of the previous effective measurement cycle. Compare the currently collected real-time raw wind speed data with the dynamic threshold window; If the real-time raw wind speed data exceeds the dynamic threshold window, it is determined to be an abnormal data point and is removed; if the real-time raw wind speed data is within the dynamic threshold window, it is determined to be a valid data point and is retained. The retained valid data points are connected in chronological order of collection time to form a continuous purified wind speed data stream.

[0006] Furthermore, the step of using timestamps to align each purified wind speed data point in the purified wind speed data stream to the same time slice, forming a spatiotemporally synchronized wind speed dataset, includes: Define a unified time axis that divides the entire measurement period into multiple time slices of equal length; Obtain the original timestamp carried by each purification wind speed data in the purification wind speed data stream; Map the original timestamp of each purification wind speed data point to the center point of the nearest standard time slice on the time axis. All purification wind speed data mapped to the same standard time slice are grouped together to form a snapshot of wind speed data corresponding to the standard time slice; By integrating wind speed data snapshots from all standard time slices, a spatiotemporally synchronized wind speed dataset aligned in the time dimension is constructed.

[0007] Furthermore, the spatiotemporally synchronized wind speed dataset is spatially structured to construct a wind speed matrix corresponding to the layout of the wind speed sensor array, including: A two-dimensional coordinate system is established to represent the measurement surface, and the precise coordinate position of each sensor in the wind speed sensor array in the two-dimensional coordinate system is determined. Extract a standard time-slice wind speed data snapshot from the spatiotemporally synchronized wind speed dataset; Each wind speed value in the wind speed data snapshot is filled into the corresponding position of a matrix corresponding to the discrete grid of the measurement surface, according to the coordinate position of its sensor in the two-dimensional coordinate system. If there is no corresponding sensor at a certain discrete grid location, then fill in the empty value mark in the corresponding position of the matrix; Traverse all standard time slices, repeat the wind speed data snapshot extraction and numerical input operations to form a wind speed matrix sequence arranged in time series, with each matrix being the wind speed matrix of the corresponding time slice.

[0008] Further, based on the wind speed matrix, the mean wind speed of the measurement surface of the test area of ​​the environmental test chamber under each time slice and the absolute deviation of the wind speed at each sensor location from the mean wind speed are calculated, including: For a wind speed matrix corresponding to a time slice, after excluding the null value markers, the arithmetic mean of all valid wind speed values ​​in the matrix is ​​calculated to obtain the mean wind speed of the measurement surface under the corresponding time slice. For each effective wind speed value in the wind speed matrix, the absolute value of the difference between the effective wind speed value and the mean wind speed of the corresponding time slice is calculated to obtain the absolute deviation of the corresponding sensor position in the current time slice. Record the absolute deviation value of each sensor position in each time slice.

[0009] Furthermore, the statistical analysis of the wind speed deviation sequence at each sensor location, and the extraction of statistical features of the wind speed deviation sequence, include: For a given sensor location, the absolute deviation of that location across all time slices within the measurement period is obtained, forming a wind speed deviation sequence for that location. Calculate the arithmetic mean of the wind speed deviation sequence to obtain the mean absolute deviation; Calculate the standard deviation of the wind speed deviation sequence to obtain the degree of deviation fluctuation; Calculate the extreme values ​​of the wind speed deviation sequence to obtain the maximum absolute deviation; The average absolute deviation, the degree of deviation fluctuation, and the maximum absolute deviation are collectively used as the statistical characteristic quantity of the corresponding sensor position.

[0010] Further, the calculation of the arithmetic mean of the wind speed deviation sequence to obtain the mean absolute deviation includes: Sum all the absolute deviation values ​​in the wind speed deviation sequence; Count the total number of absolute deviation values ​​contained in the wind speed deviation sequence; Dividing the summation result by the total number of values ​​yields the average absolute deviation of the corresponding sensor position.

[0011] Further, the step of spatially interpolating the statistical characteristic quantities to generate a wind speed uniformity characteristic field covering the measurement surface includes: Statistical features are obtained at each sensor location, including mean absolute deviation, deviation fluctuation, and maximum absolute deviation. Using the coordinates of the sensor position as interpolation nodes and the mean absolute deviation as the interpolation target value, a spatial interpolation algorithm is used to interpolate the discrete grid of the entire measurement surface to generate the mean absolute deviation distribution field. Using the coordinates of the sensor position as the interpolation node and the degree of deviation fluctuation as the interpolation target value, the same spatial interpolation algorithm is used to perform interpolation calculations on the discrete grid of the entire measurement surface to generate a deviation fluctuation distribution field. Using the coordinates of the sensor position as interpolation nodes and the maximum absolute deviation as the interpolation target value, the same spatial interpolation algorithm is used to perform interpolation calculations on the discrete grid of the entire measurement surface to generate the maximum absolute deviation distribution field. The average absolute deviation distribution field, the deviation fluctuation degree distribution field, and the maximum absolute deviation distribution field are integrated to form the wind speed uniformity feature field, which is a multi-dimensional characterization of wind speed differences and fluctuations.

