Aerosol particle spectrum distribution inversion method based on intelligent swarm algorithm

Abstract

The invention discloses an aerosol particle spectrum distribution inversion method based on an intelligent swarm algorithm, and the method comprises the steps: selecting double logarithmic normal distribution as a function form of an aerosol particle spectrum, and taking aerosol particle number concentration, geometric standard deviation and median radius as unknown parameters; the method comprises the following steps: setting a numerical range of unknown parameters, and inputting the unknown parameters into a Red-tailed Hawk (RTH) algorithm in an intelligent group algorithm; according to the relation between the aerosol optical thickness and the aerosol particle spectrum, the RTH algorithm is used for optimizing and generating parameters to calculate the aerosol optical thickness, the aerosol optical thickness is compared with the aerosol optical thickness measured through the sun photometer, and when the error value reaches the minimum value, the parameters at the moment are adopted for inversion of aerosol particle spectrum distribution. Compared with the prior art, the method has the advantages that appropriate parameters of the aerosol particle spectrum function are selected through the RTH algorithm, accurate inversion of atmospheric aerosol particle spectrum distribution is achieved, calculation is simple, and the application range is wide.

Description

Technical Field

[0001] The present invention relates to the field of atmospheric optics technology, and in particular to an aerosol particle spectrum distribution inversion method based on an intelligent swarm algorithm. Background Art

[0002] Aerosols are an important part of the earth-atmosphere system and have a profound impact on the earth's climate, ecosystem and human health. Its main sources are divided into two categories: anthropogenic and natural. Anthropogenic sources include pollutants emitted by vehicle exhaust and factories; natural sources include sandstorms and volcanic eruptions. In the atmosphere, aerosols affect the radiation balance by scattering and absorbing radiation and change the properties of clouds. Aerosols have an important impact on climate, ecosystems and human health. The aerosol particle spectrum is one of the important parameters to describe the optical properties of aerosols. Accurate measurement and inversion through relevant instruments can provide a scientific basis for understanding the composition, size distribution and dynamic changes of aerosols in the atmosphere, which is helpful for air pollution source analysis and climate impact assessment.

[0003] There are currently many methods for inverting aerosol particle spectra based on solar photometer data. The most common one is the AERONET (Atmospheric Radiation Measurement Network) algorithm. The AERONET network provides solar photometer observation data worldwide, and uses a special inversion algorithm to invert aerosol particle spectra and other parameters based on multi-wavelength optical thickness and scattering angle information. The regularization algorithm uses the first-kind Fredholm integral equation to build an inversion formula, handles the pathological problems that may arise in the inversion process, and introduces additional prior information or constraints to suppress high-frequency oscillations or instabilities in the inversion process. These methods are more complex in processing optical data and require advanced mathematical and computational techniques to implement and optimize the inversion algorithm, which increases the implementation difficulty and computational cost of the method. Summary of the invention

[0004] The purpose of the present invention is to provide an aerosol particle spectrum distribution inversion method based on an intelligent swarm algorithm to solve the problems of complex aerosol particle spectrum inversion formula, difficult implementation and high calculation cost.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is:

[0006] An aerosol particle spectrum distribution inversion method based on intelligent swarm algorithm includes the following steps:

[0007] Step S1: Obtain the optical thickness of the bottom atmosphere obtained by inverting the sun photometer data as a measured value;

[0008] Step S2: using the double log-normal distribution function as a function for inverting the aerosol particle spectrum distribution;

[0009] Step S3: calculating the aerosol optical depth according to the aerosol particle spectrum function;

[0010] Step S4: Select the fluctuation range of unknown parameters in the aerosol particle spectrum function as the initial population of the RTH algorithm, use the RTH algorithm to optimize the parameters, and bring them into the aerosol particle spectrum function to calculate the aerosol optical thickness;

[0011] Step S5: According to the minimum error value between the calculated aerosol optical thickness data and the measured value, the aerosol particle spectrum parameters are obtained, and then the distribution of the aerosol particle spectrum is determined by inversion.

