Particulate matter concentration detection system based on physical separation and light scattering

Abstract

The invention relates to the field of environment monitoring, in particular to a particulate matter concentration detection system based on physical separation and light scattering. Through multi-level cooperation of the air inlet pretreatment unit, the physical separation unit and the light scattering detection unit, high-precision, low-cost and high-reliability detection of particulate matter concentration and type is realized. Wherein the cyclone separator effectively eliminates interference of large particles, and the signal-to-noise ratio of optical detection is remarkably improved; few-angle detection is combined with a variable polaroid, and a forward detector and a lateral detector are matched, so that a sufficient judgment basis is provided for a machine learning model. The above technical characteristics support each other and are linked with each other to jointly form an integrated monitoring system with scientific preciseness and engineering practicability.

Description

Technical Field

[0001] This invention relates to the field of environmental monitoring, specifically to a particulate matter concentration detection system based on physical separation and light scattering. Background Technology

[0002] Light scattering, with its advantages of fast response, relatively low cost, and ease of online monitoring, has become one of the mainstream technologies for monitoring particulate matter in ambient air. Currently, it mainly falls into two technical approaches:

[0003] Firstly, there is single-angle light scattering technology. This technology has a simple structure and the lowest cost, but its measurement accuracy is greatly affected by various factors such as particle size distribution, complex refractive index (composition), and environmental humidity. In complex and variable atmospheric environments, its measurement error is significant, and it can usually only be used for trend judgment and rough estimation, failing to meet the needs of high-precision monitoring. For example, Chinese patent CN109477777A discloses a particulate matter measuring device that uses a cyclone device with a fixed cutting point in conjunction with a single-angle (115°-140°) light scattering detector. It can only achieve a rough measurement of the concentration of particles of a single size and has no particle type identification function. Its physical separation unit is an integrated fixed structure without self-cleaning design, and long-term operation is prone to blockage due to particle accumulation, resulting in extremely high maintenance costs.

[0004] Secondly, there is multi-angle light scattering technology. This technology collects scattered light signals from multiple angles and uses complex algorithms to retrieve particulate matter information, improving measurement accuracy compared to single-angle technology. However, this system has a complex optical structure (requiring more than five detectors), is difficult to assemble and adjust, resulting in high costs (3-5 times that of single-angle systems). Furthermore, its algorithm heavily relies on model assumptions, and it still has significant limitations when dealing with particulate matter with unknown or complex chemical compositions, making it difficult to achieve industrial-scale promotion.

[0005] Furthermore, neither of the aforementioned technologies effectively addresses the particle size range issue at the physical level, requiring optical sensors to constantly handle signal interference from wide particle size distributions. This is a key reason why their measurement accuracy is difficult to fundamentally improve. Additionally, Chinese patents CN112755661B and its accompanying invention CN215931601U employ a dual-path cyclone separation structure combined with a radioactive source-paper tape detection technology. While achieving multi-particle size concentration measurement, the detection principle is based on ray attenuation, not light scattering. This results in slow response speed (relying on time-division sampling cycles of over 30 minutes), high radioactive source maintenance costs, and a lack of type identification functionality. Moreover, its dual-path separation structure also lacks anti-clogging design, limiting its applicability to various scenarios.

[0006] Among existing light scattering detection technologies, no solution can achieve accurate identification of particulate matter types through simple and reliable optical design. Some solutions mention multi-dimensional light scattering detection, but they lack clarity regarding the acquisition method of polarization parameters and their correlation with particulate matter characteristics, resulting in a lack of feasibility support for type identification. Therefore, there is an urgent need in this field for a new particulate matter monitoring technology that can improve measurement accuracy, control hardware costs, possess good engineering reliability, and achieve particulate matter type identification through clear and feasible technical means. This would address the core pain points of existing technologies, such as insufficient accuracy, limited functionality, poor adaptability, high maintenance costs, and vague descriptions of type identification techniques. Summary of the Invention

[0007] To address the above problems, the present invention proposes the following technical solution:

[0008] A particulate matter concentration detection system based on physical separation and light scattering includes a physical separation unit and a light scattering detection unit connected in sequence.

[0009] The physical separation unit includes a cyclone separator; the cyclone separator is used to separate particles larger than a preset particle size from the bottom of the gas.

[0010] The light scattering detection unit includes a laser emitter, a forward photodetector, and a lateral photodetector. The laser emitter emits laser light into the gas from the cyclone separator. The forward and lateral photodetectors receive the laser light after it interacts with the gas. The angle between the forward photodetector and the laser light is 3°-15°. The angle between the lateral photodetector and the laser light is 90°-120°. A variable-angle polarizer is also provided at the front end of the lateral photodetector. The laser light emitted by the laser emitter is polarized light.

