JPH1137935A - Method and apparatus for determining a plurality of components in sewage - Google Patents

Abstract

PROBLEM TO BE SOLVED: To measure the concentration of water quality corrupting factors quickly and conveniently by calculating the concentration of water quality corrupting factors from the spectroscopic spectrums of a light transmitted through a solvent to be measured and a light scattered by the solvent using multiple regression analysis. SOLUTION: A solvent 4 to be measured in an optical cell 3 is irradiated with a light 2 from a light source 1 and a transmitted light 5 is passed through a spectrometer to produce a spectrometric spectrum from which the absorbance is operated at an operating section 7 for a plurality of wavelengths. The calculated absorbances are delivered to an operating section 8. The operating section 8 is prestored with the working curves for total organic carbon, total nitrogen, ammonia nitrogen and total phosphorus having the water quality factor concentration and the absorbance as objective variable and explanation variable, respectively, and mixed components are subjected to quantitative analysis by multiple regression analysis. As a second method, quantitative analysis is performed by pattern recognition method employing the neural network of absorbance of a plurality of wavelengths obtained from the spectrometric spectrum. As a third method, quantitative analysis is performed utilizing a transportation equation.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and an apparatus for quantitatively analyzing an amount of organic substances, nitrogen, phosphorus and the like which are used as an index of water pollution in sewage treatment.

[0002]

2. Description of the Related Art Water quality factors such as total organic carbon, total nitrogen, ammonia nitrogen, and total phosphorus in sewage are used as numerical values indicating the properties of sewage. Quantitative analysis of these water quality factors is required. Conventionally, the measurement is carried out according to a sewage test method (issued by the Japan Sewerage Association), and a specific measurement method is used for each water quality factor. In addition, any of the measurement methods requires preparation of a calibration curve by measuring a standard substance in advance, and pretreatment such as addition of a prescribed reagent, heating, and removal of suspended components.

The general method of measuring the four water quality factors of total organic carbon, total nitrogen, ammonia nitrogen and total phosphorus will be described below. (1) Total organic carbon (TOC): First, as pretreatment,
The suspended components in the sampling water are uniformly dispersed by an ultrasonic crusher.

Next, a small amount of the sample water subjected to the pretreatment is
Along with the oxygen of the carrier gas, it is sent to an oxidation catalyst filling tube (TC furnace) for measuring all carbon at a high temperature of 950 ° C., and the carbon in the organic substance and the carbon in the inorganic substance are converted into carbon dioxide gas. Is measured with a non-dispersive infrared gas analyzer, and the total carbon content (TC) is determined from a TC calibration curve.

Further, in another route, the sample is sent to an oxidation catalyst filled tube (IC furnace) for measuring inorganic carbon which is maintained at a temperature (150 ° C.) at which the organic matter of the sample is not decomposed, and the generated gas is dehumidified. After the dust removal treatment, the carbon dioxide gas decomposed and generated by the above operation is measured, and the amount of inorganic carbon (IC) is determined by an IC calibration curve. Finally, the total organic carbon content is calculated by subtracting the inorganic carbon content (IC) from the total carbon content (TC).

Before the sample measurement, the TC / IC standard solution is measured in advance by the same operation as described above, and the TC / I
Create a calibration curve for C and use it for calculating measured values. For total organic carbon, an example of a measurement flow using a TOC analyzer is extracted from the sewage test method and is shown in FIG. (2) Total nitrogen (T-N): a devarda alloy was added to sample water, alkalized with sodium hydroxide, and then distilled.
Ammoniacal nitrogen, nitrite nitrogen, nitrate nitrogen, and organic nitrogen in sample water (reduced and decomposed by alkali)
Of ammonia corresponding to the total amount of Next, after the organic nitrogen in the remaining liquid is thermally decomposed by the Kjeldahl nitrogen method, it is again distilled under alkalinity, and the distilled ammonia is measured by the ion electrode method in combination with the above-mentioned ammonia to determine the total nitrogen. I do. (3) Ammonia nitrogen (NH ₄ —N): The sample water was made alkaline, the ammonia distilled off during heating distillation was absorbed by sulfuric acid, and the distillate was adjusted to pH 11 to 13,
The ammonium ion is changed to ammonia, and this ammonia is measured using an ion electrode (ammonia electrode).
Ammonia nitrogen is determined. (4) Total phosphorus (TP): Add nitric acid and sulfuric acid to sample water, heat and evaporate to near dryness to decompose, and measure the phosphoric acid ion-form phosphorus generated based on molybdenum blue absorption spectrophotometry. By doing so, the total phosphorus in the sample water is determined.

