A multi-sensor fusion wind measuring instrument and method based on Kalman filter

Abstract

The application discloses a kind of multi-sensor fusion wind measuring instrument and wind measuring method based on Kalman filter, the wind measuring method includes: main control circuit obtains the data of heating rod and 36 temperature sensors, according to the principle of thermal wind measurement, wind speed and direction are measured, and the initial value of wind speed and direction is obtained;The main control circuit obtains the data of wind guide pipe and micro pressure sensor, and according to the principle of wind pressure measurement, wind speed and direction are measured, and the observation value of wind speed and direction is obtained;The main control circuit uses preset Kalman filtering algorithm to process initial value and observation value, and obtains wind speed and direction.The application provides wind measuring instrument that can be suitable for field harsh scene, and provides wind measuring method based on wind measuring instrument, which can accurately measure wind speed and direction.

Description

Technical Field

[0001] This invention relates to a multi-sensor fusion anemometer and wind measurement method based on a Kalman filter, belonging to the field of meteorological detection technology. Background Technology

[0002] Solar radiation causes heat to be absorbed everywhere on Earth. Due to uneven heating in different places, the air becomes warmer or colder. Warm air expands and rises, while cold air cools and sinks. This air movement creates wind.

[0003] Wind speed and direction measurements are of great significance to our lives, playing a crucial role in aviation, agriculture, industry, and meteorological monitoring. In aviation, they affect flight safety; high-altitude turbulence can cause aircraft to shake, lose control, or even stall and crash. Low-altitude wind shear is known as a killer of aircraft, and it can easily cause accidents during takeoff and landing. In agriculture, wind has a significant impact on the growth, development, reproduction, morphology, and behavior of crops. In industry, for example, coal mining is a high-risk industry in my country; excessively high or low wind speeds can affect the physical and mental health of workers, and in severe cases, can cause explosions. Statistics show that nearly 50% of coal mines in my country are high-gas mines, and gas outburst mines account for as much as 17.6%.

[0004] Currently, all sectors of society are developing rapidly towards refinement and intensive operations. Precise monitoring of meteorological elements at small scales is of great significance to human production and daily life. Therefore, considering the above points, accurate measurement of wind speed and direction remains a crucial issue that my country needs to pay close attention to, in order to reduce the risk of coal dust explosions in the industrial sector and to guide aerospace, industrial and agricultural production, environmental monitoring, and meteorological disaster early warning.

[0005] Currently, commonly used anemometers include mechanical anemometers, hot-wire anemometers, and ultrasonic anemometers. However, they all have certain technical drawbacks. For example, mechanical anemometers, due to their rotating mechanical structure, experience wear on the rotating shaft over prolonged use, affecting measurement accuracy. Furthermore, mechanical anemometers are affected by static friction at the contact surface of the rotating shaft, resulting in a limited starting wind speed and poor measurement accuracy in low-wind-speed environments. Hot-wire anemometers, due to their internal fragile heating wire, have poor reliability and cannot be used extensively in the field. Ultrasonic anemometers are more expensive, and the surface of the ultrasonic transducer is relatively fragile; scratches and impacts can cause sound wave scattering, leading to errors. Additionally, ultrasonic anemometers are subject to shadowing effects, resulting in higher measurement errors at specific angles. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a multi-sensor fusion anemometer and wind measurement method based on a Kalman filter, applicable to harsh field environments. To achieve the above objective, this invention employs the following technical solution:

[0007] In a first aspect, the present invention provides a multi-sensor fusion anemometer based on a Kalman filter, comprising:

[0008] The anemometer body, the thermal anemometer panel located on the upper part of the anemometer body, and the anemometer top cover located on the upper part of the thermal anemometer panel;

[0009] A heating rod and 36 temperature sensors are fixed on the thermal wind measurement panel through holes, and the 36 temperature sensors are evenly distributed around the heating rod.

[0010] The upper surface of the anemometer's top cover is provided with four air guide pipes, and the lower surface of the anemometer's top cover is provided with a micro-pressure sensor. The air guide pipes are connected to the micro-pressure sensor.

[0011] The anemometer body contains a main control circuit, which is connected to the heating rod and 36 temperature sensors. It measures wind speed and direction based on the thermal anemometer principle to obtain initial values ​​of wind speed and direction. The main control circuit is also connected to the air duct and a micro-pressure sensor, which measures wind speed and direction based on the wind pressure anemometer principle to obtain observed values ​​of wind speed and direction. The main control circuit processes the initial and observed values ​​to obtain wind speed and direction.

