Method for detection of foreign matters in waste paper pulp

Abstract

To detect existence of contaminated foreign matters and kinds and sizes thereof with high probability in a paper making step using a waste paper raw material contaminated with foreign matters including taboo articles.SOLUTION: A near-infrared spectrum of a waste paper pulp slurry obtained by using a waste paper raw material or a pulp mat obtained from a waste paper pulp slurry is acquired by using a hyper-spectral camera, and this spectrum is analyzed to detect foreign matters in the waste paper pulp slurry.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a method for detecting foreign matter in a process for preparing waste paper pulp from waste paper raw materials. [Background technology]

[0002] In today's paper mills, waste paper is an important raw material, and new paper products are often manufactured using the waste paper pulp made from it. However, the recovered waste paper raw material inevitably contains foreign matter. While some foreign matter does not cause any problems in the papermaking process even if it is present in waste paper pulp, there are also taboo substances that can lead to a decrease in quality even in small amounts. For example, in the case of laminated paper, which is likely to be entrained in paper during the recycling process, the laminated film is a taboo substance. If these taboo substances are not properly removed, they can increase the burden on the waste paper pulp manufacturing process and cause a decrease in the quality of the manufactured products.

[0003] Therefore, various methods for detecting and removing foreign matter in the collection and sorting processes of waste paper used as a papermaking raw material, or in the papermaking process using waste paper, have been studied.Patent Document 1 proposes a method for detecting foreign matter, which is an impurity in bleached pulp used as a papermaking raw material, by irradiating light onto a pulp sheet obtained from the pulp, converting the transmitted or reflected light into the area of ​​defects, and comparing it with a preset control value to classify and detect the occurrence of defects and remove the defective products.

[0004] Patent document 2 proposes a device that connects an inspection tube to the flow path of waste paper pulp slurry that flows while still containing foreign matter, shines a fluorescent lamp onto the tube, receives the transmitted light, and then images the foreign matter with a linear CCD camera.

[0005] Patent Document 3 proposes a method for producing waste paper pulp by crushing waste paper packaging containing prohibited materials and determining whether the cellulose component is dominant or not based on the absorbance spectrum measured on the crushed material using a hyperspectral camera. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Publication No. 03-140849 [Patent Document 2] Japanese Patent Application Publication No. 06-43104 [Patent Document 3] Patent No. 7105632 Summary of the Invention [Problem to be solved by the invention]

[0007] However, although Patent Document 1 uses transmitted or reflected light, and Patent Document 2 uses transmitted light to detect the number and area of ​​foreign matter, neither method can identify the type of foreign matter detected. As a result, regardless of the type of foreign matter found, the final step is to remove it, which unnecessarily reduces manufacturing efficiency.

[0008] However, the method of Patent Document 2 detects foreign matter in the flow path of the waste paper pulp slurry, and the main foreign matter assumed to be carbon particles that were not removed during the deinking process of the waste paper, and it was difficult to detect other components.

[0009] The method in Patent Document 3 aims for detection upstream in the manufacturing process compared to the inspection tube in Patent Document 2. Since detection is performed at a very upstream stage, including the addition of a process not typically performed, namely, disintegration of collected waste paper packaging, if waste paper containing foreign matter can be separated at this stage, downstream impacts can be prevented. However, at the disintegration stage, the loose paper bundles are spread out on a conveyor (Patent Document 3, paragraphs

[0033] to

[0050] ). Even if the paper bundles are aligned on the conveyor without overlapping at all, only the side facing the hyperspectral camera can be distinguished; if the opposite side is laminated, detection is not possible. While the side facing the camera can distinguish between those containing predominantly cellulose components and those not, it is not possible to detect the reverse side in the case of laminated film, etc. Furthermore, near-infrared spectra are measured under conditions with little influence from moisture, and the waveforms can only distinguish between raw materials containing predominantly cellulose components and those not, making it difficult to identify the type of foreign matter.