[0012] Furthermore, the step of using a spatial interpolation algorithm to perform interpolation calculations on the discrete grid of the entire measurement surface includes: The inverse distance weighted interpolation algorithm is selected as the spatial interpolation algorithm; For each discrete grid point to be interpolated, search for all sensor nodes within its preset radius. The reciprocal of the distance from each sensor node to the discrete grid point is calculated as the weight; The interpolation target values ​​of the corresponding sensor nodes are weighted and averaged using the weights, and the result is used as the interpolation result of the discrete grid points. Traverse all discrete grid points to complete the interpolation calculation for the entire measurement surface.

[0013] Further, the step of quantifying the wind speed uniformity feature field into a wind speed uniformity level distribution map based on a preset uniformity evaluation threshold includes: Multiple levels are defined for wind speed uniformity, and a corresponding threshold range is set for each level. The threshold range is determined based on the numerical range of the mean absolute deviation distribution field and the maximum absolute deviation distribution field. Each discrete grid point in the wind speed uniformity feature field is compared with the threshold range based on its average absolute deviation value and maximum absolute deviation value. Determine the wind speed uniformity level for each discrete grid point; Using different visual identifiers, the wind speed uniformity level to which each discrete grid point belongs is marked in the two-dimensional coordinate system of the measurement surface, generating a wind speed uniformity level distribution map that intuitively displays the spatial distribution of wind speed uniformity.

[0014] Compared with the prior art, the beneficial effects of the present invention are: Using timestamps as a benchmark, the purified wind speed data, after anomaly removal, is aligned to the same time slice to form a spatiotemporally synchronized dataset. This technique merges each purified wind speed data point into a fixed-duration time window according to its timestamp, ensuring that the same time slice contains synchronous observations from all sensors at that moment, eliminating spatiotemporal misalignments caused by differences in sensor sampling clocks or data transmission delays. Compared to the misaligned datasets formed by asynchronous acquisition or splicing of single-time-point data in conventional methods, this technique ensures that subsequent spatial domain structuring is based on the actual wind speed distribution at the same moment, avoiding confusion of spatial differences due to time differences. This provides a reliable synchronous data foundation for accurately calculating the wind speed mean of each time slice and the absolute deviation of each sensor from the mean, improving the spatiotemporal consistency of the data.

[0015] The absolute deviations of all time slices are aggregated to generate deviation sequences for each sensor. Statistical analysis is then performed to extract statistical features from these sequences. A spatial interpolation algorithm is then used to extend these features to generate a continuous wind speed uniformity feature field covering the measurement surface. This technique transforms the cumulative deviation information of each sensor across multiple time slices into statistical features. By utilizing spatial interpolation, it overcomes the limitations of discrete sensor points, forming a continuous uniformity representation of the entire measurement surface. Compared to conventional evaluations based solely on discrete deviations at a single time point, this process integrates temporal fluctuation information, elevating local point deviations to regional continuous features. This comprehensively reflects subtle uniformity differences across the measurement surface, improves the spatial resolution of the evaluation, and makes the final quantified wind speed uniformity level distribution map more closely match the actual regional uniformity distribution, avoiding the local limitations of discrete point evaluations. Attached Figure Description

[0016] Figure 1This is a flowchart illustrating the steps of a wind speed uniformity measurement method based on a wind speed sensor according to the present invention. Figure 2 A flowchart for generating a time-slice alignment and spatiotemporal synchronization dataset; Figure 3 The flowchart shows the calculation of mean and absolute wind speed. Figure 4 This is a field map showing the distribution of the average absolute deviation of wind speed. Figure 5 This is a time series plot of wind speed from multiple sensors. Detailed Implementation

[0017] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0018] See Figure 1 A wind speed sensor array is deployed on the measurement surface of the test area in the environmental test chamber. All sensors synchronously collect raw wind speed data. The collected raw data stream is inspected and anomalies are removed to obtain a purified wind speed data stream. Using the timestamps carried in the purified wind speed data, each purified wind speed data point is aligned to a unified time slice, thus forming a wind speed dataset that is synchronous in time and has a corresponding relationship in space. The spatiotemporally synchronized wind speed dataset is spatially structured, and a wind speed matrix is ​​constructed according to the physical layout of the wind speed sensor array. The elements of this matrix are the wind speed values ​​of each sensor in a specific time slice. Based on the wind speed matrix in each time slice, the mean wind speed of the entire measurement surface in that time slice is calculated, and the absolute deviation of the wind speed value at each sensor location from the mean is calculated one by one. The absolute deviations of all time slices within the measurement period are summarized, and for each sensor location, a wind speed deviation sequence of that location over time is generated throughout the entire measurement period. Statistical analysis is performed on the wind speed deviation sequence of each sensor location to extract statistical features that characterize the deviation at that location. By using the statistical characteristics of all sensor locations as known nodes, a wind speed uniformity feature field continuously covering the entire measurement surface is generated through a spatial interpolation algorithm. Based on a preset uniformity evaluation threshold, the wind speed uniformity feature field is quantified into a wind speed uniformity level distribution map with different level labels, intuitively displaying the spatial distribution of wind speed uniformity within the measurement surface.