[0011] During the work process, by Input the pre-trained machine learning model to obtain particulate matter concentration;

[0012] in, This indicates the intensity of the scattered light received by the forward photodetector;

[0013] This represents the intensity of scattered light received by the lateral photodetector when the polarization direction of the variable-angle polarizer is parallel to the laser beam.

[0014] This represents the intensity of scattered light received by the lateral photodetector when the polarization direction of the variable-angle polarizer is perpendicular to the laser beam.

[0015] Indicates gas temperature;

[0016] Indicates the humidity of the gas.

[0017] Preferably, the parameters input to the pre-trained machine learning model also include ;in, The calculation formula is:

[0018] in, Indicates the degree of depolarization.

[0019] Preferably, the machine learning model uses particulate matter concentration and particulate matter type distribution as supervision signals during the training process.

[0020] Preferably, during operation, the pre-trained machine learning model also outputs a particulate matter type distribution.

[0021] Preferably, the bottom of the cyclone separator is also provided with a particle collection box, and the outer shell of the particle collection box is made of transparent material.

[0022] Preferably, the angle between the forward photodetector and the laser is 8°; the angle between the lateral photodetector and the laser is 100°.

[0023] Preferably, a light trap is provided on the opposite side of the laser emitter to absorb the laser beam that is not scattered by the particles, thereby avoiding stray light interference caused by laser reflection.

[0024] Preferably, an intake pretreatment unit is connected before the physical separation unit;

[0025] The air intake pretreatment unit includes an air intake, a mixing chamber, and a heating module connected in sequence; a mixing fan is also connected to the mixing chamber via a pipe; the heating module is used to heat the gas when the relative humidity of the gas is greater than a threshold, so as to form a low relative humidity area.

[0026] Preferably, the air inlet is funnel-shaped, and the diameter of the air inlet gradually decreases along the direction of gas flow.

[0027] Preferably, it also includes a purging system, the purging system comprising:

[0028] The third switching valve is installed on the pipeline between the physical separation unit and the polarized light scattering detection unit.

[0029] A purge blower, the outlet of which is connected to the pipeline between the third switching valve and the polarized light scattering detection unit;

[0030] The first switching valve is installed on the pipeline where the purge fan is located, and is used to control the purge fan to purge the pipeline between the third switching valve and the polarized light scattering detection unit.

[0031] The purge blower has two inlets: one is connected to the air via a filter, and the other is directly connected to the air via a second switching valve.

[0032] Beneficial Effects: This invention achieves high-precision, low-cost, and high-reliability detection of particulate matter concentration and type through multi-level synergy of an air intake pretreatment unit, a physical separation unit, and a small-angle polarized light scattering detection unit. Specifically, the cyclone separator effectively eliminates interference from large particles, significantly improving the signal-to-noise ratio of optical detection. High-precision particulate matter concentration measurement can be achieved with only a few angle detectors (two angles in this embodiment). The heating module ensures no interference in high-humidity environments, guaranteeing the effectiveness of polarization measurements. The combination of laser and variable polarizer, along with forward and lateral dual detectors, constructs a multi-dimensional feature vector including particle size, composition, morphology, and polarization response, providing sufficient discrimination criteria for machine learning models. These technical features support each other and are interconnected, collectively forming an integrated monitoring system that combines scientific rigor with engineering practicality. Attached Figure Description

[0033] Figure 1 This is a simplified structural diagram of the present invention;

[0034] Reference numerals: 11-Dust filter, 12-Air inlet, 13-Mixing chamber, 14-Mixing fan, 15-Heating module, 21-Cyclone separator, 22-Particle collection box, 31-Laser emitter, 32-Forward photodetector, 33-Side photodetector, 34-Variable angle polarizer, 4-Flow control module, 5-Control unit, 61-Filter, 62-Purge fan, 63-First switching valve, 64-Second switching valve, 65-Third switching valve. Detailed Implementation

[0035] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0036] like Figure 1 As shown, a particulate matter concentration detection system based on physical separation and low-angle polarized light scattering includes an air intake pretreatment unit, a physical separation unit, and a polarized light scattering detection unit connected in sequence.

