JP2002090299A - Method for distinguishing grade of high-molecular material - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for more easily and speedily distinguishing the grade of a high-molecular material than a conventional method for measuring mid-infrared spectra. SOLUTION: The near-infrared spectra of the high-molecular material of which the grade is to be distinguished are measured in the wavelength range of near-infrared rays. Data analysis techniques (a neural network, main component analysis method, PLS analysis method, hierarchical cluster analysis method, SIMCA method, etc.), in chemometrics are applied to the near-infrared spectrum data to distinguish the grade of the high-molecular material on the basis of information acquired by the data analysis techniques. In the case of distinguishing the grade (HDPE, LDPE) of polyethylene, near-infrared spectra are measured in the range of 1.6-2.0 μm. The range of 1.6912-1.7783 μm is especially preferable.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for identifying a grade of a polymer material such as plastic, fiber, rubber, paint, adhesive or the like based on near infrared spectrum data. is there.

[0002]

2. Description of the Related Art In Japan, more than 15 million tons of polymer materials are produced every year as plastics, fibers, rubbers, paints, adhesives, etc., but in recent years, recycling has become an essential issue due to environmental problems. In the recycling of polymer materials, in order to improve the properties of recycled materials,
In addition to sorting polymer materials by type, it is also required to sort polymer materials of the same type according to their properties, that is, grades. This is because the price of recycled materials varies greatly depending on the grade, for example, high-density polyethylene (HDPE) is more than twice as expensive as low-density polyethylene (LDPE).

[0003] Conventionally, methods for identifying the grade of a polymer material at the laboratory level include gel permeation chromatography (GPC), pyrolysis gas chromatography (Pyrosis-GC, PyGC), light scattering, osmotic pressure, and the like. Vapor pressure, ultracentrifugation, electron microscope, X-ray and neutral diffraction, NM
R spectrum (13C-NMR and solid-state NMR), infrared spectrum (IR), elemental analysis, atomic absorption analysis, X-ray photoelectron spectroscopy, secondary ion mass spectrometry, thermogravimetry (TG), differential thermal analysis (DTA), An analysis method that combines destructive and nondestructive measurement such as differential scanning calorimetry (DSC) is used. However, these methods, although capable of identifying the grade of the polymer material, lack the quickness of measurement and identification of the grade, and thus have reached a practical level that can be used at the time of actual recycling at present. There was nothing.

[0004]

The most promising method among the above-mentioned methods for identifying the grade of a polymer material, which is closest to the practical use level, is considered to be a method utilizing infrared rays. When the polymer material is irradiated while changing the wavelength of the infrared light, the ratio of infrared light absorbed or reflected by the polymer material changes depending on the wavelength. When the absorption or reflectance of the infrared ray is measured while changing the wavelength of the irradiation light, an infrared spectrum is obtained. Since the infrared spectrum changes depending on the type and grade of the monomer used for the polymer material, the type and grade of the polymer material can be identified from the difference in the infrared spectrum.

The infrared spectrum is a near-infrared spectrum (wavelength range: 1.0 to
2.5 μm), mid-infrared spectrum (wavelength range: 2.5
~ 25μm), far infrared spectrum (wavelength range: 25 ~
100 μm). Among these,
Since the absorptance of the mid-infrared ray differs depending on the grade of the polymer material, it is considered that the grade of the polymer material can be distinguished relatively accurately based on the mid-infrared spectrum. However, in the mid-infrared wavelength range, since the light absorption of the polymer material is considerably high, it is necessary to perform a transmission method measurement after performing a pretreatment to reduce the thickness of the sample (object to be measured). . Therefore, the method of measuring the infrared spectrum in the wavelength range of the mid-infrared rays is a destructive inspection, and requires a complicated work of pre-treating the sample, which is a simple and quick grade identification method. Did not.

[0006] The problem of identifying polymer grades in a simple and rapid manner is the problem of efficient recycling of all polymer materials such as plastics, fibers, rubbers, paints and adhesives. It can also occur.

The present invention has been made in view of the above problems, and an object of the present invention is to provide a simpler and faster method for identifying the grade of a polymer material as compared with a conventional method for measuring a mid-infrared spectrum. To provide.

