Aerosol particle size distribution sensor and aerosol particle size distribution measurement method

Abstract

The present invention belongs to the field of particle measurement and sensing, specifically to an aerosol particle size distribution sensor and a method for measuring aerosol particle size distribution. The sensor comprises a particle holding chamber, a spectral sensor chip, and multiple white light LED light sources at different scattering angles. Each scattering angle is the angle at which the light source's beam propagation direction rotates from the particle holding chamber center to the spectral sensor chip. The multiple light sources are used to sequentially emit white light into the particle holding chamber, and the spectral sensor chip is used to collect scattered light generated by the particles to be measured in the particle holding chamber based on the white light, obtaining an aerosol scattering spectrum signal to analyze the aerosol particle size distribution. The scattering angle combination formed by each scattering angle is optimized to minimize the deviation in aerosol parameters between different aerosols, with the difference in aerosol scattering spectrum signal being less than a threshold. The present invention achieves accurate measurement of aerosol particle size distribution using a sensor with a simple structure.

Description

Technical Field

[0001] The present invention belongs to the field of particle measurement sensing, and more specifically, relates to an aerosol particle size distribution sensor and an aerosol particle size distribution measurement method. Background Art

[0002] Aerosols are solid or liquid particles suspended in the air, such as smoke, dust, mist, pollen, and microorganisms. These particles can have a significant impact on the atmospheric environment, human health, and industrial production. Specific applications include monitoring atmospheric aerosol parameters for PM10 (mass concentrations of particles less than 10μm), PM2.5 (mass concentrations of particles less than 2.5μm), and PM1 (mass concentrations of particles less than 1μm). These aerosols are closely related to human health. PM10 particles can directly enter the human chest through the throat, PM2.5 particles can reach the bronchi, and PM1 particles are absorbed by the alveoli. These particles can enter the human body and cause cardiovascular, cerebrovascular, and respiratory diseases. According to relevant reports, 70% of industrial production is related to aerosols, and the global aerosol monitoring market is expected to increase from US$1.3 billion in 2022 to US$2.2 billion in 2027.

[0003] Common methods for measuring aerosol particle size distribution include differential mobility, aerodynamics, and laser diffraction. These technologies require complex and bulky instruments, primarily used in laboratory settings. These instruments, such as aerodynamic particle spectrometers, differential mobility particle spectrometers, and laser particle size analyzers, are used in increasingly diverse applications. Passive monitoring needs, such as fire smoke, atmospheric particulate matter, and particle concentrations in semiconductor production workshops, require multiple devices to operate online in real time for extended periods. Large laboratory instruments are subject to high installation and maintenance costs, while small, portable, and low-cost sensors can effectively meet the needs of distributed monitoring.

[0004] Chinese invention patent specification CN216247609U discloses a method for measuring the particle size distribution of a multi-wavelength particle system. This method utilizes the scattered light intensity of particles at three different wavelengths to achieve real-time measurement of particle size. Chinese patent publication number CN220751904U uses a Mini / Micro-LED multi-wavelength array light source and a three-color CCD or CMOS camera chip to detect the scattering spectrum of smoke and implement a smoke fire alarm using a threshold method. This patent installs LED lights of different wavelengths at different positions on the detector structure and controls the LED lights to emit light at different times. The CCD or CMOS chip collects the scattered light signals, thereby obtaining scattered light signals at multiple wavelengths. Although this patent achieves the collection of scattered light intensity data at multiple wavelengths, it does not discuss the relationship between smoke characteristics and scattered light intensity. Furthermore, the multiple LED array design is complex and difficult to miniaturize. Summary of the Invention

[0005] In response to the defects of the existing technology and the need for improvement, the present invention provides an aerosol particle size distribution sensor and an aerosol particle size distribution measurement method, the purpose of which is to achieve accurate measurement of aerosol particle size distribution through a sensor with a simple structure.

[0006] To achieve the above object, according to one aspect of the present invention, an aerosol particle size distribution sensor is provided, comprising: a particle receiving cavity, a spectral sensing chip, and a plurality of white light LED light sources located at different scattering angles;

[0007] Among them, each scattering angle is the rotation angle of the light beam propagation direction of the white light LED light source at the particle holding cavity to the direction from the center of the particle holding cavity to the spectral sensor chip; multiple white light LED light sources are used to emit white light into the particle holding cavity in sequence, and the spectral sensor chip is used to collect the scattered light generated by the particles to be tested located in the particle holding cavity based on the white light, to obtain an aerosol scattering spectrum signal to analyze the aerosol particle size distribution; the scattering angle combination formed by each scattering angle is optimized with the goal of minimizing the aerosol parameter deviation between different aerosols whose aerosol scattering spectrum signal difference is less than a threshold.

