High-precision wind speed measuring system

Abstract

The invention relates to a high-precision wind speed measurement system which realizes accurate measurement of wind speed by combining advanced sensor technology, data processing capability and an innovative data fusion algorithm. The system comprises a data acquisition module used for acquiring wind speed data in real time; the data processing module is used for amplifying, filtering and digitally processing the collected original wind speed electric signal; the data fusion algorithm module adopts an advanced data fusion algorithm comprising the steps of normalization processing, optimal fusion set determination, fusion degree matrix calculation, consistent fusion degree and distribution equilibrium calculation, weight coefficient distribution and the like to fuse wind speed data from a plurality of data sources so as to improve the accuracy and stability of measurement; and the data analysis module is used for carrying out statistical analysis on the fused wind speed data, and carrying out deep analysis on the fused data based on the combination of an ANOWREEMD algorithm and an RPSEMD algorithm so as to obtain a high-precision wind speed result.

Description

Technical Field

[0001] The invention relates to the technical field of meteorological detection and data processing, and in particular to a high-precision wind speed measurement system. Background Art

[0002] With the intensification of global climate change and the rapid development of renewable energy technology, the demand for accurate measurement of wind speed is growing. Traditional wind speed measurement systems often rely on a single measurement device and simple data processing methods, and cannot provide high-precision and high-stability wind speed data, which has caused significant limitations in the fields of wind energy development, weather forecasting, and environmental monitoring.

[0003] With the continuous progress of sensor technology, data processing technology and cloud computing technology, new possibilities have been provided for the upgrade and optimization of high-precision wind speed measurement systems. By integrating multiple sensors, applying advanced data processing algorithms and utilizing cloud computing platforms, real-time collection, precise analysis and intelligent processing of wind speed data can be achieved, thereby improving the accuracy and reliability of wind speed measurement.

[0004] In the existing technology, the EMD modal algorithm has good adaptability for meteorological measurement, but due to the existence of the mode mixing problem, measurement errors are difficult to avoid, which is particularly prominent under complex meteorological conditions, resulting in inaccurate wind speed data, which in turn affects the effective utilization of wind energy resources and the accuracy of meteorological forecasts. It is a direction worthy of research to improve the mode mixing problem and further improve the upper limit of wind speed measurement accuracy through algorithm improvement and combination. Summary of the invention

[0005] Purpose of the invention: In order to solve the problems mentioned in the background technology, the present invention proposes a high-precision wind speed measurement system, which collects wind speed data in real time through modular structure processing, and further solves the modal mixing problem in EMD by improving the ANOWREEMD algorithm and the RPSEMD algorithm through data fusion and combination, and accurately calculates the wind speed data, providing a feasibility reference for subsequent meteorological research or resource utilization.

[0006] Technical solution:

[0007] A high-precision wind speed measurement system, the system comprising:

[0008] Wind speed acquisition module, with built-in wind speed sensor, is used to collect raw wind speed data;

[0009] The original signal processing module is connected to the wind speed acquisition module and is used to amplify, filter and digitize the collected original wind speed data;

[0010] The data fusion algorithm module is connected to the original signal processing module to normalize the processed electrical signals, establish the optimal fusion set to calculate the wind speed sensor weight coefficient and perform the final fusion to obtain the preliminary wind speed measurement value;

[0011] The data analysis module is connected to the data fusion algorithm module, integrating the ANOWREEMD algorithm based on the ICEEMDAN algorithm fusion adaptive noise addition strategy optimization and residual term processing method and the RPSEMD algorithm to process EMD mode mixing, which is used to further refine the preliminary wind speed measurement value and output the final IMF final and the residual signal R new , used for statistical analysis and forecasting.

[0012] Furthermore, the wind speed sensor built into the wind speed acquisition module includes a hot wire wind speed sensor, an ultrasonic wind speed sensor and a laser wind speed sensor, and the original wind speed data includes:

[0013] The wind speed change data collected from the hot wire wind speed sensor is represented by x 热线 ;

[0014] The wind speed change data collected from the ultrasonic wind speed sensor is expressed as x 超声波 ;

[0015] The wind speed change data collected from the laser wind speed sensor is represented by x 激光 ;

[0016] The above data are input into the data fusion algorithm module as basic data after being preprocessed by the original signal processing module.

