A method and system for evaluating the measurement uncertainty of wind direction calibration results

Abstract

The application provides a wind direction calibration result measurement uncertainty evaluation method and system, signal transmission time between two groups of probes in a two-dimensional plane is acquired, wind direction indication values are calculated according to the relationship among the distance between the probes, the sound propagation speed and the signal transmission time; a measurement model is established based on external environmental factors and internal factors that affect the inaccuracy of the wind direction indication values and the standard wind direction values; the standard uncertainty is determined and calculated according to the measurement model; the sensitivity coefficient is determined according to the function partial derivative of each input parameter in the measurement model; the expanded uncertainty is determined according to the probability synthesis method based on the Bayesian network and the confidence interval prediction method of machine learning; specifically, starting from the measurement working principle of the ultrasonic wind direction and speed sensor, the factors affecting the wind direction measurement indication values are comprehensively analyzed, the sources of the uncertainty in the measurement process can be more accurately identified and evaluated, and the accuracy of the measurement uncertainty evaluation is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wind direction measurement uncertainty evaluation, and particularly relates to a wind direction calibration result measurement uncertainty evaluation method and system. BACKGROUND

[0002] In order to determine the measurement accuracy of an ultrasonic wind direction and speed sensor, the ultrasonic wind direction and speed sensor is usually calibrated according to the requirements of JJF 1934-2021 Ultrasonic Wind Direction and Speed Measuring Instrument Calibration Specification. In the wind direction calibration process, the deviation between the wind direction indication value of the ultrasonic wind direction and speed sensor and the indication value (i.e. standard wind direction value) of the wind direction standard device is calculated by comparing the two, and the deviation is usually expressed as an indication error. The indication error can reflect the accuracy of the wind direction measurement value of the ultrasonic wind direction and speed sensor, and thus determine its usability.

[0003] Measurement uncertainty is an important parameter for evaluating the reliability of measurement results, which represents the dispersion of the measured value and the degree to which the measurement result represents the true value of the measured value. In order to ensure the accuracy and reliability of the indication error of the wind direction of the ultrasonic wind direction and speed sensor, the measurement uncertainty of the indication error needs to be evaluated.

[0004] At present, the evaluation of the measurement uncertainty of the wind speed sensor generally adopts the GUM method published by the ISO international organization in 1993. However, the evaluation of the wind direction indication value of the ultrasonic wind direction and speed sensor still has some deficiencies. Usually, the measured wind direction indication value and the standard wind direction indication value are repeatedly measured, and then the A-type standard uncertainty is calculated according to the Bessel formula to obtain the final evaluation result. The factors that may affect the measurement accuracy of the ultrasonic sensor itself are not analyzed in depth from the structure and working principle of the ultrasonic sensor. Therefore, the existing evaluation method may cause deviation in the uncertainty evaluation result of the indication error of the wind direction of the ultrasonic wind direction and speed sensor. SUMMARY

[0005] Therefore, the application provides a wind direction calibration result measurement uncertainty evaluation method and system, which aims to improve the accuracy of the final wind direction measurement uncertainty evaluation result of the ultrasonic wind direction and speed sensor.

[0006] The first aspect of the embodiment of the present application provides a wind direction calibration result measurement uncertainty evaluation method, which is applied to an ultrasonic wind direction and speed sensor and an angle encoder, the ultrasonic wind direction and speed sensor includes two groups of probes which are orthogonal to each other and are used to collect signal transmission time between the probes, and the angle encoder is used to collect a standard wind direction value, and the method comprises the following steps:

[0007] acquiring signal transmission time between the two groups of probes in a two-dimensional plane, and calculating a wind direction indication value according to a relationship among a distance between the probes, a sound propagation speed and the signal transmission time;

[0008] establishing a measurement model based on external environmental factors and internal factors affecting inaccuracy of the wind direction indication value and the standard wind direction value, wherein the external environmental factors include airflow disturbance, obstacles, air density change and electromagnetic interference, the internal factors affecting inaccuracy of the wind direction indication value include sensitivity difference of a signal module in the ultrasonic wind direction and speed sensor and quantization error of an AD converter in the ultrasonic wind direction and speed sensor, and the internal factors affecting inaccuracy of the standard wind direction value include device hardware and performance defect influence and installation error influence;

[0009] determining and calculating a standard uncertainty according to the measurement model, wherein the standard uncertainty includes a standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane and a standard uncertainty of the angle encoder corresponding to the wind direction indication value;

[0010] determining a sensitivity coefficient of the corresponding standard uncertainty according to a function partial derivative of each input parameter in the measurement model;

[0011] calculating a combined standard uncertainty according to parameters of the measurement model and a Bayesian network-based probability synthesis method, and determining an expanded uncertainty by a machine learning confidence interval prediction method according to the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty and a corresponding actual expanded uncertainty.

[0012] Further, a calculation formula of the measurement model is as follows:

[0013] ;

[0014] is a wind direction measurement error value, μ1 is an airflow disturbance influence factor, μ2 is an obstacle influence factor, μ3 is an air density change influence factor, μ4 is an electromagnetic interference influence factor, is a standard wind direction value measured by the angle encoder, D a is a sensitivity difference of the signal module in the ultrasonic wind direction and speed sensor, D bFor the quantization error of the AD converter in the ultrasonic wind direction and speed sensor, D1 is the wind direction influence value of the angle encoder bearing rotation hysteresis, D2 is the wind direction change influence value caused by the mechanical wind direction rotation inertia of the angle encoder, D3 is the wind direction influence value corresponding to the installation reference deviation of the angle encoder, D4 is the wind direction influence value corresponding to the insufficient fixing stability of the angle encoder, t ab For the ultrasonic wind direction and speed sensor a probe generating an ultrasonic signal, the time experienced by the ultrasonic signal received by the ultrasonic wind direction and speed sensor b probe, t ba For the ultrasonic wind direction and speed sensor b probe generating an ultrasonic signal, the time experienced by the ultrasonic signal received by the ultrasonic wind direction and speed sensor a probe, t cd For the ultrasonic wind direction and speed sensor c probe generating an ultrasonic signal, the time experienced by the ultrasonic signal received by the ultrasonic wind direction and speed sensor d probe, t dc For the ultrasonic wind direction and speed sensor d probe generating an ultrasonic signal, the time experienced by the ultrasonic signal received by the ultrasonic wind direction and speed sensor c probe.

[0015] Further, in the step of determining and calculating the standard uncertainty according to the measurement model, the standard uncertainty is divided into A-type standard uncertainty and B-type standard uncertainty according to the repeatability measurement and the non-repeatability measurement, wherein in the two-dimensional plane, the standard uncertainty introduced by the signal transmission time between the two groups of probes in the repeatability measurement is the A-type standard uncertainty, the standard uncertainty introduced by the wind direction indication value of the angle encoder in the repeatability measurement is the A-type standard uncertainty, the airflow disturbance influence factor, the obstacle influence factor, the air density change influence factor, the electromagnetic interference influence factor, the angle encoder bearing rotation hysteresis wind direction influence value, the wind direction change influence value caused by the mechanical wind direction rotation inertia of the angle encoder, the wind direction influence value corresponding to the installation reference deviation of the angle encoder, and the wind direction influence value corresponding to the insufficient fixing stability of the angle encoder are the B-type standard uncertainty.

[0016] Further, the A-type standard uncertainty is calculated by using the Bessel formula, and the calculation formula is:

[0017] ;

[0018] ;

[0019] wherein x i is the i-th input parameter, is the average value, n is the corresponding measurement number when the A-type standard uncertainty is evaluated, u A is the A-type standard uncertainty.