[0008] Furthermore, the scattering angle combination is optimized in the following manner:

[0009] Determine the number of aerosol species contained in the industrial scene to which the particles to be measured belong; preset multiple candidate scattering angle combinations;

[0010] Calculating the relative deviation between the aerosol scattering spectrum signals corresponding to each two aerosols for each scattering angle combination; using each aerosol as a reference aerosol, determining the relative deviation less than a preset threshold among all relative deviations corresponding to the reference aerosol, and screening out the other aerosols corresponding to all relative deviations less than the preset threshold that are paired with the reference aerosol to form an indistinguishable aerosol set of the reference aerosol; calculating the average value of the deviation between any parameter of each reference aerosol and the corresponding parameter of each indistinguishable aerosol in the indistinguishable aerosol set of the reference aerosol, and adding the average values ​​corresponding to each reference aerosol as the value deviation of the parameter between the reference aerosol and the indistinguishable aerosol for the scattering angle combination, thereby obtaining the value deviation of the various parameters for the scattering angle combination;

[0011] The value deviations of various parameters under each scattering angle combination are synthesized to obtain the parameter deviation corresponding to the scattering angle combination; and the scattering angle combination with the smallest parameter deviation is selected as the final scattering angle combination.

[0012] Furthermore, the relative deviation between the aerosol scattering spectrum signals is calculated by the following calculation method:

[0013]

[0014] Where, is the scattering angle combination [θ1,θ2,…,θ n ]The aerosol scattering spectrum signal of aerosol y under j represents the number of wavelengths after the wavelength discretization of the aerosol spectrum signal, and y represents aerosols A and A′; represents the average relative deviation between the aerosol scattering spectral signals of aerosols A and A′, p A_k,g and p A′_k,g They represent the light intensity values ​​of two aerosols A and A′ at the same wavelength k and the same scattering angle g.

[0015] Furthermore, the deviation SME of each parameter x under each candidate scattering angle combination is calculated as follows:

[0016]

[0017] Where θ1, θ2,…, θ n represents the scattering angle combination, K represents the number of aerosol species, x i represents the value of the parameter x of the current reference aerosol i, N represents the number of indistinguishable aerosols in the corresponding indistinguishable aerosol set when various aerosols are used as reference aerosols, x j ′ represents the value of parameter x of indistinguishable aerosol j in the indistinguishable aerosol set corresponding to the current reference aerosol i, μ, σ, and m represent the median particle size and standard deviation of the aerosol particle size distribution and the refractive index of the aerosol, respectively.

[0018] Furthermore, the refractive index m includes a real part of the refractive index and an imaginary part of the refractive index.

[0019] Furthermore, the system further includes: a processor, wherein the processor is configured to:

[0020] According to the discrete acquisition wavelength of the spectral sensor chip, the preset discrete particle size value of the particle to be measured and the initially determined refractive index m, the scattered light matrix Q is obtained based on the Mie scattering theory. [λ,d,m] ;

[0021] According to the scattered light matrix Q [λ,d,m] and the initially determined particle size distribution vector F [μ,σ] , the formula P [λ] =Q [λ,d,m] ×F [μ,σ] The calculated theoretical aerosol scattering spectrum signal P[λ] Compare with the aerosol scattering spectrum signal collected by the sensor to update the refractive index m and particle size distribution vector F [μ,σ] , continue to iterate until the theoretical aerosol scattering spectrum signal P [λ] The deviation from the aerosol scattering spectrum signal collected by the sensor meets the threshold requirement;

[0022] According to the theoretical aerosol scattering spectrum signal P [λ] The aerosol parameters [μ, σ, m] of the particles to be measured are obtained by analysis, thereby obtaining the aerosol particle size distribution.

[0023] Furthermore, the processor uses an ant colony algorithm of swarm intelligence to perform the analysis of the aerosol particle size distribution, specifically: the processor uses an ant colony algorithm of swarm intelligence to perform the analysis of the aerosol particle size distribution, specifically: after executing the first iterative update, select L ants from ME ants as learning ants, M is the number of ants in the ant colony corresponding to each update iteration, and E is the number of elite ants in M; L is an integer multiple of E, and the initial value of each parameter of each learning ant is the value of an elite ant in the elite ant set of the first iteration. The parameter values ​​of each learning ant correspond to different elite ants, and the parameters of each learning ant start from the second iteration, and each time according to the learning speed v l =[x1(t)-x l (t)]·rand·c, randomly approaching the elite ant with the smallest F in the elite ant set in the previous (t) iteration; where x1(t) is the value of the parameter x of the elite ant with the smallest F in the t iteration, c is the preset update parameter, x l (t) is the value of the learning ant l in t iterations, rand is a random number between 0 and 1, and the parameter x of the updated learning ant l is x′ l (t+1)=x l (t)+v l , F is the theoretical aerosol scattering spectrum signal P of each ant [λ] The error between the aerosol scattering spectrum signal collected by the sensor; in each iteration starting from the second iteration, the ants in the ant colony except the elite ants and the learning ants are updated in the same way as the elite ants.

[0024] Furthermore, the white light LED light source is a white light LED lamp bead.

[0025] The present invention also provides an aerosol particle size distribution measurement method, which uses the aerosol particle size distribution sensor as described above to achieve aerosol particle size distribution measurement.