As described above, the conventional measurement methods for the above-mentioned four water quality factors use processing and measurement methods centering on chemical analysis techniques.

[0008]

In recent years, from the viewpoint of preserving the quality of sewage effluent and reusing treated water, it is necessary to enhance the sewage treatment using the amount of organic substances, nitrogen, phosphorus, etc. as water pollution factors as indicators. Therefore, a method for quickly and easily measuring the concentration of the water pollution factor has been desired.

However, as described above, the conventional analysis method is as follows.
The measurement method differs for each water quality factor, and pretreatment is required.
It takes a long time to obtain a measurement result. In addition, TOC analyzers and the like also exist due to the development of instrumental analysis. However, since the measurement principle is the same as that of the above-described method, a different analyzer is required for each measurement item, and the purchase price and running cost increase.

The above problems are summarized as follows. (1) The measurement principle is different for each water quality factor to be measured, and complex measurement is impossible. (2) When performing the measurement, many operations including a pretreatment and a long time are required. The present invention has been made to solve these problems, and an object of the present invention is to reduce the concentration of total organic carbon, total nitrogen, ammonia nitrogen, and total phosphorus, which are water pollution factors, in sewage. It is an object of the present invention to provide a method and an apparatus for quantifying a plurality of components in sewage which are simple and can be quantitatively determined in a complex manner without treatment.

[0011]

In order to solve the above problems, the present invention utilizes the principle that the wavelength absorption characteristics of light differ depending on water quality factors in order to measure different substances according to the same principle. That is, the solvent to be measured containing the suspension component is irradiated with light, the light transmitted through and scattered from the solvent to be measured is spectrally analyzed, absorbances at a plurality of wavelengths are calculated from the spectral spectrum, and the obtained absorbances are used. Thus, a plurality of components in the solvent to be measured are determined.

Here, there are the following three methods for calculating the concentrations of a plurality of contained components in the solvent to be measured from the obtained plurality of absorbances. First, the first method is to perform quantitative analysis of mixed components by multiple regression analysis on absorbances at a plurality of wavelengths obtained from a spectrum. In this method, a calibration curve for total organic carbon, total nitrogen, ammonia nitrogen, and total phosphorus is prepared in advance by multiple regression analysis using the water quality factor concentration as the target variable and the absorbance as the explanatory variable, Calculate by substituting multiple absorbances into the line. This method uses the absorbance at a plurality of wavelengths obtained from the spectral spectrum as an explanatory variable, performs a multiple regression analysis, aggregates information contained in the spectrum, and eliminates interference of many other variables. Thus, quantitative analysis of the mixed component can be performed by associating the target water quality factor concentration with the spectral data.

Next, the second method is to quantitatively analyze the absorbance at a plurality of wavelengths obtained from the spectrum by a pattern recognition method using a neural network. In this method, first, a plurality of obtained absorbances are normalized to 0 to 1, and the plurality of normalized absorbances are learned in advance by a back propagation method to determine a weight coefficient, an input layer, an intermediate layer, and an output layer. The input is input to the input layer of the neural network composed of layers, and the output from the output layer is calculated by performing a predetermined operation. Here, in determining the weighting factor of the neural network, the absorbance at a plurality of wavelengths λ _{1 to} λ _n is measured for a plurality of m samples of known concentrations containing all water quality factors for quantitative purposes,
The obtained absorbances of n wavelengths and the corresponding m sets of four water quality factor concentrations are learned as input and output layers of a neural network, respectively, to determine weighting factors. Note that the various calculations described above are performed using the above-described calibration curves,
This can be performed by a device including an arithmetic unit that stores a neural network.

