Livestock and poultry breeding house pm2.5 detection and distribution characterization method, system and equipment

Abstract

This invention discloses a method, system, and equipment for PM2.5 detection and distribution characterization in livestock and poultry farms, relating to the field of environmental pollution element technology. The method includes PM2.5-based detection and distribution characterization. 2.5 Spectral data of the sample, PM 2.5 The true values ​​of mass concentration and the true values ​​of mass concentration and relative content of each metal element are used to determine the relationship with PM. 2.5 The carbon fingerprint peak and its corresponding signal intensity, which are highly linearly correlated with the true mass concentration, were used to determine the dynamic correction spectral matrix of the carbon fingerprint peak and the multivariate PM. 2.5 A quantitative analysis model for mass concentration was established; carbon fingerprint peaks were extracted from the above-mentioned calibrated spectral matrix to correct the characteristic peaks of metal elements, thereby determining a quantitative analysis model for the relative content of multiple metal elements; based on the determined model, the PM2.5 concentration at the sampling sites to be analyzed was predicted. 2.5 The invention assesses the mass concentration, mass concentration, and relative content of metallic elements, and characterizes the distribution of metallic elements. This invention improves PM... 2.5 Efficiency, accuracy, and generalization performance of metal element detection.

Description

Technical Field

[0001] This invention relates to the field of environmental pollution element technology, and in particular to a method, system and equipment for detecting and characterizing the distribution of PM2.5 in livestock and poultry farms. Background Technology

[0002] Livestock and poultry farming is an important component of the agricultural economy. With the continuous development of modern farming management techniques and the Internet of Things (IoT) technology, the livestock and poultry farming industry is also showing a rapid development trend. Environmental conditions are an important factor affecting the output efficiency and economic benefits of livestock and poultry farming, and intensive livestock and poultry farming is also one of the major sources of air pollution.

[0003] Particulate matter (PM) has become one of the most concerning air pollutants in recent years. Studies show that PM generated during livestock farming accounts for approximately 50% of agricultural PM emissions across Europe. Inhalable fine particulate matter (PM2.5) 2.5 (Dp≤2.5μm) can reach the lungs of humans and animals through the respiratory system, and even enter the bloodstream, directly affecting the health of the human respiratory and cardiovascular systems, and impacting the health and production performance of animals. 2.5 It is also a carrier of many pollutants, including PM from animals. 2.5 They typically contain heavy metals, various organic odor components, and different bioactive ingredients, which can seriously affect human and animal health. Long-term exposure to high PM2.5 levels in livestock and poultry sheds... 2.5 The environment, where heavy metals accumulate, can be harmful to animals and humans. Therefore, PM2.5 in livestock and poultry farms... 2.5 Rapid detection and distribution characterization methods for heavy metal elements are of great significance for improving air quality in and around livestock sheds, protecting the health of staff and surrounding residents, and achieving welfare-oriented livestock farming and efficient and healthy production.

[0004] Currently, the national standard specifies PM 2.5 Methods for detecting metal elements mainly include atomic absorption spectrometry (AAS), inductively coupled plasma optical emission spectrometry (ICP-OES), and inductively coupled plasma mass spectrometry (ICP-MS). These methods require complex pretreatment such as digestion and may generate environmental pollution byproducts, resulting in drawbacks such as complex operation and long detection cycles. 2.5 Element concentration characteristics exhibit spatiotemporal variability, changing with meteorological conditions. These long-period chemical detection methods have a time lag and cannot meet the requirements for PM2.5 concentration. 2.5 Timeliness testing. Summary of the Invention

[0005] The purpose of this invention is to provide a method, system, and equipment for detecting and characterizing PM2.5 distribution in livestock and poultry farms, in order to solve the above-mentioned problems.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] Firstly, this invention provides a PM for livestock and poultry breeding houses. 2.5 Detection and distribution characterization methods include:

[0008] PM2.5 at different sampling sites inside and outside livestock and poultry houses 2.5 sample;

[0009] Get each PM 2.5 Spectral data of the sample and PM 2.5 True mass concentration and each PM 2.5 The true mass concentration and true relative content of each metal element in the sample;

[0010] Extracting from spectral data related to PM 2.5 The carbon element fingerprint peak and corresponding signal intensity are highly linearly correlated with the true value of mass concentration.

[0011] Based on the extracted carbon fingerprint peaks and their corresponding signal intensities, a dynamic calibration spectral matrix for carbon fingerprint peaks is constructed, and the characteristic peaks of the metal element to be analyzed are extracted from the carbon fingerprint peaks of the metal element based on the dynamic calibration spectral matrix for carbon fingerprint peaks.

[0012] According to each PM 2.5 Spectral data of the sample and PM 2.5 True mass concentration values, establishing multivariate PM 2.5 Mass concentration quantitative analysis model; the multivariate PM 2.5 Mass concentration quantitative analysis model is used to detect PM in the air. 2.5 mass concentration;

[0013] Based on the carbon fingerprint peak of the metal element to be analyzed, the characteristic peaks of the metal element and each PM are corrected. 2.5 The true relative content of each metal element in the sample was used to establish a quantitative analysis model for the relative content of multiple metal elements; this model was then used to detect PM. 2.5 The relative content of each metallic element in the solution;

[0014] Obtain PM at the target sampling site 2.5 Samples, and based on the multivariate PM 2.5 The mass concentration quantitative analysis model and the relative content quantitative analysis model of the multi-metal elements determine the PM at the target sampling site. 2.5 PM of the sample 2.5 Mass concentration and relative content of each metal element, and based on PM at the target sampling site. 2.5 PM of the sample2.5 Estimate PM by mass concentration and relative content of each metal element. 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded; the target sampling site is the sampling site to be analyzed.

