Multi-material waste plastic hyperspectral and ultrasonic coupling intelligent sorting and purification method

Abstract

The application discloses a multi-material waste plastic hyperspectral and ultrasonic coupling intelligent sorting and purification method, which sequentially carries out vibration uniform material paving and conveying of mixed waste plastics; short-wave infrared hyperspectral and RGB visual multi-modal detection are combined with deep learning to identify materials, so that eight types of single plastics, mixed materials, harmful plastics and hard impurities such as PET and HDPE are separated and collected; single-material plastics are subjected to material separation crushing and shaping and air screening and impurity removal; then, corresponding gradient three-stage ultrasonic cleaning is matched; extrusion solid-liquid separation and low-temperature drying are carried out; and the second hyperspectral reexamination is carried out to remove trace heterogeneous impurities, so that high-purity single-material regenerated plastics are obtained.

Description

Technical Field

[0001] This invention patent relates to the field of control technology, specifically a method for intelligent sorting and purification of multi-material waste plastics using hyperspectral and ultrasonic coupling. Background Technology

[0002] Resource utilization of waste plastics is the core path to solving white pollution and achieving solid waste reduction and resource recovery. Currently, industrial recycled raw materials generally exhibit complex characteristics such as mixed materials, similar colors, heavy surface pollution, soft and hard materials mixed together, and composite layers. Existing industrial processes have long adopted a fixed process of "crushing first, then sorting, and then washing," which has many insurmountable technical bottlenecks, resulting in low-grade recycled plastics, high losses, and poor stability.

[0003] Traditional water washing processes lack material compatibility, are incomplete in removing impurities, and damage materials. Friction washing and hot alkaline washing can only remove surface dust and oil, but cannot remove grease embedded in plastic micropores, printing primer, adhesive residue, and VOC adsorbents. The uniform high-temperature strong alkaline process will cause PET to hydrolyze and turn white, PE / PP to soften and collapse at the edges, and PVC to dechlorinate and age due to heat. The material loss rate is generally high, which seriously restricts the production of high-end recycled materials. Summary of the Invention

[0004] The purpose of this invention is to provide a method for intelligent sorting and purification of multi-material waste plastics using hyperspectral and ultrasonic coupling, in order to solve the problems mentioned in the background art.