[0026] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects:

[0027] (1) The present invention proposes an aerosol particle size distribution sensor, which obtains aerosol scattering spectrum signals by using a spectral sensing chip and a white light LED light source. Since the spectral signals collected by the spectral sensing chip are scattering spectrum signals of discrete wavelengths, and the white light LED light source contains a wide continuous visible light spectrum, the wavelength range of the measured scattering spectrum signal is wide, and the multi-dimensional scattering spectrum signal contains rich aerosol characteristics, with the help of these characteristics, the particle size distribution parameters can be accurately sensed. In addition, the present invention adopts multiple white light LED light sources, each of which is located at a different scattering angle. It is a sensor with a multi-scattering angle structure, thereby integrating the different spectral characteristics at different scattering angles, further enriching the aerosol characteristics, and realizing accurate sensing of aerosol particle size distribution. The sensor of the present invention has a simple structure, and uses a relatively small number of scattering angles to obtain multi-dimensional scattering spectrum signals to achieve precise aerosol particle size distribution measurement. It has the advantages of simple structure and accurate measurement.

[0028] (2) The processor in the sensor of the present invention adopts a new ant colony algorithm to perform aerosol particle size analysis. Specifically, the learning ants will search for the best solution near the elite ants, that is, each learning ant will search for the best update path in its surroundings, thereby avoiding the rapid convergence of the ant colony and adding diversity to the entire ant colony, enabling the ant colony to effectively escape the local optimal solution trap and further converge to the global optimal solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 A schematic diagram of an aerosol particle size distribution sensor provided by an embodiment of the present invention;

[0030] Figure 2 A schematic diagram of the spectrum discretization principle provided by an embodiment of the present invention;

[0031] Figure 3 The scattering spectra under different aerosol parameters at a scattering angle of 45° provided in the embodiment of the present invention;

[0032] Figure 4 Result graphs of median particle size μ and refractive index m derived from scattering spectrum data under ±5% and ±10% noise using the ant colony algorithm provided in an embodiment of the present invention; wherein A is a result graph of median particle size μ derived from scattering spectrum data under ±5% noise, B is a result graph of refractive index m derived from scattering spectrum data under ±5% noise, C is a result graph of median particle size μ derived from scattering spectrum data under ±10% noise, and D is a result graph of refractive index m derived from scattering spectrum data under ±10% noise;

[0033] Figure 5 A diagram showing the difference between the scattering spectrum signals of the same aerosol at different scattering angles provided by an embodiment of the present invention;

[0034] Figure 6 SME distribution results for different aerosol parameters under dual scattering angle combinations provided by embodiments of the present invention; wherein A is the SME distribution diagram of the median particle size distribution of the aerosol under dual scattering angle combinations, B is the SME distribution diagram of the standard deviation of the aerosol particle size distribution under dual scattering angle combinations, C is the SME distribution diagram of the real part of the aerosol refractive index under dual scattering angle combinations, and D is the SME distribution diagram of the imaginary part of the aerosol refractive index under dual scattering angle combinations;

[0035] Figure 7 The distribution diagram of SME' under the double scattering angle combination structure provided by the embodiment of the present invention;

[0036] Figure 8 A structural design diagram of an aerosol particle size distribution sensor provided in an embodiment of the present invention;

[0037] Figure 9 A graph showing the aerosol scattering spectrum signal results of smoldering smoke measured by the aerosol particle size distribution sensor provided by an embodiment of the present invention;

[0038] Figure 10 A graph showing the aerosol scattering spectrum signal results of open fire smoke measured by the aerosol particle size distribution sensor provided in an embodiment of the present invention;

[0039] Figure 11 Based on Figure 9 and 10 Comparison chart of the aerosol particle size distribution analyzed by the results and the measurement results of the reference aerosol particle spectrometer. DETAILED DESCRIPTION

[0040] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0041] Example 1

[0042] An aerosol particle size distribution sensor, such as Figure 1 As shown, it includes: a particle holding cavity, a spectral sensing chip, and multiple white light LED light sources located at different scattering angles;

[0043] Among them, each scattering angle is the rotation angle of the light beam propagation direction of the white light LED light source at the particle holding cavity to the direction from the center of the particle holding cavity to the spectral sensor chip; multiple white light LED light sources are used to emit white light into the particle holding cavity in sequence, and the spectral sensor chip is used to collect the scattered light generated by the particles to be tested located in the particle holding cavity based on the white light, to obtain an aerosol scattering spectrum signal to analyze the aerosol particle size distribution; the scattering angle combination formed by each scattering angle is optimized with the goal of minimizing the aerosol parameter deviation between different aerosols whose aerosol scattering spectrum signal difference is less than a threshold.

[0044] This embodiment uses a spectral sensor chip and a white light LED light source to obtain an aerosol scattering spectrum signal. Since the spectral signal collected by the spectral sensor chip is a scattering spectrum signal of discrete wavelengths, and the white light LED light source contains a wide continuous visible light spectrum, the wavelength range of the measured scattering spectrum signal is wide, and the multi-dimensional scattering spectrum signal contains rich aerosol characteristics, with the help of these characteristics, the particle size distribution parameters can be accurately sensed. In addition, this embodiment uses multiple white light LED light sources, each of which is located at a different scattering angle. It is a sensor with a multi-scattering angle structure, which integrates the different spectral characteristics at different scattering angles, further enriches the aerosol characteristics, and realizes accurate sensing of aerosol particle size distribution. The sensor of this embodiment has a simple structure and uses fewer scattering angles to collect multiple scattering spectrum signals, thereby realizing accurate aerosol particle size distribution measurement. It has the advantages of simple structure and accurate measurement.