[0010] Therefore, the object of this invention is to enable the detection of the presence of contaminated foreign matter, including the type and size, with a high probability in papermaking processes that use waste paper raw materials that have contaminated foreign matter, including prohibited items. [Means for solving the problem]

[0011] This invention solves the above problem by providing a first solution, which is a foreign matter detection method that uses a hyperspectral camera to obtain a near-infrared spectrum of a waste paper pulp slurry obtained using waste paper raw materials or a pulp mat obtained from the waste paper pulp slurry, and analyzes the near-infrared spectrum to detect foreign matter in waste paper pulp.

[0012] In the foreign object detection method according to the present invention, in addition to the first solution, a second solution can be selected in which, as the analysis, discrimination is performed after dimensionality reduction.

[0013] Furthermore, the papermaking method of the present invention can implement a third solution, which is a papermaking method in which foreign matter contained in the waste paper pulp slurry is detected using the foreign matter detection method, which is the first or second solution, and the foreign matter is removed from the waste paper pulp slurry according to the composition of the detected foreign matter, and paper is made. [Effects of the Invention]

[0014] The foreign matter detection method of this invention is applied to wastepaper pulp slurry or the pulp mat obtained therefrom, so it is less likely that foreign matter will be hidden in the shadow and remain undetected. In other words, because the wastepaper packaging is once disintegrated, the measurement is representative of the entire amount and allows for real-time measurement. Because the measurement target is the dispersed wastepaper pulp slurry or the pulp mat obtained therefrom, even a portion of the sample represents the entire amount. If foreign matter is sufficiently dispersed within the slurry, even a small amount of moisture can be detected by analyzing the near-infrared spectrum obtained with a hyperspectral camera, i.e., even if the pulp concentration (= solids content) is quite low. Of course, in this case, it is preferable to concentrate the sample to further increase the detection rate.

[0015] Furthermore, if measurements can be made in real time, it is particularly responsive, and if any foreign matter that is a taboo item is detected, it can be removed before the papermaking process, which is highly useful for quality control in the papermaking process, so it is preferable to do so in real time or as close to it as possible.As an analysis method, if discrimination is performed after dimensional compression, the amount of data becomes significantly smaller due to dimensional compression, making high-speed processing possible at the time of final discrimination, and real-time processing is fully possible even on terminals or servers with general computing power. [Brief explanation of the drawings]

[0016] [Figure 1] FIG. 1 is a flow chart illustrating an embodiment of a foreign object detection method according to the present invention. [Figure 2] An example diagram of an embodiment in which a hyperspectral camera is used to measure waste paper pulp slurry. [Figure 3](a) A graph showing an example of the near-infrared spectrum of a waste paper pulp slurry having a solid content of 1% by mass, (b) A graph showing an example of the near-infrared spectrum of a waste paper pulp slurry having a solid content of 4% by mass in an example, (c) A graph showing an example of the near-infrared spectrum of a wet paper having a solid content of 10% by mass before dewatering in an example, (d) A graph showing an example of the near-infrared spectrum of a wet paper having a solid content of 25% by mass after dewatering in an example. [Figure 4] Hyperspectral camera images of wet paper containing polyethylene film pieces in the example [Figure 5] (a) A graph showing an example of a near-infrared spectrum used as training data, and (b) a graph showing an example of a second-order derivative spectrum obtained by smoothing (a) and performing second-order differentiation. [Figure 6] A graph that can be classified by performing principal component analysis to compress the dimensions of three types of training data into two components. [Figure 7] A copy of Figure 4 and an image that can be clearly distinguished using linear discriminant analysis [Figure 8] Images showing foreign matter detection for each example where the solid content of the sample was changed DETAILED DESCRIPTION OF THE INVENTION

[0017] The present invention provides a method for detecting foreign matter by analyzing near-infrared spectra of a wastepaper pulp slurry or a pulp mat obtained therefrom using a hyperspectral camera during a papermaking process in which wastepaper pulp is prepared from wastepaper raw materials.

[0018] An embodiment of a papermaking process including the foreign matter detection method according to the present invention is shown in Fig. 1. A collection of collected waste paper raw material 11 (waste paper packing) is broken into small pieces by adding water in a pulping process S12, and becomes a slurry, which becomes waste paper pulp slurry 12.