[0012] In the above technical solution, in step S1, the step of obtaining the sunphotometer data includes: obtaining the output voltage values ​​of the sunphotometer in six bands of 340nm, 440nm, 500nm, 675nm, 870nm and 1020nm; and calculating the aerosol optical thickness according to the Beer-Lambert-Bouguer law.

[0013] In the above technical solution, in step S2, the formula of the function for inverting the aerosol particle spectrum distribution is:

[0014] (1)

[0015] Among them, N i is the aerosol particle number concentration, is the geometric standard deviation, r mi is the median radius and r is the radius of the aerosol particle.

[0016] In the above technical solution, in step S3, the relationship between the aerosol particle spectrum and the aerosol optical thickness is:

[0017] (2)

[0018] Among them, Q ext (λ, r, m) is the extinction efficiency factor, λ is the wavelength, m is the complex refractive index, r m is the upper limit of the aerosol particle radius, r 0 is the lower limit of the aerosol particle radius, n(r, z) is the aerosol particle spectrum distribution at a certain height, n(r, z)dz is the number of aerosol particles with a radius range from r to r+dr at height z, and z m is the height of the top of the atmospheric column.

[0019] In the above technical solution, in step S4, the step of optimizing parameters using the RTH algorithm includes:

[0020] Select N in the aerosol particle spectrum function i , 、r mi as unknown parameters;

[0021] Select the fluctuation range of unknown parameters as the initial population of the RTH algorithm;

[0022] The RTH algorithm is divided into three stages to optimize parameters: the high-altitude soaring stage, the red-tailed hawk will fly very far to find the best position for prey; the low-altitude circling stage, the red-tailed hawk flies lower to the ground in a spiral to surround the prey; the stooping and diving stage, the red-tailed hawk moves to the best hunting position within a selected area around the prey, and attacks the prey from the best position during the low-flying stage.

[0023] In the above technical solution, in step S5, the step of calculating the minimum error to obtain the aerosol particle spectrum distribution includes: i , 、r mi Substitute it into the aerosol particle spectrum function, and use the relationship formula between the aerosol particle spectrum and the aerosol optical thickness to calculate the aerosol optical thickness; use the calculated aerosol optical thickness and the true value as the error, and under the optimization of the RTH algorithm, obtain the N corresponding to the minimum error value. i , 、r mi To determine the spectral distribution, the actual atmospheric aerosol particle spectral distribution can be obtained.

[0024] Beneficial effects:

[0025] The present invention discloses an aerosol particle spectrum distribution inversion method based on an intelligent swarm algorithm, wherein the method selects a double logarithmic normal distribution as the function form of the aerosol particle spectrum, wherein the aerosol particle number concentration, geometric standard deviation and median radius are used as unknown parameters; by setting the numerical range of the unknown parameters, the RTH algorithm in the intelligent swarm algorithm is input; according to the relationship between the aerosol optical thickness and the aerosol particle spectrum, the RTH algorithm is used to optimize the generated parameters to calculate the aerosol optical thickness, and the parameters are compared with the aerosol optical thickness measured by the sun photometer, and when the error value reaches the minimum, the parameters at this time are used to invert the aerosol particle spectrum distribution. Compared with the prior art, the present invention selects appropriate parameters of the aerosol particle spectrum function through the RTH algorithm, realizes the accurate inversion of the atmospheric aerosol particle spectrum distribution, and the calculation is simple and has a wide range of applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a flow chart of an aerosol particle spectrum distribution inversion method based on intelligent swarm algorithm.

[0027] Figure 2 This is the flow chart of the RTH algorithm in the intelligent swarm algorithm.

[0028] Figure 3 The figure shows the comparison between the aerosol optical thickness (AOD) inversion result and the measured data described in the specific implementation manner of the present invention. DETAILED DESCRIPTION

[0029] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. However, the following embodiments are limited to explaining the present invention, and the protection scope of the present invention should include the entire contents of the claims, and through the description of the following embodiments, those skilled in the art can fully implement the entire contents of the claims of the present invention.