[0005] To address the aforementioned technical problems, this invention provides the following technical solution: a method for intelligent sorting and purification of multi-material waste plastics using hyperspectral and ultrasonic coupling, comprising the following steps: S1: Mixed waste plastic raw materials of various materials are conveyed in a single layer through a vibrating material leveling system; S2: It adopts short-wave infrared hyperspectral combined with RGB visual multimodal detection method, identifies material material through deep learning model, and divides raw materials into four categories based on the identification results: single material plastic, mixed materials, harmful plastic and hard impurities, and collects them into eight categories for storage. The eight categories of raw materials include PET, HDPE, LDPE, PP, PS, ABS, PC, and PVC. S21: During the raw material transportation process, spectral cubic data is continuously collected by short-wave infrared hyperspectral camera through push-scan, and material morphology, texture and stain features are collected simultaneously by industrial camera to form dual-modal fusion recognition basic data of spectral molecular features and visual morphological features. The short-wave infrared hyperspectral camera is equipped with a stable infrared supplementary light source. S22: The collected spectral data are sequentially processed by dark current correction, whiteboard correction, SG smoothing filtering, baseline correction and wavelet feature extraction to eliminate detection interference caused by light fluctuations, surface stains and dust. S23: Input the preprocessed feature data into a lightweight one-dimensional CNN model to complete fast inference, accurately distinguish the material type of the raw material, and identify foreign objects such as metal, glass, sand, and rubber mixed in the material. S24: Based on the model recognition results, the pneumatic sorting valve group sorts the raw materials by air and diverts them to the corresponding conveyor branch line, and then transports them to the corresponding storage area by the elevator to complete the classification and collection. S3: The collected single-material plastic material streams are shaped by the appropriate crushing process. After the material particle size is uniform, the free light impurities are removed by air screening and loosening and stabilizing treatment. S31: The single-material raw material in the storage area is conveyed to the crusher via the conveyor line to complete homogeneous crushing and shaping; S32: The crushed raw materials are conveyed to the negative pressure air classifier dust collector, where free dust, paper scraps and light floating impurities are removed by negative pressure airflow to avoid impurities from contaminating the cleaning solution later. S33: The dust-removed raw material is conveyed to the vibrating screen and uniform material machine to complete the vibration loosening and particle size sorting, ensuring that the material in the subsequent process has uniform particle size and consistent state. S4: Based on the material material identified by the pre-sorting, the corresponding ultrasonic cleaning parameters are automatically matched. Through a three-stage gradient ultrasonic cleaning process of low-frequency coarse cleaning, medium-frequency adaptive fine cleaning, and high-frequency fine rinsing, the precise removal of impurities by material is achieved. S41: Low-frequency ultrasonic coarse cleaning uses 20kHz low-frequency ultrasound and 40℃ water temperature, combined with low-foaming neutral environmentally friendly cleaning agent to clean the raw materials. S42: Clean the raw materials by matching exclusive medium-frequency cleaning parameters according to different plastic materials; S43: High-frequency ultrasonic fine rinsing uses an 80kHz ultrasonic frequency and room temperature water to clean the raw materials. The cleaning parameters for different plastic materials are as follows: The PET material is subjected to an ultrasonic frequency of 40kHz and a water temperature of 50℃. The PP, HDPE, and LDPE materials are all manufactured using a 60kHz ultrasonic frequency and a 35℃ water temperature. The PS material is produced using a 40kHz ultrasonic frequency and a 40℃ water temperature. The ABS material is made using a 40kHz ultrasonic frequency and a 45℃ water temperature. The PC material is subjected to an ultrasonic frequency of 40kHz and a water temperature of 45℃. The PVC material is made using an 80kHz ultrasonic frequency and a 30℃ water temperature. S5: After the wet material has been cleaned, solid-liquid separation is completed by extrusion; S6: The dehydrated material is dried and dehumidified by low-temperature hot air drying, and then the dried material is fully covered by a high-spectral equipment for secondary sorting to remove trace amounts of heterogeneous impurities and missorted materials, resulting in high-purity single-material recycled plastic.

[0006] The beneficial effects achieved by this invention patent are as follows: The system employs a three-stage gradient ultrasonic cleaning approach: low-frequency coarse washing, medium-frequency adaptive fine washing, and high-frequency fine rinsing. It can also match exclusive cleaning parameters according to the physical and surface characteristics of eight different plastic materials. The low-frequency coarse washing, combined with environmentally friendly cleaning agents, removes large areas of stubborn stains. The medium-frequency wash uses customized ultrasonic frequencies and water temperatures for different materials, and the high-frequency rinse completes the fine cleaning of residual impurities. At the same time, it sets exclusive temperature and frequency parameters for special P materials, balancing the impurity removal effect with material protection.

[0007] By employing short-wave infrared hyperspectral and RGB visual multimodal fusion detection technology, combined with a constant infrared supplementary light source, it can simultaneously collect the spectral molecular characteristics of materials and their visual characteristics of morphology, texture, and stains. This breaks through the technical bottleneck of single visual detection, which can only identify appearance and cannot distinguish between homogeneous and heterogeneous plastics.

[0008] Through multiple data preprocessing techniques such as dark current correction, whiteboard correction, SG smoothing filtering, baseline correction, and wavelet feature extraction, external interference such as light fluctuations, surface stains, and dust is effectively eliminated.

[0009] A dual quality control system was constructed, consisting of pre-process intelligent sorting and post-process full-coverage secondary spectral sorting, forming a closed-loop purification quality control throughout the entire process. Attached Figure Description

[0010] Figure 1 The flowchart shows the method for intelligent sorting and purification of multi-material waste plastics using hyperspectral and ultrasonic coupling in the embodiments. Detailed Implementation

[0011] The present invention will be further described below with reference to the accompanying drawings.