[0020] Further, the B-class standard uncertainty supposes the value interval [x-w, x+w] of the measured value according to the measurement information, wherein w represents the interval half-width, supposes the probability distribution of the measured value, and determines k according to the probability distribution. Specifically, when the probabilities of each point in the value interval of the measured value are equal, the uniform distribution is satisfied, k= ; when the probability of the measured value appearing at the interval endpoint value is greater than the probability of appearing at the intermediate value, the inverse sine distribution is satisfied, k= ; and when the probability of the measured value appearing at the interval endpoint value is less than the probability of appearing at the intermediate value, the normal distribution is satisfied, k=2.

[0021] Further, the calculation formula of the B-class standard uncertainty is as follows:

[0022] u B =w / k;

[0023] wherein u B is the B-class standard uncertainty, w is the interval half-width, and k is the probability distribution parameter.

[0024] Further, the step of calculating the combined standard uncertainty according to the parameters of the measurement model and the probability synthesis method based on the Bayesian network, and determining the expanded uncertainty according to the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty, and the corresponding actual expanded uncertainty by the confidence interval prediction method of machine learning includes:

[0025] constructing each input and output of the measurement model into a Bayesian network structure, wherein each node represents a variable, and the directed edge between the nodes represents the causal relationship between the variables;

[0026] determining the probability distribution of each node, and using the inference algorithm of the Bayesian network to infer the posterior probability distribution of the output according to the probability distribution of each input;

[0027] determining the combined standard uncertainty by calculating the standard deviation of the posterior probability distribution of the output;

[0028] obtaining historical data under different measurement conditions, wherein the historical data includes the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty, and the corresponding actual expanded uncertainty, and constructing a machine learning model, wherein the standard uncertainty, the sensitivity coefficient, and the combined standard uncertainty of each input are used as features, and the actual expanded uncertainty is used as a target variable for training;

[0029] inputting the standard uncertainty, the sensitivity coefficient, and the combined standard uncertainty of each input obtained by the current measurement into the trained machine learning model, and outputting the expanded uncertainty.

[0030] The second aspect of the embodiment of the present application provides a wind direction calibration result measurement uncertainty evaluation system, the system comprising:

[0031] The first calculation module is configured to obtain the signal transmission time between the two groups of probes in the two-dimensional plane, and calculate the wind direction indication value according to the relationship among the distance between the probes, the sound propagation speed and the signal transmission time.

[0032] The model establishment module is configured to subtract the wind direction indication value from the standard wind direction value based on the internal factors affecting the inaccuracy of the standard wind direction value, and establish a measurement model.

[0033] The second calculation module is configured to determine and calculate the standard uncertainty according to the measurement model, wherein the standard uncertainty comprises the standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane and the standard uncertainty of the angle encoder corresponding to the wind direction indication value.

[0034] The first determination module is configured to determine the sensitivity coefficient of the corresponding standard uncertainty according to the partial derivative of each input parameter in the measurement model.

[0035] The second determination module is configured to calculate the combined standard uncertainty according to the parameters of the measurement model and the Bayesian network-based probability synthesis method, and determine the expanded uncertainty by the machine learning confidence interval prediction method according to the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty and the corresponding actual expanded uncertainty.

[0036] The third aspect of the embodiment of the present application provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the wind direction calibration result measurement uncertainty evaluation method provided by the first aspect.

[0037] The fourth aspect of the embodiment of the present application provides an electronic device, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor realizes the wind direction calibration result measurement uncertainty evaluation method provided by the first aspect when executing the program.

[0038] This invention provides a method and system for evaluating the measurement uncertainty of wind direction calibration results. The method involves acquiring the signal transmission time between two sets of probes in a two-dimensional plane, calculating the wind direction indication based on the relationship between the probe spacing, sound propagation speed, and signal transmission time; establishing a measurement model based on external environmental factors and internal factors affecting the inaccuracy of the wind direction indication compared to the standard wind direction value; determining and calculating the standard uncertainty based on the measurement model, including the standard uncertainty introduced by the signal transmission time between the two sets of probes in the two-dimensional plane, and the standard uncertainty of the angle encoder in the corresponding wind direction indication; and determining the standard uncertainty based on the input parameters in the measurement model. The partial derivatives of the function are used to determine the sensitivity coefficient of the corresponding standard uncertainty. Based on the parameters of the measurement model and the probability synthesis method based on Bayesian networks, the combined standard uncertainty is calculated. Then, based on the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty, and the corresponding actual expanded uncertainty, the expanded uncertainty is determined by the confidence interval prediction method of machine learning. Specifically, by starting from the measurement working principle of ultrasonic wind direction and speed sensors, a comprehensive analysis of the factors affecting their wind direction measurement values ​​can more accurately identify and evaluate the sources of uncertainty in the measurement process, which helps to reduce the deviation of the evaluation results and improve the accuracy of measurement uncertainty evaluation. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of an ultrasonic wind direction and speed sensor accuracy measurement system.

[0040] Figure 2 A schematic diagram illustrating the working principle of an ultrasonic wind direction and speed sensor;

[0041] Figure 3 A schematic diagram of the ultrasonic wind direction and speed sensor;

[0042] Figure 4 This is a flowchart illustrating the implementation of a method for evaluating the measurement uncertainty of wind direction calibration results according to Embodiment 1 of the present invention.

[0043] Figure 5 This is a structural block diagram of a wind direction calibration result measurement uncertainty evaluation system provided in Embodiment 3 of the present invention;

[0044] Figure 6 This is a schematic diagram of the electronic device in Embodiment 4 of the present invention. Detailed Implementation

[0045] For the purpose of promoting an understanding of the principles of the application, reference will now be made to the embodiments illustrated in the drawings. There is shown in the drawings several embodiments of the application. It is expressly understood that the drawings are only for the purpose of illustration and are not intended to limit the present application in any in any way. In the drawings:

[0046] It should be noted that when an element is referred to as being "on" another element, it can be directly on the other element or intervening elements can also be present. In addition, it should be noted that when an element is referred to as being "connected", "coupled", or "linked" to another element, it can be directly connected, coupled, or linked to the other element or intervening elements can also be present. As used herein the terms "vertical", "horizontal", "left", "right", and the like are merely used for the purpose of illustration and are not intended to limit the present application in any way.

[0047] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description of the application herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0048] Reference will now be made to the drawings, wherein: Figure 1 , it should be noted that the ultrasonic wind direction and speed sensor is a two-dimensional ultrasonic wind direction and speed sensor. According to the requirements of the calibration specification, the ultrasonic wind direction and speed sensor and the angle encoder are usually placed in a stable wind speed environment (wind tunnel) during the design of the experimental scheme. By comparing the values of the ultrasonic wind direction and speed sensor and the angle encoder (standard wind direction value), the error value of the wind direction measurement value of the ultrasonic wind direction and speed sensor can be calculated.

[0049] Before calibration, the ultrasonic wind direction and speed sensor and the angle encoder are fixed on the horizontal placement table of the wind tunnel base, and the wind direction of the sensor is aligned with the orientation of the angle encoder. When a certain wind speed is set in the wind tunnel, the horizontal placement table is rotated at the same time. The wind direction values of the sensor and the angle encoder are read at the same time, and the deviation between the two is compared to obtain the corresponding sensor wind direction value error. It can be understood that the measurement model established in the prior art is:

[0050] (1);

[0051] wherein, is the wind direction measurement error value, D is the wind direction value of the ultrasonic wind direction and speed sensor, D sFor the standard wind direction value, it should be noted that the measurement model of the wind direction indication error of the ultrasonic wind direction and speed sensor in the prior research ignores the internal factors affecting the inaccurate indication of the ultrasonic wind direction and speed sensor.