[0017] Furthermore, the x 热线 , x 超声波 , x 激光 It reflects the wind speed detected by different sensors at a certain moment or time period, including the wind speed and direction parameters. After data preprocessing by the original signal processing module, it is input into the data fusion algorithm module for normalization processing. The formula is as follows:

[0018]

[0019] Where x represents the raw data x of any sensor 热线 , x 超声波 , x 激光 ; x′ is the normalized data value, x min is the minimum value among all sensor data sets, x max is the maximum value among all sensor data sets.

[0020] Furthermore, the detailed steps of the data fusion algorithm module to establish the optimal fusion set, calculate the wind speed sensor weight coefficient and perform the final fusion are as follows:

[0021] Determine the optimal fusion set:

[0022] Based on the normalized data, a threshold is defined to determine the optimal fusion set, exclude outliers, and use the data in the optimal fusion set for subsequent data processing and analysis to reduce the impact of outliers on the final results. The formula for determining the optimal fusion set is as follows:

[0023]

[0024] Where: S opt is the optimal fusion set, x i is the observation value of the ith sensor at time t, is the average observation value of all sensors at time t, τ is the preset threshold, the i-th sensor refers to any sensor in the data acquisition module, specifically a hot wire, ultrasonic or laser wind speed sensor, depending on their number or identification in the data acquisition module, the hot wire wind speed sensor is 1, the ultrasonic wind speed sensor is 2, and the laser wind speed sensor is 3;

[0025] Calculate the fusion degree matrix:

[0026] Based on the optimal fusion set data, calculate the fusion degree matrix:

[0027]

[0028] Among them, F ij is the fusion degree between sensors i and j at time t, x j is the observation value of the jth sensor at time t, σ is the standard deviation, which is used to control the sensitivity of the fusion degree;

[0029] Calculate the consistent fusion degree and distribution balance degree:

[0030] Calculate the consistent fusion degree and distribution balance of each sensor data for the allocation of weight coefficients. The calculation formulas for consistent fusion degree and distribution balance are as follows:

[0031]

[0032] Among them, C i is the consistent fusion degree of the i-th sensor at time t, E i is the distribution balance of the i-th sensor at time t. According to the consistent fusion degree and distribution balance, a weight coefficient is assigned to each sensor. The weight coefficient allocation formula is as follows:

[0033]

[0034] Among them, w i is the weight coefficient of the ith sensor at time t, which will be used to calculate the preliminary wind speed value;

[0035] Fuse and output preliminary wind speed measurements:

[0036] The formula for the final fusion estimate is as follows:

[0037]

[0038] in, is the estimated wind speed at time t, and n is the total number of sensors, i.e., the sum of hot wire wind speed sensors, ultrasonic wind speed sensors, and laser wind speed sensors.

[0039] Furthermore, the data analysis module is deployed in detail as follows:

[0040] This module receives the output data from the data fusion algorithm module, and further processes the output data using the adaptive noise optimization of the integrated regenerative phase-shifted sine-assisted empirical mode decomposition algorithm RPSEMD algorithm and the ensemble empirical mode decomposition algorithm ANOWREEMD with wavelet residual enhancement. The ANOWREEMD algorithm integrates the adaptive noise addition strategy optimization and the improvement of the residual term processing method on the basis of the ICEEMDAN algorithm. The preliminary wind speed measurement value is decomposed by the RPSEMD and ANOWREEMD combined algorithm to obtain the intrinsic mode function IMFs, and the standard deviation, maximum and minimum parameters of each IMF are used for statistical analysis to confirm the final accurate wind speed value.

[0041] Furthermore, the specific steps of processing the output data by combining the RPSEMD algorithm with the ANOWREEMD algorithm are as follows:

[0042] Step 1: Initialize algorithm parameters

[0043] Set the initial noise amplitude N0, the number of iterations M, the number of IMFs to be selected K, and the wavelet transform parameters, including the wavelet basis ψ and the number of decomposition levels J;

[0044] Step 2: Data Preprocessing

[0045] Detrending the wind speed signal x(t) yields x′(t):

[0046] x′(t)=x(t)-detrend(x(t))

[0047] Where detrend(x(t)) means removing the trend component of the signal x(t) to reduce the modal aliasing in the subsequent decomposition;