[0008]

In order to achieve the above object, the invention of claim 1 measures a near-infrared spectrum in a near-infrared wavelength range for a polymer material to be graded, and obtains the near-infrared spectrum. A grade identification method for a polymer material characterized by applying a data analysis technique in chemometrics to data and identifying a grade of the polymer material based on information acquired by the data analysis technique. Here, "chemometrics" is intended to apply mathematical and statistical methods to plan and select optimal procedures and optimal experimental plans, and to maximize the amount of chemical information obtained from chemical data. (See, for example, "Chemometrics-New Analytical Chemistry", Maruzen, p. 1) by Tetsuro Aijima. In addition, the above "Data analysis method in chemometrics"
In addition to neural networks, principal component analysis, PLS (Partial Least Squares Regression),
PCR (Principal Components Regression) method, hierarchical cluster analysis method, SIMCS (Soft Independent M
odeling of Class Analogy method, KNN (k nearest ne)
ibors) method (see, for example, Yoshikatsu Miyashita and Shinichi Sasaki, "Chemometrics Chemical Pattern Recognition and Multivariate Analysis", Kyoritsu Shuppan).

In the method for identifying a grade of a polymer material according to the first aspect, a near-infrared ray (wavelength range: 1.0) including a wavelength having a different light absorptance or reflectance depending on the grade of the polymer material.
２．５2.5 μm), it is possible to obtain near-infrared spectrum data including information on the grade of the polymer material. Moreover, the near infrared rays
Unlike mid-infrared rays (wavelength range: 2.5-25 μm)
Since it is not excessively absorbed by the polymer material and has transmitted light and reflected light of moderate intensity, it is not necessary to perform pretreatment on the polymer material to be measured, and the infrared spectrum of sufficient intensity Data can be obtained. Then, by applying a data analysis method in chemometrics to the data of the near-infrared spectrum, information on the grade of the polymer material is extracted from the data of the near-infrared spectrum. The grade of the polymer material is identified based on the information obtained by this data analysis technique. Since data processing using a computer can be easily applied to the data analysis technique in chemometrics, simple and quick grade identification can be performed.

According to a second aspect of the present invention, in the method for identifying a grade of a polymer material according to the first aspect, the data analysis method in the chemometrics includes sharpening of peaks, correction of a baseline inclination, and correction of the near-infrared spectrum data. It is characterized in that pre-processing for performing processing including normalization and data processing by a neural network for near-infrared spectrum data subjected to the pre-processing are performed. Here, the "neural network" is a mathematical model that mimics information processing performed by the brain, and is a network that performs data processing performed by replacing data processing performed by nerve cells with processing elements called units and interconnecting them. It is. The above-mentioned near-infrared spectrum data processing is a neural network having a three-layer structure including an input layer, an intermediate layer, and an output layer, in which a hierarchical perseprotron model is adopted, and an error-back propagation method. The neural network learned in (1) is preferable.

In the method for identifying a grade of a polymer material according to a second aspect, pre-processing including second-order differentiation processing and normalization processing is performed on the near-infrared spectrum data.
In addition to sharpening peaks in the near-infrared spectrum, variation in data among a plurality of polymer materials to be graded is suppressed, and the accuracy of final grade identification is increased. The pre-processed near-infrared spectrum data is input to a neural network, and grade identification is performed based on the data output from the neural network. In the case of the analysis method using the neural network, by optimizing the parameters of the neural network, the accuracy of the grade identification is improved, and even when the polymer material is classified into the subdivided grades, each grade can be identified. Becomes possible. Even when the near-infrared spectrum data has strong nonlinearity, the grade of the polymer material can be identified with high accuracy.

According to a third aspect of the present invention, in the method for identifying a grade of a polymer material according to the first aspect, as the data analysis method in the chemometrics, sharpening of a peak, correction of a slope of a baseline, and correction of the near-infrared spectrum data are performed. Pre-processing for performing processing including normalization, and principal component analysis, PLS analysis, hierarchical cluster analysis, and SI
And performing data processing by at least one of the MCA method.