[0030] Example

[0031] like Figure 1 As shown in FIG. 1 , an aerosol particle spectrum distribution inversion method based on an intelligent swarm algorithm is provided, which includes the following steps:

[0032] Step S1: Obtain the optical thickness of the bottom atmosphere obtained by inverting the sun photometer data as a measured value;

[0033] Step S2: using the double logarithmic normal distribution function as a function for inverting the aerosol particle spectrum distribution; assuming that the aerosol is a superposition of multiple modes, the formula is:

[0034] (1)

[0035] Where N i , 、r mi are the aerosol particle number concentration, geometric standard deviation and median radius, respectively, and r is the aerosol particle radius.

[0036] Step S3: Calculate the aerosol optical thickness according to the aerosol particle spectrum function. The relationship between the aerosol particle spectrum and the aerosol optical thickness is as follows:

[0037] (2)

[0038] Among them, Q ext (λ, r, m) is the extinction efficiency factor, λ is the wavelength, m is the complex refractive index, r m and r 0 They represent the upper and lower limits of the aerosol particle radius, n(r, z) represents the aerosol particle spectrum distribution at a certain height, n(r, z)dz represents the number of aerosol particles with a radius range of r to r+dr at height z, and z m is the height of the top of the atmospheric column.

[0039] Step S4: Select the fluctuation range of unknown parameters in the aerosol particle spectrum function as the initial population of the RTH algorithm, and use the RTH algorithm to optimize the parameters. Substitute the optimized parameters into the aerosol particle spectrum function to calculate the aerosol optical thickness. Figure 2 As shown, the RTH algorithm is used to continuously optimize the parameters. According to the unknown parameters (N i , 、r mi ) possible range, randomly generate a set of initial parameter values ​​as the initial population of the RTH algorithm. RTH algorithm optimization process specific optimization process:

[0040] During the high-altitude soaring stage, the red-tailed hawk will fly very far to find the best location for its prey. The mathematical model is as follows:

[0041] (3)

[0042] , (4)

[0043] (5)

[0044] Where X(t) represents the position of the red-tailed hawk at iteration t; t is the number of iterations; X best is the best position to obtain; X mean is the average value of the position; Levy(dim) represents the Levy flight distribution function that can be calculated according to formula (4), s and β are constants, dim represents the dimension, u, v are random values ​​between (0,1); TF(t) represents the transition factor function, which is calculated according to formula (5), where T max Indicates the maximum number of iterations.

[0045] The key to this stage is to detect the location of the prey. When the eagle approaches the prey, the step size will gradually decrease. Due to the hybrid evolutionary group update mechanism, the eagle spends a small number of iterations in the exploration process. The high-altitude soaring stage based on the Levy function avoids falling into the local minimum.

[0046] In the low-altitude hovering phase, the eagle flies lower to the ground in a spiral line to surround the prey. The mathematical model is as follows:

[0047] (6)

[0048] Where x and y represent direction coordinates, which can be calculated as follows:

[0049] (7)

[0050] Among them, R 0represents the initial value of the radius (0.5-3), A represents the angle gain (5-15), rand represents the random gain (0-1), and r represents the control gain (1,2). This stage improves the convergence speed of the development stage by gradually reducing the flight altitude and surrounding the prey. Spiral flight and low-altitude dive further enhance the accuracy of the RTH algorithm.

[0051] During the stooping phase, the Red-tailed Hawk moves to the best hunting position within a selected area around the prey, and attacks the prey from the best position during the low-flying phase. The mathematical model is as follows:

[0052] (8)

[0053] Among them, each step size can be calculated as follows:

[0054] (9)

[0055] Among them, α and G are acceleration and gravity factors respectively.

[0056] Step S5: In the final population of the RTH algorithm, find the parameter combination with the smallest error with the measured optical thickness. This parameter combination is determined as the parameter of the aerosol particle spectrum, that is, N i , and r mi Using the determined parameter values, the spectral distribution of aerosol particles is inverted through the double log-normal distribution function.