[0048] Step 3: Adaptive Noise Addition

[0049] Dynamically adjust the amplitude and frequency of the noise according to the frequency and amplitude characteristics of the wind speed signal:

[0050] N(t)=N0·f(frequency(x′(t)),amplitude(x′(t)))

[0051] Where f is a function that adjusts the noise according to the frequency and amplitude characteristics of the signal, frequency(x′(t)) and amplitude(x′(t)) represent the frequency and amplitude characteristics of the signal x′(t), respectively;

[0052] Step 4: RPSEMD decomposition

[0053] For each intrinsic mode IM k , generating a sine wave S k (t):

[0054] S k (t) = A k sin(2πf k t+φ k )

[0055] In the formula, A k is the amplitude of the sine wave, f k is the frequency, φ k It is the phase;

[0056] The generated sine wave S k (t) is added to the original signal x′(t) to form a new signal to be decomposed x″(t):

[0057] x″(t)=x′(t)+S k (t)

[0058] Perform EMD decomposition on x″(t) to obtain a series of IMF components {IMF1,IMF2,...,IMF n} and residual component R;

[0059] Step 5: Phase Shift Processing

[0060] For each added sine wave S k (t) Phase shifting is performed to preserve the details of each IM and eliminate the additional sine wave:

[0061] φ k ′=φ k +Δφ

[0062] Where Δφ is the phase shift, which can be expressed as:

[0063] Step 6: Loop decorrelation operation

[0064] Calculate the correlation coefficient between the first IMF component IMF1 and the remaining components, and perform a cyclic decorrelation operation to ensure that the correlation coefficient between IMF1 and the remaining components is less than the given threshold ρ:

[0065]

[0066] Step 7: IMF Reconstruction

[0067] From the IMF components obtained in all iterations, the IMF with the highest correlation is selected as the final result;

[0068] Step 8: Residual item processing

[0069] Use wavelet transform to analyze and process the residual term R more finely:

[0070] R wavelet =wavelet transform(R,ψ,J)

[0071] Wherein, wavelet transform represents the wavelet transform function;

[0072] Step 9: Ensemble Averaging

[0073] The final IMFs are obtained by averaging the same order IMFs of all noisy signals:

[0074]

[0075] In the formula, IMF m ′ represents the IMF component obtained at the mth iteration;

[0076] Step 10: Reconstruct the signal

[0077] Subtracting the obtained IMFs from the original signal gives a new residual signal:

[0078] R new = x′(t)-IMF final

[0079] Step 11: Iterative Updates

[0080] Repeat steps 3-9 until the residual signal meets the stopping criterion or reaches the maximum number of iterations M;

[0081] Step 12: Output the results

[0082] Output the final IMF final and the residual signal R new, used for subsequent statistical analysis and prediction.

[0083] Furthermore, the system also includes a user module, and the data analysis module also includes a decision-making system;

[0084] The decision system is used to compare the final wind speed accuracy value obtained by the data analysis module with the estimated target wind speed, and update the decision information of the measuring device or related equipment according to the comparison result. If the actual wind speed is consistent with the target wind speed, the current operation is maintained; if there is a deviation, the module will prompt the possible adjustment measures;

[0085] The user module is connected to the data analysis module. After the decision system makes a decision, the decision support information is output to the user module and the information is visualized, including real-time wind speed data, historical wind speed trends, warning signals and adjustment suggestions, so that users can understand the status of the wind speed measurement system in a timely manner and respond.

[0086] Furthermore, the data analysis module is connected to the wind speed collection module via a communication interface, and the output result is fed back to the wind speed collection module and the data fusion algorithm module to iterate the wind speed collection strategy and data fusion, and utilize the feedback mechanism to enable the system to self-optimize.

[0087] Beneficial effects:

[0088] 1. The present invention obtains more comprehensive original wind speed input data by setting up various types of wind speed sensors, and can obtain more stable final wind speed data in different meteorological environments, effectively improving the robustness and stability of the system.

[0089] 2. In the data analysis module, the present invention combines the ANOWREEMD algorithm based on the ICEEMDAN algorithm, which integrates the adaptive noise addition strategy optimization and the residual term processing method, with the RPSEMD algorithm to effectively solve the modal mixing problem in the EMD, improve the accuracy of signal decomposition, and thus further improve the accuracy of the output wind speed.