[0012] In conjunction with the first aspect, further, 32 of the 36 temperature sensors are evenly distributed in a 4×8 ring array, with the included angle between adjacent sensors in each ring being 45°, and the rings from the inside out being the first, second, third, and fourth rings respectively; 4 temperature sensors are located outside the fourth ring to form the fifth ring, with the included angle between adjacent sensors in the fifth ring being 90°.

[0013] In conjunction with the first aspect, the air ducts are arranged in a cross shape, wherein two air ducts at 180° to each other are connected to the same micro-pressure sensor.

[0014] Secondly, the present invention provides a multi-sensor fusion wind measurement method based on a Kalman filter as described in the first aspect, comprising:

[0015] The main control circuit acquires data from the heating rod and 36 temperature sensors, and measures wind speed and direction based on the principle of thermal wind measurement to obtain the initial values ​​of wind speed and direction.

[0016] The main control circuit acquires data from the air duct and the micro-pressure sensor, and measures the wind speed and direction according to the wind pressure measurement principle to obtain the observed values ​​of wind speed and direction;

[0017] The main control circuit uses a preset Kalman filter algorithm to process the initial and observed values ​​to obtain wind speed and wind direction.

[0018] In conjunction with the second aspect, the aforementioned thermal wind measurement principle further includes:

[0019] When the sensors are evenly distributed around the heat source, wind blows towards the ring array from any direction. The central constant-temperature heat source carries away some of the heat, causing changes in the temperature values ​​of the sensors in each ring, creating a temperature difference. The temperature difference is particularly pronounced in the first ring, exhibiting a clear Gaussian distribution. Therefore, for calculating the wind direction angle, only the eight temperature sensors in the innermost ring need to be used for evaluation. However, for calculating the wind speed, the average value of the 32 temperature sensors on the array must be used to evaluate the flow velocity.

[0020]

[0021] In equation (1), T n Let n be the temperature values ​​of each temperature sensor, n = 1, 2, 3, ..., 32;

[0022] The ambient temperatures were set to -10℃, 0℃, 10℃, 20℃, 30℃, and 40℃, and the wind speeds were set to 0.5m / s, 1m / s, 1.5m / s, 2m / s, 2.5m / s, 3m / s, 3.5m / s, 4m / s, 4.5m / s, 5m / s, 5.5m / s, and 6m / s, respectively. The average temperature values ​​from 32 temperature sensors measured in multiple experiments were fitted into a curve. Let the ambient temperature be x, the average temperature value from the temperature sensors be y, and the wind speed be z. The fitting function is:

[0023]

[0024] In equation (2), p1 = 10.8945, p2 = 0.6475, p3 = -1.019, p4 = 0.0055, p5 = -0.2575, p6 = 0.0012, p7 = 0.227, p8 = 0.0002, p9 = -2.5708e -5 ;

[0025] When the sensors are evenly distributed around the heat source, the wind blows towards the ring array from any direction. The central constant-temperature heat source carries away some of the heat from the heat source, causing the temperature values ​​of the temperature sensors in each ring to change, forming a temperature difference. The temperature difference in the first ring is particularly obvious. The temperature difference shows a clear Gaussian distribution. The temperature is higher in the wind direction. Detecting the peak of the Gaussian curve can determine the direction of wind flow.

[0026] When wind blows across the anemometer from any direction, the relationship between the temperature sensor readings at different angles of the innermost ring and the corresponding temperature sensor values ​​can be expressed by a Gaussian function as follows:

[0027]

[0028] In equation (3), (θ) i ,y i (i = 1, 2, 3, ..., 8) represents the temperature sensor θ at different angles. i The corresponding temperature reading y i y max θ represents the peak value of the Gaussian curve, corresponding to the reading of the temperature sensor with the highest reading; max The peak position of the Gaussian curve represents the angular position of the temperature sensor with the highest value; S represents the half-width information of the Gaussian curve.

[0029] Taking the natural logarithm of both sides of equation (3), we get:

[0030]

[0031] make:

[0032]

[0033] Equation (5) can be expressed in matrix form as follows:

[0034]

[0035] Equation (6) is denoted as Z = XB. According to the least squares principle, the generalized least squares solution of the matrix B is:

[0036] B = (X) T X) -1 X T Z (7)

[0037] Finally, according to equation (7), the parameter y is obtained. max and θ max To determine the direction of wind flow.