Both the first and second treatment methods do not require a pretreatment of the solvent to be measured, and are simply and complexly composed of four organic carbon, total nitrogen, ammonia nitrogen and total phosphorus. Water quality factor concentration can be determined. For the multiple regression analysis and the neural network analysis method described above, the concentration of the four water quality factors can be instantaneously obtained by using a commercially available neural network system or multivariate analysis software and a personal computer.

Further, the third method can be applied to the case where suspended components are mixed in the solvent to be measured, and the absorbance from the transmission and scattering light intensities of the solvent to be measured is determined by the absorption equation obtained by using the transport equation. From the coefficient and the scattering coefficient, quantitative analysis of the mixed component is performed by a conversion table. The above first and second methods are based on the premise that the absorbance can be accurately measured. However, when the solvent to be measured contains a lot of suspended components, the measured value of one light intensity detector is used. The measurement results of all the absorbances have a large measurement error. Therefore, this third
The method comprises a first light detector for measuring transmitted light intensity, and one or more second light detectors for measuring scattered light intensity. The absorption cross section per unit area (absorption coefficient) and the value converted into the scattering cross section per unit volume of the solvent to be measured (scattering coefficient) are used. These conversion methods can be performed by an apparatus having an operation unit storing a conversion table in which the light intensity is calculated by solving the transport equation, and a reliable absorbance can be obtained.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the method of the present invention will be described based on three embodiments. [Embodiment 1] FIG. 1 is a schematic diagram showing a main configuration of an apparatus to which a first method of the present invention is applied.
Also in FIG. 3, common parts are represented by the same reference numerals.

In FIG. 1, a light 2 from a light source 1 is converged by a lens 12 and is irradiated on a solvent 4 to be measured in sewage held in an optical cell 3, and the transmitted light 5 is separated by a spectroscope 6a. , A plurality of wavelengths λ ₁ to
The calculation unit 7 calculates the absorbances E _{W1 to} E _Wn of λ _n . The calculated absorbances E _{W1 to} E _Wn are sent to the calculation unit 8, and the calculation unit 8 preliminarily calculates total organic carbon, total nitrogen, ammonia nitrogen, using water quality factor concentration as an objective variable and absorbance as an explanatory variable. Each calibration curve for all phosphorus is stored. In this calculation unit 8, the absorbances E _{W1 to} E _W are added to the stored calibration curves.
_The above four water quality factor concentrations are calculated by substituting _Wn , and the calculation results are output to the output unit 9, the display unit 10, and the storage device 1
Sent to 1.

Here, the above four water quality factors are actually detected.
A quantitative curve was prepared and quantitative analysis was performed using the apparatus of FIG.
Here is an example. Specific procedures and results are as follows. Is
First, the preparation of the calibration curve included the above four water quality factors.
Measure the spectrum of multiple samples with known concentrations
And the correlation coefficient of the absorbance at the wavelength versus the wavelength shown in FIG.
Are created in a two-wavelength correlation matrix. further,
Referring to the two-wavelength correlation matrix, the components of sewage are discussed.
Grouping of independent wavelengths
Then, narrow down the wavelengths that are significant as water quality indicators. In FIG.
In the example, the wavelength (λ₁₉₀, Λ₁₉₁), (Λ₁₉₂, Λ₁₉₃)
Are highly correlated, and the correlation between the two is
Λ₁₉₀And λ₁₉₁Or λ₁₉₂When
λ ₁₉₃Can be considered as the same group, and
Are independent in describing sewage components.
You. Therefore, the absorbance data of 12 samples of actual sewage
A two-wavelength correlation matrix is created based on
Eight wavelengths λ₁~ Λ₈= 190, 194, 20
0, 211, 224, 232, 271, 353 nm
Issued. Next, target 30 new samples of actual sewage
And the wavelength λ of the sample₁~ Λ₈Absorbance of the explanatory variable, below
Quantitative values of four water quality factors in water test method as objective variables
Multiple regression analysis was performed to obtain an optimized multiple regression equation. this
The multiple regression equation, that is, the calibration curve is expressed by equations (1) to (4).
You.