[0015] According to PM 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded to characterize the distribution of each metal element in the air of livestock and poultry breeding houses.

[0016] Secondly, this invention provides a PM for livestock and poultry breeding houses. 2.5 Detection and distribution characterization systems, including:

[0017] The sample acquisition module is used to acquire PM at different sampling sites inside and outside livestock and poultry houses. 2.5 sample;

[0018] The spectral data and true value acquisition module is used to acquire the data for each PM. 2.5 Spectral data of the sample and PM 2.5 True mass concentration and each PM 2.5 The true mass concentration and true relative content of each metal element in the sample;

[0019] The carbon fingerprint peak extraction module is used to extract peaks related to PM from spectral data. 2.5 The carbon element fingerprint peak and corresponding signal intensity are highly linearly correlated with the true value of mass concentration.

[0020] The carbon fingerprint peak correction and metal element characteristic peak extraction module is used to construct a dynamic correction spectral matrix of carbon fingerprint peaks based on the extracted carbon fingerprint peaks and corresponding signal intensities, and to extract the carbon fingerprint peak correction and metal element characteristic peaks of the metal element to be analyzed based on the dynamic correction spectral matrix of carbon fingerprint peaks.

[0021] Multivariate PM 2.5 The mass concentration quantitative analysis model building module is used to establish models based on each PM. 2.5 Spectral data of the sample and PM 2.5 True mass concentration values, establishing multivariate PM 2.5 Mass concentration quantitative analysis model; the multivariate PM 2.5 Mass concentration quantitative analysis model is used to detect PM in the air. 2.5 mass concentration;

[0022] The module for establishing quantitative analysis models of relative content of multiple metal elements is used to correct the characteristic peaks of metal elements and each PM based on the carbon fingerprint peak of the metal element to be analyzed. 2.5The true relative content of each metal element in the sample was used to establish a quantitative analysis model for the relative content of multiple metal elements; this model was then used to detect PM. 2.5 The relative content of each metallic element in the solution;

[0023] The estimation module is used to obtain the PM at the target sampling point. 2.5 Samples, and based on the multivariate PM 2.5 The mass concentration quantitative analysis model and the relative content quantitative analysis model of the multi-metal elements determine the PM at the target sampling site. 2.5 PM of the sample 2.5 Mass concentration and relative content of each metal element, and based on PM at the target sampling site. 2.5 PM of the sample 2.5 Estimate PM by mass concentration and relative content of each metal element. 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded; the target sampling site is the sampling site to be analyzed.

[0024] Distribution characterization module, used to characterize PM 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded to characterize the distribution of each metal element in the air of livestock and poultry breeding houses.

[0025] Thirdly, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor runs the computer program to cause the electronic device to perform a PM for a livestock and poultry breeding shed according to the first aspect. 2.5 Detection and distribution characterization methods.

[0026] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0027] This invention provides a method for detecting PM in livestock and poultry houses based on carbon peak dynamic correction spectroscopy. 2.5 A method, system, and equipment for determining the relative content and mass concentration of multiple metallic elements in air, and for visualizing the distribution characteristics of these metallic elements in the air of livestock sheds, wherein the method includes collecting PM2.5 concentration data. 2.5 Spectroscopy, extraction of carbon fingerprint peaks, construction of dynamic calibration spectra of carbon peaks, and establishment of PM 2.5 The mass concentration detection model and the elemental relative content detection model based on carbon peak correction spectrum were finally used to estimate PM2.5 in the air of livestock houses. 2.5 The mass concentration of the metal-containing element is determined, and its distribution characteristics are visualized. This invention can effectively weaken PM from different sources. 2.5 The influence of matrix effects caused by elemental carbon / organic carbon ratio, particle size distribution, and elemental composition on spectral signals enables heterogeneous PM... 2.5Simultaneous detection of multiple elements improves PM efficiency. 2.5 The efficiency, accuracy, and generalization performance of metal element detection provide a reference for further rapid detection and characterization of the content and distribution characteristics of metal elements in regional particulate matter. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 A PM for livestock and poultry breeding houses provided in this embodiment of the invention 2.5 A flowchart illustrating the detection and distribution characterization methods;

[0030] Figure 2 This is a top view of the chicken coop and a distribution diagram of sampling sites provided in an embodiment of the present invention;

[0031] Figure 3 The 12 PMs provided in the embodiments of the present invention 2.5 Distribution map of metal elements inside and outside the chicken house. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0034] Example 1

[0035] The focus of this embodiment is to protect the extraction and PM. 2.5 The carbon fingerprint peak with high linear correlation to mass concentration was identified, and a dynamic correction spectrum of the carbon fingerprint peak was constructed. A multivariate regression model based on the carbon fingerprint peak correction spectrum was established for the simultaneous detection of heterogeneous PM in livestock and poultry houses. 2.5 A method for visualizing and characterizing the distribution characteristics of elements, including their mass concentration, relative content, and mass concentration.

[0036] like Figure 1 As shown, this embodiment provides a PM for livestock and poultry breeding houses.2.5 The detection and distribution characterization method includes the following steps:

[0037] Step 100: Obtain PM at different sampling sites inside and outside the livestock and poultry breeding sheds. 2.5 The samples, specifically, were collected using a quartz membrane as the sampling filter and a medium-flow particulate matter sampler at different sampling sites inside and outside the livestock and poultry sheds to obtain PM2.5 concentrations. 2.5 sample.