[0012] like Figure 1 As shown, this invention patent discloses a method for intelligent sorting and purification of multi-material waste plastics using hyperspectral and ultrasonic coupling, comprising the following steps: S1: The raw material is conveyed in a single layer through a vibrating uniform material system; Specifically, the raw materials are vibrated and combed by a frequency-controlled vibrating feeder and then evenly spread onto the belt conveyor line, thereby avoiding the stacking of raw materials. S2: It adopts short-wave infrared hyperspectral combined with RGB visual multimodal detection, identifies material material through deep learning model, and stores four types of material streams: single material plastic, mixed materials, harmful plastic and hard impurities through rear pneumatic sorting valve group. S21: The belt conveyor line transports raw materials and continuously scans and collects spectral cubic data using a short-wave infrared hyperspectral camera. Simultaneously, an industrial camera is mounted to collect material morphology, texture, and stain characteristics, forming a dual-modal fusion recognition of spectral molecular features and visual morphological features. S22: The spectral data undergoes dark current correction, whiteboard correction, SG smoothing filtering, baseline correction, and wavelet feature extraction to eliminate light fluctuations, surface stains, and dust interference. S23: Input a lightweight one-dimensional CNN model to complete fast reasoning, accurately distinguish raw material categories, and identify foreign objects such as metal, glass, sand, and rubber. The raw material categories include PET, HDPE, LDPE, PP, PS, ABS, PC, and PVC; S24: Based on the identification results, the pneumatic sorting valve group sorts the raw materials by air and diverts them to the corresponding conveyor branch line, and then transports them to the corresponding storage area by the elevator. Preferably, the infrared hyperspectral camera is equipped with a constant infrared supplementary light source. The constant infrared supplementary light source is used to eliminate ambient light interference, avoid spectral baseline drift and reflectivity fluctuation caused by natural light, lamp flicker, and uneven brightness in the workshop, and ensure that the spectral data collected each time is consistent and comparable. At the same time, it is used to enhance the imaging of dirty and black materials. Ordinary white light has extremely low contrast in recognizing plastics covered with oil stains, carbon black coloring, and film covering. The infrared light source can penetrate thin layers of floating dust and light oil stains and stably extract the molecular spectral fingerprint of the plastic body. S3: For single-material material flows, appropriate crushing processes are used to shape them, and after unifying the particle size, impurities are removed by air screening, and the material is loosened and stabilized. S31: Raw materials in the storage area are conveyed to the crusher via a conveyor line for homogenization and crushing; S32: After crushing, the raw material is conveyed to the negative pressure air classifier dust collector, where the negative pressure airflow removes free dust, paper scraps, and light floating impurities, preventing impurities from being carried into the cleaning tank and causing contamination of the cleaning solution. S33: After dust removal, the raw materials are conveyed to a vibrating screen and homogenizer for sorting and classifying the loosened material, ensuring uniform particle size and consistent state of the material in subsequent processes. S4: Based on the material material identified by the pre-sorting, the three-level gradient ultrasonic cleaning parameters are automatically matched to achieve precise impurity removal according to material type; S41: Low-frequency ultrasonic coarse cleaning removes surface oil, mud, large pieces of residual adhesive, and thick layers of ink, quickly removing heavy-duty contaminants and protecting the cleanliness of the subsequent fine cleaning tank. Specifically, the low-frequency ultrasonic frequency is 20kHz, combined with a water temperature of 40℃, and a low-foaming, neutral, environmentally friendly cleaning agent is used. S42: Adaptive medium-frequency fine washing based on the current raw material material; The PET material is subjected to an ultrasonic frequency of 40kHz and a water temperature of 50℃ to enhance the microporous oil stains, beverage residues, and peeling of printing primer. Preferably, the water temperature is ≤50℃ to avoid high-temperature hydrolysis and whitening of PET, as well as crystallization and atomization. The PP / PE material is treated with ultrasonic frequency of 60kHz and water temperature of 35℃, using low-temperature gentle cavitation to remove impurities and prevent the soft film from softening, deforming, curling, or sticking. The HDPE / LDPE material uses an ultrasonic frequency of 60kHz and a water temperature of 35℃. HDPE / LDPE material is relatively soft and easily deformed by heat. The use of high frequency and low excitation parameters can thoroughly remove surface oil and printing residues, while effectively avoiding problems such as film curling and sheet softening and adhesion. The PS material is ultrasonically tested at a frequency of 40kHz and a water temperature of 40℃. PS material is brittle, prone to cavitation stress whitening and breakage. Using a moderate frequency and low temperature parameter ensures the cleaning effect while protecting the integrity of the substrate and preventing breakage and powdering. The ABS engineering plastic is suitable for cleaning at 40kHz, 45℃, and 12min. The ABS surface easily adsorbs dust and adhesive residue. This parameter can efficiently remove attached impurities and avoid micro-deformation of the shell and loss of surface gloss caused by high temperature. The PC material uses an ultrasonic frequency of 40kHz and a water temperature of 45℃. The PC material is extremely sensitive to surface fine lines, stains, and residues. The moderate frequency and gentle cavitation can remove oil and dust embedded in the micropores, eliminate surface fogging and micro-scratches, and ensure the transparency of the transparent material. The PVC material is subjected to an ultrasonic frequency of 80kHz, a water temperature of 30℃, and a low-temperature, low-vibration mode to prevent PVC from decomposing, dechlorinating, yellowing, and aging due to heat. It is discharged and collected separately. S43: High-frequency ultrasonic fine rinsing of raw materials, using an ultrasonic frequency of 80kHz and a water temperature of room temperature, small-scale cavitation micro-jet to remove residual agents, micro-particle dust and trace amounts of colloids from the surface, significantly reducing the ash content and residual odor of the finished product. The three-stage gradient ultrasonic cleaning consists of three ultrasonic cleaning tanks connected in series by a continuous material conveyor chain. S5: Dehydration treatment of purified materials; Specifically, the wet material after cleaning first enters a screw extruder for extrusion to achieve solid-liquid separation; S6: The material is dried by low-temperature hot air and then sorted again by high-spectral equipment to remove foreign impurities, resulting in high-purity single-material recycled plastic. S61: After solid-liquid separation, the raw material is fed into a closed drying duct at a constant temperature of 30℃~50℃ by a continuous conveying mechanism for drying and dehumidification. The internal air velocity in the sealed drying duct is 1.2 m / s to 1.5 m / s. S62: The dried finished raw materials pass through the secondary hyperspectral re-inspection station at a uniform speed. The desktop high-precision hyperspectral equipment is used to perform full-coverage material verification on the single material, accurately identifying trace amounts of heterogeneous materials and misclassified impurities remaining from the pre-sorting.