[0052] When the internal factors of the inaccurate indication of the ultrasonic wind direction and speed sensor, the working principle of the ultrasonic wind direction and speed sensor should be analyzed first, please refer to Figure 2 and Figure 3 , Figure 2 is a schematic diagram of the working principle of the ultrasonic wind direction and speed sensor, Figure 3 is a schematic diagram of the structure of the ultrasonic wind direction and speed sensor, specifically, the ultrasonic wind direction and speed sensor is divided into two groups of probes, a probe 1 and b probe 2, c probe 3 and d probe 4, the ab direction is the x axis, the cd direction is the y axis, and the two groups of probes are orthogonal to each other, wherein four probes are installed with ultrasonic transducers, each transducer has the functions of generating and receiving ultrasonic waves, and the two groups of ultrasonic wind direction and speed sensor probes measure the wind speed in two different directions by sending and receiving ultrasonic signals, and then the time difference of the two ultrasonic wave transmissions is used to measure the wind speed in two different directions, and the wind speed vector in the x axis is v x , the corresponding wind speed vector in the y axis is v y , then the wind speed in two different directions is synthesized to obtain the corresponding wind speed:

[0053] (2)

[0054] The actual wind speed is v, the distance between a probe 1 and b probe 2 and the distance between c probe 3 and d probe 4 are both L, and the sound speed is c.

[0055] Example one

[0056] Please refer to Figure 4 , Figure 4 shows a flow chart of a wind direction calibration result measurement uncertainty evaluation method provided by the embodiment one of the application, which is applied to an ultrasonic wind direction and speed sensor and an angle encoder, the ultrasonic wind direction and speed sensor includes two groups of probes, and the two groups of probes are orthogonal to each other, and is used to collect the signal transmission time between the probes, the angle encoder is used to collect the standard wind direction value, and the method specifically includes steps S01 to S05.

[0057] Step S01, obtaining the signal transmission time between the two groups of probes in a two-dimensional plane, calculating the wind direction indication value according to the relationship among the distance between the probes, the sound propagation speed and the signal transmission time.

[0058] It should be noted that the time experienced by the ultrasonic wind direction and speed sensor a probe generating an ultrasonic signal and the ultrasonic wind direction and speed sensor b probe receiving the ultrasonic signal is t ab :

[0059] (3)

[0060] wherein L is the distance between the a probe and the b probe, and the unit is m;

[0061] c is the sound propagation speed, and the unit is m / s;

[0062] v x is the wind speed vector value of the wind speed v in the x axis, and the unit is m / s.

[0063] The time experienced by the ultrasonic wind direction and speed sensor b probe generating an ultrasonic signal and the ultrasonic wind direction and speed sensor a probe receiving the ultrasonic signal is t ba :

[0064] (4)

[0065] According to formula (3) and formula (4), the corresponding wind speed vector of the wind speed v in the x axis is v x :

[0066] (5)

[0067] Similarly, the corresponding wind speed vector of the wind speed v in the y axis is v y :

[0068] (6)

[0069] According to the above formula, the corresponding ultrasonic wind direction and speed sensor indication v can be obtained:

[0070] (7)

[0071] According to the above formula, and considering that the ultrasonic wind speed sensor belongs to an electronic product, the sensitivity difference of the signal module and the quantization error of the AD converter will cause linear deviation of the wind direction data, and finally the wind direction indication D of the ultrasonic wind direction and speed sensor is:

[0072] (8)

[0073] t ab is the time experienced by the ultrasonic wind direction and speed sensor a probe generating an ultrasonic signal and the ultrasonic wind direction and speed sensor b probe receiving the ultrasonic signal, t baThe time t experienced by the ultrasonic wind direction and speed sensor a probe to receive the ultrasonic signal generated by the ultrasonic wind direction and speed sensor b probe cd The time t experienced by the ultrasonic wind direction and speed sensor d probe to receive the ultrasonic signal generated by the ultrasonic wind direction and speed sensor c probe dc The time t experienced by the ultrasonic wind direction and speed sensor d probe to receive the ultrasonic signal generated by the ultrasonic wind direction and speed sensor c probe a The sensitivity difference of the signal module in the ultrasonic wind direction and speed sensor b The quantization error of the AD converter in the ultrasonic wind direction and speed sensor.

[0074] Step S02, based on external environmental factors, and internal factors affecting the inaccuracy of the wind direction value and the standard wind direction value, a measurement model is established.

[0075] Wherein, the external environmental factors include airflow disturbance, obstacles, air density changes and electromagnetic interference, the internal factors affecting the inaccuracy of the wind direction value include the sensitivity difference of the signal module in the ultrasonic wind direction and speed sensor and the quantization error of the AD converter in the ultrasonic wind direction and speed sensor, the internal factors affecting the inaccuracy of the standard wind direction value include the influence of device hardware and performance defects and installation errors, it should be noted that the calculation formula of the measurement model is:

[0076] (9);

[0077] Wherein, the internal factors affecting the inaccuracy of the standard wind direction value mainly come from three aspects, first from the device hardware and performance defects (mainly including the accuracy of the angle encoder core component, the response speed of the device), second from external environmental interference (airflow disturbance, obstacle influence, air density change, electromagnetic interference), and second from installation error (installation reference deviation, insufficient fixing stability), in detail, The wind direction measurement error value, μ1 is the airflow disturbance influence factor, μ2 is the obstacle influence factor, μ3 is the air density change influence factor, μ4 is the electromagnetic interference influence factor, The standard wind direction value measured by the angle encoder, D1 is the wind direction influence value of the angle encoder bearing rotation lag, D2 is the wind direction change influence value caused by the mechanical wind direction rotation inertia of the angle encoder, D3 is the wind direction influence value corresponding to the angle encoder installation reference deviation, D4 is the wind direction influence value corresponding to the insufficient fixing stability of the angle encoder.

[0078] S03, determining and calculating the standard uncertainty according to the measurement model, the standard uncertainty including the standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane and the standard uncertainty of the angle encoder corresponding to the wind direction indication value.

[0079] It should be noted that the standard uncertainty includes the standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane and the standard uncertainty of the angle encoder corresponding to the wind direction indication value (including the standard uncertainty introduced by the repetitive measurement of the wind direction indication value of the angle encoder and the standard uncertainty introduced by the measurement error of the angle encoder).

[0080] Specifically, according to the repetitive measurement and the non-repetitive measurement, the standard uncertainty is divided into A-type standard uncertainty and B-type standard uncertainty, the repetitive measurement is divided into A-type standard uncertainty, and the non-repetitive measurement is divided into B-type standard uncertainty, wherein the standard uncertainty introduced by the repetitive measurement of the signal transmission time between the two groups of probes in the two-dimensional plane is A-type standard uncertainty, the standard uncertainty introduced by the repetitive measurement of the wind direction indication value of the angle encoder is A-type standard uncertainty, and the air flow disturbance influence factor, the obstacle influence factor, the air density change influence factor, the electromagnetic interference influence factor, the wind direction influence value of the bearing rotation hysteresis of the angle encoder, the wind direction change influence value caused by the mechanical wind direction rotation inertia of the angle encoder, the wind direction influence value corresponding to the installation reference deviation of the angle encoder, and the wind direction influence value corresponding to the insufficient fixing stability of the angle encoder are B-type standard uncertainty.