[0019]

(Equation 1)

The calibration curves (1) to thus created
The equation (4) is stored in the calculation unit 8 of the apparatus shown in FIG. 1, and a new quantitative analysis of 30 samples of actual sewage is performed. 5 to 8. From these results, all the correlation coefficients were 0.9 or more. From these results, it is understood that the present invention, in which a plurality of absorbances are analyzed by multiple regression analysis, is a quantitative method having sufficient accuracy consistent with the conventional method. For the multiple regression analysis performed here, commercially available software capable of performing multivariate analysis on a computer can also be used. The above is a description of the method and apparatus having the greatest feature in that the quantitative analysis of multiple components of sewage is realized by combining the absorbance at multiple wavelengths of a sample with the multiple regression equation (calibration curve) created by multiple regression analysis. . [Embodiment 2] Next, FIG. 2 is a schematic diagram showing the main configuration of an apparatus to which the second method of the present invention is applied.

In this figure, a light 2 from a light source 1 is converged by a lens 12 and irradiated on a solvent 4 to be measured in sewage held in an optical cell 3, and the transmitted light 5 is separated by a spectroscope 6a. A plurality of wavelengths λ ₁ to
The calculation unit 7 calculates the absorbances E _{W1 to} E _Wn of λ _n . The absorbances E _{W1 to} E _Wn are sent to the operation unit 13,
Reference numeral 3 stores a neural network including an input layer, an intermediate layer, and an output layer, which has been learned in advance by a back propagation method and has determined a weight coefficient. The arithmetic unit 13 normalizes the plurality of absorbances E _{W1 to} E _Wn to 0 to ₁ , and outputs the n normalized absorbances E _W1 ′ to E _Wn ′ to the input layer of the neural network. A predetermined calculation is performed on each of the results, and the total organic carbon,
Calculate the concentration of total nitrogen, ammonia nitrogen and total phosphorus.
This final calculation result is sent to the output unit 9, the display unit 10, and the storage device 11.

Here, the configuration of the neural network stored in the arithmetic unit 13 is shown in FIG. 9, and an example of the construction procedure will be specifically described. The configuration in FIG. 9 includes an input unit 14, a conversion unit neural network 15, an output unit 16, and a learning unit 17 for learning the conversion unit neural network 15. The input unit 14 has eight wavelengths λ _{1 to} λ ₈ = 190, ₁ determined in the same manner as in the method described in the previous embodiment.
94, 200, 211, 224, 232, 271, 35
The result of the absorbance at 3 nm was calculated from 0 to 1 according to the following equation (5).
To the real number signal (E _Wn ′), and inputs the converted signal to the conversion unit neural network 15.

[0023]

(Equation 2)

The conversion unit neural network 15 converts the input signal into total organic carbon, total nitrogen, ammonia nitrogen,
It is converted into a concentration signal (0 to 1) for each water quality factor of all phosphorus, and output to the output unit 16. Here, the conversion unit neural net 15 is constituted by the three-layer back propagation neural network shown in FIG.
Units, the middle layer has five units, and the output layer has four water quality factors (A to D). The output unit 16 calculates and outputs true densities of the signals A to D from the conversion unit neural network 15 by the following equations (6) to (9), respectively. These formulas are inverse calculation formulas of conversion of teacher signal 0 to 1 for each water quality factor concentration in pattern learning described later.

[0025]

(Equation 3)

In the back propagation method, a neural network is constructed by learning an input pattern to be recognized (determining a weight coefficient of the neural network).
I do. Therefore, for the twelve actual sewage samples containing the four water quality factors described above, the eight wavelengths of absorbance are input to the input unit 14 of the neural network of FIG. 9 for learning, and twelve sets of outputs are output. A teacher pattern was given with each water quality factor concentration determined by the sewage test method as 0 to 1 signal, and the learning unit 17 made initial learning. At this time, the convergence was achieved by approximately 1000 learning operations, and the maximum error was 0.05 or less. The neural network used here is an emulator having a ring-coupled parallel architecture using four coprocessors, and implements the back propagation method by software.