[0019] The foreign matter detection process S21, which is the core of the foreign matter detection method of this invention, is inserted and executed in any of the processes following the dust removal process S31, which mainly removes heavy foreign matter from the waste paper pulp slurry 12, and before the papermaking process S35, which uses the slurry to make paper. Figure 1 shows multiple examples of moving from any of the processes to the foreign matter detection process S21 and then returning to the original process. In other words, the types of lines leading to and exiting the foreign matter detection process S21 each represent a different embodiment.

[0020] In this foreign matter detection step S21, a hyperspectral camera 21 acquires near-infrared spectra of the waste paper pulp slurry 12 prepared in the pulping step S12 or the area where the slurry or a pulp mat obtained by concentrating the slurry flows in any of the steps. The hyperspectral camera 21 is a spectroscopic camera that can record light of different wavelengths at individual observation points. The hyperspectral camera 21 used in this invention is preferably capable of dispersing and recording light in any range in the near-infrared region (wavelengths of 700 nm to 2500 nm), and more preferably capable of covering the 900 nm to 2000 nm range. This is because important components, which are prohibited in the papermaking process and will be described later, are easily detected by spectra in this wavelength range. However, to enable analysis by wavelength, a hyperspectral camera 21 capable of dispersing and recording in 100 or more bands in the measurement wavelength range is desirable. However, depending on the level of foreign matter contamination, a multispectral camera with fewer bands may also be used.

[0021] Obtaining near-infrared spectra for the entire waste paper pulp slurry 12 or the pulp mat is desirable from the standpoint of accuracy. However, in this case, the amount of data becomes so large that real-time processing cannot keep up, and the amount of data to be stored may become too large. In addition to an embodiment in which the entire waste paper pulp slurry 12 or the pulp mat is subjected to the foreign matter detection step S21 in each route of Figure 1, an embodiment in which a portion of the waste paper pulp slurry 12 or the pulp mat is extracted and used as a representative for foreign matter detection may also be used. If foreign matter is sufficiently dispersed in the waste paper pulp slurry 12 or the pulp mat during the process up to detection, foreign matter detection can be performed with sufficient accuracy even by extracting a portion. Furthermore, if a portion is extracted, and the extracted amount is sufficiently small compared to the total, the waste paper pulp slurry 12 or the pulp mat used for detection does not need to be returned between the original processes. Furthermore, information that foreign matter has been detected in the foreign matter detection step S21, and information on the type, size and amount of foreign matter contained may be used to adjust the execution conditions from the dust removal step S31 to the papermaking step S35.

[0022] An example of the situation in which the hyperspectral camera 21 acquires a near-infrared spectrum of the wastepaper pulp slurry 12 is shown in the schematic diagram of Figure 2. The area where the wastepaper pulp slurry 12 and pulp mat flow is set within the angle of view of the hyperspectral camera 21. In this way, a near-infrared spectrum is acquired in real time for the flowing wastepaper pulp slurry 12 and pulp mat. The near-infrared spectrum thus acquired is data in a wavelength range that includes at least a portion of the near-infrared range described above. In particular, to detect particularly significant contraindications, such as waxed or laminated products, it is preferable to include a wavelength range of 1000 nm to 1400 nm or 1500 nm to 1800 nm, and even more preferably 1000 nm to 1400 nm. These ranges suppress the influence of the water content, which accounts for the majority of the wastepaper pulp slurry, even in wastepaper pulp slurries with a low solids content, making it easier to detect peaks specific to contraindications. This tendency is particularly pronounced in the 1000 to 1400 nm range. By analyzing near-infrared spectra that cover these wavelength ranges, the influence of moisture can be further reduced, making it possible to classify each prohibited component in more detail than with conventional visible light cameras, which were only able to distinguish between pulp and non-pulp.In addition, because the waste paper is first made into a pulp slurry before observation, it is possible to observe and capture hot melt adhesives and other substances in the slurry, which do not easily come to the surface when the paper is in its original state.

[0023] The obtained near-infrared spectrum is analyzed to determine the presence of foreign matter. The analysis is performed automatically using a computer. Specific analysis methods may include statistical techniques that perform multivariate analysis, or determination using a classifier that has learned machine learning on information about the presence of foreign matter. These analyses may be performed by an analysis device 22 connected to the hyperspectral camera 21 directly or via a network. The analysis device 22 may be a general computer with the necessary programs installed.