[0038] In conjunction with the second aspect, the wind pressure measurement principle further includes:

[0039] The four air ducts are numbered sequentially as duct A, duct C, duct B, and duct D in a clockwise direction when viewed from above. A coordinate system is established with the angle bisector of the angle between duct A and duct C as the Y-axis. The airflow velocity V in the direction of duct A and duct B is... AB The airflow velocity V in the direction of pipe C and pipe D CD And wind speed V, expressed by the following formula:

[0040]

[0041]

[0042]

[0043] In equations (8) and (9), K is the correction coefficient for the actual wind speed, ρ is the fluid density, and P is the fluid density. A P is the internal air pressure at the opening of pipe A. B P is the internal air pressure at the opening of pipe B. C P is the internal air pressure at the opening of pipe C. D The internal air pressure at the opening of pipe D;

[0044] The wind direction angle α is the angle between the wind direction and the Y-axis, expressed by the following formula:

[0045]

[0046] In conjunction with the second aspect, the preset Kalman filter algorithm further includes:

[0047] Step 1: The prediction process, expressed by the following formula:

[0048]

[0049] P k∣k-1 =FP k-1∣k-1 F T +Q k (13)

[0050] Based on the wind speed v at the previous moment and the wind speed deviation using the thermal anemometer principle. Describe the state, that is Establish a time series model of the state, expressed by the following formula:

[0051]

[0052] Since there is no control process, substituting equation (14) into equation (12) yields:

[0053]

[0054] In the formula, That is, the state transition matrix;

[0055] The error covariance matrix P is represented by a 2×2 matrix. k∣k-1 : Estimating the covariance Q from the state representation of wind speed v Covariance of deviation from estimated speed The process noise covariance Q represents k :

[0056] Then the error covariance matrix P at time k k∣k-1 It can be expressed by the following formula:

[0057]

[0058] Step 2: The correction process is expressed by the following formula:

[0059] S k =HP - k∣k-1 H T +R (17)

[0060]

[0061]

[0062] P k∣k =(IK k H)P k∣k-1 (20)

[0063] Observation z k If the wind speed is measured using the wind pressure anemometer principle, then let H = [1 0], and from z k and The residual values ​​can be obtained together:

[0064]

[0065] The error covariance matrix P at time k in equation (16) k∣k-1 Substituting into equation (16), we can obtain the residual covariance:

[0066]

[0067] Observation z k It is the wind speed value measured by the wind pressure measurement principle, so R in equation (17) is equal to the variance of the observed value;

[0068] H, P k∣k-1 and S k Substituting into equation (18), we obtain the Kalman coefficients:

[0069]

[0070] Combining equations (19), (21), and (23), we obtain the system state at time k after Kalman filtering:

[0071]

[0072] Formula (23), H, P k∣k-1 Substituting into equation (20), we obtain the updated error covariance matrix:

[0073]

[0074] Wind speed and wind direction are obtained by combining equations (22)-(25).

[0075] Compared with the prior art, the beneficial effects achieved by the multi-sensor fusion anemometer and wind measurement method based on Kalman filter provided in this embodiment of the invention include:

[0076] This invention provides a main body of an anemometer, a thermal anemometer panel located on the upper part of the main body, and an anemometer top cover located on the upper part of the thermal anemometer panel. A heating rod and 36 temperature sensors are fixed to the thermal anemometer panel through holes, and the 36 temperature sensors are evenly distributed around the heating rod. A main control circuit is provided inside the main body of the anemometer, and the main control circuit is connected to the heating rod and the 36 temperature sensors. Based on the thermal anemometer principle, it measures wind speed and direction to obtain initial values ​​for wind speed and direction. Four air guide pipes are provided on the upper surface of the anemometer top cover, and a micro-pressure sensor is provided on the lower surface of the anemometer top cover. The air guide pipes are connected to the micro-pressure sensors. The main control circuit is connected to the air guide pipes and the micro-pressure sensors, and measures wind speed and direction based on the wind pressure anemometer principle to obtain observed values ​​for wind speed and direction.