[0045] As a preferred embodiment, the white light LED light source is a white light LED lamp bead. In practice, white light LED lamp beads have relatively high power and generate strong scattered light signals, which can achieve the sensitivity of the spectral sensor chip at an aerosol concentration of 0.2dB / m.

[0046] As a preferred embodiment, the above scattering angle combination is optimized by the following method:

[0047] Determine the number of aerosol types contained in the industrial scene to which the particles to be measured belong (each aerosol parameter value combination corresponds to an aerosol); preset multiple candidate scattering angle combinations;

[0048] Calculating the relative deviation between the aerosol scattering spectrum signals corresponding to each two aerosols for each scattering angle combination; using each aerosol as a reference aerosol, determining the relative deviation less than a preset threshold among all relative deviations corresponding to the reference aerosol, and screening out the other aerosols corresponding to all relative deviations less than the preset threshold that are paired with the reference aerosol to form an indistinguishable aerosol set of the reference aerosol; calculating the average value of the deviation between any parameter of each reference aerosol and the corresponding parameter of each indistinguishable aerosol in the indistinguishable aerosol set of the reference aerosol, and adding the average values ​​corresponding to each reference aerosol as the value deviation of the parameter between the reference aerosol and the indistinguishable aerosol for the scattering angle combination, thereby obtaining the value deviation of the various parameters for the scattering angle combination;

[0049] The value deviations of various parameters under each scattering angle combination are synthesized to obtain the parameter deviation corresponding to the scattering angle combination; and the scattering angle combination with the smallest parameter deviation is selected as the final scattering angle combination.

[0050] This embodiment aims to measure aerosol parameter values ​​and thereby obtain aerosol particle size distribution. To ensure more accurate parameter measurement, it is considered that under any scattering angle combination, the scattering spectral signals of different aerosols may be similar, or that the measured scattering spectral signals may deviate from the actual values ​​due to the presence of noise and other influences in actual measurements. To address this problem, this embodiment proposes to optimize a scattering angle combination so that even if the scattering spectral signals deviate, the analyzed parameters are closest to the actual parameter values.

[0051] Therefore, in a specific implementation, this preferred embodiment first selects any aerosol from all aerosols as a reference aerosol, then calculates the relative deviation of the scattering spectrum under the same scattering angle combination between the reference aerosol and all aerosols, compares the relative deviation with a preset deviation threshold, and screens aerosols with a relative deviation less than the threshold to form a set. This set is all aerosols that are difficult to distinguish from the reference aerosol in the scattering spectrum, and is defined as an indistinguishable aerosol set.

[0052] As mentioned above, the aerosol parameters in the indistinguishable aerosol set are the possible values ​​of the theoretically analyzed reference aerosol under the scattering spectrum. That is, in an actual environment, when the object to be measured is the above-mentioned reference aerosol, the scattering spectrum obtained not only characterizes the reference aerosol parameters, but also contains the values ​​of all aerosol parameters in the indistinguishable aerosol set in the case of misjudgment. The deviation is calculated between the same aerosol parameter of the reference aerosol and the same parameter of all aerosols in the indistinguishable aerosol set, and the above deviations are summed and averaged. The average deviation of all reference aerosols is summed and divided by the total number of aerosols to obtain the deviation of the aerosol parameters measured by the scattering spectrum under the scattering angle combination. The scattering angle combination with the smallest parameter deviation is selected from all candidate scattering angle combinations with the best measurement performance.

[0053] As a preferred embodiment, the relative deviation between aerosol scattering spectrum signals is calculated by the following calculation method:

[0054]

[0055] Where, is the scattering angle combination [θ1,θ2,…,θ n ]The aerosol scattering spectrum signal of aerosol y under j represents the number of wavelengths after the wavelength discretization of the aerosol spectrum signal, and y represents aerosols A and A′; represents the average relative deviation between the aerosol scattering spectral signals of aerosols A and A′, p A_k and p A′_k They represent the light intensity values ​​of two aerosols A and A′ at the same wavelength k and the same scattering angle g.

[0056] As a preferred embodiment, the deviation SME of each parameter x under each candidate scattering angle combination is calculated by the following method:

[0057]

[0058] Where θ1, θ2,…, θ n represents the scattering angle combination, K represents the number of aerosol species, x i represents the value of the parameter x of the current reference aerosol i, N represents the number of indistinguishable aerosols in the corresponding indistinguishable aerosol set when various aerosols are used as reference aerosols, x j′ represents the value of parameter x for indistinguishable aerosol j in the set of indistinguishable aerosols corresponding to the current reference aerosol i. μ, σ, and m represent the median diameter and standard deviation of the aerosol particle size distribution, as well as the aerosol's refractive index, respectively. Preferably, the refractive index m comprises a real part and an imaginary part. The real part represents the degree of change in the propagation velocity of light within the aerosol, while the imaginary part represents the aerosol's absorption of light energy. These parts have different physical meanings. Therefore, this embodiment further modifies the refractive index m into a real part and an imaginary part.