[0090] 3. Compared with the conventional EMD algorithm, the ANOWREEMD and RPSEMD combined algorithm constructed by the present invention has a smaller algorithm volume due to the suppression effect of false components caused by the adaptive design of sinusoidal signals and the reduction of reconstruction errors. In the process of wind speed measurement, it reduces the deployment and computing power resources to a certain extent and reduces the measurement cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] Figure 1 It is a schematic diagram of the overall structure of the system of the present invention;

[0092] Figure 2 This is a schematic diagram of the processing steps of the RPSEMD algorithm combined with the ANOWREEMD algorithm of the present invention. DETAILED DESCRIPTION

[0093] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0094] like Figure 1-2 As shown, the present invention discloses a high-precision wind speed measurement system, the system comprising:

[0095] Wind speed acquisition module, with built-in wind speed sensor, is used to collect raw wind speed data;

[0096] The wind speed sensors built into the wind speed acquisition module include hot wire wind speed sensors, ultrasonic wind speed sensors and laser wind speed sensors. The original wind speed data includes:

[0097] The wind speed change data collected from the hot wire wind speed sensor is represented by x 热线 ;

[0098] The wind speed change data collected from the ultrasonic wind speed sensor is expressed as x 超声波 ;

[0099] The wind speed change data collected from the laser wind speed sensor is represented by x 激光 ;

[0100] The above data are input into the data fusion algorithm module as basic data after being preprocessed by the original signal processing module.

[0101] The original signal processing module is connected to the wind speed acquisition module and is used to amplify, filter and digitize the collected original wind speed data.

[0102] The data fusion algorithm module is connected to the original signal processing module to normalize the processed electrical signals, establish the optimal fusion set to calculate the wind speed sensor weight coefficient and perform the final fusion to obtain the preliminary wind speed measurement value. The following is the data processing process of the data fusion algorithm module:

[0103] x 热线 , x 超声波 , x 激光It reflects the wind speed detected by different sensors at a certain moment or time period, including the wind speed and direction parameters. After data preprocessing by the original signal processing module, it is input into the data fusion algorithm module for normalization processing. The purpose of normalization is to eliminate the differences in the dimensions and numerical ranges of different sensor data, ensure that the data is at the same level, and make the data collected by different sensors or different scenarios comparable and consistent. The formula is as follows:

[0104]

[0105] Where x represents the raw data x of any sensor 热线 , x 超声波 , x 激光 ; x′ is the normalized data value, x min is the minimum value among all sensor data sets, x max is the maximum value among all sensor data sets.

[0106] Determine the optimal fusion set:

[0107] Based on the normalized data, a threshold is defined to determine the optimal fusion set and exclude outliers. The determination of the optimal fusion set helps to improve the accuracy and stability of wind speed measurement. The data in the optimal fusion set is used for subsequent data processing and analysis to reduce the impact of outliers on the final results. The formula for determining the optimal fusion set is as follows:

[0108]

[0109] Where: S opt is the optimal fusion set, x i is the observation value of the ith sensor at time t, is the average observation value of all sensors at time t, τ is the preset threshold, and the i-th sensor refers to any sensor in the data acquisition module, specifically a hot wire, ultrasonic or laser wind speed sensor, depending on their number or identification in the data acquisition module, the hot wire wind speed sensor is 1, the ultrasonic wind speed sensor is 2, and the laser wind speed sensor is 3.

[0110] Calculate the fusion degree matrix:

[0111] Calculate the fusion matrix to evaluate the consistency between sensor data. The calculation of the fusion matrix is ​​based on the data in the optimal fusion set to ensure the reliability of the data. The calculation formula of the fusion matrix is ​​as follows:

[0112]

[0113] Among them, F ij is the fusion degree between sensors i and j at time t, xj is the observation value of the jth sensor at time t, and σ is the standard deviation, which is used to control the sensitivity of the fusion degree.