The neural network thus constructed is stored in the operation unit 13 of the apparatus shown in FIG. 2, and 50 actual sewage samples are measured by the apparatus shown in FIG. FIGS. 11 to 14 show the comparison results. From these figures, it can be seen that there is a good correlation with a correlation coefficient of 0.9 or more. This is a result showing the effectiveness of the constructed neural network analysis method, and can be said to be a quantitative method with sufficient accuracy as in the first method. that's all,
A method and apparatus for realizing quantitative analysis of multiple components of sewage by combining spectral spectra of sewage samples and back propagation neural networks were described. [Embodiment 3] This embodiment shows the third method. In the apparatus of FIGS. 1 and 2 described above, it is premised that the absorbance can be accurately measured. However, when the solvent to be measured contains a large amount of suspended components, the optical system of the apparatus in FIGS. In a system, detection of light intensity can include measurement errors. Thus, a method for measuring the absorbance without being affected by the mixed suspended components will be described below. First, given the scattering coefficient and the absorption coefficient of the medium, and given the shape of the medium and the light source, the amount of transmitted and scattered output light can be obtained by solving the transport equation of the following equation (10) described in the reference. Can be (References: MSPatterson, Th
e Propagation of Optical Radiation inTissue I. Mod
els of Radiation Transport and their Application.
Lasers in Medical Science 1991, 6: 155-168).

[0028]

(Equation 4)

Here, μ _a : light absorption coefficient of the solvent to be measured having _a suspension component, μ _s : light scattering coefficient of the solvent to be measured having a suspension component, L (r, Ω): unit area at position r , Unit time, energy of light going in direction Ω per unit solid angle dΩ dμ _s (r, Ω ′ → Ω): scattering probability of light going from Ω ′ to Ω at position r s (r, Ω): light source is there.

This equation can be generally solved using the Monte Carlo method. FIG. 3 shows the third embodiment of the present invention.
FIG. 2 is a schematic diagram showing a configuration of a main part of an apparatus to which the method is applied. The process of obtaining the components contained in the solvent to be measured from the absorbance is the same as that in the apparatus shown in FIG. 1 or FIG. In FIG. 3, light 2 from a light source 1 is converged by a lens 12 and is irradiated on a solvent 4 to be measured in sewage held in an optical cell 3 so that the intensity of the transmitted light 5 is first.
, And the intensity of the scattered light 18 is measured by the second photodetector 6c. The measured data of the transmitted light intensity and the scattered light intensity are sent to the arithmetic unit 19, and the arithmetic unit 19 actually uses the Monte Carlo method (10)
The conversion tables shown in Tables 1 and 2 in which the transport equations of the equations are solved are stored.

[0031]

[Table 1]

[0032]

[Table 2]

Here, Table 1 shows the transmitted light intensity corresponding to the scattering coefficient μ _s written in the vertical direction and the absorption coefficient μ _a written in the horizontal direction, and Table 2 shows the same. Represents the scattered light intensity. Here, in the Monte Carlo method, light (photons) is incident on a measurement system to scatter light,
This is a method of calculating absorption stochastically, in which the trajectory of each incident photon is simulated using a random number and an optical constant value of a measurement system. Therefore, since light scattering and absorption are stochastic phenomena, complete physical phenomena can be reproduced by using random numbers.

The data of the transmitted light intensity and the scattered light intensity transmitted to the arithmetic unit 19 as described above can be obtained by searching the conversion tables of Tables 1 and 2 to obtain the absorption coefficient, that is, the absorbance with high accuracy. Can be. Finally, a specific search method of the conversion table will be described. FIG. 15, FIG.
Is a schematic diagram of a conversion table, in which transmitted light intensity and scattered light intensity are respectively shown. Here, the transmitted light intensity obtained by the apparatus of FIG.
Assume that it was 5. It is confirmed that the transmitted light intensity is 0.5 in FIG. 15 and the scattered light intensity is 0.05 in FIG. In this example, it can be seen that the group with the transmitted light intensity of 0.5 is located on the diagonal line that rises to the right, and the group with the scattered light intensity of 0.05 is located on the diagonal line that declines to the right ( All are gray coated parts). Therefore, when comparing the two tables, attention is focused on the overlapping portion of the previously confirmed positional relationship, and the absorbance for which the absorption coefficient corresponding to the overlapping position is desired to be obtained. Therefore, in this example, when comparing the positions of the two groups of the detected transmitted and scattered light intensities, they overlap at the intersections of the diagonal lines in the figure, and the absorption coefficient at this time, that is, the absorbance is 0.00.
3 is found.