[0059] When performing scattering angle combination optimization, the preset threshold ΔM of the relative deviation between aerosol scattering spectrum signals is, for example, 0.1. In addition, the set of indistinguishable aerosols A′ of the reference aerosol A can be expressed as:

[0060] Since the imaginary part of the refractive index mI has a value of 0, in order to measure the deviation between aerosol parameters and optimize the performance of the sensor, this embodiment calculates the value deviation between aerosol parameters through the Symmetric Mean Absolute Percentage Error (SME).

[0061] As a preferred embodiment, the present invention further includes: a processor: wherein the processor is configured to:

[0062] According to the discrete acquisition wavelength of the above-mentioned spectral sensor chip, the preset discrete particle size value of the particles to be measured and the initially determined refractive index m, the scattered light matrix Q is obtained based on the Mie scattering theory. [λ,d,m] ;

[0063] According to the scattered light matrix Q [λ,d,m] and the initially determined particle size distribution vector F [μ,σ] , the formula P [λ] =Q [λ,d,m] ×F [μ,σ] The calculated theoretical aerosol scattering spectrum signal P [λ] Compare with the aerosol scattering spectrum signal collected by the sensor to update the refractive index m and particle size distribution vector F [μ,σ] , continue to iterate until the theoretical aerosol scattering spectrum signal P [λ] The deviation from the aerosol scattering spectrum signal collected by the sensor meets the threshold requirement;

[0064] According to the theoretical aerosol scattering spectrum signal P [λ] The aerosol parameters [μ, σ, m] of the particles to be measured are obtained by analysis, thereby obtaining the aerosol particle size distribution.

[0065] According to Mie scattering theory, aerosols are composed of particles of different sizes, with a particle size distribution parameter in the particle size dimension. Usually, the secondary scattering of particles can be ignored. Under this condition, the scattered light signal of the aerosol can be regarded as the sum of the scattered light of all particles. The aerosol scattering spectrum can be expressed as:

[0066]

[0067] Where d represents the particle size, d min d max are the lower and upper limits of particle size, p(λ) is the scattering spectrum function, λ is the wavelength of scattered light, f(d) is the particle size distribution of aerosol, is the single-particle scattering spectrum obtained based on Mie scattering theory.

[0068] It can be seen from formula (1) that the aerosol scattering spectrum is the result of the superposition of single-particle scattering spectra of different particle sizes. Since the single-particle scattering spectrum is the result of the joint coupling of aerosol parameters and optical parameters, the aerosol scattering spectrum is also affected by aerosol parameters and optical parameters.

[0069] Since the lognormal distribution has a high degree of consistency with the smoke particle size distribution, the aerosol particle size distribution is set to the lognormal distribution in this embodiment, which is consistent with the actual aerosol distribution. The calculation formula is:

[0070]

[0071] Where μ is the median particle size and σ is the standard deviation.

[0072] Setting the particle size distribution to a log-normal distribution can simplify the aerosol scattering spectrum model, making the scattering spectrum a function related to [μ, σ, m], which is defined as:

[0073]

[0074] Since the actual spectral data obtained is discrete, equation (3) can be rewritten as:

[0075] P [λ] =Q [λ,d,m] ×F [μ,σ] (4)

[0076] Among them, P [λ] is j wavelengths The scattering spectrum signal is Q [λ,d,m] is j wavelengths and g particle size The j×g-order single-particle scattered light matrix composed of F [μ,σ]is the preset particle size g The particle size distribution vector after discretization is specifically

[0077] from Figure 2 It can be seen that the discretized aerosol scattering spectrum is a discrete single particle scattering spectrum. With discrete particle size distribution Perform relevant transformation operations on the particle size dimension. Figure 3 The scattering spectra under different aerosol parameters have significant differences, which is beneficial to the sensing of aerosol parameters [μ, σ, m]. Figure 3 is the aerosol scattering spectrum with a scattering angle of 45° ("spectrum" is continuous, "spectral signal" is discrete), sample 1 is the aerosol scattering spectrum with a median particle size μ of 150 nm, variance σ of 1.3, and a refractive index m of 1.30+0i; sample 2 is the aerosol scattering spectrum with a median particle size μ of 150 nm, variance σ of 1.3, and a refractive index m of 1.55+0.5i; sample 3 is the aerosol scattering spectrum with a median particle size μ of 550 nm, variance 1.3, and a refractive index m of 1.3+0i; sample 4 is the aerosol scattering spectrum with a median particle size μ of 150 nm, variance σ of 1.7, and a refractive index m of 1.30+0i.

[0078] To accurately decouple the scattering spectrum P [λ] Aerosol particle size distribution parameter F [μ,σ] In this embodiment, the ant colony algorithm of swarm intelligence is preferably used, which relies on the mutual relationship between individuals in the group and can effectively find the correct optimization direction. It has the characteristics of high stability and strong robustness. [λ] As input, combined with the scattered light matrix Q [λ,d,m] and the particle size distribution vector F [μ,σ] , continue to iterate according to formula (4), and finally converge to P [λ] The optimal solution of the corresponding aerosol parameters [μ, σ, m]. Figure 4 The median particle size μ and refractive index m obtained by the ant colony algorithm from the scattering spectrum data with ±5% and ±10% noise are shown. It can be seen that the ant colony algorithm has excellent aerosol parameter decoupling ability.