[0038] Step 200: Obtain each PM 2.5 Spectral data of the sample and PM 2.5 True mass concentration and each PM 2.5 The true values ​​of the mass concentration and relative content of each metal element in the sample are as follows:

[0039] (1) Laser-induced breakdown spectroscopy (LIBS) was used to determine the PM levels. 2.5 Spectral data of the sample.

[0040] (2) Weigh each sample filter membrane before and after sampling using a 1 / 100,000 balance, determine the mass difference before and after sampling, and determine the mass of each PM based on the mass difference and the volume of sampled air. 2.5 PM of the sample 2.5 True value of mass concentration C PM2.5 (μgm -3 ).

[0041] (3) PM was determined using inductively coupled plasma mass spectrometry. 2.5 The mass of metal elements in the sample, and based on PM 2.5 Calculate the mass and constant volume for each PM 2.5 The true mass concentration C of each metal element in the sample metal (ng m -3 ) and the true relative content C rel (ngμg -1 ).

[0042] Step 300: Extract the data related to PM from the spectral data. 2.5 The carbon fingerprint peak and corresponding signal intensity are highly linearly correlated with the true value of mass concentration.

[0043] Based on the atomic spectroscopy database of the National Institute of Standards and Technology (NIST), multiple carbon fingerprint peaks were extracted from LIBS spectral data, and PM2.5 was fitted based on these carbon fingerprint peaks. 2.5The true mass concentration is used to form a univariate linear regression model, a calibration curve is established, and the coefficient of determination (R²) is selected. 2 The carbon fingerprint peak corresponding to the largest calibration curve is the carbon fingerprint peak with the highest linear correlation, and its wavelength is λ. C n samples (i.e., PM) 2.5 In the sample), the carbon fingerprint peak signal intensity corresponding to the i-th sample is I. λC (i) .

[0044] Step 400: Based on the extracted carbon fingerprint peak and the corresponding signal intensity, construct the dynamic correction spectrum of the carbon fingerprint peak, and extract the characteristic peaks of the corrected metal elements based on the dynamic correction spectrum of the carbon fingerprint peak.

[0045] In n samples, the original LIBS spectral intensities of the m bands corresponding to the i-th sample are I. raw (i) =[I λ1 (i) , ..., I λm (i) ] m The corresponding carbon fingerprint peak signal intensity is I. λC (i) .

[0046] The original LIBS spectral matrix of n samples is X raw =[I raw (1) , ..., I raw (n) ] T n*m The corresponding carbon fingerprint peak matrix is ​​C = [I λC (1) , ..., I λC (n) ] T n .

[0047] The LIBS spectral matrix for carbon peak dynamic correction (CPDC) is as follows:

[0048] I CPDC (i) =I raw (i) / I λC (i) =[I λ1 (i) , ..., I λm (i) ] m / I λC(i) (1);

[0049] X CPDC =[I CPDC (1) ,…,I CPDC (n) ] T n*m (2);

[0050] In the formula, I CPDC (i) To dynamically correct the LIBS spectrum of the carbon fingerprint peak corresponding to the i-th sample, X CPDC The LIBS spectral matrix is ​​dynamically corrected for the carbon fingerprint peaks of n samples.

[0051] Based on the NIST atomic spectroscopy database, the carbon fingerprint peaks of the metal element to be analyzed are extracted from the dynamic calibration spectral matrix of the carbon fingerprint peaks to correct the characteristic peaks of the metal element.

[0052] Step 500: Based on each PM 2.5 Spectral data of the sample and PM 2.5 True mass concentration values, establishing multivariate PM 2.5 Mass concentration quantitative analysis model; the multivariate PM 2.5 Mass concentration quantitative analysis model is used to detect PM in the air. 2.5 Mass concentration.

[0053] By combining random frog (RF) and successive projection algorithm (SPA), the original LIBS spectral lines were extracted from the PM. 2.5 For the characteristic bands related to mass concentration, multiple linear and nonlinear regression models were established using partial least squares (PLS) and extreme learning machine (ELM) methods, respectively, and the results were based on the coefficient of determination (R²). 2 The root mean squared error (RMSE) is used to select the optimal model for detecting PM2.5 in the air. 2.5 Mass concentration.

[0054] Step 600: Establish a multi-element relative content quantitative analysis model based on the dynamic correction of LIBS spectral lines using carbon fingerprint peaks, specifically as follows:

[0055] Univariate linear regression models for the relative abundance of elements were established by dynamically correcting characteristic peaks from the carbon fingerprint peaks of the elements to be analyzed. Characteristic bands highly correlated with each element in the LIBS spectrum were extracted using RF and SPA to dynamically correct the carbon fingerprint peaks, and multivariate linear and nonlinear regression models were established using PLS and ELM, respectively. Based on R... 2 Using RMSE to select the optimal model for rapid PM detection 2.5 The relative content of each element in the sample.

[0056] Step 700: Obtain PM at the target sampling site 2.5 Samples, and based on the multivariate PM 2.5 The mass concentration quantitative analysis model and the relative content quantitative analysis model of the multi-metal elements determine the PM at the target sampling site. 2.5 PM of the sample 2.5 Mass concentration and relative content of each metal element, and based on PM at the target sampling site. 2.5 PM of the sample 2.5 Estimate PM by mass concentration and relative content of each metal element. 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded; the target sampling site is the sampling site to be analyzed, specifically: through PM2.5... 2.5 The detection results from the mass concentration analysis model and the results from the quantitative analysis of relative elemental content will be used to calculate the PM2.5 concentration using LIBS. 2.5 mass concentration C PM2.5-LIBS (μg m -3 The relative content C of metallic element i i-rel-LIBS (ngμg -1 Multiplying them together allows us to estimate PM2.5. 2.5 The mass concentration C of metallic element i in air i-LIBS (ngm -3 ).