[0081] More specifically, the A-type standard uncertainty is calculated by using the Bessel formula, and the calculation formula is:

[0082] ;

[0083] ;

[0084] wherein x i is the i-th input parameter, is the average value, n is the corresponding measurement number when the A-type standard uncertainty is evaluated, and u A is the A-type standard uncertainty.

[0085] It should be noted that the introduction of the Bessel formula to calculate the experimental standard deviation is essentially the core goal of the "unbiased estimation" of the A-type uncertainty, which is to fully characterize the fluctuation characteristics of each measurement by using the "residual information" (the deviation of the single measurement value from the arithmetic mean value) of all measurement data to avoid the information loss of the range method. The core value of the formula is to integrate the fluctuation information of all measurement data through "residual sum of squares", and then correct the underestimation of the finite sample to the population by "n-1", so that the calculated experimental standard deviation is closer to the true random error dispersion degree. Especially for meteorological sensors and other "measurement value fluctuation is subtle (such as 0.1℃ level temperature deviation)" scenes, the Bessel formula can more accurately capture the tiny random error and avoid the neglect of details by the range method.

[0086] The B-type standard uncertainty supposes the value interval [x-w, x+w] of the measured value according to the measurement information, wherein w represents the interval half-width, assumes the probability distribution of the measured value, and determines k according to the probability distribution. Specifically, when the probabilities of each point in the supposed value interval of the measured value are equal, the uniform distribution is satisfied, k = 1. When the probability of the measured value appearing at the interval endpoint value is greater than the probability of appearing at the intermediate value, the inverse sine distribution is satisfied, k = 3. When the probability of the measured value appearing at the interval endpoint value is less than the probability of appearing at the intermediate value, the normal distribution is satisfied, k = 2. It should be noted that the calculation formula of the B-type standard uncertainty is:

[0087] u B = w / k.

[0088] Wherein, u B is the B-type standard uncertainty, w is the interval half-width, and k is the probability distribution parameter.

[0089] Step S04, according to the function partial derivative of each input parameter in the measurement model, the sensitivity coefficient of the corresponding standard uncertainty is determined.

[0090] Specifically, the function partial derivative of each input parameter is calculated, and the expression of the sensitivity coefficient is:

[0091] ;

[0092] Wherein, c i is the sensitivity coefficient of the i-th input parameter.

[0093] Step S05, according to the parameters of the measurement model and the Bayesian network-based probability synthesis method, calculate the combined standard uncertainty, and according to the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty and the corresponding actual expanded uncertainty, determine the expanded uncertainty by the machine learning confidence interval prediction method.

[0094] Specifically, each input and output of the measurement model is constructed into a Bayesian network structure, wherein each node represents a variable, and the directed edges between the nodes represent the causal relationship between the variables. For example, the input quantities include signal transmission time t ab , t ba , t cd , and t dc , sensitivity difference D a , quantization error of AD converter D b , etc., are defined as root nodes, and the output quantity is the wind direction measurement error value , which is defined as a leaf node, the wind direction indication value D of the ultrasonic wind direction and speed sensor, the standard wind direction value D s , and the environmental factors (air flow disturbance influence factor, obstacle influence factor, air density change influence factor, and electromagnetic interference influence factor) are defined as intermediate nodes. Further, the above root nodes, intermediate nodes and leaf nodes are associated by the measurement model formula. For example, the signal transmission time affects the calculation of the wind direction indication value, and further affects the wind direction measurement error value.

[0095] The probability distribution of each node is determined, and the inference algorithm of the Bayesian network is used to infer the posterior probability distribution of the output quantity according to the probability distribution of each input quantity. It should be noted that for input quantities with known probability distribution (such as the probability distribution of signal transmission time obtained by a large number of experiments, the probability distribution of sensor parameters, etc.), the probability distribution is directly input; for uncertain relationships, the conditional probability table is learned through historical measurement data or expert knowledge. Further, Gibbs sampling is used to start from the prior distribution of the root node, iteratively sample the values of the intermediate nodes and leaf nodes until convergence.

[0096] The combined standard uncertainty is determined by calculating the standard deviation of the posterior probability distribution of the output quantity. It can be understood that the standard deviation, i.e. the combined standard uncertainty, is calculated from the converged wind direction measurement error value sample.

[0097] Obtain historical data under different measurement conditions, the historical data including the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty and the corresponding actual expanded uncertainty, and construct a machine learning model to take the standard uncertainty, the sensitivity coefficient and the combined standard uncertainty of each input quantity as features, and to take the actual expanded uncertainty as a target variable for training, in the embodiment of the present application, the machine learning model adopts gradient boosting tree;

[0098] Input the standard uncertainty, the sensitivity coefficient and the combined standard uncertainty of each input quantity obtained by the current measurement into the trained machine learning model, output the expanded uncertainty, in addition, the model can also output the predicted confidence level, and adjust the expanded uncertainty according to the reliability of the prediction result (if the confidence level is low, the range of the expanded uncertainty is appropriately enlarged; if the confidence level is high, the expanded uncertainty is determined according to the predicted value).

[0099] It should be noted that the advantages of adopting the probability combination method based on the Bayesian network and the confidence interval prediction method of machine learning to determine the combined standard uncertainty and the expanded uncertainty are that the traditional uncertainty combination is based on the error transfer formula, and it is assumed that the input quantities are independent of each other and are normally distributed, but in actual measurement, there may be complex dependent relationships between the input quantities (such as air flow disturbance affecting signal transmission time and air density, and then affecting sound propagation speed), and the distribution may not be normal. The Bayesian network can well handle the dependent relationship between variables, intuitively represent the cause and effect and association between variables through the probability graph model, and describe the uncertainty of the input quantity by using the probability distribution instead of the simple variance, which is more in line with the nature of the uncertainty of actual measurement. The posterior probability distribution of the measurement result obtained by reasoning can more comprehensively and accurately reflect the uncertainty of the measurement result, and compared with the traditional method, it can capture more complex uncertainty transmission and combination. In addition, the determination of the expanded uncertainty depends on the fixed inclusion factor, and the change rule of the uncertainty under different measurement conditions in the historical data is not fully utilized. The machine learning model can learn the complex nonlinear relationship between each factor (standard uncertainty, sensitivity coefficient, etc.) and the expanded uncertainty from a large amount of historical data, and such relationship is difficult to be accurately modeled by the traditional method. By predicting the expanded uncertainty through the trained model, the expanded uncertainty can be dynamically and accurately determined according to the specific situation of the current measurement (the uncertainty of each input quantity), and the judgment of the confidence level is combined to make the evaluation of the expanded uncertainty more adaptive and accurate, and the mechanicalness and limitations of the traditional fixed inclusion factor method are avoided.

[0100] In summary, the wind direction calibration result measurement uncertainty evaluation method in the above embodiments of the present application, by obtaining the signal transmission time between the two groups of probes in the two-dimensional plane, according to the relationship between the distance between the probes, the sound propagation speed and the signal transmission time, the wind direction indication value is calculated; based on external environmental factors, and the internal factors affecting the inaccuracy of the wind direction indication value and the standard wind direction value, a measurement model is established; according to the measurement model, the standard uncertainty is determined and calculated, which includes the standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane, and the standard uncertainty of the angle encoder corresponding to the wind direction indication value; according to the function partial derivative of each input parameter in the measurement model, the sensitivity coefficient of the corresponding standard uncertainty is determined; according to the parameters of the measurement model and the probability synthesis method based on the Bayesian network, the combined standard uncertainty is calculated, and according to the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty and the corresponding actual expanded uncertainty, the expanded uncertainty is determined through the confidence interval prediction method of machine learning, Specifically, by starting from the measurement principle of the ultrasonic wind direction and speed sensor, comprehensively analyzing the factors affecting the wind direction measurement indication value, the uncertainty source in the measurement process can be more accurately identified and evaluated, which helps to reduce the deviation of the evaluation result and improve the accuracy of the measurement uncertainty evaluation.