[0079] As a preferred embodiment, the processor uses an ant colony algorithm of swarm intelligence to perform the analysis of the aerosol particle size distribution, specifically:

[0080] The median particle size μ and variance σ of the aerosol particle size distribution of M ants in the ant colony are randomly set, as well as the aerosol refractive index m. The above ants represent a set of aerosols. The discrete aerosol scattering spectrum formula P is used. [λ] =Q [λ,d,m] ×F[μ,σ] Calculate the aerosol scattering spectrum signal P of M ants in the ant colony [λ] The deviation is calculated by comparing the aerosol scattering spectrum signal measured by the aerosol particle size distribution sensor. where p k,g is the light intensity value of the scattered spectrum signal measured by the aerosol sensor at wavelength k and scattering angle g, p k,g ′ is the aerosol scattering spectrum signal calculated by the ant according to the discrete aerosol scattering spectrum formula. For the entire ant colony, there are M preset ants, so there are also M calculated deviations F. The M deviations F are arranged from small to large, and the top E ants are selected as elite ants to form an elite ant set. Under the M set of aerosol parameter [μ, σ, m] values, the values ​​of the elite ants are closer to the scattering spectrum signal of the sensor. The standard deviation of each elite ant in the elite ant set between the value of each parameter and the value of the same parameter of other elite ants is calculated. x h (t) is the value of the elite ant h on the parameter x, o (t) is the value of the same parameter x taken by other ants in the elite ant set except the elite ant h, t is the number of iterations, and β is the preset weight parameter to facilitate the subsequent adjustment of the width of the normal distribution. In addition, the parameter values ​​x calculated by each elite ant are h (t) and standard deviation s h (t) is used as the parameter of the normal distribution, and the normal distribution of each parameter of each elite ant is calculated. That is, the probability density function of the parameter value under the elite ant is used to measure the relationship between each elite ant and the entire elite ant set. Finally, the normal distribution of each parameter of each elite ant is calculated according to the preset weights to calculate W={w1,w2,…,w E} Perform weighted summation to obtain the normal distribution used to describe the elite ant set The next generation (t+1) updates the parameter values ​​of ants other than the elite ants of the current generation. Specifically, it is based on the normal distribution of the entire elite ant set. The parameter values ​​of ME ants in the current generation, excluding the elite ants of the current generation, are updated in a random manner using a roulette wheel. However, since the update method using a normal distribution converges faster and is closely related to the number of elite ants E, when E is too small, it is easy to converge to incorrect aerosol parameter values. Therefore, this embodiment selects L ants from ME ants as learning ants, where L is an integer multiple of E. The parameter values ​​of each learning ant in the second iteration are the values ​​of an elite ant in the elite ant set of the first iteration. The parameter values ​​of each learning ant correspond to different elite ants, and they will learn at a certain learning speed v. l =[x1(t)-xl (t)]·rand·c randomly approaches the elite ant with the smallest F in the elite ant set in the previous iteration (t), where x1(t) is the value of the parameter x of the elite ant with the smallest F in the t iteration, c is the preset update parameter, and x l (t) is the value of the learning ant l in t iterations, rand is a random number between 0 and 1, and the parameter x of the updated learning ant l is x′ l (t+1)=x l (t)+v l This makes the ant colony search for possible parameter values ​​around the elite ants, increasing their chances of escaping from incorrect aerosol parameter values. Therefore, in each iteration, the elite ants will update the aerosol parameter values ​​of the MEL ants other than the elite ants and the learning ants, while the learning ants will update their own values.

[0081] For the above formula (4), the scattering angle θ0 is an important parameter for the calculation of the scattering spectrum. It changes the aerosol characteristics by affecting the scattering spectrum characteristics, resulting in spatial differences in the scattering spectrum. Figure 5 As shown in the figure, the normalized aerosol scattering spectrum distribution with a refractive index of 1.45+0i, a median particle size of 250nm, and a variance of 1.2 at different scattering angles has significant differences in the scattering spectrum characteristics under the same aerosol parameters. The prominent characteristic differences have richer aerosol characteristics, that is, they can characterize more diverse and complex aerosol parameters. Therefore, in order to pursue the outstanding sensing performance of the sensor, the sensor structure will be optimized through different scattering angle combinations. The aerosols in Table 1 are common aerosol samples in fire smoke detection scenarios, with standard references and repeated tests, and are the main test objects of this embodiment. The typical aerosol parameters covered in Table 1 have differences in median particle size, variance and refractive index due to different components and generation methods. These differences will be further reflected in the scattering spectrum, so that the scattering spectrum can sense aerosol parameters.

[0082] Table 1 Typical aerosol parameters

[0083]

[0084] When optimizing the scattering angle combination, the scattering angle combination optimization problem can be expressed as:

[0085]

[0086] Since this embodiment focuses on the simple structure of the sensor but highlights the performance, for the fire smoke detection scenario, the optimized structure of this embodiment is a double scattering angle, that is, a combination of θ1 and θ2. The candidate scattering angle range is 40°-140°, with a value of every 5°. The SME distribution obtained by solving is as follows: Figure 6 As shown, Figure 5 The minimum relative deviation of μ is about 27.14%, and the best scattering angle combination is 100° and 140°; the minimum relative deviation of σ is 11.00%, and the best scattering angle combination is 40° and 140°; the minimum absolute deviation of mR is 0.045%, and the best scattering angle combination is 40° and 105°; the minimum absolute deviation of mI is 1.24%, and the best scattering angle combination is 40° and 95°.