In the method for classifying a polymer material according to a third aspect, as described above, the pretreatment is performed to sharpen peaks in a near-infrared spectrum and to perform a plurality of polymer material classification targets. To reduce the variation in data between the two, and improve the accuracy of final grade identification. The pre-processed near-infrared spectrum data is subjected to data processing by at least one of the above-described principal component analysis method, PLS method, hierarchical cluster analysis method, and SIMCA method, and grade identification is performed based on the output result. I do. In particular, by combining two or more of the principal component analysis method, the PLS method, the hierarchical cluster analysis method, and the SIMCA method, and comparing the analysis results, it is possible to improve the accuracy of the grade identification of the polymer material.

According to a fourth aspect of the present invention, in the method for identifying a grade of a polymer material according to the first, second or third aspect, the wavelength range of the near-infrared spectrum data is 1.6 to 2.0 μm. The target polymer material is polyethylene.

According to the fourth aspect of the present invention, there is provided a method for identifying a grade of a polymer material, wherein a wavelength range of 1.6 to 2.0 μm in which a portion having significantly different optical characteristics due to a difference in density of polyethylene exists.
By using the data of the near-infrared spectrum of the above, the grade due to the difference in the density of polyethylene is identified.

According to a fifth aspect of the present invention, in the method for identifying a grade of a polymer material according to the fourth aspect, the wavelength range of the near-infrared spectrum is 1.6912 to 1.7773 μm.

In the above-mentioned wavelength range of 1.6 to 2.0 μm, especially in the wavelength range of 1.6912 to 1.7778 μm, the change in optical characteristics due to the difference in density of polyethylene is large. Therefore, in the method for classifying a polymer material according to claim 5, the wavelength range of 1.6912 to 1.7778 μm is used.
By performing near-infrared spectrum measurement and data processing limited to m, the grade due to the difference in density of polyethylene is identified.

[0018]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment in which the present invention is applied to a grade identification method for identifying a grade of polyethylene will be described. Polyethylene is mass-produced as one of general-purpose plastics, which is a polymer material. Polyethylene is light (specific gravity: 0.91 to 0.97), transparent to translucent, and cold-resistant (usable temperature range: -6
0-80 ° C), and the electrical insulation is particularly excellent.
High withstand voltage and good high frequency characteristics. It does not absorb or permeate water, but has the property of passing air. In addition, polyethylene is excellent in water resistance, oil resistance, acid resistance (excluding concentrated nitric acid), alkali resistance, and organic solvent (except at high temperatures), and is low in cost. Although polyethylene is inferior in strength, synthesis, surface hardness, and heat resistance, it has properties exceeding these properties and is therefore used in large quantities. At the same time, it is true that they are being mass-consumed and discarded.

On the other hand, the full-scale enforcement of the Containers and Packaging Recycling Law and the Home Appliance Recycling Law has pushed the need to recycle plastic. Under such circumstances, there is a pressing need to recycle polyethylene that has been consumed or discarded in large quantities. However, polyethylene having different physical properties and chemical properties depending on applications, that is, polyethylene having different grades is used.
The reason why polyethylenes of different grades are present is that, depending on the method of producing polyethylene, branching may occur, and the degree of crystallinity may decrease accordingly. This decrease in crystallinity leads to a decrease in rigidity and an improvement in transparency. As described above, appearance characteristics such as rigidity (flexibility) and transparency of polyethylene greatly vary depending on crystallinity. Using density as a guide for crystallinity, high density polyethylene (HDPE: High Density PolyEthylene) and low density polyethylene (LDPE: Low Density PolyEthyle)
ne). More recently,
A new type of engineering plastic, polyethylene, has been developed, and there are some aspects that cannot be specified simply by density alone. Since polyethylene of different grades having different densities have different properties, as described above,
When recycling, it is necessary to separate polyethylene for each grade.

In this embodiment, in order to separate the polyethylene by grade, the near-infrared spectrum of polyethylene is measured, and a data analysis method in chemometrics is applied to near-infrared spectrum data.
A grade identification method for identifying the grade of polyethylene based on the information obtained by this data analysis technique is used.

Here, the wavelength in the near infrared region is used due to the optical characteristics of polyethylene. CH
The bands of the first, second, third, etc. of the stretching vibrations are all doublets, but the longer wavelength side is due to the methylene group and the shorter wavelength side is due to the methyl group. It is. Although the intensity becomes weaker as the order of the overtones increases to 1, 2, and 3, the resolving power of the band improves, so that the near infrared region corresponding to the higher order vibration is extremely effective for identifying the compound. Another advantage of using this near-infrared region is that only vibrations with large anharmonic constants remain in this region, and those with small anharmonic constants are not observed, making the spectrum relatively simple. There are advantages.