[0077] The main control circuit of this invention uses a preset Kalman filter algorithm to process the initial and observed values ​​to obtain wind speed and direction. This invention integrates the results of thermal anemometer and wind pressure anemometer principles, offering the advantage of accurate wind speed and direction measurement. The anemometer provided by this invention has a stable structure and an accurate wind measurement method. It overcomes the shortcomings of commonly used mechanical anemometers, which are prone to damage and affect measurement accuracy; it overcomes the shortcomings of hot-wire anemometers, which are prone to damage and cannot be used on a large scale in the field; and it overcomes the shortcomings of ultrasonic anemometers, which have high errors. This invention provides a high-precision, easy-to-maintain anemometer and wind measurement method that can be used in harsh field environments. Attached Figure Description

[0078] Figure 1 This is a structural diagram of a multi-sensor fusion anemometer based on a Kalman filter provided in Embodiment 1 of the present invention;

[0079] Figure 2 This is a top view of the top-cover-free structure of a multi-sensor fusion anemometer based on a Kalman filter provided in Embodiment 1 of the present invention;

[0080] Figure 3 This refers to the temperature of the innermost sensor probe under different wind speeds and ambient temperatures in a multi-sensor fusion wind measurement method based on a Kalman filter provided in Embodiment 2 of the present invention.

[0081] Figure 4 This is a distribution diagram of the heating rod temperature on a thermal anemometer panel under different wind directions in a multi-sensor fusion wind measurement method based on a Kalman filter provided in Embodiment 2 of the present invention.

[0082] Figure 5 This is a temperature distribution map around the heating source in a multi-sensor fusion wind measurement method based on a Kalman filter provided in Embodiment 2 of the present invention.

[0083] Figure 6 This is a schematic diagram of establishing a coordinate system in a multi-sensor fusion wind measurement method based on a Kalman filter provided in Embodiment 2 of the present invention.

[0084] In the diagram: 1. Anemometer body; 2. Thermal anemometer panel; 3. Anemometer top cover; 4. Temperature sensor; 5. Heating rod; 6. Air duct; 7. Micro-pressure sensor. Detailed Implementation

[0085] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0086] Example 1:

[0087] like Figure 1 As shown, this embodiment provides a multi-sensor fusion anemometer based on a Kalman filter, including an anemometer body, a thermal anemometer panel disposed on the upper part of the anemometer body, and an anemometer top cover disposed on the upper part of the thermal anemometer panel.

[0088] The thermal air sensor panel is a non-thermal conductive nylon structure with a diameter of 150mm and a thickness of 4mm. The thermal air sensor panel has holes for the installation of temperature sensors and heating plates.

[0089] A heating rod and 36 temperature sensors are fixed to the thermal anemometer panel through holes, and the 36 temperature sensors are evenly distributed around the heating rod.

[0090] The heating rod is a 150W ceramic heating rod with a diameter of 3mm. It is installed in the center of the thermal air measurement panel through the central hole, and its top is 50mm higher than the thermal air measurement panel.

[0091] The 36 temperature sensors are PT100A grade temperature sensors with a diameter of 3mm and a surface covered with pure copper. They are evenly distributed on the thermal anemometer panel through the holes on the panel, with their tops 15mm away from the panel.

[0092] like Figure 2As shown, 32 of the 36 temperature sensors are evenly distributed in a 4×8 ring array. The angle between adjacent sensors in each ring is 45°. From the inside out, the rings are designated as the first, second, third, and fourth rings. The distance between the innermost ring's eight temperature sensors and the central heating source is 3mm. The distances between the first and second rings, the second and third rings, and the third and fourth rings are all 5mm. The distance between the fourth ring and the edge of the base is 8mm. The remaining four temperature sensors, used to measure ambient temperature, are located outside the fourth ring, forming the fifth ring. The angle between adjacent sensors in the fifth ring is 90°, and the distance between the fifth and fourth rings is 5mm.

[0093] The top surface of the anemometer cover is equipped with four air guide tubes, and the bottom surface of the anemometer cover is equipped with a micro-pressure sensor. The air guide tubes are connected to the micro-pressure sensor.

[0094] The air ducts are arranged in a cross shape, with two ducts at 180° to each other connected to the same micro-pressure sensor. The inner diameter of the air duct is 2mm and the wall thickness is 1mm. The micro-pressure sensor is an SM9541 micro-differential pressure sensor.

[0095] The anemometer's main body contains a main control circuit, which is connected to a heating rod and 36 temperature sensors. Based on the principle of thermal anemometer measurement, it measures wind speed and direction to obtain initial values. The main control circuit is also connected to an air duct and a micro-pressure sensor, which measures wind speed and direction based on the principle of wind pressure measurement to obtain observed values. Finally, the main control circuit processes the initial and observed values ​​to obtain the final wind speed and direction.