[0019] See Figure 2In one embodiment of the present invention, for a uniformity test section used in a wind tunnel experiment, the measurement surface is a rectangular area. Sixteen wind speed sensors are deployed within this rectangular area, arranged in a 4x4 matrix to form a wind speed sensor array. The measurement period is set to 600 seconds. During this period, all 16 wind speed sensors in the array simultaneously activate and collect real-time raw wind speed data. The sampling frequency of all wind speed sensors is set to 10 Hz. The synchronously collected real-time raw wind speed data converges to form a raw wind speed data stream. Real-time verification is performed on each piece of real-time raw wind speed data in the raw wind speed data stream to remove abnormal data points and obtain a purified wind speed data stream. During this process, a dynamic threshold window is established for the real-time raw wind speed data of each wind speed sensor. The upper and lower boundaries of the dynamic threshold window are dynamically determined by the wind speed statistics of the previous valid measurement period. The currently collected real-time raw wind speed data is compared with the dynamic threshold window established for the corresponding wind speed sensor to determine whether the real-time raw wind speed data is within the dynamic threshold window. If the currently collected real-time raw wind speed data exceeds the dynamic threshold window, it is determined to be an abnormal data point and marked for removal. If the currently collected real-time raw wind speed data is within the dynamic threshold window, it is determined to be a valid data point and retained. All retained valid data points are connected in the order of collection time to form a continuous purified wind speed data stream.

[0020] In some embodiments, the upper boundary threshold of the dynamic threshold window and lower boundary threshold Calculated using the formula: ; ; in: This represents the average value of all valid wind speed data from the corresponding anemometer within the previous valid measurement period. This represents the standard deviation of all valid wind speed data from the corresponding anemometer within the previous valid measurement period. It is a preset coefficient.

[0021] In implementation, a unified timeline is defined with a total length of 600 seconds. The entire 600-second measurement period is divided into 6000 equal-length time slices, each 0.1 seconds long. The raw timestamp of each purified air velocity data point in the purified air velocity data stream is acquired, recording the precise moment the data was collected. The raw timestamp of each purified air velocity data point is mapped to the center point of the nearest standard time slice on the unified timeline. For example, if the raw timestamp of a purified air velocity data point is 123.456 seconds, its nearest standard time slice is the one centered at 123.45 seconds; therefore, this data is mapped to the standard time slice centered at 123.45 seconds. All purified air velocity data points mapped to the same standard time slice are grouped together to form a snapshot of the air velocity data corresponding to that standard time slice. This snapshot contains the air velocity values ​​collected by each wind speed sensor within the 0.1-second time slice. By integrating wind speed data snapshots corresponding to all 6,000 standard time slices, a spatiotemporally synchronized wind speed dataset that is fully aligned in the time dimension is constructed.

[0022] It is understandable that the length of the time slice determines the granularity of time alignment; shorter time slices provide finer time synchronization. In some embodiments, considering the frequency of wind speed changes and sensor response characteristics, the length of the time slice can be set to 0.2 seconds or 0.05 seconds. For the wind speed sensor with a sampling frequency of 10 Hz in the example above, each sensor should theoretically generate one data point in each standard time slice. The mapping process handles time deviations caused by minor differences in sensor clocks or data transmission delays. Optionally, the mapping rule can employ rounding down, rounding up, or rounding to the nearest standard time slice center point. By mapping all data to a unified time grid, wind speed data from different sensors within the same time slice become comparable, providing a time reference for subsequent calculation of the average wind speed on the measurement surface.

[0023] See Figure 3In one embodiment of the present invention, a two-dimensional coordinate system is established to represent the measurement surface. The origin of the coordinate system is set at the lower left corner of the measurement surface, the positive X-axis points to the right, and the positive Y-axis points upward. The measurement surface is 4 meters long and 3 meters wide. The wind speed sensor array contains 16 wind speed sensors arranged in a 4x4 matrix, determining the precise coordinate position of each wind speed sensor in the two-dimensional coordinate system. The precise coordinates of the wind speed sensor located in the first row and first column are (0.5 m, 0.5 m); the precise coordinates of the wind speed sensor located in the first row and second column are (1.5 m, 0.5 m); the precise coordinates of the wind speed sensor located in the first row and third column are (2.5 m, 0.5 m); the precise coordinates of the wind speed sensor located in the first row and fourth column are (3.5 m, 0.5 m); the precise coordinates of the wind speed sensor located in the second row and first column are (0.5 m, 1.5 m); and following this pattern, the precise coordinates of the wind speed sensor located in the fourth row and fourth column are (3.5 m, 2.5 m). A wind speed data snapshot corresponding to a standard time slice is extracted from the spatiotemporally synchronized wind speed dataset. This wind speed data snapshot contains 16 wind speed values, each corresponding to one wind speed sensor. Each wind speed value in the wind speed data snapshot is filled into a matrix according to the coordinate position of its corresponding wind speed sensor in a two-dimensional coordinate system.

[0024] In practice, a matrix corresponding to the discrete grid of the measurement surface is constructed for structured storage of wind speed data. The density of the discrete grid can be higher than the actual deployment density of the wind speed sensors. For example, a measurement surface 4 meters long and 3 meters wide can be divided into a 40-row, 30-column grid, with each grid cell representing a 0.1-meter × 0.1-meter area. When filling the matrix with wind speed values, the row and column indices of the wind speed sensor in the discrete grid matrix are determined based on the precise coordinates of the sensor, and the wind speed values ​​are filled into the corresponding positions in the matrix. For positions in the discrete grid matrix that do not correspond to wind speed sensors, such as grid points with coordinates (0.7 meters, 0.7 meters), a null value marker is filled into the corresponding position in the matrix. The null value marker can be a specific numerical value. The operation of extracting wind speed data snapshots and filling the matrix with values ​​is repeated for all standard time slices. For 6000 standard time slices, a wind speed matrix sequence containing 6000 matrices is formed, arranged in a time sequence. Each matrix in the sequence is the wind speed matrix for that time slice.