The near-infrared spectrum of polyethylene in this embodiment was measured using a near-infrared spectrum measuring device (trade name: PlaScan) manufactured by Opto Giken. This near-infrared spectrum measuring device uses an acousto-optic modulation filter (AOTF) spectroscopy, and is in a near-infrared region (1.0 to 2.5 μm).
At a high resolution (0.
0005 μm). Using this near-infrared spectrum measuring apparatus, the intensity of diffuse reflection light at 2400 points in a wavelength range of 1.1 to 2.2 μm was measured for a ceramic as a standard sample at a cumulative number of 20 times. Next, for a plurality of polyethylenes to be grade-identified, the same wavelength range of 1.1 to 2.2 μm 24
The intensity of the diffuse reflection light at the 00 point was measured at 20 times of integration. Then, for each measurement wavelength point, the logarithm of the relative reflectance, which is the ratio of the reflected light intensity of the polyethylene to the reflected light intensity of the ceramic, is calculated as the absorbance, whereby the near-infrared spectrum in the wavelength region of 1.1 to 2.2 μm is calculated. Data was obtained. The measurement of the near-infrared spectrum was performed on 70 samples of films and sheets of polyethylene (HDPE: 6 types, LDPE: 8 types) such as plastics used in daily life.

Next, before performing data processing such as a neural network on the measurement data of the near-infrared spectrum indicating the relationship between the wavelength and the absorbance, smoothing for removing noise, sharpening of a peak, and normalization are performed. As data processing, the following pre-processing of data (1) to (4) was performed. (1) The minimum value is 0 and the maximum value is 1 for the measured spectrum data at 2400 points in the wavelength range of 1.1 to 2.2 μm.
It was standardized so that (2) The normalized spectrum of 2400 points was averaged every 10 points, and data of 240 points was created. (3) Using the data at 240 points, a second derivative spectrum was calculated. (4) Only data at 89 points in the wavelength range of 1.6 to 2.0 μm were taken out, and normalization was performed again so that the maximum absolute value was 1.

The reason why the data averaging process is performed is to remove noise. The reason why the second derivative processing is performed is to sharpen the peak and correct the inclination of the baseline. As other preprocessing of this kind of data, there is also an MSC (Multiplicative Scatter Correction) method and the like, but in the present embodiment, a second derivative method that is simple in calculation is mainly used. In addition, in the pre-processing of the above data,
Computer as data processing device (Fujitsu FM
V-6233N / A, OS: "Microsoft Excel", Microsoft software installed on Microsoft Windows95)
98 "was used.

FIGS. 1 and 2 respectively show the wavelength region 1.1.
It is a graph of the average near-infrared spectrum of high-density polyethylene (HDPE) and low-density polyethylene (LDPE) in -2.2 micrometers, and its second derivative spectrum. Also,
FIGS. 3 and 4 are graphs in which the spectra of FIGS. 1 and 2 are plotted in the wavelength region of 1.6 to 2.0 μm, respectively. As can be seen from these graphs, the significant difference in the near infrared spectrum between high density polyethylene (HDPE) and low density polyethylene (LDPE) is 1.6-2.0 μm.
Seen in the area. High-density polyethylene (HDPE) and low-density polyethylene (LDPE) were identified using the difference in near-infrared spectrum in this wavelength region.

In the present embodiment, a neural network employing a hierarchical perceptron model is used as one of the data analysis methods in chemometrics for analyzing the near infrared spectrum data. The data processing by this neural network is performed by the above computer (FMV-6233N / A by Fujitsu, OS: Microsoft
Windows95) software (trade name: NEUROSIM / L) manufactured by Fujitsu Limited was used.