[0057] like Figure 3 As shown, 8 groups of AOD calculated values ​​and measured values ​​at different time points on April 30, 2020 were selected at a wavelength of 500nm at the Shouxian Observatory in Anhui Province, where the black broken line is the actual solar optical thickness value obtained by the Shouxian Observatory in Anhui Province, and the red broken line is the solar optical thickness value obtained by the RTH algorithm of the present invention. The average value of the relative deviation between the calculated value and the measured value is 3.47%. It can be concluded that the calculated value of the solar optical thickness obtained by the present invention is close to the measured value.

[0058] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features applied for herein.

Claims

1. A method for inverting aerosol particle spectrum distribution based on intelligent swarm algorithm, characterized in that: The following steps are involved: Step S1: Obtain the optical thickness of the bottom atmosphere obtained by inverting the sun photometer data as a measured value; Step S2: using the double log-normal distribution function as a function for inverting the aerosol particle spectrum distribution; Step S3: calculating the aerosol optical depth according to the aerosol particle spectrum function; Step S4: Select the fluctuation range of unknown parameters in the aerosol particle spectrum function as the initial population of the RTH algorithm, use the RTH algorithm to optimize the parameters, and bring them into the aerosol particle spectrum function to calculate the aerosol optical thickness; Step S5: According to the minimum error value between the calculated aerosol optical thickness data and the measured value, the aerosol particle spectrum parameters are obtained, and then the distribution of the aerosol particle spectrum is determined by inversion.

2. According to claim 1, the method for inverting aerosol particle spectrum distribution based on intelligent swarm algorithm is characterized by: In step S1, the step of obtaining the sun photometer data includes: obtaining the output voltage values ​​of the sun photometer in six bands of 340nm, 440nm, 500nm, 675nm, 870nm and 1020nm; and calculating the aerosol optical thickness according to the Beer-Lambert-Bouguer law.

3. The aerosol particle spectrum distribution inversion method based on intelligent swarm algorithm according to claim 1 is characterized by: In step S2, the formula of the function for inverting the aerosol particle spectrum distribution is: (1) Among them, N i is the aerosol particle number concentration, is the geometric standard deviation, r mi is the median radius, and r is the radius of the aerosol particle.

4. The aerosol particle spectrum distribution inversion method based on intelligent swarm algorithm according to claim 1 is characterized by: In step S3, the relationship between the aerosol particle spectrum and the aerosol optical thickness is: (2) Among them, Q ext (λ, r, m) is the extinction efficiency factor, λ is the wavelength, m is the complex refractive index, r m is the upper limit of the radius of the aerosol particles, r0 is the lower limit of the radius of the aerosol particles, n(r, z) is the aerosol particle spectrum distribution at a certain height, n(r, z)dz is the number of aerosol particles with a radius range of r to r+dr at height z, and z m is the height of the top of the atmospheric column.

5. The aerosol particle spectrum distribution inversion method based on intelligent swarm algorithm according to claim 1 is characterized by: In step S4, the step of optimizing parameters using the RTH algorithm includes: Select N in the aerosol particle spectrum function i , 、r mi as unknown parameters; Select the fluctuation range of unknown parameters as the initial population of the RTH algorithm; The RTH algorithm is divided into three stages to optimize parameters: the high-altitude soaring stage, in which the red-tailed hawk will fly far away to find the best position for prey; the low-altitude circling stage, in which the hawk flies lower to the ground in a spiral to surround the prey; the stooping and diving stage, in which the red-tailed hawk moves to the best hunting position within a selected area around the prey, and attacks the prey from the best position during the low-flying stage.

6. The aerosol particle spectrum distribution inversion method based on intelligent swarm algorithm according to claim 1 is characterized by: In step S5, the step of calculating the minimum error to obtain the aerosol particle spectrum distribution includes: i , 、r mi Substitute it into the aerosol particle spectrum function, and use the relationship formula between the aerosol particle spectrum and the aerosol optical thickness to calculate the aerosol optical thickness; use the calculated aerosol optical thickness and the true value as the error, and under the optimization of the RTH algorithm, obtain the N corresponding to the minimum error value. i , 、r mi To determine the spectral distribution, the actual atmospheric aerosol particle spectral distribution can be obtained.

Images