[0087] Since the characteristics of different aerosol parameters on the scattering spectrum are different, the optimization results of different parameters are also different. In order to comprehensively evaluate and optimize the aerosol parameter characteristics, this embodiment preferably averages the optimization results of the four parameters, which are described in detail as follows:

[0088]

[0089] Among them, SME μ SME σ SME mR and SMEs mI are the deviations of aerosol parameters μ, σ, mR and mI respectively. The final result of averaging the deviations of all typical aerosol parameters is SME′. Figure 7 The best structure is the combination of scattering angles 70° and 140°, and the average SME value of all parameter deviations is 10.85%. μ The deviation is 28.12%, with the smallest SME μ The difference is about 0.98%; the structure SME σ 13.89%, with the smallest SME σ The difference is about 2.89%; the structure SME mR 0.050%, with the smallest SME mR The difference is about 0.005%; the structure SME mI 1.33%, with the smallest SME mI The difference is about 0.09%. It can be seen that the difference between the combined structure of scattering angles 70° and 140° and the optimal characterization structure of each parameter is small, and can achieve comprehensive characterization of aerosol multi-parameters.

[0090] Figure 8 This is a prototype design of the scattering spectrum principle, including a sensor structure diagram and a sensor optical structure diagram. It should be noted that although the preferred scattering angle combination is 70° and 140°, in order to facilitate the installation of LED lamp beads, this embodiment selects a suboptimal scattering angle combination of 45° and 140°. The scattering angle difference in this suboptimal structure is large, so there is no need to increase the volume of the entire sensor. In addition, this structure has good aerosol particle size distribution parameter characteristics, which can ensure the accuracy of aerosol parameter inversion. Figure 6 and Figure 7 It can be seen that the SME′ of the 45° and 140° combination structures is 11.72%, SME μ 33.47%, SME σ 11.62%, SME mR 0.13%, SME mI The SME' of the 45° and 140° structures differs by approximately 0.87% compared to the 70° and 140° scattering angle combinations. Furthermore, since the light from LEDs is divergent, it generates a lot of stray light, which increases the background value and affects the sensor's response range and measurement accuracy. The 45° and 140° structures can reduce the interference of stray light.

[0091] In addition, to enhance the sensitivity and stability of the sensor, e.g. Figure 8 As shown, this embodiment adopts the following measures in structural design:

[0092] (1) To increase the sensitivity and anti-interference capability of the prototype, a multi-lens and aperture combination was used to focus the LED light in the prototype's light channel design. Similarly, a 3mm diameter aperture was also set at the front end of the AS7341 spectral sensor chip to ensure that the AS7341 receives less scattered light.

[0093] (2) In the structural design, the intersection point of light emission and light reception is set to about 7 mm, which not only ensures that there is no interference between the backward light emission channel and the receiving channel, but also ensures that the sensor has strong sensitivity;

[0094] (3) A lens with a focal length of 4mm is installed in front of the AS7341 sensor to increase the photosensitive surface of the AS7341 spectral sensor chip and thus increase sensitivity;

[0095] (4) To ensure constant illumination of the LED lamp, a heat sink and fan are installed behind the LED lamp and above the PCB board for heat dissipation;

[0096] During the measurement process, the specific working process of the sensor is as follows:

[0097] (1) The LED lights at 45° and 140° work alternately;

[0098] (2) The light emitted by the LED lamp produces scattered light at the intersection point, which is received by the spectrum sensor chip to obtain spectral data, such as Figure 9 and Figure 10 shown.

[0099] Finally, the ant colony algorithm was combined to decouple the aerosol particle size distribution parameters, and the measurement results of the scanning mobility particle sizer SMPS (Scanning Mobility Particle Sizer) 3936 of TSI Company in the United States were used as a reference. Figure 11 As shown), it can be seen that the present invention shows good performance in actual measurement.

[0100] Example 2

[0101] The present invention also provides an aerosol particle size distribution measurement method, which uses the aerosol particle size distribution sensor as described above to achieve aerosol particle size distribution measurement.

[0102] The relevant technical solutions are the same as those in Example 1 and will not be described again here.