[0024] Among the above analyses, the multivariate analysis method is not particularly limited as long as it can effectively extract information contained in the near-infrared spectrum. For example, principal component analysis and partial least squares regression, which are dimension reduction methods, are used. The variables after dimension reduction may be one or two components, or three or more components if many types of contaminants are expected. Furthermore, dimension reduction is not particularly necessary if calculation speed can be ensured, but it is desirable to perform dimension reduction using principal component analysis or the like to the extent necessary for real-time measurement. The results obtained in this way may be combined with a discriminant analysis method such as linear discriminant analysis to determine whether or not there is a contaminant that requires action.

[0025] When using a machine-learned classifier in the above analysis, it is possible to distinguish between materials with very similar chemical structures, such as polyethylene and wax, by using discrimination data under the same conditions as training data for the classifier. Machine learning methods include support vector machines and K-nearest neighbor methods, and deep learning such as neural networks can also be used.

[0026] The analyzer 22 or a papermaking process control device (not shown) connected to it issues an alarm or issues instructions to each process in the papermaking process based on the results of the determination. From the rough selection process S32 to the beating process S34, operating conditions are adjusted according to the type, size, and amount of foreign matter detected. To appropriately reduce the amount of foreign matter that reaches the next process or to refine the foreign matter in the paper product to an acceptable level, for example, the amount of pulp (reject) containing a large amount of foreign matter that is screened and discharged outside the system is adjusted in the rough selection process S32 and the refining process S33, and the processing time and load are adjusted in the beating process S34.

[0027] Heavy foreign matter such as metal fragments is removed in the dust removal process S31 from the waste paper pulp slurry 12. If the foreign matter detection process S21 is inserted before the dust removal process S31 at this time, any detected prohibited foreign matter that can be removed is also removed.

[0028] Smaller foreign matter that could not be removed in the dust removal process S31, hot melt adhesive, resin fragments, etc., are removed using a screen or the like in the rough selection process S32 and the refinement process S33. If necessary, a deinking and bleaching process may be inserted into this waste paper pulp slurry to remove ink, paint, etc. contained in the waste paper. However, the rough selection process S32 may be omitted depending on the condition of the raw material used and the type of paper to be made. The pulp fibers are then beaten in the beating process S34, and paper is made in the papermaking process S35 to obtain paper made from waste paper raw materials.

[0029] Paper manufactured using the foreign matter detection method of this invention can be made so that the presence of foreign matter is properly and precisely identified and identified during the manufacturing process, and then these foreign matter is properly removed. This reduces the number of products that do not meet specifications compared to conventional methods, and also reduces overall manufacturing costs.

[0030] Another embodiment of the foreign matter detection method according to the present invention will be described. In this embodiment, a portion of the waste paper pulp slurry 12 obtained in the macerating process S12 shown in FIG. 1 is extracted and temporarily paper-made to prepare a sample in a form that can be observed with high precision using a hyperspectral camera 21 before any of the processes from the dust removal process S31 to the papermaking process S35. This sample is then subjected to a foreign matter detection process S21 in which the hyperspectral camera 21 observes the near-infrared spectrum. Because paper made from the slurry is not a laminate in which foreign matter wraps around to the back side as viewed from the camera, the entire sample can be observed with high precision using the hyperspectral camera 21. The results of analyzing the near-infrared spectrum thus obtained are reflected in the rough screening process S32, refining process S33, and beating process S34 of the extracted original waste paper pulp slurry 12, and the subsequent papermaking process is carried out. However, since it takes time to make paper after extraction, this method is suitable for temporary waiting before any other process. However, if equipment that automatically prepares samples is introduced, it will be possible to analyze the paper while it is being made in near real time.

[0031] In all embodiments, after the waste paper pulp slurry 12 is obtained in the defibration step S12, in each step prior to the papermaking step S35, the waste paper pulp slurry 12 or pulp mat is often handled with a solids content of 3% to 30% by mass, but by measuring at a higher solids content, the influence of moisture can be reduced and the detection rate of foreign matter by the hyperspectral camera 21 can be further increased. By setting the solids content to 10% by mass or more, the improvement in detection rate due to this concentration can be suitably exhibited.