[0025] In some embodiments, the mean wind speed and absolute deviation are calculated. For a wind speed matrix with time slice sequence number T, all positions marked as null values ​​are excluded. Only 16 positions in the wind speed matrix have valid wind speed values. The arithmetic mean of these 16 valid wind speed values ​​is calculated to obtain the mean wind speed of the measurement surface at the corresponding time slice T. For each valid wind speed value in the wind speed matrix (where i identifies the wind speed sensor location index), calculate the effective wind speed value. Wind speed mean with time slice T The absolute value of the difference between them yields the absolute deviation of the corresponding wind speed sensor position in time slice T. Record the absolute deviation of each wind speed sensor location in each time slice. For the wind speed sensor with location index i, its absolute deviation over the entire measurement period is recorded as a sequence { }

[0026] It is understandable that the construction of the wind speed matrix combines wind speed data from spatially discrete points with time series data to form structured data. The handling of null value marking ensures that subsequent calculations are based only on valid observation points. Absolute bias The calculation formula quantifies the instantaneous deviation between the wind speed at a specific location at a specific time and the overall average wind speed of the measurement surface. Optionally, in calculating the average wind speed... If some sensor data is missing in a given time slice due to abnormal removal (marked as null values), the averaging calculation will automatically exclude these null values ​​and only use the actual valid wind speed sensor data for the current time slice. For the example above, the wind speed value of the wind speed sensor located at coordinates (1.5 m, 1.5 m) at time slice T=100... The mean surface wind speed was 5.2 m / s, while the mean surface wind speed measured at time slice T=100 was... If the speed is calculated to be 5.0 m / s, then the absolute deviation at this position at this moment is... The value is |5.2-5.0| = 0.2 m / s. In some embodiments, the data structure for recording the absolute deviation can be a two-dimensional array, where one-dimensional index represents the time slice number and the other-dimensional index represents the wind speed sensor location identifier, or a separate list of absolute deviation sequences that change over time can be maintained for each wind speed sensor location.

[0027] In one embodiment of the present invention, for a wind speed sensor located at a specific coordinate position in a wind speed sensor array, such as a wind speed sensor at coordinate position (1.5 m, 1.5 m), the absolute deviation value of this coordinate position is obtained for all time slices within the measurement period. The measurement period includes 6000 time slices, from time slice T=1 to time slice T=6000. The absolute deviation value of this coordinate position in each time slice is recorded to form a wind speed deviation sequence for this coordinate position. The wind speed deviation sequence can be represented in list form { },in This represents the absolute deviation of the position (1.5 meters, 1.5 meters) at time slice T=1. The absolute deviation represents the time slice T=6000. The arithmetic mean of this wind speed deviation sequence is calculated to obtain the mean absolute deviation. The standard deviation of this wind speed deviation sequence is calculated to obtain the degree of deviation fluctuation. The extreme values ​​of this wind speed deviation sequence are calculated, i.e., the maximum values ​​in the sequence are found, to obtain the maximum absolute deviation. The mean absolute deviation, the degree of deviation fluctuation, and the maximum absolute deviation together serve as statistical characteristics of the coordinate position (1.5 meters, 1.5 meters).

[0028] In practice, the specific process of calculating the arithmetic mean of the wind speed deviation sequence involves summing all absolute deviation values ​​in the sequence. The wind speed deviation sequence contains 6000 absolute deviation values; these 6000 values ​​are summed to obtain the result. The total number of absolute deviation values ​​in the wind speed deviation sequence is counted; the total number is 6000. The summation is then divided by the total number of values ​​(6000), and the quotient is the average absolute deviation of the corresponding sensor location. The calculation of the average absolute deviation can be expressed by the formula: ,in Represents the mean absolute deviation. This represents the absolute deviation value corresponding to the k-th time slice in the wind speed deviation sequence. This represents the total number of absolute deviation values ​​in the wind speed deviation sequence, in this example... , Indicates the sequence from arrive All Perform summation.

[0029] In some embodiments, the degree of deviation fluctuation is calculated using the standard deviation formula, which is:

[0030] in: This represents the degree of deviation fluctuation (i.e., the sample standard deviation). This represents the absolute deviation value corresponding to the k-th time slice in the wind speed deviation sequence. This represents the calculated mean absolute deviation. This represents the total number of data points in the sequence. The maximum absolute deviation is calculated by finding the wind speed deviation sequence {...} The maximum value in}, i.e. , This represents the maximum absolute deviation. For each wind speed sensor location in the wind speed sensor array, the above process is repeated to obtain the wind speed deviation sequence for each location, and the corresponding average absolute deviation is calculated. Degree of deviation fluctuation and maximum absolute deviation For an array containing 16 wind speed sensors, 16 sets of statistical characteristics will be obtained, each set containing three values: mean absolute deviation. Degree of deviation fluctuation and maximum absolute deviation .