[0096] Example 2:

[0097] This embodiment provides a multi-sensor fusion wind measurement method based on a Kalman filter, including:

[0098] The main control circuit acquires data from the heating rod and 36 temperature sensors, and measures wind speed and direction based on the principle of thermal wind measurement to obtain the initial values ​​of wind speed and direction.

[0099] The main control circuit acquires data from the air duct and the micro-pressure sensor, and measures the wind speed and direction according to the wind pressure measurement principle to obtain the observed values ​​of wind speed and direction;

[0100] The main control circuit uses a preset Kalman filter algorithm to process the initial and observed values ​​to obtain wind speed and wind direction.

[0101] The principle of thermal anemometers is as follows: Thermal anemometers operate under the direct interaction of thermal and wind fields, and their working principle is a comprehensive theory involving multiple disciplines such as thermodynamics and fluid mechanics. Before designing a thermal anemometer, it is necessary to analyze the direct heat transfer between the fluid and the heat source. When the fluid and the heated solid wall are in direct contact and in relative motion, this heat transfer process is called convective heat transfer.

[0102] When the sensors are evenly distributed around the heat source, wind blows towards the ring array from any direction. The central constant-temperature heat source carries away some of the heat, causing changes in the temperature values ​​of the sensors in each ring, creating a temperature difference. The temperature difference is particularly pronounced in the first ring, exhibiting a clear Gaussian distribution. Therefore, for calculating the wind direction angle, only the eight temperature sensors in the innermost ring need to be used for evaluation. However, for calculating the wind speed, the average value of the 32 temperature sensors on the array must be used to evaluate the flow velocity.

[0103]

[0104] In equation (1), T n Let n be the temperature values ​​of each temperature sensor, n = 1, 2, 3, ..., 32;

[0105] The ambient temperatures were set to -10℃, 0℃, 10℃, 20℃, 30℃, and 40℃, and the wind speeds were set to 0.5m / s, 1m / s, 1.5m / s, 2m / s, 2.5m / s, 3m / s, 3.5m / s, 4m / s, 4.5m / s, 5m / s, 5.5m / s, and 6m / s, respectively. The average temperature values ​​from 32 temperature sensors measured in multiple experiments were fitted into a curve, as shown in the figure. Figure 3 As shown. Let the ambient temperature be x, the average temperature value from the temperature sensor be y, and the wind speed be z. The fitting function is:

[0106]

[0107] In equation (2), p1 = 10.8945, p2 = 0.6475, p3 = -1.019, p4 = 0.0055, p5 = -0.2575, p6 = 0.0012, p7 = 0.227, p8 = 0.0002, p9 = -2.5708e -5 .

[0108] The temperature distribution of the heating rod on the thermal anemometer panel under different wind directions is as follows: Figure 4 As shown, the temperature distribution around the central heating rod of the circular array anemometer resembles a Gaussian distribution function from the inside out, with higher temperatures further up the wind direction. Therefore, as long as the peak value of the Gaussian curve is detected, such as... Figure 5 As shown, the direction of wind flow can be determined.

[0109] When wind blows across the anemometer from any direction, the relationship between the temperature sensor readings at different angles of the innermost ring and the corresponding temperature sensor values ​​can be expressed by a Gaussian function as follows:

[0110]

[0111] In equation (3), (θ)i ,y i (i = 1, 2, 3, ..., 8) represents the temperature sensor θ at different angles. i The corresponding temperature reading y i y max θ represents the peak value of the Gaussian curve, corresponding to the reading of the temperature sensor with the highest reading; max The peak position of the Gaussian curve represents the angular position of the temperature sensor with the highest value; S represents the half-width information of the Gaussian curve.

[0112] Taking the natural logarithm of both sides of equation (3), we get:

[0113]

[0114] make:

[0115]

[0116] Equation (5) can be expressed in matrix form as follows:

[0117]

[0118] Equation (6) is denoted as Z = XB. According to the least squares principle, the generalized least squares solution of the matrix B is:

[0119] B = (X) T X) -1 X T Z (7)

[0120] Finally, according to equation (7), the parameter y is obtained. max and θ max To determine the direction of wind flow.

[0121] Wind pressure measurement principle: Assuming that air is an incompressible gas, according to Bernoulli's equation for an ideal incompressible gas: In an ideal flow field, the sum of the kinetic energy, gravitational potential energy, and pressure potential energy of the fluid at any two points on the same streamline is a constant, as shown in formula (8):

[0122]

[0123] In equation (8), ρ is the fluid density, v is the fluid velocity, g is the gravitational acceleration, h is the height of the plumb bob, P is the pressure potential energy, and C is a constant.