The method of the present invention has been described based on the embodiments. In each case, the time required for all treatments depends on the capabilities of the spectrometer, but is generally on the order of several minutes, as well as continuous water quality analysis, as well as the addition of a continuous sampling device and optical contamination countermeasures. Applicable to automatic analyzers and advanced sewage treatment (control) equipment. In the above description, as the water quality factors of sewage, total organic carbon, total nitrogen,
Although four types of ammonia nitrogen and total phosphorus have been described, as can be seen from the measurement principle of the present invention, the solvent is not limited to water.
It is apparent that the measuring method and the measuring device can be widely applied to components having different absorbances by selecting the wavelength, without being limited to the kind.

[0036]

As described above, conventionally, the measuring method differs for each water quality factor, and many pretreatment operations are required and complicated, and much labor is required to obtain the measurement results. According to the present invention, as described in the three embodiments, a neural network that measures absorbances of a plurality of wavelengths of a sewage sample and calculates a plurality of component concentrations by pattern recognition of a spectral spectrum (analysis by a nonlinear model). Or, by using the method of multiple regression analysis (linear model analysis) to obtain the component concentration from the linear combination of the absorbances of multiple wavelengths, the organic matter, nitrogen, and phosphorus that can serve as sewage water quality maintenance index are removed before turbidity removal, etc. It has become possible to perform complex quantification without any treatment. In addition, even when there are many suspended components in the sample, it became possible to obtain a more accurate absorbance from the creation of the conversion table obtained by solving the transport equation and the measurement of the transmitted light intensity and the scattered light intensity. .

[Brief description of the drawings]

FIG. 1 is a schematic diagram showing a main configuration of an apparatus to which a first method of the present invention is applied.

FIG. 2 is a schematic diagram showing a main configuration of an apparatus to which a second method of the present invention is applied.

FIG. 3 is a schematic diagram showing a main configuration of an apparatus to which a third method of the present invention is applied.

FIG. 4 is a schematic diagram illustrating a two-wavelength correlation matrix of absorbance.

FIG. 5: Comparison between TOC analysis values using the first method and analysis results of the conventional method

FIG. 6: Comparison between TN analysis values using the first method and analysis results of the conventional method

FIG. 7: Comparison diagram of NH ₄ -N analysis values using the first method and analysis results of the conventional method

FIG. 8: Comparison between TP analysis values using the first method and analysis results of the conventional method.

FIG. 9: Schematic diagram of the neural network used in the second method

FIG. 10: Basic structure diagram of the transforming unit neural network used in the second method

FIG. 11: Comparison between TOC analysis values using the second method and analysis results of the conventional method

FIG. 12: Comparison between TN analysis values using the second method and analysis results of the conventional method

FIG. 13: Comparison of NH ₄ —N analysis values using the second method with analysis results of the conventional method

FIG. 14: Comparison between TP analysis values using the second method and analysis results of the conventional method

FIG. 15 is a schematic diagram of a conversion table for obtaining transmitted light intensity from an absorption coefficient and a scattering coefficient.

FIG. 16 is a schematic diagram of a conversion table for obtaining scattered light intensity from an absorption coefficient and a scattering coefficient.

FIG. 17: Flow chart of an example of a conventional TOC analysis method

[Explanation of symbols]

 1: Light source 2: Light 3: Optical cell 4: Solvent to be measured 5: Transmitted light 6a: Spectroscope 6b: Photodetector 6c: Photodetector 7: Operation unit 8: Operation unit 9: Output unit 10: Display unit 11 : Storage device 12: Lens 13: Operation unit 14: Input unit 15: Conversion unit neural network 16: Output unit 17: Learning unit 18: Scattered light 19: Operation unit

Claims (10)