[0031] Understandable, mean absolute deviation This characterizes the degree of deviation of the wind speed at this location from the average wind speed of the measurement surface, and the extent of the deviation fluctuation. This characterizes the dispersion of the absolute deviation of the position over time, with the maximum absolute deviation being... This characterizes the most severe instantaneous deviation ever observed at this location. Three statistical features describe the non-uniformity of wind speed at this location from different dimensions. Optionally, in calculating the mean absolute deviation... If, during a certain time slice, data at a specific location is marked as null due to anomalies and assigned a specific value (such as 0 or ignored) in previous absolute deviation calculations, then corresponding consistency processing is required during summation and counting. For some embodiments, the total number of data points in the wind speed deviation sequence may not be 6000, depending on the measurement period length and the granularity of the time slice division, but the calculation principle remains consistent. For a wind speed sensor at coordinate position (0.5 m, 0.5 m), the average absolute deviation calculated from its wind speed deviation sequence... It could be 0.15 m / s, with varying degrees of deviation. It could be 0.08 m / s, with a maximum absolute deviation. It could be 0.40 m / s.

[0032] In one embodiment of the present invention, calculated statistical features are obtained at each sensor location in the wind speed sensor array. These statistical features include the mean absolute deviation, the degree of deviation fluctuation, and the maximum absolute deviation. For the 16 wind speed sensors deployed in this embodiment, the position coordinates (X, Y) of each sensor and its corresponding statistical features constitute known data points. Referring to Table 1, example data for some sensor locations are shown.

[0033] Table 1: Examples of Statistical Characteristics of Some Sensor Positions

[0034] Using the two-dimensional coordinates of the sensor location as interpolation nodes and the mean absolute deviation M as the interpolation target value, a selected spatial interpolation algorithm is used to interpolate the discrete grid of the entire measurement surface, generating a mean absolute deviation distribution field. The division of the discrete grid of the measurement surface is consistent with the example, for example, divided into 40 rows and 30 columns, totaling 1200 grid points. Using the same sensor coordinates as interpolation nodes and the deviation fluctuation degree S as the interpolation target value, the same spatial interpolation algorithm is used to interpolate the discrete grid of the entire measurement surface, generating a deviation fluctuation degree distribution field. Using the same sensor coordinates as interpolation nodes and the maximum absolute deviation Dmax as the interpolation target value, the same spatial interpolation algorithm is used to interpolate the discrete grid of the entire measurement surface, generating a maximum absolute deviation distribution field. The mean absolute deviation distribution field, the deviation fluctuation degree distribution field, and the maximum absolute deviation distribution field are integrated to form a multi-dimensional wind speed uniformity feature field that characterizes wind speed differences and fluctuations. The wind speed uniformity feature field contains three interpolated values ​​at each grid point: interpolated mean absolute deviation, interpolated deviation fluctuation degree, and interpolated maximum absolute deviation.

[0035] In practical implementation, the inverse distance weighted interpolation algorithm is selected as the spatial interpolation algorithm. The formula for the inverse distance weighted interpolation algorithm is expressed as follows:

[0036] in: The interpolation result represents the grid point p to be interpolated. This represents the interpolated target value at the j-th known sensor node. This represents the Euclidean distance between the grid point p to be interpolated and the j-th known sensor node. The parameter representing the power of the distance is usually set to 2. This represents the number of known sensor nodes participating in this interpolation calculation. The summation sign is applied to all n known sensor nodes within a preset search radius. For each discrete grid point p to be interpolated, all sensor nodes within its preset radius are searched. The preset radius is set to, for example, 2 meters. The distance from each sensor node to the discrete grid point p is calculated. reciprocal As weights. The calculated weights are used to interpolate the target value for the corresponding sensor node. A weighted average is performed, and the result of the weighted average is used as the interpolation result for the discrete grid point p. The process iterates through all 1200 discrete grid points on the measurement surface, performing the inverse distance weighted interpolation calculation described above for each point. This completes the interpolation calculation for the entire measurement surface, generating the mean absolute deviation distribution field, the deviation fluctuation degree distribution field, and the maximum absolute deviation distribution field, respectively.

[0037] In some embodiments, for a grid point to be interpolated located at coordinates (2.0 m, 1.0 m), a search is performed on known sensor nodes within a 2-meter radius. This may find four sensor nodes with coordinates (1.5, 0.5), (2.5, 0.5), (1.5, 1.5), and (2.5, 1.5). The distances from the grid point to these four nodes are then calculated. The interpolation average absolute deviation value of the grid point is calculated using the average absolute deviation values ​​at these four nodes, based on the inverse distance weighting formula. It can be understood that the inverse distance weighting interpolation algorithm assumes that the value of the point to be interpolated is influenced by nearby known points, with the degree of influence inversely proportional to the distance. The interpolated distribution field provides an estimate of the wind speed uniformity characteristics of continuous spatial locations on the measurement surface, compensating for data gaps caused by sparse sensor deployment. Optionally, other algorithms such as Kriging interpolation or radial basis function interpolation can also be selected for spatial interpolation. The preset search radius setting must ensure that each grid point to be interpolated can find at least one known sensor node. In some embodiments, if no sensor node is found within the search radius, the interpolation result of that grid point can be marked as invalid or recalculated using a larger search radius. The average absolute deviation distribution field reflects the spatial distribution pattern of wind speed deviation from the average level, the deviation fluctuation degree distribution field reflects the spatial distribution of wind speed deviation instability, and the maximum absolute deviation distribution field reflects the spatial region where the most severe instantaneous deviation may occur. The wind speed uniformity characteristic field combining these three provides comprehensive uniform spatial information.