[0124] If there are two parallel, oppositely opening thin tubes, tube A and tube B, within the flow path, then according to equation (8):

[0125]

[0126] The air at the opening of pipe A can be considered stationary due to the stagnation point, i.e., v A =0m / s, its internal wind pressure P A The total air pressure at that altitude; the air velocity v at the opening of pipe B. B It can be approximated as the incoming flow velocity v, and its internal wind pressure P B This refers only to the static pressure of air at that location. Furthermore, since the openings of pipe A and pipe B are at the same height (h),... A =h B Therefore, the flow velocity v can be obtained by solving formula (9):

[0127]

[0128] However, in actual measurement processes, due to the influence of fluid viscosity, it is usually... A correction factor K is introduced to correct the actual wind speed, as shown in formula (11):

[0129]

[0130] Based on the above principle, by setting four thin tubes orthogonally to each other in a plane, the velocity and angle of the flow parallel to the two-dimensional plane in any direction can be measured.

[0131] like Figure 6 A coordinate system is established as shown. The four air ducts are numbered sequentially as duct A, duct C, duct B, and duct D in a clockwise direction when viewed from above. The Y-axis is the angle bisector of the angle between duct A and duct C. The airflow velocity V in the direction of duct A and duct B is also shown. AB The airflow velocity V in the direction of pipe C and pipe D CD And wind speed V, expressed by the following formula:

[0132]

[0133]

[0134]

[0135] In equations (12) and (13), K is the correction coefficient for the actual wind speed, ρ is the fluid density, and P is the fluid density. A P is the internal air pressure at the opening of pipe A. B P is the internal air pressure at the opening of pipe B. C P is the internal air pressure at the opening of pipe C. D The internal air pressure at the opening of pipe D;

[0136] The wind direction angle α is the angle between the wind direction and the Y-axis, expressed by the following formula:

[0137]

[0138] The preset Kalman filtering algorithm includes:

[0139] Step 1: The prediction process, expressed by the following formula:

[0140]

[0141] P k∣k-1 =FP k-1∣k-1 F T +Q k (17)

[0142] Based on the wind speed v at the previous moment and the wind speed deviation using the thermal anemometer principle. Describe the state, that is Establish a time series model of the state, expressed by the following formula:

[0143]

[0144] Since there is no control process, substituting equation (18) into equation (16) yields:

[0145]

[0146] In the formula, That is, the state transition matrix;

[0147] The error covariance matrix is ​​represented by a 2×2 matrix.

[0148] Estimating the covariance Q from the state representation of wind speed v Covariance of deviation from estimated speed The process noise covariance Q represents k :

[0149] Then the error covariance matrix P at time k k∣k-1 It can be expressed by the following formula:

[0150]

[0151] Step 2: The correction process is expressed by the following formula:

[0152] S k =HP - k∣k-1 H T +R (21)

[0153]

[0154]

[0155] P k∣k=(IK k H)P k∣k-1 (twenty four)

[0156] Observation z k If the wind speed is measured using the wind pressure anemometer principle, then let H = [1 0], and from z k and The residual values ​​can be obtained together:

[0157]

[0158] The error covariance matrix P at time k in equation (20) k∣k-1 Substituting into equation (21), we can obtain the residual covariance:

[0159]

[0160] Observation z k R is the wind speed value measured by the wind pressure measurement principle, therefore R in equation (21) is equal to the variance of the observed value;

[0161] H, P k∣k-1 and S k Substituting into equation (22), we obtain the Kalman coefficients:

[0162]

[0163] Combining equations (23), (25), and (27), we obtain the system state at time k after Kalman filtering:

[0164]

[0165] Formula (28), H, P k∣k-1 Substituting into equation (24), we obtain the updated error covariance matrix:

[0166]

[0167] Wind speed and wind direction are obtained by combining equations (26)-(29).