[0037] The air intake pretreatment unit includes an air inlet 12, a mixing chamber 13, and a heating module 15 connected in sequence. A mixing fan 14 is also connected to the mixing chamber 13 via a pipe. In this embodiment, a physical volume (mixing chamber) is introduced, allowing the sampling airflow to enter this chamber first. During this process, the high-speed airflow is buffered, its kinetic energy is reduced, and it is fully mixed and homogenized with the existing air in the chamber. Subsequently, this "tamed" and stable airflow enters the optical detection area in a controlled and uniform manner. This is equivalent to adding an "airflow stabilizer" and a "sample homogenizer" at the front end of the sensor.

[0038] The air inlet 12 is funnel-shaped, with its diameter gradually decreasing along the gas flow direction. The minimum inner diameter of the air inlet 12 is 10mm, and the expansion angle is 30°. This shape of the air inlet 12 can reduce turbulence interference when the airflow is drawn in. A dust filter 11 is also provided at the front end of the air inlet 12 to preliminarily filter large debris such as leaves and insects. The heating module 15 is used to heat the gas when the relative humidity of the gas is greater than a threshold value to form a low relative humidity area. In this embodiment, the relative humidity threshold value is 35%RH.

[0039] When gas enters the working area of ​​the heating module 15, the system will monitor the relative humidity of the gas in real time. When the humidity exceeds the preset threshold, the heating module 15 will automatically start and accurately reduce the relative humidity of the area to below the threshold. In this low-humidity controlled environment, the bound water adsorbed on the surface and pores of the particles will fully evaporate, so that the particles will be completely restored to their inherent form and internal structure in a dry state, eliminating the effect of moisture on the optical properties of the particles.

[0040] The physical separation unit includes a cyclone separator 21 and a particle collection box 22 disposed at the bottom of the cyclone separator 21. The cyclone separator 21 is used to separate particles larger than a preset particle size from the bottom of the gas. For example, when this system is to detect PM2.5 (aerodynamic diameter ≤ 2.5 µm), the cyclone separator 21 is controlled to separate particles with a particle size of 2.5 µm. The outer shell of the particle collection box 22 is made of transparent material for easy observation. The preset particle size can be adjusted by replacing the corresponding cyclone separator 21, etc. How to adjust the preset particle size is existing technology and will not be described in detail here.

[0041] The light scattering detection unit includes a laser emitter 31, a forward photodetector 32, and a lateral photodetector 33. The laser emitter 31 emits laser light into the gas from the cyclone separator 21. The forward photodetector 32 and the lateral photodetector 33 receive the laser light after it interacts with the gas. The angle between the forward photodetector 32 and the laser light is 3°-15° (preferably 8°). At this angle, the scattered light signal is sensitive to changes in particle size and is used to obtain particle size-related information. The angle between the lateral photodetector 33 and the laser light is 90°-120° (preferably 100°). At this angle, the scattered light signal is sensitive to changes in particle composition and morphology and is used to obtain composition / morphology-related information. A variable-angle polarizer 34 is also provided at the front end of the lateral photodetector 33. The laser light emitted by the laser emitter 31 is polarized light. In this embodiment, the wavelength of the laser emitter 31 is 635 nm-870 nm.

[0042] In a preferred embodiment, a light trap is also provided on the opposite side of the laser emitter 31 to absorb the laser beam that is not scattered by the particles and avoid stray light interference caused by laser reflection.

[0043] In conventional light scattering particulate matter concentration detection systems, the impact of stray light interference is usually within an acceptable range. However, this system relies on high-precision acquisition of forward (3°-15°) and lateral (90°-120°) dual-angle scattering signals, and the construction of multi-dimensional feature vectors based on depolarization degree measurement using a variable polarizer to provide a basis for discrimination in pre-trained machine learning models. Stray light can significantly interfere with polarization state recognition and accurate acquisition of scattered light intensity, thus directly affecting the accuracy of particulate matter concentration and type detection. The inventors keenly recognized this unique pain point of this system and, by adding a light trap on the opposite side of the laser source, can efficiently absorb the laser beam that has not been scattered by particles, fundamentally avoiding stray light interference caused by laser reflection. This significantly improves the signal purity and signal-to-noise ratio of the optical detection module, ensuring the authenticity and reliability of the dual-angle scattering data and depolarization degree parameters. This further strengthens the core advantages of high precision and high reliability of this system, making the particulate matter concentration and type distribution results output by the machine learning model more accurate and reliable.

[0044] During the work process, by Input the pre-trained machine learning model to obtain particulate matter concentration;

[0045] in, This indicates the intensity of the scattered light received by the forward photodetector 32;

[0046] This represents the intensity of scattered light received by the lateral photodetector 33 when the polarization direction of the variable-angle polarizer is parallel to the laser.