[0038] See Figure 4 This is a map showing the distribution of the mean absolute deviation of wind speed. It's a spatial distribution visualization generated based on inverse distance weighted interpolation, visually illustrating the spatial distribution pattern of wind speed deviations from the average level within the measurement surface, clearly identifying "hot spots" and "cold spots" in wind speed uniformity. Low deviation areas are concentrated in the central region of the measurement surface (X≈1.5, Y≈1.5–2.5), with deviations of approximately 0.075–0.135, where the wind speed is closest to the regional average. High deviation areas are mainly distributed on the right side of the measurement surface (X>2.5), especially in the upper right corner (X≈3.5, Y≈3.5), with deviations close to 0.315, where the wind speed deviation from the average level is most pronounced. Figure 4 The red triangles in the center mark the deployment locations of the 16 wind speed sensors, and the interpolation results at these points are completely consistent with the original statistical values. The mean absolute deviation distribution field directly reflects the spatial pattern of wind speed deviation from the regional average level and is one of the core indicators for assessing wind speed uniformity.

[0039] In one embodiment of the invention, multiple levels are defined for wind speed uniformity, such as "highly uniform," "generally uniform," "somewhat non-uniform," and "non-uniform." A corresponding threshold range is set for each level, determined based on the numerical ranges of the mean absolute deviation distribution field and the maximum absolute deviation distribution field generated in the embodiment. In a specific implementation, the interpolated mean absolute deviation values ​​of all grid points in the mean absolute deviation distribution field are analyzed to obtain the numerical range of the interpolated mean absolute deviation across the entire measurement surface, for example, from 0.05 m / s to 0.40 m / s. The interpolated maximum absolute deviation values ​​of all grid points in the maximum absolute deviation distribution field are analyzed to obtain the numerical range of the interpolated maximum absolute deviation across the entire measurement surface, for example, from 0.15 m / s to 0.85 m / s. Combining these two distribution field numerical ranges, a comprehensive threshold range including the mean absolute deviation component and the maximum absolute deviation component is set for each wind speed uniformity level. The threshold range for the "highly uniform" level can be set as follows: mean absolute deviation ≤ 0.10 m / s and maximum absolute deviation ≤ 0.30 m / s. The threshold range for the "Generally Uniform" level can be set as follows: 0.10 m / s < mean absolute deviation ≤ 0.20 m / s and 0.30 m / s < maximum absolute deviation ≤ 0.50 m / s. The threshold range for the "Slightly Non-Uniform" level can be set as follows: 0.20 m / s < mean absolute deviation ≤ 0.30 m / s and 0.50 m / s < maximum absolute deviation ≤ 0.70 m / s. The threshold range for the "Non-Uniform" level can be set as: mean absolute deviation > 0.30 m / s or maximum absolute deviation > 0.70 m / s.

[0040] In practice, each discrete grid point in the wind speed uniformity feature field is compared one by one with the set threshold ranges for each level, based on its interpolated average absolute deviation and maximum absolute deviation. This determines the wind speed uniformity level to which each discrete grid point belongs. In the two-dimensional coordinate system of the measurement surface, different visual identifiers are used to label the wind speed uniformity level to which each discrete grid point belongs, generating a wind speed uniformity level distribution map that visually displays the spatial distribution of wind speed uniformity. Different colors, fill patterns, or symbols can be used for the visual identifiers. For example, in the generated color-mapped level distribution map, dark green can be used to represent "highly uniform" level areas, light green to represent "generally uniform" level areas, yellow to represent "relatively non-uniform" level areas, and red to represent "non-uniform" level areas.

[0041] In some embodiments, the threshold range can be set based on a weighted evaluation function:

[0042] in: Represents the evaluation index value. The interpolation mean absolute deviation representing the grid points The maximum absolute deviation of the interpolation at grid points. and They are and Weighting coefficients, weighting coefficients and The determination is based on the degree of emphasis placed on average deviation and maximum deviation in practical applications. It is based on the evaluation index values. The numerical range is divided into level thresholds, for example, setting... For "high uniformity", "Generally uniform" "Somewhat uneven" The result is "non-uniform". This can be understood as the evaluation function... By merging the two features into a single evaluation index, the logic for grade determination is simplified. For a grid point with coordinates (1.8 meters, 1.2 meters), its interpolated mean absolute deviation... The maximum absolute deviation of the interpolation is 0.09 m / s. The speed is 0.28 m / s, and this grid point meets the dual threshold condition for the "high uniformity" level. and Therefore, it is classified as "highly uniform". In the generated wind speed uniformity level distribution map, the location of this grid point will be filled or marked with a visual identifier representing the "highly uniform" level (such as dark green). For the grid point at coordinates (3.0 m, 0.8 m), its interpolated average absolute deviation is... The maximum absolute deviation of the interpolation is 0.25 m / s. At 0.58 m / s, this grid point falls within the threshold range of the "less uniform" level. and Therefore, it is classified as "relatively uneven" and filled in yellow on the distribution map. Optionally, the choice of visual identifiers can be adjusted according to the output device or display requirements. For example, in a black and white printed report, grid lines of different densities or dot patterns can be used to distinguish the levels. In some embodiments, in addition to color filling, the level text can also be directly labeled on each grid point or key area of ​​the distribution map. The final generated wind speed uniformity level distribution map presents the spatial distribution of wind speed uniformity levels in different areas within the measurement surface.