As shown in FIG. 5, this neural network has a three-layer structure in which an input layer, an intermediate layer, and an output layer are each one layer. The learning of the neural network uses an error back propagation method. Was. The near-infrared spectrum (second derivative spectrum) data of each pre-processed polyethylene is input to the input layer of the neural network, and the polyethylene identification code (HDPE: 1, LDPE: 0) is output to the output layer. Learning was performed by inputting as teacher data. Further, in the present embodiment, since data analysis was performed to identify either HDPE or LDPE, the output layer was set to one unit. As other conditions, the input layer has 89 units,
The intermediate layer and three units are set, and e = 5.
Using 0 and a = 0.1, the initial random number seed was set to 1, and the convergence determination condition was set to 0.1. FIG. 6 is a schematic diagram of the neural network after learning under the above conditions.

HDPE and LDPE of the above neural network
Was determined by the leave-one-out method. Specifically, five samples were extracted from all 70 samples, and after learning 65 samples, the discriminating ability was determined for the five samples. This is repeated 14 times in total, and the HDPE and LD of the neural network are
The ability to identify PE. As a result, the discrimination ability of the neural network was 100%, and it was found that HDPE and LDPE could be completely discriminated.

In the present embodiment, the principal component analysis method, the PLS analysis method, the hierarchical cluster analysis method, and the SIMCA method are used as other data analysis methods in chemometrics for analyzing the data of the near-infrared spectrum. Data processing by these data analysis methods used software (trade name: Pirouette Ver. 2.6) manufactured by InfoMetrix, which was installed in the computer. Using this software, data analysis was performed on 70 polyethylene samples as in the case of the neural network.

FIG. 7 shows a first principal component and a second principal component when the above 70 polyethylene samples were subjected to principal component analysis using 89 data in the second derivative spectrum in the wavelength range of 1.6 to 2.0 μm. FIG. 8 is a graph showing the relationship between the principal components and FIG. 8 is a graph showing the relationship between the second principal component and the third principal component in the same principal component analysis. According to the results of FIGS. 7 and 8, the analysis data of HDPE and the analysis data of LDPE can be classified into independent clusters.
It can be seen that LDPE and LDPE are sufficiently distinguishable.

FIG. 9 shows the first principal component and the first principal component when the 70 polyethylene samples were subjected to PLS analysis using 89 data in the second derivative spectrum in the wavelength range of 1.6 to 2.0 μm. FIG. 10 is a graph showing a relationship between two principal components, and FIG. 10 is a graph showing a relationship between a second principal component and a third principal component in a similar PLS analysis. According to the results of FIGS. 9 and 10, similarly to the result of the principal component analysis, the analysis data of HDPE and the analysis data of LDPE can be classified into independent clusters, and HDPE and LDPE can be sufficiently distinguished. You can see that there is.

FIG. 11 shows the results of a hierarchical cluster analysis of the above 70 polyethylene samples, using 89 data in the second derivative spectrum in the wavelength range of 1.6 to 2.0 μm. It is a dendrogram. FIG. 12 shows a high-density polyethylene sample (HDP
FIG. 13 is a graph showing the result of analyzing the E) using the SIMCA method using 89 data in the second derivative spectrum in the wavelength region of 1.6 to 2.0 μm. FIG. 13 shows the results for the low density polyethylene sample (LDPE). , A similar SIMCA
6 is a graph showing the results of analysis by the method. These FIG.
In addition, from the results of FIGS. 12 and 13, similarly to the result of the above-described principal component analysis, the analysis data of HDPE and the analysis data of LDPE can be classified into independent clusters.
It can be seen that LDPE and LDPE are sufficiently distinguishable.

Next, in the present embodiment, the above-mentioned wavelength region 1.
In order to further narrow down the wavelength range of the data input to the neural network for the purpose of further narrowing down 89 points of data of 6 to 2.0 μm and further improving the ability to identify the grade of polyethylene, the wavelength region 1 Dispersion statistical processing was performed on the second derivative spectrum of 0.6 to 2.0 μm. As a result, as shown in the graph of FIG. 14, there is a main peak of dispersion in the wavelength range of 1.6912 to 1.778 μm. .77
It turns out that it is sufficient to use data of the near-infrared spectrum of 83 μm.

Therefore, the wavelength range from 1.6912 to 1.77.
Only data of 20 points of 83 μm were taken out and normalized again so that the maximum value of the absolute value was 1. Then, as shown in FIG. 15, these 20 points of data are input to the input layer of the neural network, and learning is performed under the same conditions as when the number of input data is 89, and HDPE and LD
The discrimination ability of PE was judged. As a result, even when these 20 data were used, the discrimination ability of the neural network was 100%, and it was possible to completely discriminate between HDPE and LDPE. FIG. 16 is a schematic diagram of a neural network after learning when the number of input data is 20.