[0057] Step 800: According to PM 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded to characterize the distribution of each metal element in the air of livestock and poultry farms. Specifically, based on PM2.5 concentration... 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded, and the distribution of each metal element in the air of the livestock and poultry breeding house is calculated and characterized using the Kriging interpolation method.

[0058] Compared with the prior art, this embodiment has the following advantages:

[0059] (1) Excavation and PM 2.5 The carbon fingerprint peak, which is highly correlated with mass concentration, was used to construct a dynamic correction spectrum of the carbon fingerprint peak, effectively weakening heterogeneous PM. 2.5The matrix effect interferes with the LIBS signal intensity, improving the stability and generalization of the detection.

[0060] (2) Based on the dynamic correction spectrum of carbon fingerprint peaks, random frog jumping and random projection algorithms are used to extract important and non-redundant feature bands of the target element. Linear and nonlinear multivariate PMs are established by PLS and ELM respectively. 2.5 The relative elemental content detection model improves the accuracy of multi-element analysis.

[0061] (3) Simultaneous PM detection in livestock and poultry houses based on LIBS detection 2.5 Mass concentration, relative content of multiple elements, mass concentration detection and distribution visualization, improving PM2.5 levels in livestock and poultry farms. 2.5 Overall analysis efficiency.

[0062] Example 2

[0063] Laser-induced breakdown spectroscopy (LIBS) is an atomic spectroscopy technique that analyzes elemental information in samples by collecting the plasma transition spectra generated during laser ablation and combining them with mathematical methods. Due to its advantages of requiring no complex preprocessing, being fast, minimally invasive, environmentally friendly, and capable of simultaneously analyzing multiple elements online, LIBS has been widely applied in various fields such as industry, agriculture, and medicine in recent years.

[0064] Due to different types of livestock and poultry houses and the PM inside and outside the livestock and poultry houses 2.5 Due to their different primary sources, LIBS exhibits significant differences in elemental composition, organic / elemental carbon distribution, and particle size distribution. The matrix effect of LIBS leads to heterogeneous PM... 2.5 Elements of the same mass concentration produce LIBS signals of varying intensities, affecting modeling and analysis, and impacting LIBS detection of PM. 2.5 The generalization performance of medium element content presents challenges.

[0065] Therefore, in order to weaken heterogeneous PM 2.5 The matrix effect interferes with the LIBS signal, improving the stability of detection and providing a method for PM2.5 detection in livestock and poultry houses based on dynamic carbon peak correction spectrum. 2.5 Multi-element content detection and distribution visualization methods, aiming to explore the relationship between PM2.5 and PM2.5. 2.5 Characteristic spectral lines with high correlation between concentration and the content of each element, and the improvement of PM in chicken houses through dynamic correction of carbon peaks. 2.5 Accuracy, timeliness, and generalization of element detection.

[0066] This embodiment provides a PM (particulate matter) method for livestock and poultry farms based on dynamic correction spectra of carbon fingerprint peaks. 2.5 The detection and distribution characterization method includes the following steps:

[0067] S1: Obtain PM at different sampling sites inside and outside livestock and poultry houses. 2.5 sample.

[0068] The sampling site was a typical H-type cage-raised layer hen house in Zhejiang Province, with a total length of 54m and a width of 12m. The house contained four rows of cages and five aisles, housing approximately 17,000 Jinghong layer hen chickens. Eleven sampling sites (numbers 1-11) were set up inside the house, and three sampling sites (numbers 12-14) were set up outside the house. Figure 2 As shown. The sampling height was 1.5m, which is the average breathing height of a person. The sampling instrument was a Zhongrui ambient air particulate matter comprehensive sampler (ZR-2933, Qingdao, China), with a sampling flow rate of 100 L / min. -1 Through PM 10 Cutter and PM 2.5 The cutter cut particles with a diameter of 2.5 μm, and the sampling time for a single session was 3 hours (9:00-12:00 / 13:00-15:00). The filter membrane used was an 81 mm diameter quartz membrane, which was wrapped in aluminum foil before sampling and fired at 450℃ for 10 hours to remove initial organic matter. After natural cooling, it was equilibrated in a constant temperature and humidity chamber (HWS-250, China) for 24 hours, with the temperature set at 25±1℃ and the humidity at 50±5%. During the sampling period, sampling sites 1, 12, and 13 were sampled twice, sampling site 14 was sampled three times, and the remaining sampling sites were sampled once. After sampling, a total of 19 images containing PM2.5 were obtained. 2.5 The filter membrane of the sample (i.e., PM) 2.5 (Samples) used for PM during sampling at each sampling site 2.5 Mass concentration and elemental content analysis.

[0069] S2: Acquire LIBS spectrum to obtain PM 2.5 The true values ​​of mass concentration and multi-element content.

[0070] S2.1: Will carry PM 2.5 The filter membranes of the samples were equilibrated in a constant temperature and humidity chamber for 24 hours (temperature set at 25±1℃, humidity set at 50±5%). Before and after sampling, the filter membranes were weighed using a 0.0001 g balance (Mettler Toledo XS105DU, Switzerland) and the weight was recorded. Each filter membrane was weighed three times and the average weight was taken. 2.5 The mass concentration was calculated by dividing the difference in filter membrane mass before and after sampling by the volume of the sampled air. After weighing, the PM2.5-containing filter membrane was removed from the filter. 2.5 The sample filter membrane was placed in a polystyrene filter membrane storage box and stored in a -4°C freezer.