[0101] Embodiment two

[0102] The wind direction calibration result measurement uncertainty evaluation method provided by the embodiment two of the present application is explained by examples, Specifically, the model of the ultrasonic wind direction and speed sensor used is HY-UWS11, the measurement range of the angle encoder is 0°~360°, the graduation error is not greater than 0.1°, the experimental environment: temperature is 20℃, air pressure is 1010hPa, humidity is 55%RH, and the calibration data is shown in Table 1.

[0103] Table 1 Calibration data

[0104]

[0105] According to the measurement model:

[0106] ;

[0107] Determine each input parameter, including is the wind direction measurement error value, μ1 is the airflow disturbance influence factor, μ2 is the obstacle influence factor, μ3 is the air density change influence factor, μ4 is the electromagnetic interference influence factor, is the standard wind direction value measured by the angle encoder, D a is the sensitivity difference of the signal module in the ultrasonic wind direction and speed sensor, D bFor the quantization error of the AD converter in the ultrasonic wind direction and speed sensor, D1 is the wind direction influence value of the angle encoder bearing rotation hysteresis, D2 is the wind direction change influence value caused by the mechanical wind direction inertia of the angle encoder, D3 is the wind direction influence value corresponding to the installation reference deviation of the angle encoder, D4 is the wind direction influence value corresponding to the insufficient fixing stability of the angle encoder, t ab For the ultrasonic wind direction and speed sensor a probe to generate an ultrasonic signal, the time experienced by the ultrasonic signal received by the ultrasonic wind direction and speed sensor b probe, t ba For the ultrasonic wind direction and speed sensor b probe to generate an ultrasonic signal, the time experienced by the ultrasonic signal received by the ultrasonic wind direction and speed sensor a probe, t cd For the ultrasonic wind direction and speed sensor c probe to generate an ultrasonic signal, the time experienced by the ultrasonic signal received by the ultrasonic wind direction and speed sensor d probe, t dc For the ultrasonic wind direction and speed sensor d probe to generate an ultrasonic signal, the time experienced by the ultrasonic signal received by the ultrasonic wind direction and speed sensor c probe.

[0108] It should be noted that the uncertainty is divided into two categories, A (i.e. uncertainty introduced by direct measurement) and B (uncertainty introduced by non-direct measurement). In order to more accurately obtain the uncertainty introduced by direct measurement, it is usually necessary to repeatedly measure the directly measured quantity. The measured value is the uncertainty component introduced by multiple repeated measurements. As shown in Table 1, t ab The measurement corresponding experimental standard deviation is 4.0×10 -8 s, which belongs to the uncertainty introduced by repeated measurement, so according to the A-type standard uncertainty calculation formula, it is obtained that u ( t ab )=4.0×10 -8 s / =1.1×10 -8 s; as shown in Table 1, t ba The measurement corresponding experimental standard deviation is 7.0×10 -8 s, which belongs to the uncertainty introduced by repeated measurement, so according to the A-type standard uncertainty calculation formula, it is obtained that u ( t ba )=7.0×10 -8 s / =2.3×10 -8 s; as shown in Table 1, t dc The measurement corresponding experimental standard deviation is 8.0×10 -8s represents the uncertainty introduced by repeatability measurement. Therefore, according to the formula for calculating Type A standard uncertainty, we obtain... u ( t cd ) = 8.0 × 10 -8 s / =2.6×10 -8 s; As can be seen from Table 1, t dc The standard deviation of the measurement corresponding to the experiment was 1.0 × 10⁻⁶. -7 s represents the uncertainty introduced by repeatability measurement. Therefore, according to the formula for calculating Type A standard uncertainty, we obtain... u ( t dc ) = 1.0 × 10 -7 s / =3.2×10 -8 s; As shown in Table 1, the standard deviation of the angle encoder measurement is 0.16°, which is an uncertainty introduced by repeatability measurement. Therefore, according to the formula for calculating Type A standard uncertainty, the following is derived: u ( D S1 )=0.16° / .

[0109] For Type B standard uncertainty:

[0110] (1) External environmental interference

[0111] 1) Airflow disturbance introduces u(μ1)

[0112] By consulting the test reports of the wind tunnel test section, when there is no airflow disturbance, the airflow disturbance coefficient is 1.0, and the deflection error of the airflow introduced by the disturbance is 10%, which is introduced by non-repeatable measurement. Assuming a uniform distribution, according to the formula for calculating Type B standard uncertainty, the introduced standard uncertainty is:

[0113] ;

[0114] 2) The influence of obstacles introduces u(μ2)

[0115] In the wind tunnel calibration environment, obstacles such as the internal support structure and sensor mounting brackets can cause local airflow around or turbulence, thus interfering with wind direction measurements. By reviewing wind tunnel design documents and similar calibration experimental data, the obstacle influence coefficient is generally taken as 1.0 when there are no obstacles. When conventional obstacles (such as fixing bolts with a diameter ≤ 5mm or support plates with a thickness ≤ 2mm) exist in the wind tunnel test section, and the maximum wind direction deflection error caused by these obstacles has been verified through multiple tests to be ±8%.

[0116] ;

[0117] 3) Air density change introduces u (μ3)

[0118] Air density p is related to temperature T and air pressure P, and air density change will affect the speed of ultrasonic wave propagation c, and then indirectly affect the wind direction calculation result. When there is no change in air density, the air density change coefficient is generally taken as 1.0. The environmental parameters of this calibration experiment are: temperature 20℃ (fluctuation range ±0.5℃), air pressure 1010hPa (fluctuation range ±2hPa). Through sensitivity analysis, the maximum wind direction deviation corresponding to the fluctuation of this environmental parameter is calculated as ±6%. The influence of air density change on wind direction is a continuous random distribution, and the deviation fluctuates uniformly within the allowable range, which conforms to the uniform distribution.

[0119] ;

[0120] 4) Electromagnetic interference introduces u (μ4)

[0121] The signal module of the ultrasonic wind direction and speed sensor (such as ultrasonic transducer driving circuit, AD conversion circuit) is easily affected by external electromagnetic interference, resulting in noise in the measurement signal. When there is no electromagnetic interference, the electromagnetic interference coefficient is generally taken as 1.0. In this calibration environment, the electromagnetic radiation generated by the wind tunnel control system and data acquisition equipment is tested by electromagnetic compatibility test, and the maximum measurement deviation of wind direction caused by it is ±5%, which conforms to the arcsine distribution.