[0114] Calculate the consistent fusion degree and distribution balance degree:

[0115] The consistent fusion degree and distribution balance degree of each sensor data are calculated for the allocation of weight coefficients to ensure the accuracy of the final wind speed estimation. The calculation formulas for consistent fusion degree and distribution balance degree are as follows:

[0116]

[0117] Among them, C i is the consistent fusion degree of the i-th sensor at time t, E i is the distribution balance of the i-th sensor at time t. According to the consistent fusion degree and distribution balance, a weight coefficient is assigned to each sensor. The weight coefficient allocation formula is as follows:

[0118]

[0119] Among them, w i is the weight coefficient of the ith sensor at time t, which will be used to calculate the preliminary wind speed value.

[0120] Fuse and output preliminary wind speed measurements:

[0121] The weighted average is calculated as the final wind speed estimate to obtain a more accurate and stable wind speed measurement result. The formula for the final fusion estimate is as follows:

[0122]

[0123] in, is the estimated wind speed at time t, and n is the total number of sensors, i.e., the sum of hot wire wind speed sensors, ultrasonic wind speed sensors, and laser wind speed sensors.

[0124] The above is the data processing process of the data fusion algorithm module. The data analysis module is connected to the data fusion algorithm module, integrating the ANOWREEMD algorithm based on the ICEEMDAN algorithm fusion adaptive noise addition strategy optimization and residual term processing method and the RPSEMD algorithm to process EMD mode mixing, which is used to further refine the preliminary wind speed measurement value and output the final IMFsIMF final and the residual signal R new For statistical analysis and prediction,

[0125] The specific steps of data processing in the data analysis module are as follows:

[0126] Step 1: Data Reception

[0127] This step involves receiving the output data from the data fusion algorithm module, which has been normalized, outliers excluded, and multi-sensor data fused to form a set of reliable preliminary wind speed measurements. These measurements are the basis for subsequent analysis.

[0128] Step 2: Application of ANOWREEMD algorithm integrated with RPSEMD

[0129] On the basis of step 1, the fused data is further processed by the adaptive noise optimization of the integrated regenerative phase-shifted sine-assisted empirical mode decomposition algorithm (RPSEMD algorithm) and the ensemble empirical mode decomposition algorithm with wavelet residual enhancement (ANOWREEMD algorithm). The ANOWREEMD algorithm is based on the ICEEMDAN algorithm and integrates the optimization of the adaptive noise addition strategy and the improvement of the residual term processing method. The ANOWREEMD algorithm combined with the RPSEMD algorithm can effectively solve the mode mixing problem in EMD and improve the accuracy of signal decomposition. The data and results of each step provide input for the subsequent steps to form a coherent data processing flow. Compared with the traditional EMD, this method has a smaller computational scale and can provide more reasonable decomposition results. The preliminary measured wind speed signal is decomposed by the algorithm to obtain a series of intrinsic mode functions (IMFs).

[0130] Step 3: IMFs statistical analysis

[0131] Based on the IMFs obtained in step 2, statistical analysis is performed, including calculating statistical parameters such as the standard deviation, maximum value, and minimum value of each IMF. These parameters reflect the characteristics of wind speed signals at different frequency scales.

[0132] Step 4: Determine the exact value

[0133] The statistical parameters obtained in step 3 are combined with the advantages of the ANOWREEMD algorithm integrated with RPSEMD to determine the exact value of the wind speed. The ANOWREEMD algorithm integrated with RPSEMD can effectively reduce noise interference and improve the accuracy and robustness of the decomposition by dynamically adjusting the noise level based on signal characteristics and noise intensity.

[0134] The specific steps of RPSEMD algorithm combined with ANOWREEMD algorithm to process output data are as follows:

[0135] Step 1: Initialize algorithm parameters

[0136] Set the initial noise amplitude N0, the number of iterations M, the number of IMFs to be selected K, and the wavelet transform parameters, including the wavelet basis ψ and the number of decomposition levels J;

[0137] Step 2: Data preprocessing

[0138] Detrending the wind speed signal x(t) yields x′(t):

[0139] x′(t)=x(t)-detrend(x(t))

[0140] Where detrend(x(t)) means removing the trend component of the signal x(t) to reduce the modal aliasing in the subsequent decomposition;

[0141] Step 3: Adaptive noise addition

[0142] Dynamically adjust the amplitude and frequency of the noise according to the frequency and amplitude characteristics of the wind speed signal:

[0143] N(t)=N0·f(frequency(x′(t)),amplitude(x′(t)))

[0144] Where f is a function that adjusts the noise according to the frequency and amplitude characteristics of the signal, and frequency(x′(t)) and amplitude(x′(t)) represent the frequency and amplitude characteristics of the signal x′(t), respectively.