[0013] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent sorting and purification of multi-material waste plastics using hyperspectral and ultrasonic coupling, characterized in that: Includes the following steps: S1: Mixed waste plastic raw materials of various materials are conveyed in a single layer through a vibrating material homogenizing system; S2: It adopts short-wave infrared hyperspectral combined with RGB visual multimodal detection method, identifies material material through deep learning model, and divides raw materials into four categories based on the identification results: single material plastic, mixed materials, harmful plastic and hard impurities, and collects them into eight categories for storage. The eight categories of raw materials include PET, HDPE, LDPE, PP, PS, ABS, PC, and PVC. S3: The collected single-material plastic material streams are shaped using appropriate crushing processes. After unifying the particle size, the material is then removed by air screening, loosening and stabilizing the pressure to remove free light impurities. S4: Based on the material material identified by the pre-sorting, the corresponding ultrasonic cleaning parameters are automatically matched. Through a three-stage gradient ultrasonic cleaning process of low-frequency coarse cleaning, medium-frequency adaptive fine cleaning, and high-frequency fine rinsing, the precise removal of impurities by material is achieved. S5: After the wet material has been cleaned, solid-liquid separation is completed by extrusion; S6: The dehydrated material is dried and dehumidified by low-temperature hot air drying, and then the dried material is fully covered by a high-spectral equipment for secondary sorting to remove trace amounts of heterogeneous impurities and missorted materials, resulting in high-purity single-material recycled plastic.