Further, the above wavelength range 1.6912 to 1.7.
Data analysis was performed on the 20-point data of 783 μm using the principal component analysis method, the PLS analysis method, the hierarchical cluster analysis method, and the SIMCA method as described above.
As shown in FIGS. 17, 18, 19, 20, 21, 22 and 23, the analysis data of LDPE and the analysis data of HDPE can be classified into independent clusters,
HDPE and LDPE were found to be sufficiently distinguishable. In these cases, the pre-processing of the above data was partially changed so that MSC (Multiplicative Scatter Correctio
n) Even if the same data analysis including the method is performed, HDPE and LDP
It was found that E could be identified.

Further, in the present embodiment, the above wavelength region 1..
At 6912 to 1.7773 μm, the number of input data used for data analysis is reduced to 13, 10, 6, 5, and 4 sequentially, and a parameter scan is performed using the neural network after learning.
Data analysis such as the neural network and principal component analysis was performed. As a result, as shown in FIGS. 24 and 25, the number of input data (the number of units in the input layer) is finally reduced to 4 by using a neural network that has learned similarly to the case where the number of input data is 89. (1.6
9785 μm, 1.70702 μm, 1.74368 μ
m, 1.74827 μm), as in the case where the number of input data is 89 and 20, the Leave-on
As a result of examining the performance of discrimination ability by e-out method,
Was obtained. Also, the principal component analysis, the PLS analysis, and the hierarchical cluster analysis are shown in FIGS.
7, As shown in FIGS. 28, 29 and 30, even when the number of input data is finally reduced to four, the analysis data of HDPE and the analysis data of LDPE can be classified into independent clusters. It was found that HDPE and LDPE were sufficiently distinguishable.

Here, when a parameter scan was performed in a neural network with four units of the input layer, one point (1.711) was used to discriminate between HDPE and LDPE.
6 μm) is not considered to be involved. Each input value is 0
From the behavior in the case of, it can be considered that the value is shifted upward as a whole by the value of the constant term. Alternatively, it can be considered that the threshold value of the sensitivity is increasing. Therefore, it is considered necessary to introduce a bias to reduce the threshold value and to retrain the neural network.

Therefore, as shown in FIG. 31, a unit for bias is added to the input layer, and a neural network of 5 units in total is constructed. After learning with this neural network and examining the discrimination performance of the neural network by the leave-one-out method described above, 100%
Was obtained. FIG. 32 is a schematic diagram of a neural network after learning when units for the above-described bias are added.

FIGS. 33 and 34 are graphs respectively showing the actual dispersion of the second derivative spectrum in the wavelength region identified by the neural network when the input layer is divided into five units including the above bias. , And actual average spectra of HDPE and LDPE. 33 and 34, reference numerals P1 to P
The wavelength indicated by 4 corresponds to the input data. From this result, the neural network recognizes the largest variance and the second largest variance. The remaining two are regions where the degree of dispersion is not necessarily large. However, when reviewing the actual average spectra of HDPE and LDPE, a clear difference is observed in the portion falling to the right from the maximum peak. This part can be considered to be recognized by the neural network.

As described above, according to the present embodiment, by combining near-infrared reflection spectrum measurement and various data analysis techniques in chemometrics, polyethylene 2
As a result of examining whether it is possible to discriminate between the two grades (HDPE, LDPE), it was found that linear analysis such as principal component analysis in chemometrics was sufficient. Also,
Data analysis using a neural network, which is a non-linear analysis, was able to be sufficiently identified, and the hit rate was found to be 100%. In particular, in data analysis using a neural network, the actual spectral range (1.6958 to 1. .6) is determined so that a human looks at the near-infrared spectrum to identify two grades of polyethylene (HDPE and LDPE).
6999 μm, 1.7050 to 1.7091 μm, 1.
7416 to 1.7457 μm, 1.7462 to 1.75
The average spectrum of (03 μm) was identified. This can be considered as observing the state of the stretching movement of the terminal of the hydrocarbon unit in which polyethylene is used. As described above, it is possible to sufficiently identify the grade of polyethylene by using the data at the specific four wavelengths in the near infrared spectrum. Therefore, it is possible to more easily and quickly identify the grade of polyethylene. When a system for identifying polyethylene of different grades is configured by combining the above-mentioned near-infrared spectrum measuring device and computer, the identification system can be reduced in size.