[0145] Step 4: RPSEMD decomposition

[0146] For each intrinsic mode IM k , generating a sine wave S k (t):

[0147] S k (t) = A k sin(2πf k t+φ k )

[0148] In the formula, A k is the amplitude of the sine wave, f k is the frequency, φ k It is the phase;

[0149] The generated sine wave S k (t) is added to the original signal x′(t) to form a new signal to be decomposed x″(t):

[0150] x″(t)=x′(t)+S k (t)

[0151] Perform EMD decomposition on x″(t) to obtain a series of IMF components {IMF1,IMF2,...,IMF n} and residual component R;

[0152] Step ⑤: Phase shift processing

[0153] For each added sine wave S k (t) Phase shifting is performed to preserve the details of each IM and eliminate the additional sine wave:

[0154] φ k ′=φ k +Δφ

[0155] Where Δφ is the phase shift, which can be expressed as:

[0156] Step 6: Circular decorrelation operation

[0157] Calculate the correlation coefficient between the first IMF component IMF1 and the remaining components, and perform a cyclic decorrelation operation to ensure that the correlation coefficient between IMF1 and the remaining components is less than the given threshold ρ:

[0158]

[0159] Step 7: IMF reconstruction

[0160] From all the IMF components obtained in the iterations, select the IMF with the highest correlation as the final result; Step ⑧: Residual term processing

[0161] Use wavelet transform to analyze and process the residual term R more finely:

[0162] R wavelet =wavelet transform(R,ψ,J)

[0163] Wherein, wavelet transform represents the wavelet transform function;

[0164] Step 9: Ensemble average

[0165] The final IMFs are obtained by averaging the same order IMFs of all noisy signals:

[0166]

[0167] In the formula, IMF m ′ represents the IMF component obtained in the mth iteration.

[0168] Step ⑩: Reconstruct the signal

[0169] Subtracting the obtained IMFs from the original signal gives a new residual signal:

[0170] R new = x′(t)-IMF final

[0171] step Iterative Updates

[0172] Repeat steps 3-9 until the residual signal meets the stopping criterion or reaches the maximum number of iterations M;

[0173] step Result Output

[0174] Output the final IMF final and the residual signal R new , used for subsequent statistical analysis and prediction.

[0175] The data analysis module also has a decision system for updating the decision information of the measuring device or related equipment according to the output final result. The following steps are performed by the decision system:

[0176] Step 5: Target Wind Speed ​​Comparison

[0177] Compare the precise wind speed value determined in step 4 with the preset target wind speed. This step is the core of the decision-making process and determines whether action is needed to adjust the wind speed measurement system or related equipment.

[0178] Step 6: Decision support

[0179] Based on the comparison results in step 5, the data analysis module generates decision support information. If the actual wind speed matches the target wind speed, the current operation is maintained; if there is a deviation, the module will prompt the possible adjustment measures.

[0180] The user module is connected to the data analysis module. After the decision system makes a decision, it outputs the decision support information to the user module and visualizes the information, including real-time wind speed data, historical wind speed trends, warning signals and adjustment suggestions, so that users can understand the status of the wind speed measurement system in a timely manner and respond.

[0181] During the operation of the system, the data analysis module is connected to the wind speed collection module via the communication interface. The output results are fed back to the wind speed collection module and the data fusion algorithm module to iterate the wind speed collection strategy and data fusion, and use the feedback mechanism to enable the system to self-optimize.

[0182] The above description of the embodiments enables professionals and technicians in the field to implement or use the present invention. Various modifications to the embodiments will be apparent to professionals and technicians. The general principles of the present invention can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention should not be limited to the embodiments shown herein, but should cover the widest range consistent with the principles and novel features disclosed in the present invention.