[0043] See Figure 5This is a multi-sensor wind speed time series plot, showing the real-time wind speed changes of five wind speed sensors from 16:06:30 to 16:08:00. It is the core visualization result of the "raw wind speed data stream acquisition" stage in the wind speed uniformity measurement method. All sensors exhibit obvious fluctuations in wind speed, without a continuous upward or downward unidirectional trend, consistent with the randomness of natural wind fields. Sensor 13 shows the highest overall wind speed, mainly concentrated between 4.5–5.8 m / s, with the largest fluctuation range. Sensors 7 and 19 show the next highest wind speeds, roughly between 4.0–5.2 m / s, highly synchronized with the trend of sensor 13. Sensors 1 and 25 show the lowest overall wind speeds, mainly between 2.5–3.5 m / s, with relatively small fluctuation ranges. The wind speed curves of sensors 13, 7, and 19 fluctuate almost synchronously, while sensors 1 and 25 form another relatively independent fluctuation pattern, directly reflecting the spatial non-uniformity of wind speed within the measurement area.

[0044] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for measuring wind speed uniformity based on a wind speed sensor, characterized in that, The method includes: An array of wind speed sensors is deployed on the measurement surface of the test area in the environmental test chamber to synchronously collect raw wind speed data. After inspection and anomaly removal, the purified wind speed data stream is obtained. The timestamps are used to align each piece of purified wind speed data in the purified wind speed data stream to the same time slice, forming a spatiotemporally synchronized wind speed dataset. The spatiotemporally synchronized wind speed dataset is structured in the spatial domain to construct a wind speed matrix corresponding to the layout of the wind speed sensor array. The elements of the wind speed matrix are the wind speed values ​​of each sensor in the corresponding time slice. Based on the wind speed matrix, the mean wind speed of the measurement surface of the test area of ​​the environmental test chamber under each time slice and the absolute deviation of the wind speed at each sensor location from the mean wind speed are calculated. Summarize the absolute deviations across all time slices, and generate a wind speed deviation sequence for each sensor location. Statistical analysis is performed on the wind speed deviation sequence at each sensor location to extract the statistical features of the wind speed deviation sequence; Spatial interpolation of the statistical features is performed to generate a wind speed uniformity feature field covering the measurement surface; Based on a preset uniformity evaluation threshold, the wind speed uniformity feature field is quantified into a wind speed uniformity level distribution map.

2. The method for measuring wind speed uniformity based on a wind speed sensor according to claim 1, characterized in that, The purified wind speed data stream obtained after inspection and anomaly removal includes: Multiple wind speed sensors are deployed on the measurement surface of the area to be tested in the environmental test chamber to form a wind speed sensor array; A measurement period is set, and within the measurement period, real-time raw wind speed data from all sensors in the wind speed sensor array are collected simultaneously to generate a raw wind speed data stream. Anomalies are removed from each real-time raw wind speed data in the raw wind speed data stream to obtain a purified wind speed data stream, including: A dynamic threshold window is established for the real-time raw wind speed data of each sensor, and the boundary of the dynamic threshold window is dynamically determined by the wind speed statistics of the previous effective measurement cycle. Compare the currently collected real-time raw wind speed data with the dynamic threshold window; If the real-time raw wind speed data exceeds the dynamic threshold window, it is determined to be an abnormal data point and is removed; if the real-time raw wind speed data is within the dynamic threshold window, it is determined to be a valid data point and is retained. The retained valid data points are connected in chronological order of collection time to form a continuous purified wind speed data stream.

3. The method for measuring wind speed uniformity based on a wind speed sensor according to claim 2, characterized in that, The use of timestamps aligns each purified wind speed data point in the purified wind speed data stream to the same time slice, forming a spatiotemporally synchronized wind speed dataset, including: Define a unified time axis that divides the entire measurement period into multiple time slices of equal length; Obtain the original timestamp carried by each purification wind speed data in the purification wind speed data stream; Map the original timestamp of each purification wind speed data point to the center point of the nearest standard time slice on the time axis. All purification wind speed data mapped to the same standard time slice are grouped together to form a snapshot of wind speed data corresponding to the standard time slice; By integrating wind speed data snapshots from all standard time slices, a spatiotemporally synchronized wind speed dataset aligned in the time dimension is constructed.

4. The method for measuring wind speed uniformity based on a wind speed sensor according to claim 3, characterized in that, The spatiotemporally synchronized wind speed dataset is structured in the spatial domain to construct a wind speed matrix corresponding to the layout of the wind speed sensor array, including: A two-dimensional coordinate system is established to represent the measurement surface, and the precise coordinate position of each sensor in the wind speed sensor array in the two-dimensional coordinate system is determined. Extract a standard time-slice wind speed data snapshot from the spatiotemporally synchronized wind speed dataset; Each wind speed value in the wind speed data snapshot is filled into the corresponding position of a matrix corresponding to the discrete grid of the measurement surface, according to the coordinate position of its sensor in the two-dimensional coordinate system. If there is no corresponding sensor at a certain discrete grid location, then fill in the empty value mark in the corresponding position of the matrix; Traverse all standard time slices, repeat the wind speed data snapshot extraction and numerical input operations to form a wind speed matrix sequence arranged in time series, with each matrix being the wind speed matrix of the corresponding time slice.