[0071] S2.2: PM2.5 is collected using a self-built laser-induced breakdown spectroscopy (LIBS) detection system.2.5 Spectral data of the sample. A 532 nm pulsed laser was generated using an Nd:YAG pulsed solid-state laser (Vlite-200, Beamtech, Beijing, China), with a pulse duration of 8 ns, energy of 60 mJ, and a pulse frequency of 1 Hz. The pulsed laser was transmitted through an optical path system and delivered to the sample (i.e., the sample containing PM). 2.5 A lens (f = 100 mm) above the sample's filter membrane focuses on and ablates the sample surface. The distance between the focusing lens and the sample surface is 98 mm. Plasma spectral signals are acquired using a combination of two spectrometers and detectors. The echelle grating spectrometer (ME5000, Andor, Belfast, UK) has a detection wavelength range of 199-1031 nm and a resolution of 0.03 nm; the monochromator (SR500i, Andor, Belfast, UK) has a detection wavelength range of 189-210 nm and a resolution of 0.02 nm. The echelle grating spectrometer is used to acquire signals for metallic elements, and the monochromator is used to acquire signals for carbon elements. Signals from both spectrometers are acquired using an ICCD photodetector (iStarDH334T-18F-03, Andor, Belfast, UK). Timing control between the laser and the ICCD detector is achieved using a digital delay generator (DG645, Stanford Research Systems, California, USA). The sample was moved using an XYZ three-dimensional motorized translation stage, with a 3×3 array spaced 2 mm apart as the ablation path. Half of the sample was loaded with PM... 2.5 The sample's filter membrane was used for LIBS detection, and PM was loaded onto each half of the membrane using a punch. 2.5 Six evenly distributed small discs, each 8 mm in diameter, were cut from the filter membrane of the sample, resulting in a total of 114 small disc samples. Nine sites were tapped on each small disc sample, with each site tapped three times and the taps accumulated to ensure the PM adsorbed on the filter membrane was effectively contained. 2.5 The sample was completely ablated, and the average spectrum of nine sites was used as the spectrum of the small disc, resulting in a total of 114 sets of spectral data. After optimization, the LIBS parameters were set as follows: energy 60 mJ, middle step delay 2 μs, integration time 18 μs, monochromator delay 1.16 μs, and integration time 1 ms.

[0072] S2.3: Carry 1 / 4 of the PM 2.5The filter membrane of the sample was cut into small pieces with ceramic scissors, added to a polytetrafluoroethylene digestion tube, followed by 5 mL of 65% concentrated nitric acid and 1 mL of 30% hydrogen peroxide. The mixture was then placed in a microwave digester (MARS6, CEM, USA) for digestion. After digestion, the digestion tube was placed on a hot plate at 180°C to remove the acid down to 1 mL, then diluted to volume with a 10 mL volumetric flask and filtered through a 0.45 μm aqueous filter membrane. Finally, inductively coupled plasma mass spectrometry (ICP-MS) was used to detect the elements Al, Ti, Cr, Mn, Fe, Cu, Zn, As, Sr, Cd, Ba, and Pb in the filtered solution. PM2.5 was calculated based on the sample mass and the volume adjusted. 2.5 The mass concentration and relative content of each metal element in the sample.

[0073] S3: Extraction and PM 2.5 The carbon fingerprint peak and corresponding signal intensity are highly linearly correlated with the true value of mass concentration.

[0074] Based on the atomic spectroscopy database of the National Institute of Standards and Technology (NIST), carbon fingerprint peaks at CI 165.7008 nm, CI 193.0905 nm, and CI 247.856 nm were extracted from the sample spectra using LIBS spectral data. PM2.5 fingerprint peaks were then fitted based on these carbon fingerprint peaks. 2.5 A univariate linear regression model of mass concentration was used to establish a calibration curve, and the peak with the largest coefficient of determination (C193.0905nm) was selected. 2 =0.8522) is the carbon fingerprint peak with the highest linear correlation. Among the 114 small disc samples, the carbon fingerprint peak signal intensity corresponding to the i-th sample is I. 193nm (i) .

[0075] S4: Construct a dynamic calibration spectral matrix of carbon fingerprint peaks and extract carbon fingerprint peaks to correct the characteristic peaks of metal elements.

[0076] Among the 114 small disc samples, the i-th small disc sample corresponds to 28683 original LIBS spectral intensities I. raw (i) =[I 199nm (i) , ..., I 1031nm (i) ] 28683 The corresponding carbon fingerprint peak signal intensity is I 193nm (i) .

[0077] The original LIBS spectral matrix X of n small circular samples raw =[I raw(1) , ..., I raw (114) ] T 114*28683 The corresponding carbon fingerprint peak matrix C = [I 193nm (1) , ..., I 193nm (114) ] T .

[0078] The constructed carbon peak dynamic correction (CPDC) LIBS spectral matrix is ​​as follows:

[0079] I CPDC (i) =I raw (i) / I 193nm (i) =[I 199nm (i) , ..., I 1031nm (i) ] / I 193nm (i) (3).

[0080] X CPDC =[I CPDC (1) , ..., I CPDC (114) ] T 114*26863 (4).

[0081] In the formula, I CPDC (i) To dynamically correct the LIBS spectrum of the carbon fingerprint peak corresponding to the i-th small disc sample, X CPDC The LIBS spectral matrix was dynamically corrected for the carbon fingerprint peaks of 114 samples.

[0082] Based on the NIST atomic spectroscopy database, the characteristic peaks of the metal elements of the element to be analyzed are extracted from the dynamic correction spectral matrix of the carbon fingerprint peak.

[0083] S5: Establishing a multivariate PM 2.5 Mass concentration quantitative analysis model.