Claims

1. A high-precision wind speed measurement system, characterized in that: The system comprises: Wind speed acquisition module, with built-in wind speed sensor, is used to collect raw wind speed data; The original signal processing module is connected to the wind speed acquisition module and is used to amplify, filter and digitize the collected original wind speed data; The data fusion algorithm module is connected to the original signal processing module to normalize the processed electrical signals, establish the optimal fusion set to calculate the wind speed sensor weight coefficient and perform the final fusion to obtain the preliminary wind speed measurement value; The data analysis module is connected to the data fusion algorithm module, integrating the ANOWREEMD algorithm based on the ICEEMDAN algorithm fusion adaptive noise addition strategy optimization and residual term processing method and the RPSEMD algorithm to process EMD mode mixing, which is used to further refine the preliminary wind speed measurement value and output the final IMF final and the residual signal R new , used for statistical analysis and forecasting.

2. The high-precision wind speed measurement system according to claim 1, characterized in that: The wind speed sensors built into the wind speed acquisition module include hot wire wind speed sensors, ultrasonic wind speed sensors and laser wind speed sensors. The original wind speed data includes: The wind speed change data collected from the hot wire wind speed sensor is represented by x 热线 ; The wind speed change data collected from the ultrasonic wind speed sensor is expressed as x 超声波 ; The wind speed change data collected from the laser wind speed sensor is represented by x 激光 ; The above data are input into the data fusion algorithm module as basic data after being preprocessed by the original signal processing module.

3. The high-precision wind speed measurement system according to claim 2, characterized in that: The x 热线 , x 超声波 , x 激光 It reflects the wind speed detected by different sensors at a certain moment or time period, including the wind speed and direction parameters. After data preprocessing by the original signal processing module, it is input into the data fusion algorithm module for normalization processing. The formula is as follows: Where x represents the raw data x of any sensor 热线 , x 超声波 , x 激光 ; x′ is the normalized data value, x min is the minimum value in all sensor data sets, x max is the maximum value among all sensor data sets.

4. The high-precision wind speed measurement system according to claim 1 or 3, characterized in that: The detailed steps of the data fusion algorithm module to establish the optimal fusion set, calculate the wind speed sensor weight coefficient and perform the final fusion are as follows: Determine the optimal fusion set: Based on the normalized data, a threshold is defined to determine the optimal fusion set, exclude outliers, and use the data in the optimal fusion set for subsequent data processing and analysis to reduce the impact of outliers on the final results. The formula for determining the optimal fusion set is as follows: Where: S opt is the optimal fusion set, x i is the observation value of the ith sensor at time t, is the average observation value of all sensors at time t, τ is the preset threshold, the i-th sensor refers to any sensor in the data acquisition module, specifically a hot wire, ultrasonic or laser wind speed sensor, depending on their number or identification in the data acquisition module, the hot wire wind speed sensor is 1, the ultrasonic wind speed sensor is 2, and the laser wind speed sensor is 3; Calculate the fusion degree matrix: Based on the optimal fusion set data, calculate the fusion degree matrix: Among them, F ij is the fusion degree between sensors i and j at time t, x j is the observation value of the jth sensor at time t, σ is the standard deviation, which is used to control the sensitivity of the fusion degree; Calculate the consistent fusion degree and distribution balance degree: Calculate the consistent fusion degree and distribution balance of each sensor data for the allocation of weight coefficients. The calculation formulas for consistent fusion degree and distribution balance are as follows: Among them, C i is the consistent fusion degree of the i-th sensor at time t, E i is the distribution balance of the i-th sensor at time t. According to the consistent fusion degree and distribution balance, a weight coefficient is assigned to each sensor. The weight coefficient assignment formula is as follows: Among them, w i is the weight coefficient of the ith sensor at time t, which will be used to calculate the preliminary wind speed value; Fuse and output preliminary wind speed measurements: The formula for the final fusion estimate is as follows: in, is the estimated wind speed at time t, and n is the total number of sensors, i.e., the sum of hot wire wind speed sensors, ultrasonic wind speed sensors, and laser wind speed sensors.

5. The high-precision wind speed measurement system according to claim 1, characterized in that: The data analysis module deployment is detailed as follows: This module receives the output data from the data fusion algorithm module, and further processes the output data using the adaptive noise optimization of the integrated regenerative phase-shifted sine-assisted empirical mode decomposition algorithm RPSEMD algorithm and the ensemble empirical mode decomposition algorithm ANOWREEMD with wavelet residual enhancement. The ANOWREEMD algorithm integrates the adaptive noise addition strategy optimization and the improvement of the residual term processing method on the basis of the ICEEMDAN algorithm. The preliminary wind speed measurement value is decomposed by the RPSEMD and ANOWREEMD combined algorithm to obtain the intrinsic mode function IMFs, and the standard deviation, maximum and minimum parameters of each IMF are used for statistical analysis to confirm the final accurate wind speed value.