In the above embodiment, the case where the grade of polyethylene is identified has been described. However, the present invention relates to polystyrene (GP: General Purpose, H).
The present invention is also applicable to the case of identifying the grade of another polymer material such as I (High Impact) or polypropylene.

[0042]

According to the first to fifth aspects of the present invention, it is not necessary to preprocess a polymer material before measuring a near-infrared spectrum, and data processing using a computer can be easily applied. There is an excellent effect that the grade of the polymer material can be identified more easily and quickly than when the measurement is performed in the infrared wavelength region.

In particular, according to the second aspect of the present invention, it is possible to identify the finely classified grades, and the nonlinearity of near-infrared spectrum data is lower than that of a data analysis method employing a linear model. Is strong, there is an excellent effect that the grade of the polymer material can be identified with higher accuracy.

In particular, according to the fourth and fifth aspects of the invention, there is an excellent effect that the grade due to the difference in crystallinity (density) of polyethylene can be identified with high accuracy.

In particular, according to the fifth aspect of the present invention, the change in the optical characteristics due to the difference in the crystallinity (density) of the grade of polyethylene is in a wavelength range of 1.6912 to 1.7.
Since the measurement and data processing of the near infrared spectrum can be performed only at 783 μm, there is an excellent effect that it is possible to more easily and quickly identify the grade of polyethylene.

[Brief description of the drawings]

FIG. 1 is a graph of an average near-infrared spectrum of high-density polyethylene (HDPE) and low-density polyethylene (LDPE) in a wavelength region of 1.1 to 2.2 μm.

FIG. 2 is a graph of a second derivative spectrum of high-density polyethylene (HDPE) and low-density polyethylene (LDPE) in a wavelength region of 1.1 to 2.2 μm.

FIG. 3 shows a wavelength range of 1.6 to 1.6 in the near infrared spectrum of FIG.
The graph expanded about 2.0 micrometers.

FIG. 4 shows a wavelength range of 1.6 to 1.6 in the near infrared spectrum of FIG.
The graph expanded about 2.0 micrometers.

FIG. 5 is a schematic diagram of a neural network having 89 units in an input layer.

FIG. 6 is a schematic diagram of the neural network after learning.

FIG. 7 is a graph showing a relationship between a first principal component and a second principal component when principal component analysis is performed using 89 data in a second-order differential spectrum in a wavelength range of 1.6 to 2.0 μm.

FIG. 8 is a graph showing a relationship between a second principal component and a third principal component in the principal component analysis.

FIG. 9 is a graph showing a relationship between a first principal component and a second principal component when PLS analysis is performed using 89 data in the second derivative spectrum in a wavelength region of 1.6 to 2.0 μm.

FIG. 10 is a graph showing a relationship between a second principal component and a third principal component in the PLS analysis.

FIG. 11 is a dendrogram showing a result of an analysis by a hierarchical cluster analysis method using 89 pieces of data in a second derivative spectrum in a wavelength region of 1.6 to 2.0 μm.

FIG. 12 shows a high-density polyethylene sample (HDPE)
The graph which shows the result analyzed by SIMCA method using 89 data in the secondary differential spectrum of a wavelength area | region 1.6-2.0 micrometers.

FIG. 13 shows a low density polyethylene sample (LDPE)
The graph which shows the result analyzed by SIMCA method using 89 data in the secondary differential spectrum of a wavelength area | region 1.6-2.0 micrometers.

FIG. 14 is a graph showing dispersion of a second derivative spectrum in a wavelength range of 1.6 to 2.0 μm.

FIG. 15 is a schematic diagram of a neural network when the number of input data is 20;

FIG. 16 is a schematic diagram of the neural network after learning.

FIG. 17 is a graph showing a relationship between a first principal component and a second principal component when principal component analysis is performed using 20 pieces of data in a second derivative spectrum in a wavelength region of 1.6912 to 1.7778 μm.