[0084] The 114 sample spectra were divided into training and test sets in a 3:1 ratio. Multivariate PLS and ELM models were established based on the original LIBS full spectrum or random frog-jump and SPA extraction of LIBS characteristic bands. The coefficients of determination (R²) of the training and test sets were used to calculate the results. 2 c, R 2p) The root mean square error (RMSEC, RMSEP) of the training and test sets is used to evaluate the performance of different models. A comprehensive evaluation of model accuracy and analytical efficiency shows that the PLS model based on the 10 feature variables extracted using RF+SPA has the best detection capability. R 2 c and R 2 p reached 0.9816 and 0.9617 respectively for the training and test sets. 2.5 The average relative errors for the samples were 4% and 10%, respectively.

[0085] S6: Establish a quantitative analysis model for the relative content of multiple elements based on the dynamic correction of LIBS spectral lines using carbon fingerprint peaks.

[0086] S6.1: Using the carbon fingerprint peak of the element to be analyzed, a univariate linear regression model for the relative abundance of each element was established by dynamically correcting the characteristic peak. The calibration curves for each element all achieved good results (R²). 2 >0.73), especially the calibration curves R of Al, Ti, and Sr. 2 >0.85, calibration curve R of Fe 2 The value reached 0.9155, proving that the dynamic correction of the carbon fingerprint peak effectively weakened the interference of the matrix effect.

[0087] S6.2: Utilize RF and SPA to extract carbon fingerprint peaks and dynamically correct characteristic bands in LIBS spectra that are highly correlated with each element. Establish multiple linear and nonlinear regression models using PLS and ELM respectively, and use the coefficient of determination (R²) of the training and test sets. 2 c, R 2 p) The root mean square error (RMSEC, RMSEP) of the training and test sets was used to evaluate the performance of different models. The multivariate regression models based on the full spectrum for each element generally outperformed the univariate linear models based on the characteristic peaks. Furthermore, after RF and SPA feature selection, the model complexity was significantly reduced, and redundant information was decreased. For most elements, the models based on characteristic wavelengths outperformed the models based on the full spectrum. Based on model complexity and performance, the optimal prediction models for 12 elements were selected. The RMSEP of the Al, Ti, Fe, Sr, Ba, and Pb models was [missing data]. 2 When p is higher than 0.95, the R-squared values ​​for the Mn, Cu, and As models are... 2 When p is higher than 0.9, the R of Cr and Zn 2 p is higher than 0.8.

[0088] S7: PM estimation 2.5 The mass concentration of the metal element in the air.

[0089] Through PM 2.5 The detection results from the mass concentration analysis model and the results from the quantitative analysis of relative elemental content will be used to calculate the PM2.5 concentration using LIBS. 2.5 mass concentration CPM2.5-LIBS (μg m -3 The relative content C of metallic element i i-rel-LIBS (ngμg -1 Multiplying them together can estimate PM2.5. 2.5 The mass concentration C of metallic element i in air i-LIBS (ng m -3 The predicted PM2.5 concentrations were compared with the actual values ​​measured by ICP-MS, and the accuracy of the estimation was evaluated using normalized root mean square error (NRMSE) and mean relative error (MRE). The predicted NRMSE for the 12 elements' air quality concentrations ranged from 0.0475 to 0.1882, and the MRE ranged from 7.5% to 15.0%, indicating that this method can accurately estimate PM2.5 concentrations. 2.5 The mass concentration of metal elements in the air provides a reference for risk assessment of inhalable metal elements and improvement of environmental quality.

[0090] S8: Calculate and characterize the distribution of each element in the air of the breeding house using the Kriging interpolation method.

[0091] Calculate PM at 14 sampling sites during the sampling period 2.5 The mass concentration of metallic elements in the air was determined by establishing a 200*200 grid for fitting and interpolating the element concentration distribution inside and outside the chicken coop, and the results were characterized using a pseudo-color image. The results are shown below. Figure 3 As shown.

[0092] In this embodiment, the PM established based on the dynamic correction method of carbon element fingerprint peaks can be obtained from the analysis results of metal element content and mass concentration. 2.5 The multi-element relative content quantitative analysis model and element mass concentration estimation method have good accuracy and can effectively weaken heterogeneous PM. 2.5 The matrix effect interferes with detection, effectively improving the accuracy, stability, and generalization performance of the detection model.

[0093] (1) Rapid detection of PM via LIBS pulsed laser, simultaneously achieving 2.5 Concentration and multi-element content, mass concentration detection, requires no complex sample pretreatment such as constant temperature and humidity equilibration and digestion, is simple to operate, has a short detection cycle, and improves PM2.5 concentration. 2.5 Timeliness of detection and analysis.

[0094] (2) Weakening heterogeneous PM through dynamic carbon peak correction 2.5 The interference of matrix effects on the intensity of elemental characteristic spectral lines was addressed by combining RF and SPA to extract non-redundant important features. A multivariate linear and nonlinear regression model of elemental content based on dynamic carbon peak correction lines was established using PLS and ELM to improve LIBS detection of PM. 2.5 Precision, stability, and generalizability of metallic elements.

[0095] Example 3

[0096] In order to implement the method corresponding to Embodiment 1 above and achieve the corresponding functions and technical effects, a PM for livestock and poultry breeding houses is provided below. 2.5 Detection and distribution characterization system.

[0097] This embodiment provides a PM for livestock and poultry breeding sheds. 2.5 Detection and distribution characterization systems, including:

[0098] The sample acquisition module is used to acquire PM at different sampling sites inside and outside livestock and poultry houses. 2.5 sample.