[0047] This represents the intensity of scattered light received by the lateral photodetector 33 when the polarization direction of the variable-angle polarizer is perpendicular to the laser.

[0048] The depolarization degree is expressed by the following formula:

[0049] Indicates gas temperature;

[0050] Indicates the humidity of the gas;

[0051] When linearly polarized incident light irradiates particulate matter, the polarization state of the scattered light changes with the shape, composition, and internal structure of the particulate matter (i.e., depolarization effect): For spherical, homogeneous particulate matter (such as secondary inorganic salt particles), the polarization state of the scattered light is consistent with that of the incident light, and the degree of depolarization is extremely low; for non-spherical, heterogeneous particulate matter (such as sand and dust, chain-like black carbon particles), the scattered light will produce a significant polarization state deflection, and the degree of depolarization is relatively high.

[0052] The machine learning models include random forests, support vector machines, or neural networks.

[0053] During training, the machine learning model uses particulate matter concentration and particulate matter type distribution (such as the probability proportions of dust, black carbon, and secondary inorganic salts) as supervision signals. Whether the machine learning model outputs the particulate matter type distribution during inference is optional.

[0054] In this invention, the input features implicitly contain information about the particulate matter type. However, traditional models only use these features to directly regress concentration, failing to fully explore the nonlinear relationships between features. Training with type distribution as auxiliary labels is equivalent to providing the model with "feature-type-concentration" correlation supervision, enabling the model to deeply understand the true concentration differences of different types of particulate matter under the same scattered light intensity, thereby improving the accuracy of concentration prediction.

[0055] With type label supervision, the model can automatically establish a mapping relationship between "optical feature deviation - particulate matter type - concentration correction coefficient". During the inference stage, even if an unseen combination of particulate matter types is encountered, the model can adaptively adjust the concentration calculation results based on the learned correlation rules, which greatly improves the measurement accuracy and environmental adaptability of the system.

[0056] At the outlet of the light scattering detection unit, a flow control module 4 is connected, including an electronic flow meter and an adjustable fan module. The high-precision electronic flow meter provides millisecond-level real-time accurate flow rate. The control unit, combined with ambient temperature, humidity, and air pressure, calculates the operating flow rate and standard flow rate in real time and performs second-level flow control, with an average flow deviation within ±1%. The specific flow control method is existing technology and will not be elaborated here. The system also includes a control unit 5 to implement the above control functions. Its specific hardware connection method can be configured as needed and will not be elaborated here either.

[0057] This invention organically integrates a cyclone separator 21 with a light scattering detection unit to achieve a collaborative working mechanism of "physical screening first, then precise measurement." The cyclone separator 21 pre-removes coarse particles larger than the target particle size (e.g., PM2.5), significantly reducing the complexity of the particle size distribution entering the optical cavity. This allows the forward and side scattering signals to primarily reflect the characteristics of the target fine particles, thereby greatly improving the accuracy of concentration inversion and reducing reliance on complex algorithm models. This collaborative design fundamentally alleviates the signal interference problem caused by wide particle size mixing in traditional light scattering technology, achieving a technological leap from "passive correction" to "active control."

[0058] The heating module 15 and the depolarization measurement mechanism based on the variable-angle polarizer 34 form a key synergy. By actively heating to reduce the relative humidity of the gas, interference from particles on the lateral scattering signal in high humidity environments is effectively avoided, ensuring that the measured polarization state changes truly reflect the morphology and internal structural differences of solid particles. This synergy guarantees the reliability of the depolarization degree (D) as a core parameter for particle type identification, enabling accurate differentiation between spherical secondary aerosols and non-spherical dust or chain-like black carbon particles even in humid climates or rainy seasons, significantly improving the environmental adaptability and stability of type identification.

[0059] A laser emitter, along with forward / lateral dual photodetectors and a variable polarizer, constructs a multi-dimensional optical sensing system and a collaborative information acquisition mechanism integrating angle and polarization. The sensitivity of forward scattering to particle size and the sensitivity of lateral polarized scattering to morphology generate a composite signal vector containing spectral, spatial, and polarization features, providing rich and discriminative input features for subsequent machine learning models. This collaboration not only improves the measurement accuracy of individual parameters but also enables the joint identification of particulate matter's chemical composition and microstructure, overcoming the technical bottleneck of traditional small-angle systems with limited functionality.

[0060] The transparent particle collection box allows for visual management, enabling maintenance personnel to intuitively monitor the sewage discharge status and promptly replace or clean it.

[0061] The particulate matter concentration detection system further includes a purging system, which comprises:

[0062] The third switching valve 65 is installed on the pipeline between the physical separation unit and the polarized light scattering detection unit.