[0032] Furthermore, in addition to these embodiments, an exclusion process may be provided before or after any of the dust removal process S31, rough screening process S32, and fine screening process S33, in which, if foreign matter that cannot be completely removed by these processes is found, the entire waste paper pulp slurry 12 is excluded from the papermaking process. The foreign matter detection method of the present invention can even determine the type of foreign matter, making it possible to switch responses depending on the type of foreign matter found.

[0033] Furthermore, among the waste paper pulp slurries 12 thus removed, those with a relatively low foreign matter content may be mixed with other waste paper pulp slurries 12 that have been detected to have a low foreign matter content, and returned to the beating process S34 or the papermaking process S35. This is because the type, size, and amount of foreign matter that is acceptable varies depending on the application, and in some cases, a certain amount of foreign matter can be tolerated. [Example]

[0034] Next, an example will be given to confirm the discrimination performance of the foreign object detection method according to the present invention.

[0035] Example 1 Near-infrared spectra were measured for waste paper pulp (Sample 1) prepared from recycled cardboard and waste paper pulp (Sample 2) prepared from waxed recycled cardboard containing the prohibited substance "wax" as a foreign matter.

[0036] <Pulp slurry comparison test> In accordance with JIS P 8220, pulp slurries (Samples 1 and 2) were prepared from each waste paper material using a pulp disintegrator (Nippon TMC Co., Ltd.) to a solids content of 4% by mass. Near-infrared spectra (900–2500 nm) of the slurries were measured using a hyperspectral camera (JFE Techno-Research Corporation, ImSpector), and the 949–2005 nm range was used for analysis. The analysis involved principal component analysis (PCA) as a multivariate analysis, followed by dimensionality reduction and linear discriminant analysis. Since the solids content of slurries in typical papermaking processes is approximately 1% by mass, the slurries for Samples 1 and 2 were diluted with water to prepare slurries with a solids content of 1% by mass. The near-infrared spectra are shown in Figure 3(a), and the near-infrared spectra of the slurries for Samples 1 and 2 concentrated to a solids content of 4% by mass are shown in Figure 3(b). The difference in the spectra of Samples 1 and 2 increased as the solid content increased, and in particular in Figure 3(b) it was difficult to detect differences between 1500nm and 1800nm ​​due to the influence of water contained in the slurry, but at 1000nm to 1400nm, a clear increase in reflectance was observed in waxed recycled cardboard. This difference confirmed that by performing discrimination after dimensionality reduction, it is possible to distinguish cases where waxed cardboard is included in recycled paper pulp slurry.

[0037] <Wet paper contrast test> According to JIS P 8222, handsheets (10% solids by mass) were prepared from the slurries of Samples 1 and 2. The near-infrared spectra of the wet sheets before dewatering were measured under the same conditions as in the pulp slurry comparison test described above. The wet sheets were then dewatered using absorbent filter paper (Toyo Roshi Kaisha, No. 26WA, 260 mm x 260 mm) until the solids content reached approximately 25% by mass. The near-infrared spectra of the resulting wet sheets after dewatering were also measured. The near-infrared spectra are shown in Figures 3(c) and (d). Analysis of the near-infrared spectra obtained in this way under the same conditions as in the pulp slurry comparison test described above revealed that the reduction in the amount of water relative to the solids not only enhanced the overall reflectance and made the distinction easier, but also clearly revealed the difference between waxed and non-waxed recycled corrugated cardboard, even in the 1500-1800 nm wavelength range. This increased the number of discernible elements, confirming the improved ability to distinguish contaminants.

[0038] <Example of multivariate analysis of near-infrared spectral images> Regular cardboard was mixed with prohibited waxed cardboard, and pieces of polyethylene (PE) film were added to prepare pulp, which was then used to create wet paper.A hyperspectral camera (as above) was then used to confirm whether the prohibited material could be identified.