FIG. 18 is a graph showing a relationship between a second principal component and a third principal component in the principal component analysis.

FIG. 19 is a graph showing P using 20 data in the second derivative spectrum in the wavelength range of 1.6912 to 1.778 μm.
5 is a graph showing a relationship between a first principal component and a second principal component when performing LS analysis.

FIG. 20 is a graph showing a relationship between a second principal component and a third principal component in the PLS analysis.

FIG. 21 is a dendrogram showing a result of analysis by a hierarchical cluster analysis method using 20 pieces of data in a second derivative spectrum in a wavelength range of 1.6912 to 1.7778 μm.

FIG. 22 shows a high-density polyethylene sample (HDPE)
The graph which shows the result analyzed by SIMCA method using 20 data in the 2nd-order differential spectrum of a wavelength area 1.6912-1.7783 micrometers.

FIG. 23 shows a low density polyethylene sample (LDPE).
The graph which shows the result analyzed by SIMCA method using 20 data in the 2nd-order differential spectrum of a wavelength area 1.6912-1.7783 micrometers.

FIG. 24 is a schematic diagram of a neural network when the number of input data is four.

FIG. 25 is a schematic diagram of the neural network after learning.

FIG. 26 is a graph showing a relationship between a first principal component and a second principal component in principal component analysis using four pieces of data.

FIG. 27 is a graph showing a relationship between a second principal component and a third principal component in the principal component analysis.

FIG. 28 is a graph showing a relationship between a first principal component and a second principal component in a PLS analysis using four data.

FIG. 29 is a graph showing a relationship between a second principal component and a third principal component in the PLS analysis.

FIG. 30 is a dendrogram showing a result of analysis by hierarchical cluster analysis using four pieces of data.

FIG. 31 is a schematic diagram of a neural network when there is a bias input.

FIG. 32 is a schematic diagram of the neural network after learning.

FIG. 33 shows a 2 in a wavelength range of 1.68 to 1.78 μm.
5 is a graph showing the variance of the second derivative spectrum.

FIG. 34 shows high-density polyethylene (HDPE) and low-density polyethylene (LD) in a wavelength range of 1.68 to 1.78 μm.
Graph of average near infrared spectrum of PE).

Continuing from the front page (72) Inventor Takatoshi Matsumoto 1-1-1, Higashi, Tsukuba, Ibaraki Pref., National Institute of Advanced Industrial Science and Technology (72) Inventor Wako Saeki 150, Futamicho, Takaoka-shi, Toyama Pref. In-house (72) Inventor Toshio Amano 3-6-1 Nihonbashi-Honcho, Chuo-ku, Tokyo Opto Giken Co., Ltd. F-term (reference) 2G059 AA05 BB08 EE02 EE12 HH01 HH06 MM01 MM03

Claims (5)

[Claims] 1. A near-infrared spectrum in a near-infrared wavelength range of a polymer material to be grade-identified is measured, and a data analysis method in chemometrics is applied to the data of the near-infrared spectrum. A method for identifying a grade of a polymer material, wherein the grade of the polymer material is identified based on the acquired information. 2. The method of claim 1, wherein the near-infrared spectrum data includes sharpening of peaks, correction of baseline slope, and normalization as data analysis techniques in the chemometrics. And a data processing by a neural network for the near-infrared spectrum data on which the pre-processing has been performed. 3. The method for identifying a grade of a polymer material according to claim 1, wherein the near-infrared spectrum data includes sharpening of peaks, correction of baseline slope, and normalization as data analysis techniques in the chemometrics. And performing data processing on at least one of the principal component analysis method, the PLS analysis method, the hierarchical cluster analysis method, and the SIMCA method on the near-infrared spectrum data subjected to the preprocessing. Grade identification method for polymer materials. 4. The method for identifying a grade of a polymer material according to claim 1, wherein the wavelength range of the near-infrared spectrum is 1.6 to 2.0 μm, and the polymer material to be graded is polyethylene. A method for identifying a grade of a polymer material, characterized in that: 5. The method according to claim 4, wherein said near-infrared spectrum has a wavelength range of 1.6.
A method for identifying a grade of a polymer material, wherein the grade is 912 to 1.7783 μm.