5. The method for measuring wind speed uniformity based on a wind speed sensor according to claim 4, characterized in that, Based on the wind speed matrix, the mean wind speed of the measurement surface of the test area in the environmental test chamber under each time slice and the absolute deviation of the wind speed at each sensor location from the mean wind speed are calculated, including: For a wind speed matrix corresponding to a time slice, after excluding the null value markers, the arithmetic mean of all valid wind speed values ​​in the matrix is ​​calculated to obtain the mean wind speed of the measurement surface under the corresponding time slice. For each effective wind speed value in the wind speed matrix, the absolute value of the difference between the effective wind speed value and the mean wind speed of the corresponding time slice is calculated to obtain the absolute deviation of the corresponding sensor position in the current time slice. Record the absolute deviation value of each sensor position in each time slice.

6. The method for measuring wind speed uniformity based on a wind speed sensor according to claim 5, characterized in that, The statistical analysis of the wind speed deviation sequence at each sensor location, extracting statistical features of the wind speed deviation sequence, includes: For a given sensor location, the absolute deviation of that location across all time slices within the measurement period is obtained, forming a wind speed deviation sequence for that location. Calculate the arithmetic mean of the wind speed deviation sequence to obtain the mean absolute deviation; Calculate the standard deviation of the wind speed deviation sequence to obtain the degree of deviation fluctuation; Calculate the extreme values ​​of the wind speed deviation sequence to obtain the maximum absolute deviation; The average absolute deviation, the degree of deviation fluctuation, and the maximum absolute deviation are collectively used as the statistical characteristic quantity of the corresponding sensor position.

7. The method for measuring wind speed uniformity based on a wind speed sensor according to claim 6, characterized in that, The calculation of the arithmetic mean of the wind speed deviation sequence to obtain the mean absolute deviation includes: Sum all the absolute deviation values ​​in the wind speed deviation sequence; Count the total number of absolute deviation values ​​contained in the wind speed deviation sequence; Dividing the summation result by the total number of values ​​yields the average absolute deviation of the corresponding sensor position.

8. The method for measuring wind speed uniformity based on a wind speed sensor according to claim 7, characterized in that, The step of spatially interpolating the statistical characteristic quantities to generate a wind speed uniformity characteristic field covering the measurement surface includes: Statistical features are obtained at each sensor location, including mean absolute deviation, deviation fluctuation, and maximum absolute deviation. Using the coordinates of the sensor position as interpolation nodes and the mean absolute deviation as the interpolation target value, a spatial interpolation algorithm is used to interpolate the discrete grid of the entire measurement surface to generate the mean absolute deviation distribution field. Using the coordinates of the sensor position as the interpolation node and the degree of deviation fluctuation as the interpolation target value, the same spatial interpolation algorithm is used to perform interpolation calculations on the discrete grid of the entire measurement surface to generate a deviation fluctuation distribution field. Using the coordinates of the sensor position as interpolation nodes and the maximum absolute deviation as the interpolation target value, the same spatial interpolation algorithm is used to perform interpolation calculations on the discrete grid of the entire measurement surface to generate the maximum absolute deviation distribution field. The average absolute deviation distribution field, the deviation fluctuation degree distribution field, and the maximum absolute deviation distribution field are integrated to form the wind speed uniformity feature field, which is a multi-dimensional characterization of wind speed differences and fluctuations.

9. A method for measuring wind speed uniformity based on a wind speed sensor according to claim 8, characterized in that, The method of using a spatial interpolation algorithm to perform interpolation calculations on the discrete grid of the entire measurement surface includes: The inverse distance weighted interpolation algorithm is selected as the spatial interpolation algorithm; For each discrete grid point to be interpolated, search for all sensor nodes within its preset radius. The reciprocal of the distance from each sensor node to the discrete grid point is calculated as the weight; The interpolation target values ​​of the corresponding sensor nodes are weighted and averaged using the weights, and the result is used as the interpolation result of the discrete grid points. Traverse all discrete grid points to complete the interpolation calculation for the entire measurement surface.

10. A method for measuring wind speed uniformity based on a wind speed sensor according to claim 9, characterized in that, The step of quantifying the wind speed uniformity feature field into a wind speed uniformity level distribution map based on a preset uniformity evaluation threshold includes: Multiple levels are defined for wind speed uniformity, and a corresponding threshold range is set for each level. The threshold range is determined based on the numerical range of the mean absolute deviation distribution field and the maximum absolute deviation distribution field. Each discrete grid point in the wind speed uniformity feature field is compared with the threshold range based on its average absolute deviation value and maximum absolute deviation value. Determine the wind speed uniformity level for each discrete grid point; Using different visual identifiers, the wind speed uniformity level to which each discrete grid point belongs is marked in the two-dimensional coordinate system of the measurement surface, generating a wind speed uniformity level distribution map that intuitively displays the spatial distribution of wind speed uniformity.

Images