2. The method for intelligent sorting and purification of multi-material waste plastics using hyperspectral and ultrasonic coupling according to claim 1, characterized in that: Step S2 includes the following sub-steps: S21: During the raw material transportation process, spectral cubic data is continuously collected by short-wave infrared hyperspectral camera through push-scan, and material morphology, texture and stain features are collected simultaneously by industrial camera to form dual-modal fusion recognition basic data of spectral molecular features and visual morphological features. S22: The collected spectral data are sequentially processed by dark current correction, whiteboard correction, SG smoothing filtering, baseline correction and wavelet feature extraction to eliminate detection interference caused by light fluctuations, surface stains and dust. S23: Input the preprocessed feature data into a lightweight one-dimensional CNN model to complete fast inference, accurately distinguish the material type of the raw material, and identify foreign objects such as metal, glass, sand, and rubber mixed in the material. S24: Based on the model recognition results, the pneumatic sorting valve group sorts the raw materials by air and diverts them to the corresponding conveyor branch line, and then the materials are transported to the corresponding storage area by the elevator to complete the classification and collection.

3. The intelligent sorting and purification method for multi-material waste plastics using hyperspectral and ultrasonic coupling as described in claim 2, characterized in that: The short-wave infrared hyperspectral camera is equipped with a constant infrared supplementary light source.

4. The intelligent sorting and purification method for multi-material waste plastics using hyperspectral and ultrasonic coupling as described in claim 1, characterized in that: Step S3 includes the following sub-steps: S31: The single-material raw material in the storage area is conveyed to the crusher via the conveyor line to complete homogeneous crushing and shaping; S32: The crushed raw materials are conveyed to the negative pressure air classifier dust collector, where free dust, paper scraps and light floating impurities are removed by negative pressure airflow to avoid impurities from contaminating the cleaning solution later. S33: The dust-removed raw materials are conveyed to the vibrating screen and uniform material machine to complete the vibration loosening and sorting of particles, ensuring that the material particles are uniform in size and consistent in state in subsequent processes.

5. The intelligent sorting and purification method for multi-material waste plastics using hyperspectral and ultrasonic coupling as described in claim 1, characterized in that: The three-stage gradient ultrasonic cleaning in step S4 consists of three independent ultrasonic cleaning tanks connected in series. The cleaning tanks are connected by a continuous material conveyor chain to achieve continuous material transport. Step S4 includes the following sub-steps: S41: Low-frequency ultrasonic coarse cleaning uses 20kHz low-frequency ultrasound and 40℃ water temperature, combined with low-foaming neutral environmentally friendly cleaning agent to clean the raw materials. S42: Clean the raw materials by matching exclusive medium-frequency cleaning parameters according to different plastic materials; S43: High-frequency ultrasonic fine rinsing uses 80kHz ultrasonic frequency and room temperature water to clean the raw materials.

6. The method for intelligent sorting and purification of multi-material waste plastics using hyperspectral and ultrasonic coupling according to claim 5, characterized in that: The cleaning parameters for different plastic materials in step S42 are as follows: The PET material is subjected to an ultrasonic frequency of 40kHz and a water temperature of 50℃. The PP, HDPE, and LDPE materials are all manufactured using a 60kHz ultrasonic frequency and a 35℃ water temperature. The PS material is produced using a 40kHz ultrasonic frequency and a 40℃ water temperature. The ABS material is made using a 40kHz ultrasonic frequency and a 45℃ water temperature. The PC material is subjected to an ultrasonic frequency of 40kHz and a water temperature of 45℃. The PVC material is prepared using an 80kHz ultrasonic frequency and a 30℃ water temperature.

7. The intelligent sorting and purification method for multi-material waste plastics using hyperspectral and ultrasonic coupling as described in claim 6, characterized in that: The PET material is used in water with a temperature of ≤50℃.

Images