[0039] Normal recycled corrugated cardboard and waxed recycled corrugated cardboard were mixed in a weight ratio of 5:1 and disintegrated in accordance with JIS P 8220 to prepare a mixed pulp slurry with a solid content of 4% by mass. Furthermore, 10 μm thick PE films cut into 1 cm square pieces (at least 10 pieces per sheet of handmade paper to be produced) were added to this mixed pulp slurry and disintegrated, and the resulting mixture was mixed to a basis weight of 120 g / m according to JIS P 8222. 2 A handmade paper was prepared. It was then dehydrated using absorbent filter paper until the solid content reached 25% by mass, and the resulting wet paper was imaged using the hyperspectral camera. A photo taken using this spectrum is shown in Figure 4. However, because it is difficult to distinguish between the components in this image using visible light alone, the following data processing method was used for analysis.

[0040] From the near-infrared spectra (Figure 5(a)) constituting the obtained image, 10 points each were selected as training data for "pulp," "PE," and "wax." These were then standardized and preprocessed using a Savitzky-Golay filter, smoothing, and second-order differentiation. Figure 5(b) shows the second-order derivative spectra after this preprocessing stage. The characteristic behavior of "wax" around 1200 nm and 1700 nm, which was difficult to distinguish from other components before preprocessing, was now clearly distinguishable, as confirmed by visual inspection. Similar multivariate analysis was performed on each of the 10 data points using two-component principal component analysis. The results are shown in Figure 6, where each component is plotted on a separate axis. This analysis reduced the dimensions to two principal components. It was confirmed that the data for "pulp," "PE," and "wax" could all be grouped and clearly distinguished and recognized using linear discriminant analysis.

[0041] <Image analysis example using linear discriminant analysis> As described above, the discrimination process that classifies the data into three types - "pulp," "PE," and "wax" - using two-component principal component analysis was applied to the entire near-infrared spectrum image (Figure 4; for comparison, it is also shown on the left of Figure 7) to obtain an image showing the distribution of "pulp," "PE," and "wax" on the wet paper. A black and white version of this image is shown on the right of Figure 7. The entire image is "pulp," the larger bright lumps are "PE," and the scattered dark areas distributed throughout are "wax." In actual image processing, these are displayed in color to make them clearly distinguishable.

[0042] <Confirmation of detection accuracy due to increase in solid content> A mixed pulp slurry was prepared using the same procedure as the sample used in the "Example of Multivariate Analysis of Near-Infrared Spectroscopic Images" above, except that PE film was not added. This mixed pulp slurry was placed in a rectangular plastic container with a 27 cm x 35 cm base, and wet paper samples with different solid content were prepared by repeatedly absorbing water using absorbent filter paper in stages. These wet paper samples were imaged using the hyperspectral camera. Figure 8 shows images of each sample, processed using the same data processing method as in the "Example of Image Analysis Using Linear Discriminant Analysis" above. The number of detected "wax (colored portions)" was counted. As shown in Table 1 below, although the samples were prepared from the same mixed pulp slurry, the number of detectable "wax" particles increased with increasing solid content. Detection accuracy was particularly improved when the solid content exceeded 10% by mass. Furthermore, although the number of detectable particles increased significantly when the solid content exceeded 30% by mass, handling the wet paper samples became significantly more difficult.

[0043] [Table 1] [Explanation of symbols]

[0044] 11. Recycled paper raw materials 12 Waste paper pulp slurry 21 Hyperspectral Camera 22 Analyzer S12 Disintegration process S21 Foreign object detection process S31 Dust removal process S33 Selection process S34 Beating process S35 Paper making process

Claims

1. A waste paper pulp slurry obtained using a waste paper raw material or a pulp mat obtained from the waste paper pulp slurry, A foreign matter detection method for detecting foreign matter in waste paper pulp by acquiring a near-infrared spectrum using a hyperspectral camera and analyzing the near-infrared spectrum.

2. 2. The foreign object detection method according to claim 1, wherein the analysis method comprises performing discrimination after dimensionality reduction.

3. A papermaking method in which foreign matter contained in the waste paper pulp slurry is detected using the foreign matter detection method described in claim 1 or 2, and the foreign matter is removed or refined from the waste paper pulp slurry according to the composition of the detected foreign matter to produce paper.

Images