[0122] ;

[0123] (2) Introduce due to the electronic properties of the ultrasonic sensor

[0124] 1) Sensitivity difference of signal module introduces u (D a )

[0125] The sensitivity difference of the signal module of the ultrasonic sensor is mainly caused by the inconsistent frequency response of the ultrasonic transducer (such as the receiving sensitivity deviation of a / b / c / d four groups of probes). According to the sensor factory calibration report, the maximum deviation of the sensitivity of the signal module of this type of sensor corresponds to a wind direction error of 0.07°. The sensitivity difference belongs to the systematic deviation caused by production process, and its error range is statistically consistent with the uniform distribution (no obvious bias)

[0126] ;

[0127] 2) AD converter quantization error introduces u (D b )

[0128] The quantization error of the sensor AD converter is determined by its resolution. The AD resolution corresponding to the wind direction measurement of this type of sensor is 16 bits, the measurement range is 0°~360°, and the maximum absolute value of the quantization error is 360° / 2 16 ≈0.0055°. The AD quantization error is uniformly distributed within adjacent quantization intervals, which conforms to uniform distribution.

[0129] ;

[0130] (3) Internal factors affecting the accuracy of the standard wind direction value

[0131] 1) Standard uncertainty u(D1) introduced by angle encoder bearing rotation hysteresis;

[0132] The bearing rotation hysteresis of the angle encoder will cause a deviation in the indicated value corresponding to the same mechanical angle during forward and reverse rotation. According to the technical specification of the angle encoder, the maximum bearing rotation hysteresis is 0.03°. The rotation hysteresis error is symmetrically distributed during forward and reverse rotation, which conforms to uniform distribution.

[0133] ;

[0134] 2) Standard uncertainty u(D2) introduced by angle encoder mechanical wind direction inertia;

[0135] When the angle encoder rotates with the water platform, the inertia of the mechanical structure (such as the shaft and gear) will cause it to not respond to instantaneous wind direction changes in real time. Through testing, the maximum inertia deviation is 0.04° under the calibration scenario of rotation speed ≤5° / s. The influence of inertia deviation is random and uniformly distributed within the allowed range.

[0136] ;

[0137] 3) Standard uncertainty u(D3) introduced by angle encoder installation reference deviation;

[0138] The installation reference deviation is caused by the coaxial error between the encoder and the water platform. After installation and debugging, the maximum wind direction error corresponding to the coaxial error is 0.05° measured by a laser collimator. The installation reference deviation belongs to a fixed systematic deviation, and the error range conforms to uniform distribution.

[0139] ;

[0140] 4) Standard uncertainty u(D4) introduced by insufficient angle encoder fixing stability;

[0141] The lack of fixed stability can cause the encoder to produce a slight displacement during the experiment, and the maximum wind direction deviation caused by the vibration test (simulating the slight vibration during wind tunnel operation) is 0.03°. The displacement deviation is random and symmetrically distributed in the positive and negative intervals, conforming to the uniform distribution.

[0142] .

[0143] Please refer to Table 2 for the standard uncertainty value summary table.

[0144] Table 2 Standard uncertainty value summary table

[0145]

[0146] Further, according to the partial derivative of each input parameter in the measurement model, the sensitivity coefficient of the corresponding standard uncertainty is determined, and it should be noted that according to the data in Table 1, it can be known that At this time, each sensitivity coefficient is as follows:

[0147] ;

[0148] ;

[0149] ;

[0150] ;

[0151] ;

[0152] ;

[0153] ;

[0154] ;

[0155] ;

[0156] ;

[0157] ;

[0158] ;

[0159] ;

[0160] ;

[0161] ;

[0162] sensitivity coefficient of air flow disturbance, sensitivity coefficient of obstacle, sensitivity coefficient of air density change, sensitivity coefficient of electromagnetic interference, sensitivity coefficient of signal module in ultrasonic wind direction and speed sensor, sensitivity coefficient of quantization error of AD converter in ultrasonic wind direction and speed sensor, sensitivity coefficient of standard wind direction value measured by angle encoder, sensitivity coefficient of wind direction impact value of bearing rotation hysteresis of angle encoder, sensitivity coefficient of wind direction change impact value caused by mechanical wind direction rotation inertia of angle encoder, sensitivity coefficient of wind direction impact value corresponding to installation reference deviation of angle encoder, sensitivity coefficient of wind direction impact value corresponding to insufficient fixing stability of angle encoder, sensitivity coefficient of time experienced by ultrasonic wind direction and speed sensor a probe generating ultrasonic signal and ultrasonic wind direction and speed sensor b probe receiving ultrasonic signal, sensitivity coefficient of time experienced by ultrasonic wind direction and speed sensor b probe generating ultrasonic signal and ultrasonic wind direction and speed sensor a probe receiving ultrasonic signal, sensitivity coefficient of time experienced by ultrasonic wind direction and speed sensor c probe generating ultrasonic signal and ultrasonic wind direction and speed sensor d probe receiving ultrasonic signal, sensitivity coefficient of time experienced by ultrasonic wind direction and speed sensor d probe generating ultrasonic signal and ultrasonic wind direction and speed sensor c probe receiving ultrasonic signal.

[0163] sensitivity coefficient corresponding to standard wind direction measured value:

[0164] ;

[0165] Please refer to Table 3 for sensitivity coefficient table.

[0166] Table 3 Sensitivity coefficient table

[0167]

[0168] constructing each input and output of the measurement model into a Bayesian network structure, wherein each node represents a variable, and the directed edges between nodes represent the causal relationship between variables;

[0169] Determine the probability distribution of each node, and use the inference algorithm of the Bayesian network to infer the posterior probability distribution of the output quantity according to the probability distribution of each input quantity;

[0170] Determine the combined standard uncertainty by calculating the standard deviation of the posterior probability distribution of the output quantity;

[0171] Obtain historical data under different measurement conditions, including the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty, and the corresponding actual expanded uncertainty, and construct a machine learning model, taking the standard uncertainty, the sensitivity coefficient, and the combined standard uncertainty of each input quantity as features, and taking the actual expanded uncertainty as the target variable for training;

[0172] Input the standard uncertainty, the sensitivity coefficient, and the combined standard uncertainty of each input quantity into the trained machine learning model, and output the expanded uncertainty.

[0173] Embodiment Three

[0174] Please refer to Figure 5 , Figure 5 is a structural block diagram of a wind direction calibration result measurement uncertainty evaluation system provided by Embodiment Three of the present application, which comprises a first calculation module 21, a model establishment module 22, a second calculation module 23, a first determination module 24, and a second determination module 25, wherein:

[0175] The first calculation module 21 is used to obtain the signal transmission time between two groups of probes in a two-dimensional plane, and calculate the wind direction indication value according to the relationship among the distance between the probes, the sound propagation speed, and the signal transmission time.

[0176] The model establishment module 22 is used to establish a measurement model based on external environmental factors and internal factors affecting the inaccuracy of the wind direction indication value and the standard wind direction value, wherein the external environmental factors include air flow disturbance, obstacles, air density change, and electromagnetic interference, the internal factors affecting the inaccuracy of the wind direction indication value include the sensitivity difference of the signal module in the ultrasonic wind direction and speed sensor and the quantization error of the AD converter in the ultrasonic wind direction and speed sensor, and the internal factors affecting the inaccuracy of the standard wind direction value include the influence of device hardware and performance defects and installation errors, wherein the calculation formula of the measurement model is:

[0177] ;