[0168] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A multi-sensor fusion anemometer based on a Kalman filter, characterized in that, include: The anemometer body, the thermal anemometer panel located on the upper part of the anemometer body, and the anemometer top cover located on the upper part of the thermal anemometer panel; A heating rod and 36 temperature sensors are fixed on the thermal wind measurement panel through holes, and the 36 temperature sensors are evenly distributed around the heating rod. The upper surface of the anemometer's top cover is provided with four air guide pipes, and the lower surface of the anemometer's top cover is provided with a micro-pressure sensor. The air guide pipes are connected to the micro-pressure sensor. The anemometer body is equipped with a main control circuit, which is connected to the heating rod and 36 temperature sensors. It measures wind speed and direction according to the principle of thermal anemometer to obtain the initial values ​​of wind speed and direction. The main control circuit is also connected to the air duct and micro pressure sensor to measure wind speed and direction according to the principle of wind pressure anemometer to obtain the observed values ​​of wind speed and direction. The main control circuit uses a preset Kalman filter algorithm to process the initial and observed values ​​to obtain wind speed and wind direction; The wind pressure measurement principle includes: The four air ducts are numbered sequentially as duct A, duct C, duct B, and duct D in a clockwise direction when viewed from above. A coordinate system is established with the angle bisector of the angle between duct A and duct C as the Y-axis. The airflow velocity V in the direction of duct A and duct B is calculated. AB The airflow velocity V in the direction of pipe C and pipe D CD And wind speed V; wind direction angle α is the angle between the wind direction and the Y-axis, based on V AB With V CD Calculate the wind direction angle α; The preset Kalman filter algorithm includes a prediction process and a correction process. The prediction process includes the wind speed deviation based on the wind speed v at the previous moment and the thermal anemometer principle. Describe the state Establish a time series model of the state; using the state transition matrix Process noise covariance Calculate the error covariance matrix at time k; the correction process includes using the wind speed value measured by the wind pressure measurement principle as the observed value. The Kalman coefficients are calculated, and the predicted values ​​are corrected using the observation residuals. The error covariance matrix is ​​then updated to obtain the wind speed at the current moment.

2. The multi-sensor fusion anemometer based on a Kalman filter according to claim 1, characterized in that, Of the 36 temperature sensors, 32 are evenly distributed in a 4×8 ring array. The angle between adjacent sensors in each ring is 45°. From the inside out, the rings are the first, second, third, and fourth rings, respectively. Four temperature sensors are located outside the fourth ring to form the fifth ring, and the angle between adjacent sensors in the fifth ring is 90°.

3. The multi-sensor fusion anemometer based on a Kalman filter according to claim 1, characterized in that, The air ducts are arranged in a cross shape, with two air ducts at 180° to each other connected to the same micro-pressure sensor.

4. A wind measurement method based on a multi-sensor fusion anemometer using a Kalman filter, as described in any one of claims 1 to 3, characterized in that, include: The main control circuit acquires data from the heating rod and 36 temperature sensors, and measures wind speed and direction based on the principle of thermal wind measurement to obtain the initial values ​​of wind speed and direction. The main control circuit acquires data from the air duct and the micro-pressure sensor, and measures wind speed and direction based on the wind pressure measurement principle to obtain the observed values ​​of wind speed and direction; wherein, the wind pressure measurement principle includes: The four air ducts are numbered sequentially as duct A, duct C, duct B, and duct D in a clockwise direction when viewed from above. A coordinate system is established with the angle bisector of the angle between duct A and duct C as the Y-axis. The airflow velocity V in the direction of duct A and duct B is calculated. AB The airflow velocity V in the direction of pipe C and pipe D CD And wind speed V; wind direction angle α is the angle between the wind direction and the Y-axis, based on V AB With V CD Calculate the wind direction angle α; The main control circuit uses a preset Kalman filter algorithm to process the initial and observed values ​​to obtain wind speed and direction. The preset Kalman filter algorithm includes a prediction process and a correction process. The prediction process includes the wind speed deviation based on the previous moment's wind speed *v* and the principle of thermal anemometer. Describe the state Establish a time series model of the state; using the state transition matrix Process noise covariance Calculate the error covariance matrix at time k; the correction process includes using the wind speed value measured by the wind pressure measurement principle as the observed value. The Kalman coefficients are calculated, and the predicted values ​​are corrected using the observation residuals. The error covariance matrix is ​​then updated to obtain the wind speed at the current moment.