[0063] A purge fan 62, the outlet of which is connected to the pipeline between the third switching valve 65 and the polarized light scattering detection unit;

[0064] The first switching valve 63 is installed on the pipeline where the purge fan 62 is located, and is used to control the purge fan 62 to purge the pipeline between the third switching valve 65 and the polarized light scattering detection unit.

[0065] The inlet of the purge blower 62 has two paths: one path is connected to the air through the filter 61, and the other path is directly connected to the air through the second switching valve 64.

[0066] When the first switching valve 63 and the second switching valve 64 are open and the third switching valve 65 is closed, the purge fan 62 can generate a large flow of air to purge the polarization light scattering detection unit; when the first switching valve 63 is open and the second switching valve 64 and the third switching valve 65 are closed, the air is filtered by the filter 61 to generate clean air, which can be used to zero the polarization light scattering detection unit.

[0067] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A particulate matter concentration detection system based on physical separation and light scattering, characterized in that, It includes a physical separation unit and a light scattering detection unit connected in sequence; The physical separation unit includes a cyclone separator; the cyclone separator is used to separate particles larger than a preset particle size from the bottom of the gas. The light scattering detection unit includes a laser emitter, a forward photodetector, and a lateral photodetector. The laser emitter emits laser light into the gas from the cyclone separator. The forward and lateral photodetectors receive the laser light after it interacts with the gas. The angle between the forward photodetector and the laser light is 3°-15°. The angle between the lateral photodetector and the laser light is 90°-120°. A variable-angle polarizer is also provided at the front end of the lateral photodetector. The laser light emitted by the laser emitter is polarized light. During the work process, by Input the pre-trained machine learning model to obtain particulate matter concentration; in, This indicates the intensity of the scattered light received by the forward photodetector; This represents the intensity of scattered light received by the lateral photodetector when the polarization direction of the variable-angle polarizer is parallel to the laser beam. This represents the intensity of scattered light received by the lateral photodetector when the polarization direction of the variable-angle polarizer is perpendicular to the laser beam. Indicates gas temperature; Indicates the humidity of the gas.

2. The particulate matter concentration detection system based on physical separation and light scattering according to claim 1, characterized in that, The parameters input to the pre-trained machine learning model also include ;in, The calculation formula is: in, Indicates the degree of depolarization.

3. A particulate matter concentration detection system based on physical separation and light scattering according to any one of claims 1 or 2, characterized in that, During the training process, the machine learning model uses particulate matter concentration and particulate matter type distribution as supervision signals.

4. The particulate matter concentration detection system based on physical separation and light scattering according to claim 3, characterized in that, During operation, the pre-trained machine learning model also outputs the distribution of particulate matter types.

5. The particulate matter concentration detection system based on physical separation and light scattering according to claim 1, characterized in that, The bottom of the cyclone separator is also equipped with a particle collection box, the outer shell of which is made of transparent material.

6. The particulate matter concentration detection system based on physical separation and light scattering according to claim 1, characterized in that, The angle between the forward photodetector and the laser is 8°; the angle between the lateral photodetector and the laser is 100°.

7. The particulate matter concentration detection system based on physical separation and light scattering according to claim 1, characterized in that, On the opposite side of the laser emitter, a light trap is also provided to absorb the laser beam that is not scattered by the particles, so as to avoid the laser reflection forming stray light interference.

8. The particulate matter concentration detection system based on physical separation and light scattering according to claim 1, characterized in that, An intake pretreatment unit is also connected before the physical separation unit; The air intake pretreatment unit includes an air intake, a mixing chamber, and a heating module connected in sequence; a mixing fan is also connected to the mixing chamber via a pipe; the heating module is used to heat the gas when the relative humidity of the gas is greater than a threshold, so as to form a low relative humidity area.

9. The particulate matter concentration detection system based on physical separation and light scattering according to claim 8, characterized in that, The air inlet is funnel-shaped, and its diameter gradually decreases along the direction of gas flow.

10. The particulate matter concentration detection system based on physical separation and light scattering according to claim 1, characterized in that, It also includes a purging system, which comprises: The third switching valve is installed on the pipeline between the physical separation unit and the polarized light scattering detection unit. A purge blower, the outlet of which is connected to the pipeline between the third switching valve and the polarized light scattering detection unit; The first switching valve is installed on the pipeline where the purge fan is located, and is used to control the purge fan to purge the pipeline between the third switching valve and the polarized light scattering detection unit. The purge blower has two inlets: one is connected to the air via a filter, and the other is directly connected to the air via a second switching valve.

Images