[0178] is the wind direction measurement error value, μ1 is the air flow disturbance influence factor, μ2 is the obstacle influence factor, μ3 is the air density change influence factor, μ4 is the electromagnetic interference influence factor, D is a standard wind direction value measured by an angle encoder a D is a sensitivity difference of a signal module in an ultrasonic wind direction and speed sensor b D1 is an angle encoder bearing rotation hysteresis wind direction influence value, D2 is an angle encoder mechanical wind direction rotation inertia caused wind direction change influence value, D3 is an angle encoder installation reference deviation corresponding wind direction influence value, D4 is an angle encoder fixed stability deficiency corresponding wind direction influence value, t ab t is a time experienced by an ultrasonic wave signal generated by a probe a of an ultrasonic wind direction and speed sensor, and received by a probe b of the ultrasonic wind direction and speed sensor ba t is a time experienced by an ultrasonic wave signal generated by a probe b of an ultrasonic wind direction and speed sensor, and received by a probe a of the ultrasonic wind direction and speed sensor cd t is a time experienced by an ultrasonic wave signal generated by a probe c of an ultrasonic wind direction and speed sensor, and received by a probe d of the ultrasonic wind direction and speed sensor dc t is a time experienced by an ultrasonic wave signal generated by a probe d of an ultrasonic wind direction and speed sensor, and received by a probe c of the ultrasonic wind direction and speed sensor

[0179] The second calculation module 23 is configured to determine and calculate a standard uncertainty according to the measurement model, the standard uncertainty including a standard uncertainty introduced by a signal transmission time between two groups of probes in a two-dimensional plane and a standard uncertainty of an angle encoder at a corresponding wind direction indication value, and dividing the standard uncertainty into an A-type standard uncertainty and a B-type standard uncertainty according to repetitive measurement and non-repetitive measurement, wherein the standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane in the repetitive measurement is the A-type standard uncertainty, the standard uncertainty introduced by the repetitive measurement of the wind direction indication value of the angle encoder is the A-type standard uncertainty, and the air flow disturbance influence factor, the obstacle influence factor, the air density change influence factor, the electromagnetic interference influence factor, the angle encoder bearing rotation hysteresis wind direction influence value, the angle encoder mechanical wind direction rotation inertia caused wind direction change influence value, the angle encoder installation reference deviation corresponding wind direction influence value, and the angle encoder fixed stability deficiency corresponding wind direction influence value are the B-type standard uncertainty, the A-type standard uncertainty is calculated by using a Bessel formula, and a calculation formula is as follows:

[0180] ;

[0181] ;

[0182] wherein x i is an i-th input parameter, Let n be the average value, n be the number of measurements corresponding to the Type A standard uncertainty assessment, and u be the average value. A The standard uncertainty is Type A. The Type B standard uncertainty is based on measurement information to infer the value interval of the measurand [xw, x+w], where w represents the half-width of the interval. Assuming a probability distribution of the measurand's value, k is determined based on this probability distribution. Specifically, when the probability of each point occurring within the inferred value interval of the measurand is equal, a uniform distribution is satisfied, and k = When the probability of the measured value appearing at the endpoint of the interval is greater than the probability of it appearing at the middle value, then it follows an arcsine distribution, k= When the probability of the measured value appearing at the endpoint of the interval is less than the probability of it appearing at the middle value, then it follows a normal distribution, k=2, and the formula for calculating the Type B standard uncertainty is:

[0183] u B =w / k;

[0184] Among them, u B Let w be the type B standard uncertainty, w be the interval half-width, and k be the probability distribution parameter.

[0185] The first determining module 24 is used to determine the sensitivity coefficient of the corresponding standard uncertainty based on the partial derivatives of the functions of each input parameter in the measurement model.

[0186] The second determining module 25 is used to calculate the combined standard uncertainty based on the parameters of the measurement model and the probability synthesis method based on Bayesian networks, and to determine the expanded uncertainty based on the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty and the corresponding actual expanded uncertainty, using the confidence interval prediction method of machine learning.

[0187] Example 4

[0188] In another aspect, the present invention also proposes an electronic device, please refer to [link to relevant documentation]. Figure 6 The diagram shows the structure of the electronic device in Embodiment 4 of the present invention, including a memory 20, a processor 10, and a computer program 30 stored in the memory and executable on the processor. When the processor 10 executes the computer program 30, it implements the method for evaluating the measurement uncertainty of the wind direction calibration result as described above.

[0189] In some embodiments, the processor 10 may be a central processing unit (CPU), controller, microcontroller, microprocessor or other data processing chip, used to run program code stored in memory 20 or process data, such as executing access restriction programs.

[0190] The memory 20 includes at least one type of readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a magnetic memory, a magnetic disk, an optical disk, etc. The memory 20 can be an internal storage unit of the electronic device in some embodiments, such as a hard disk of the electronic device. The memory 20 can also be an external storage device of the electronic device in other embodiments, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device. Further, the memory 20 can include both the internal storage unit and the external storage device of the electronic device. The memory 20 can be used not only to store application software and various data of the electronic device, but also to temporarily store data that has been output or will be output.

[0191] It should be noted that, Figure 6 The illustrated structure does not limit the electronic device, and in other embodiments, the electronic device can include fewer or more components than illustrated, or combine certain components, or arrange the components differently.

[0192] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program. The program is executed by a processor to implement the wind direction calibration result measurement uncertainty evaluation method as described above.

[0193] Those skilled in the art can understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a list of executable instructions for implementing the logic function, which can be specifically implemented in any computer readable medium for use by or in conjunction with an instruction execution system, device or apparatus, such as a computer-based system, a system including a processor or other system that can fetch and execute instructions from the instruction execution system, device or apparatus. For the present specification, the "computer readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by or in conjunction with the instruction execution system, device or apparatus, or in conjunction with these instruction execution systems, devices or apparatus.

[0194] More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection (electronic) having one or more wires, a portable computer diskette (magnetic), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can also be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example, via an optical scanner, then compiled, interpreted, or otherwise processed, and stored in a computer memory in a form that is then employable by a computer. Examples of computer-readable media that are further within the spirit of the present application are a computer program product, a computer readable storage medium, and a computer.

[0195] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the embodiments described above, various steps or methods can be implemented, for example, by software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following techniques, which are well known in the art, can be used to implement the application: a hybrid of the above techniques, discrete logic circuit(s) having logic gates for implementing logic functions upon request pins of the discrete logic circuit(s), programmable logic array(s) (PLAs), field programmable gate array(s) (FPGAs), etc.

[0196] In the description of the specification, the description of the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the application. In the specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in one or more embodiments or examples.

[0197] The above embodiments only express several implementation manners of the application, which are described in a more specific and detailed manner, but cannot be understood as a limitation on the patent scope of the application. It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the application, which are all within the protection scope of the application. Therefore, the patent protection scope of the application should be subject to the appended claims.

Claims

1. A method of evaluating a measurement uncertainty of a wind direction calibration result, characterized by, The method is applied to an ultrasonic wind direction and speed sensor and an angle encoder, the ultrasonic wind direction and speed sensor comprises two groups of probes which are orthogonal to each other and are used to collect signal transmission time between the probes, and the angle encoder is used to collect a standard wind direction value, and the method comprises the following steps: acquiring signal transmission time between the two groups of probes in a two-dimensional plane, and calculating a wind direction indication value according to a relationship among a distance between the probes, a sound propagation speed and the signal transmission time; establishing a measurement model based on external environmental factors and internal factors affecting inaccuracy of the wind direction indication value and the standard wind direction value, wherein the external environmental factors include airflow disturbance, obstacles, air density variation and electromagnetic interference, the internal factors affecting inaccuracy of the wind direction indication value include sensitivity difference of a signal module in the ultrasonic wind direction and speed sensor and quantization error of an AD converter in the ultrasonic wind direction and speed sensor, and the internal factors affecting inaccuracy of the standard wind direction value include device hardware and performance defect influence and installation error influence; determining and calculating standard uncertainty according to the measurement model, wherein the standard uncertainty includes standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane and standard uncertainty of the angle encoder corresponding to the wind direction indication value; determining a sensitivity coefficient of the corresponding standard uncertainty according to a partial derivative of a function of each input parameter in the measurement model; calculating a combined standard uncertainty according to parameters of the measurement model and a probability synthesis method based on a Bayesian network, and determining an expanded uncertainty by a confidence interval prediction method of machine learning according to the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty and a corresponding actual expanded uncertainty.