[0103] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An aerosol particle size distribution sensor, characterized in that: include: A particle holding cavity, a spectral sensing chip, and multiple white light LED light sources at different scattering angles; Each scattering angle is the rotation angle of the light beam propagation direction of the white light LED light source at the particle holding cavity to the direction from the center of the particle holding cavity to the spectral sensor chip; multiple white light LED light sources are used to sequentially emit white light into the particle holding cavity, and the spectral sensor chip is used to collect scattered light generated by the particles to be tested located in the particle holding cavity based on the white light, to obtain an aerosol scattering spectrum signal to analyze the aerosol particle size distribution; the scattering angle combination formed by the scattering angles is optimized with the goal of minimizing the aerosol parameter deviation between different aerosols whose aerosol scattering spectrum signal difference is less than a threshold value; The scattering angle combination is optimized in the following way: Determine the number of aerosol species contained in the industrial scene to which the particles to be measured belong; preset multiple candidate scattering angle combinations; Calculating the relative deviation between the aerosol scattering spectrum signals corresponding to each two aerosols for each scattering angle combination; using each aerosol as a reference aerosol, determining the relative deviation less than a preset threshold among all relative deviations corresponding to the reference aerosol, and screening out the other aerosols corresponding to the reference aerosol group with all relative deviations less than the preset threshold to form an indistinguishable aerosol set of the reference aerosol; calculating the average value of the deviation between any parameter of each reference aerosol and the corresponding parameter of each indistinguishable aerosol in the indistinguishable aerosol set of the reference aerosol, adding the average values ​​corresponding to each reference aerosol as the value deviation of the parameter between the reference aerosol and the indistinguishable aerosol for the scattering angle combination, thereby obtaining the value deviation of the various parameters for the scattering angle combination; The value deviations of various parameters under each scattering angle combination are synthesized to obtain the parameter deviation corresponding to the scattering angle combination; and the scattering angle combination with the smallest parameter deviation is selected as the optimized scattering angle combination.

2. The aerosol particle size distribution sensor according to claim 1, characterized in that: The relative deviation between aerosol scattering spectral signals is calculated by the following calculation method: Where, is the scattering angle combination [θ1,θ2,…,θ n ]The aerosol scattering spectrum signal of aerosol y under j represents the number of wavelengths after the wavelength discretization of the aerosol spectrum signal, and y represents aerosols A and A′; represents the average relative deviation between the aerosol scattering spectral signals of aerosols A and A′, p A_k,g and p A′_k,g They represent the light intensity values ​​of two aerosols A and A′ at the same wavelength k and the same scattering angle g.

3. The aerosol particle size distribution sensor according to claim 1, characterized in that: The deviation SME of each parameter x under each candidate scattering angle combination is calculated as follows: Where θ1, θ2,…, θ n represents the scattering angle combination, K represents the number of aerosol species, x i represents the value of the parameter x of the current reference aerosol i, N represents the number of indistinguishable aerosols in the corresponding indistinguishable aerosol set when various aerosols are used as reference aerosols, x j ′ represents the value of parameter x of indistinguishable aerosol j in the indistinguishable aerosol set corresponding to the current reference aerosol i, μ, σ, and m represent the median particle size and standard deviation of the aerosol particle size distribution and the refractive index of the aerosol, respectively.

4. The aerosol particle size distribution sensor according to claim 3, characterized in that: The refractive index m is divided into two parameters: a real part of the refractive index and an imaginary part of the refractive index.

5. The aerosol particle size distribution sensor according to claim 1, characterized in that: Also includes: Processor: wherein the processor is configured to: According to the discrete acquisition wavelength of the spectral sensor chip, the preset discrete particle size value of the particle to be measured and the initially determined refractive index m, the scattered light matrix Q is obtained based on the Mie scattering theory. [λ,d,m] ; According to the scattered light matrix Q [λ,d,m] and the initially determined particle size distribution vector F [μ,σ] , the formula P [λ] =Q [λ,d,m] ×F [μ,σ] The calculated theoretical aerosol scattering spectrum signal P [λ] Compare with the aerosol scattering spectrum signal collected by the sensor to update the refractive index m and particle size distribution vector F [μ,σ] , continue to iterate until the theoretical aerosol scattering spectrum signal P [λ] The deviation from the aerosol scattering spectrum signal collected by the sensor meets the threshold requirement; According to the theoretical aerosol scattering spectrum signal P [λ] The aerosol parameters [μ, σ, m] of the particles to be measured are obtained by analysis, thereby obtaining the aerosol particle size distribution.

6. The aerosol particle size distribution sensor according to claim 5, characterized in that: The processor uses an ant colony algorithm of swarm intelligence to perform the analysis of the aerosol particle size distribution, specifically: after executing the first iterative update, L ants are selected from ME ants as learning ants, M is the number of ants in the ant colony corresponding to each update iteration, and E is the number of elite ants in M; L is an integer multiple of E, and the initial value of each parameter of each learning ant is the value of an elite ant in the elite ant set of the first iteration. The elite ant corresponding to the parameter value of each learning ant is different, and the parameters of each learning ant start from the second iteration, and each time according to the learning speed v l =[x1(t)-x l (t)]·rand·c, randomly approaching the elite ant with the smallest F in the elite ant set in the previous (t) iteration; where x1(t) is the value of the parameter x of the elite ant with the smallest F in the t iteration, c is the preset update parameter, x l (t) is the value of the learning ant l in t iterations, rand is a random number between 0 and 1, and the parameter x of the updated learning ant l is x′ l (t+1)=x l (t)+v l , F is the theoretical aerosol scattering spectrum signal P of each ant [λ] The error between the aerosol scattering spectrum signal collected by the sensor; in each iteration starting from the second iteration, the ants in the ant colony except the elite ants and the learning ants are updated in the same way as the elite ants.

7. The aerosol particle size distribution sensor according to claim 1, characterized in that: The white light LED light source is a white light LED lamp bead.

8. A method for measuring aerosol particle size distribution, characterized in that: An aerosol particle size distribution sensor as claimed in any one of claims 1 to 7 is used to achieve aerosol particle size distribution measurement.

Images