[0099] The spectral data and true value acquisition module is used to acquire the data for each PM. 2.5 Spectral data of the sample and PM 2.5 True mass concentration and each PM 2.5 The true mass concentration and the true relative content of each metal element in the sample.

[0100] The carbon fingerprint peak extraction module is used to extract peaks related to PM from spectral data. 2.5 The carbon fingerprint peak and corresponding signal intensity are highly linearly correlated with the true value of mass concentration.

[0101] The carbon fingerprint peak correction and metal element characteristic peak extraction module is used to construct a dynamic correction spectral matrix of carbon fingerprint peaks based on the extracted carbon fingerprint peaks and their corresponding signal intensities, and to extract the carbon fingerprint peak correction and metal element characteristic peaks of the metal element to be analyzed based on the dynamic correction spectral matrix of carbon fingerprint peaks.

[0102] Multivariate PM 2.5 The mass concentration quantitative analysis model building module is used to establish models based on each PM. 2.5 Spectral data of the sample and PM 2.5 True mass concentration values, establishing multivariate PM 2.5 Mass concentration quantitative analysis model; the multivariate PM 2.5 Mass concentration quantitative analysis model is used to detect PM in the air. 2.5 Mass concentration.

[0103] The module for establishing quantitative analysis models of relative content of multiple metal elements is used to correct the characteristic peaks of metal elements and each PM based on the carbon fingerprint peak of the metal element to be analyzed. 2.5 The true relative content of each metal element in the sample was used to establish a quantitative analysis model for the relative content of multiple metal elements; this model was then used to detect PM. 2.5 The relative content of each metal element in the sample.

[0104] The estimation module is used to obtain the PM at the target sampling point. 2.5 Samples, and based on the multivariate PM 2.5 The mass concentration quantitative analysis model and the relative content quantitative analysis model of the multi-metal elements determine the PM at the target sampling site. 2.5 PM of the sample 2.5 Mass concentration and relative content of each metal element, and based on PM at the target sampling site. 2.5 PM of the sample 2.5 Estimate PM by mass concentration and relative content of each metal element. 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded; the target sampling site is the sampling site to be analyzed.

[0105] Distribution characterization module, used to characterize PM 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded to characterize the distribution of each metal element in the air of livestock and poultry breeding houses.

[0106] Example 4

[0107] This invention provides an electronic device including a memory and a processor. The memory stores a computer program, and the processor runs the computer program to cause the electronic device to execute a PM for a livestock and poultry breeding shed as described in Embodiment 1. 2.5 Detection and distribution characterization methods.

[0108] Alternatively, the aforementioned electronic device may be a server.

[0109] In addition, embodiments of the present invention also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a PM for livestock and poultry breeding sheds as described in Embodiment 1. 2.5 Detection and distribution characterization methods.

[0110] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0111] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A type of PM for livestock and poultry breeding sheds 2.5 The detection and distribution characterization method is characterized by, include: PM2.5 at different sampling sites inside and outside livestock and poultry houses 2.5 sample; Get each PM 2.5 Spectral data of the sample and PM 2.5 True mass concentration and each PM 2.5 The true mass concentration and relative content of each metal element in the sample, specifically including: determining the mass difference of each sampling filter membrane before and after sampling, and determining the mass concentration and relative content of each PM based on the mass difference and the volume of sampled air. 2.5 PM of the sample 2.5 The true mass concentration value was determined using laser-induced breakdown spectroscopy to measure the concentration of each PM. 2.5 Spectral data of the sample; PM2.5 was determined using inductively coupled plasma mass spectrometry. 2.5 The mass of metal elements in the sample, and based on PM 2.5 Calculate the mass and constant volume for each PM 2.5 The true mass concentration and true relative content of each metal element in the sample; Extracting from spectral data related to PM 2.5 The carbon element fingerprint peak and corresponding signal intensity are highly linearly correlated with the true value of mass concentration. Based on the extracted carbon fingerprint peaks and their corresponding signal intensities, a dynamic calibration spectral matrix for carbon fingerprint peaks is constructed, and the characteristic peaks of the metal element to be analyzed are extracted from the carbon fingerprint peaks of the metal element based on the dynamic calibration spectral matrix for carbon fingerprint peaks. According to each PM 2.5 Spectral data of the sample and PM 2.5 True mass concentration values, establishing multivariate PM 2.5 Mass concentration quantitative analysis model; the multivariate PM 2.5 Mass concentration quantitative analysis model is used to detect PM in the air. 2.5 mass concentration; Based on the carbon fingerprint peak of the metal element to be analyzed, the characteristic peaks of the metal element and each PM are corrected. 2.5 The true relative content of each metal element in the sample was used to establish a quantitative analysis model for the relative content of multiple metal elements; this model was then used to detect PM. 2.5 The relative content of each metallic element in the solution; Obtain PM at the target sampling site 2.5 Samples, and based on the multivariate PM 2.5 The mass concentration quantitative analysis model and the relative content quantitative analysis model of the multi-metal elements determine the PM at the target sampling site. 2.5 PM of the sample 2.5 Mass concentration and relative content of each metal element, and based on PM at the target sampling site. 2.5 PM of the sample 2.5 Estimate PM by mass concentration and relative content of each metal element. 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded; the target sampling site is the sampling site to be analyzed. According to PM 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded to characterize the distribution of each metal element in the air of livestock and poultry breeding houses.

2. A PM for livestock and poultry breeding sheds according to claim 1 2.5 The detection and distribution characterization method is characterized by, PM2.5 at different sampling sites inside and outside livestock and poultry houses 2.5 Samples, specifically including: Using a quartz membrane as the sampling filter, a medium-flow particulate matter sampler was used to collect and obtain PM2.5 concentrations at different sampling sites inside and outside livestock and poultry sheds. 2.5 sample.