5. The wind measurement method according to claim 4, characterized in that, The principle of thermal wind measurement includes: When the sensors are evenly distributed around the heat source, wind blows towards the ring array from any direction. The central constant-temperature heat source carries away some of the heat, causing changes in the temperature values ​​of the sensors in each ring, creating a temperature difference. The temperature difference is particularly pronounced in the first ring, exhibiting a clear Gaussian distribution. Therefore, for calculating the wind direction angle, only the eight temperature sensors in the innermost ring need to be used for evaluation. However, for calculating the wind speed, the average value of the 32 temperature sensors on the array must be used to evaluate the flow velocity. （1） In equation (1), T n Let n be the temperature values ​​of each temperature sensor, n=1,2,3,…,32; The ambient temperatures were set to -10℃, 0℃, 10℃, 20℃, 30℃, and 40℃, and the wind speeds were set to 0.5m / s, 1m / s, 1.5m / s, 2m / s, 2.5m / s, 3m / s, 3.5m / s, 4m / s, 4.5m / s, 5m / s, 5.5m / s, and 6m / s, respectively. The average temperature values ​​from 32 temperature sensors measured in multiple experiments were fitted into a curve. Let the ambient temperature be x, the average temperature value from the temperature sensors be y, and the wind speed be z. The fitting function is: （2） In equation (2), p1=10.8945, p2=0.6475, p3=-1.019, p4=0.0055, p5=-0.2575, p6=0.0012, p7=0.227, p8=0.0002, p9=-2.5708e -5 ; When the sensors are evenly distributed around the heat source, the wind blows towards the ring array from any direction. The central constant-temperature heat source carries away some of the heat from the heat source, causing the temperature values ​​of the temperature sensors in each ring to change, forming a temperature difference. The temperature difference in the first ring is particularly obvious. The temperature difference shows a clear Gaussian distribution. The temperature is higher in the wind direction. Detecting the peak of the Gaussian curve can determine the direction of wind flow. When wind blows across the anemometer from any direction, the relationship between the temperature sensor readings at different angles of the innermost ring and the corresponding temperature sensor values ​​can be expressed by a Gaussian function as follows: （3） In equation (3), (θ) i ,y i (i=1,2,3,……,8) represents the temperature sensor θ at different angles. i The corresponding temperature reading y i y max θ represents the peak value of the Gaussian curve, corresponding to the reading of the temperature sensor with the highest reading; max The peak position of the Gaussian curve represents the angular position of the temperature sensor with the highest value; S represents the half-width information of the Gaussian curve. Taking the natural logarithm of both sides of equation (3), we get: （4） make: （5） Equation (5) can be expressed in matrix form as follows: （6） Equation (6) is denoted as Z=XB. According to the least squares principle, the generalized least squares solution of the matrix B is: （7） Finally, according to equation (7), the parameter y is obtained. max and θ max To determine the direction of wind flow.

6. The wind measurement method according to claim 4, characterized in that, The principle of wind pressure measurement includes: The airflow velocity V in the direction of pipe A and pipe B AB The airflow velocity V in the direction of pipe C and pipe D CD And wind speed V, expressed by the following formula: （8） （9） （10） In equations (8) and (9), K is the correction coefficient for the actual wind speed, ρ is the fluid density, and P is the fluid density. A P is the internal air pressure at the opening of pipe A. B P is the internal air pressure at the opening of pipe B. C P is the internal air pressure at the opening of pipe C. D The internal air pressure at the opening of pipe D; The wind direction angle α is the angle between the wind direction and the Y-axis, expressed by the following formula: （11）。 7. The wind measurement method according to claim 4, characterized in that, The preset Kalman filter algorithm includes: Step 1: The prediction process, expressed by the following formula: （12） （13） Based on the wind speed v at the previous moment and the wind speed deviation using the thermal anemometer principle. Describe the state, that is Establish a time series model of the state, expressed by the following formula: （14） Since there is no control process, substituting equation (14) into equation (12) yields: （15） In the formula, That is, the state transition matrix; The error covariance matrix is ​​represented by a 2×2 matrix. : ; Estimating covariance from the state representing wind speed Covariance of deviation from estimated speed Represents process noise covariance : ; Then the error covariance matrix at time k It can be expressed by the following formula: （16） Step 2: The correction process is expressed by the following formula: （17） （18） （19） （20） Observations If the wind speed value is measured using the principle of wind pressure measurement, then let ,Depend on and The residual values ​​can be obtained together: （21） The error covariance matrix at time k in equation (16) Substituting into equation (17), we can obtain the residual covariance: （22） Observations R is the wind speed value measured by the wind pressure measurement principle, therefore R in equation (17) is equal to the variance of the observed value; H, and S k Substituting into equation (18), we obtain the Kalman coefficients: （23） Combining equations (19), (21), and (23), we obtain the system state at time k after Kalman filtering: （24） Formula (23), H, Substituting into equation (20), we obtain the updated error covariance matrix: （25） The wind speed is obtained by combining equations (22) and (25).

Images