2. The wind direction calibration result measurement uncertainty evaluation method according to claim 1, characterized in that, A calculation formula of the measurement model is as follows: ； is the wind direction measurement error value, μ1 is the airflow disturbance influence factor, μ2 is the obstacle influence factor, μ3 is the air density change influence factor, μ4 is the electromagnetic interference influence factor, is the standard wind direction value measured by the angle encoder, D a is the sensitivity difference of the signal module in the ultrasonic wind direction and speed sensor, D b is the quantization error of the AD converter in the ultrasonic wind direction and speed sensor, D1 is the wind direction influence value of the bearing rotation hysteresis of the angle encoder, D2 is the wind direction change influence value caused by the mechanical wind direction rotation inertia of the angle encoder, D3 is the wind direction influence value corresponding to the installation reference deviation of the angle encoder, D4 is the wind direction influence value corresponding to the insufficient fixing stability of the angle encoder, t ab is the time experienced by the ultrasonic wind direction and speed sensor a probe to generate an ultrasonic signal, and the ultrasonic wind direction and speed sensor b probe to receive the ultrasonic signal, t ba is the time experienced by the ultrasonic wind direction and speed sensor b probe to generate an ultrasonic signal, and the ultrasonic wind direction and speed sensor a probe to receive the ultrasonic signal, t cd is the time experienced by the ultrasonic wind direction and speed sensor c probe to generate an ultrasonic signal, and the ultrasonic wind direction and speed sensor d probe to receive the ultrasonic signal, t dc is the time experienced by the ultrasonic wind direction and speed sensor d probe to generate an ultrasonic signal, and the ultrasonic wind direction and speed sensor c probe to receive the ultrasonic signal.

3. The wind direction calibration result measurement uncertainty evaluation method according to claim 2, characterized in that, in the step of determining and calculating the standard uncertainty according to the measurement model, the standard uncertainty is divided into an A-type standard uncertainty and a B-type standard uncertainty according to repetitive measurement and non-repetitive measurement, wherein the standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane in the repetitive measurement is the A-type standard uncertainty, the standard uncertainty introduced by the wind direction indication value in the repetitive measurement of the angle encoder is the A-type standard uncertainty, and the airflow disturbance influence factor, the obstacle influence factor, the air density variation influence factor, the electromagnetic interference influence factor, the angle encoder bearing rotation lag wind direction influence value, the wind direction change influence value caused by mechanical wind direction inertia of the angle encoder, the wind direction influence value corresponding to the angle encoder installation reference deviation and the wind direction influence value corresponding to the angle encoder fixed stability deficiency are the B-type standard uncertainty.

4. The wind direction calibration result measurement uncertainty evaluation method according to claim 3, characterized in that, the A-type standard uncertainty is calculated by using a Bessel formula, and a calculation formula is as follows: ； ； Wherein, x i is the ith input parameter, is the average value, n is the corresponding measurement number when the A-class standard uncertainty is evaluated, u A is the A-class standard uncertainty.

5. The wind direction calibration result measurement uncertainty evaluation method according to claim 4, characterized in that, The B-class standard uncertainty supposes the value interval [x-w, x+w] of the measured value according to the measurement information, wherein w represents the interval half-width, assumes the probability distribution of the measured value, and determines k according to the probability distribution. Specifically, when the probabilities of the points in the value interval of the measured value are equal, the uniform distribution is satisfied, k= ; When the probability of a measured value occurring at the interval endpoint is greater than the probability of occurring at the middle value, then the inverse sine distribution is satisfied, k = 1 ; when the probability of a measured value occurring at the interval endpoint is less than the probability of occurring at the middle value, then the normal distribution is satisfied, k = 2.

6. The wind direction calibration result measurement uncertainty evaluation method according to claim 5, characterized in that, a calculation formula of the B-type standard uncertainty is as follows: u B = w / k; where u B is the B-type standard uncertainty, w is the interval half-width, and k is the probability distribution parameter.

7. The wind direction calibration result measurement uncertainty evaluation method according to claim 6, characterized in that, the step of calculating the combined standard uncertainty according to the parameters of the measurement model and the probability synthesis method based on the Bayesian network, and determining the expanded uncertainty by the confidence interval prediction method of machine learning according to the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty and the corresponding actual expanded uncertainty comprises the following steps: constructing each input quantity and output quantity of the measurement model into a Bayesian network structure, wherein each node represents a variable, and a directed edge between nodes represents a causal relationship between variables; determining a probability distribution of each node, and using an inference algorithm of the Bayesian network to infer a posterior probability distribution of the output quantity according to the probability distribution of each input quantity; determining the combined standard uncertainty by calculating a standard deviation of the posterior probability distribution of the output quantity; obtaining historical data under different measurement conditions, the historical data including the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty, and the corresponding actual expanded uncertainty, and constructing a machine learning model to train the actual expanded uncertainty as a target variable with the standard uncertainty, the sensitivity coefficient, and the combined standard uncertainty of each input quantity as features; inputting the standard uncertainty, the sensitivity coefficient, and the combined standard uncertainty of each input quantity obtained by the current measurement into the trained machine learning model to output the expanded uncertainty.

8. A wind direction calibration result measurement uncertainty assessment system, characterized by, The system is used to implement the wind direction calibration result measurement uncertainty evaluation method according to any one of claims 1-7. The first calculation module is configured to obtain the signal transmission time between the two groups of probes in the two-dimensional plane, and calculate the wind direction indication value according to the relationship among the distance between the probes, the sound propagation speed, and the signal transmission time. The model establishment module is configured to establish a measurement model based on external environmental factors and internal factors affecting the inaccuracy of the wind direction indication value and the standard wind direction value, wherein the external environmental factors include airflow disturbance, obstacles, air density variation, and electromagnetic interference, the internal factors affecting the inaccuracy of the wind direction indication value include the sensitivity difference of the signal module in the ultrasonic wind direction and speed sensor and the quantization error of the AD converter in the ultrasonic wind direction and speed sensor, and the internal factors affecting the inaccuracy of the standard wind direction value include the influence of device hardware and performance defects and installation errors. The second calculation module is configured to determine and calculate the standard uncertainty according to the measurement model, wherein the standard uncertainty includes the standard uncertainty introduced by the signal transmission time between the two groups of probes in the two-dimensional plane and the standard uncertainty of the angle encoder at the corresponding wind direction indication value. The first determination module is configured to determine the sensitivity coefficient of the corresponding standard uncertainty according to the functional partial derivative of each input parameter in the measurement model. The second determination module is configured to calculate the combined standard uncertainty according to the parameters of the measurement model and the probability synthesis method based on the Bayesian network, and determine the expanded uncertainty by the confidence interval prediction method of machine learning according to the standard uncertainty, the corresponding sensitivity coefficient, the combined standard uncertainty, and the corresponding actual expanded uncertainty.

9. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the wind direction calibration result measurement uncertainty evaluation method according to any one of claims 1-7.

10. An electronic device, comprising: The computer program is stored in the memory and executable on the processor, and the processor implements the wind direction calibration result measurement uncertainty evaluation method according to any one of claims 1-7 when executing the program.

Images