3. A PM for livestock and poultry breeding sheds according to claim 1 2.5 The detection and distribution characterization method is characterized by, The dynamic correction spectral matrix of the carbon fingerprint peak is as follows: I CPDC (i) = I raw (i) / I λC (i) = [I λ1 (i) ,…,I λm (i) ] m / I λC (i) ; X CPDC =[I CPDC (1) ,…,I CPDC (n) ] T n*m ; Among them, I CPDC (i) For the i-th PM 2.5 Dynamically corrected spectrum of carbon fingerprint peak corresponding to the sample, X CPDC For n PMs 2.5 Dynamically corrected spectral matrix of carbon fingerprint peaks in the sample; In n PM 2.5 In the sample, the i-th PM 2.5 The original spectral intensities of the sample in the m bands are I. raw (i) =[I λ1 (i) , ..., I λm (i) ] m The corresponding carbon fingerprint peak signal intensity is I. λC (i) ; n PMs 2.5 The original spectral matrix of the sample is X raw = [I raw (1) , ..., I raw (n) ] T n*m The corresponding carbon fingerprint peak matrix is ​​C = [I λC (1) , ..., I λC (n) ] T n .

4. A PM for livestock and poultry breeding sheds according to claim 1 2.5 The detection and distribution characterization method is characterized by, According to each PM 2.5 Spectral data of the sample and PM 2.5 True mass concentration values, establishing multivariate PM 2.5 The mass concentration quantitative analysis model specifically includes: Each PM is extracted by combining random frog jumping and continuous projection algorithms. 2.5 The spectral data of the sample are related to PM 2.5 The characteristic bands related to the true value of PM were identified, and multiple linear and nonlinear regression models were established using partial least squares and extreme learning machine methods, respectively. The optimal model was selected based on the coefficient of determination and root mean square error to obtain the multivariate PM. 2.5 Mass concentration quantitative analysis model.

5. A PM for livestock and poultry breeding sheds according to claim 1 2.5 The detection and distribution characterization method is characterized by, According to PM 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded, characterizing the distribution of each metal element in the air of livestock and poultry farms, specifically including: According to PM 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded, and the distribution of each metal element in the air of the livestock and poultry breeding house is calculated and characterized using the Kriging interpolation method.

6. A type of PM for livestock and poultry breeding sheds 2.5 The detection and distribution characterization system is characterized by, include: The sample acquisition module is used to acquire PM at different sampling sites inside and outside livestock and poultry houses. 2.5 sample; The spectral data and true value acquisition module is used to acquire the data for each PM. 2.5 Spectral data of the sample and PM 2.5 True mass concentration and each PM 2.5 The true mass concentration and relative content of each metal element in the sample, specifically including: determining the mass difference of each sampling filter membrane before and after sampling, and determining the mass concentration and relative content of each PM based on the mass difference and the volume of sampled air. 2.5 PM of the sample 2.5 The true mass concentration value was determined using laser-induced breakdown spectroscopy to measure the concentration of each PM. 2.5 Spectral data of the sample; PM2.5 was determined using inductively coupled plasma mass spectrometry. 2.5 The mass of metal elements in the sample, and based on PM 2.5 Calculate the mass and constant volume for each PM 2.5 The true mass concentration and true relative content of each metal element in the sample; The carbon fingerprint peak extraction module is used to extract peaks related to PM from spectral data. 2.5 The carbon element fingerprint peak and corresponding signal intensity are highly linearly correlated with the true value of mass concentration. The carbon fingerprint peak correction and metal element characteristic peak extraction module is used to construct a dynamic correction spectral matrix of carbon fingerprint peaks based on the extracted carbon fingerprint peaks and corresponding signal intensities, and to extract the carbon fingerprint peak correction and metal element characteristic peaks of the metal element to be analyzed based on the dynamic correction spectral matrix of carbon fingerprint peaks. Multivariate PM 2.5 The mass concentration quantitative analysis model building module is used to establish models based on each PM. 2.5 Spectral data of the sample and PM 2.5 True mass concentration values, establishing multivariate PM 2.5 Mass concentration quantitative analysis model; the multivariate PM 2.5 Mass concentration quantitative analysis model is used to detect PM in the air. 2.5 mass concentration; The module for establishing quantitative analysis models of relative content of multiple metal elements is used to correct the characteristic peaks of metal elements and each PM based on the carbon fingerprint peak of the metal element to be analyzed. 2.5 The true relative content of each metal element in the sample was used to establish a quantitative analysis model for the relative content of multiple metal elements; this model was then used to detect PM. 2.5 The relative content of each metallic element in the solution; The estimation module is used to obtain the PM at the target sampling point. 2.5 Samples, and based on the multivariate PM 2.5 The mass concentration quantitative analysis model and the relative content quantitative analysis model of the multi-metal elements determine the PM at the target sampling site. 2.5 PM of the sample 2.5 Mass concentration and relative content of each metal element, and based on PM at the target sampling site. 2.5 PM of the sample 2.5 Estimate PM by mass concentration and relative content of each metal element. 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded; the target sampling site is the sampling site to be analyzed. Distribution characterization module, used to characterize PM 2.5 The mass concentration of each metal element in the air at the target sampling site is recorded to characterize the distribution of each metal element in the air of livestock and poultry breeding houses.

7. An electronic device, characterized in that, The device includes a memory and a processor, the memory storing a computer program, and the processor running the computer program to cause the electronic device to perform a PM for livestock and poultry breeding shed according to any one of claims 1 to 5. 2.5 Detection and distribution characterization methods.

Images