6. The high-precision wind speed measurement system according to claim 5, characterized in that: The specific steps of the RPSEMD algorithm combined with the ANOWREEMD algorithm to process the output data are as follows: Step 1: Initialize algorithm parameters Set the initial noise amplitude N0, the number of iterations M, the number of IMFs to be selected K, and the wavelet transform parameters, including the wavelet basis ψ and the number of decomposition levels J; Step 2: Data Preprocessing Detrending the wind speed signal x(t) yields x′(t): x′(t)=x(t)-detrend(x(t)) Where detrend(x(t)) means removing the trend component of the signal x(t) to reduce the modal aliasing in the subsequent decomposition; Step 3: Adaptive Noise Addition Dynamically adjust the amplitude and frequency of the noise according to the frequency and amplitude characteristics of the wind speed signal: N(t)=N0·f(frequency(x′(t)),amplitude(x′(t))) Where f is a function that adjusts the noise according to the frequency and amplitude characteristics of the signal, frequency(x′(t)) and amplitude(x′(t)) represent the frequency and amplitude characteristics of the signal x′(t), respectively; Step 4: RPSEMD decomposition For each intrinsic mode IM k , generating a sine wave S k (t): S k (t)=A k sin(2πf k t+φ k ) In the formula, A k is the amplitude of the sine wave, f k is the frequency, φ k It is the phase; The generated sine wave S k (t) is added to the original signal x′(t) to form a new signal to be decomposed x″(t): x″(t)=x′(t)+S k (t) Perform EMD decomposition on x″(t) to obtain a series of IMF components {IMF1,IMF2,...,IMF n } and residual component R; Step 5: Phase Shift Processing For each added sine wave S k (t) Phase shifting is performed to preserve the details of each IM and eliminate the additional sine wave: f k ′=φ k +Df Where Δφ is the phase shift, which can be expressed as: Step 6: Loop decorrelation operation Calculate the correlation coefficient between the first IMF component IMF1 and the remaining components, and perform a cyclic decorrelation operation to ensure that the correlation coefficient between IMF1 and the remaining components is less than the given threshold ρ: Step 7: IMF Reconstruction From the IMF components obtained in all iterations, the IMF with the highest correlation is selected as the final result; Step 8: Residual item processing Use wavelet transform to analyze and process the residual term R more finely: R wavelet =wavelet transform(R,ψ,J) Wherein, wavelet transform represents the wavelet transform function; Step 9: Ensemble Averaging The final IMFs are obtained by averaging the same order IMFs of all noisy signals: In the formula, IMF m ′ represents the IMF component obtained at the mth iteration; Step 10: Reconstruct the signal Subtracting the obtained IMFs from the original signal gives a new residual signal: R new =x′(t)-IMF final Step 11: Iterative Updates Repeat steps 3-9 until the residual signal meets the stopping criterion or reaches the maximum number of iterations M; Step 12: Output the results Output the final IMF final and the residual signal R new , used for subsequent statistical analysis and prediction.

7. The high-precision wind speed measurement system according to claim 1 or 5, characterized in that: The system also includes a user module, and the data analysis module also includes a decision-making system; The decision system is used to compare the final wind speed accuracy value obtained by the data analysis module with the estimated target wind speed, and update the decision information of the measuring device or related equipment according to the comparison result. If the actual wind speed is consistent with the target wind speed, the current operation is maintained; if there is a deviation, the module will prompt the possible adjustment measures; The user module is connected to the data analysis module. After the decision system makes a decision, the decision support information is output to the user module and the information is visualized, including real-time wind speed data, historical wind speed trends, warning signals and adjustment suggestions, so that users can understand the status of the wind speed measurement system in a timely manner and respond.

8. The high-precision wind speed measurement system according to claim 1, characterized in that: The data analysis module is connected to the wind speed acquisition module via a communication interface, and the output result is fed back to the wind speed acquisition module and the data fusion algorithm module to iterate the wind speed acquisition strategy and data fusion, and use the feedback mechanism to enable the system to self-optimize.

Images