[Claims] 1. A method for irradiating a solvent to be measured containing a suspended component with a light beam, separating the light transmitted and scattered through the solvent to be measured by a spectral means, calculating absorbances at a plurality of wavelengths from the spectrum, In the method for quantifying a component contained in a solvent to be measured by using the absorbance, absorbances E _{W1 to} E _Wn of _n wavelengths λ _{1 to} λ n are calculated from a spectrum of transmitted light and a scattered light of the solvent to be measured. The absorbances E _{W1 to} E _Wn are substituted into the respective calibration curves of k kinds of water quality factors prepared in advance by multiple regression analysis using the concentrations of the k kinds of water quality factors as the objective variable and the absorbance as the explanatory variable, and are calculated. A method for quantifying a plurality of components in sewage, wherein the concentration of one or more of up to k kinds of water quality factors among the kinds of water quality factors is calculated in a complex and simple manner. 2. A method for irradiating a solvent to be measured containing a suspension component with a light beam, separating the light transmitted and scattered by the solvent to be measured by a spectral means, and calculating absorbances at a plurality of wavelengths from the spectrum. In the method for quantifying a component contained in a solvent to be measured by using the absorbance, absorbances E _{W1 to} E _Wn of _n wavelengths λ _{1 to} λ n are calculated from a spectrum of transmitted light and a scattered light of the solvent to be measured. The obtained absorbances E _{W1 to} E _Wn are each calculated by an arithmetic means to be 0 to 1
And the n normalized absorbances E _W1 ′ to E _Wn ′
Is input as an input layer of an arithmetic means storing a neural network composed of an input layer, an intermediate layer, and an output layer, the weights of which are determined by learning by a back propagation method, and k types of output layers are output as a result. A predetermined operation is performed for each water quality factor for at least one of the k
A method for quantifying a plurality of components in sewage, wherein the concentration of a water quality factor up to a species is calculated in a complex and simple manner. 3. The method according to claim 1, wherein k
A method for quantifying a plurality of components in sewage in which four types of water quality factors are used: total organic carbon, total nitrogen, ammonia nitrogen, and total phosphorus. 4. The method according to claim 2, wherein, in determining the weighting factor of the neural network by the back propagation method, a plurality of m concentrations containing the k or all four water quality factors are known. For a sample, a plurality of wavelengths λ _{1 to} λ
Measure the absorbance of _n , the resulting absorbance of n wavelengths,
A method for quantifying a plurality of components in sewage, comprising using a neural network in which m sets of k or four kinds of water quality factor concentrations corresponding thereto are learned as an input layer and an output layer, and weight coefficients are determined. . 5. The method according to claim 1, wherein the means for calculating the absorbance is such that the intensity of the transmitted and scattered light is determined by an absorption cross section (absorption coefficient) per unit volume of the solvent to be measured.
And calculating the absorbance using a value converted into a scattering cross section (scattering coefficient) per unit volume of the solvent to be measured. 6. The method according to claim 5, wherein the conversion into the absorption coefficient and the scattering coefficient refers to a conversion table in which the transmission and scattering light intensities for various absorption and scattering coefficients are calculated by solving the transport equation. And a method for quantifying multiple components in sewage. 7. A solvent to be measured containing a suspended component is irradiated with a light beam, and light transmitted and scattered through the solvent to be measured is spectrally separated by a means for spectrally separating the light, and absorbances at a plurality of wavelengths are calculated from the spectral spectrum. An apparatus for quantifying a component contained in a solvent to be measured by using the absorbance, an optical cell containing the solvent to be measured, a light source for irradiating the solvent to be measured with light, and a spectroscope for separating light transmitted and scattered through the solvent to be measured. Vessel, a first calculation unit for calculating the absorbance from the obtained spectral spectrum or light intensity, each of the total organic carbon, total nitrogen, ammonia nitrogen, and total phosphorus using the water quality factor concentration as an objective variable and the absorbance as an explanatory variable. A second calculator for storing a calibration curve and calculating the four water quality factor concentrations, and an output;
An apparatus for quantifying a plurality of components in sewage, comprising a display and a storage device. 8. A solvent to be measured containing a suspension component is irradiated with a light beam, and light transmitted and scattered through the solvent to be measured is spectrally separated by a means for spectrally separating the light, and absorbances at a plurality of wavelengths are calculated from the spectral spectrum. An apparatus for quantifying a component contained in a solvent to be measured by using the absorbance, an optical cell containing the solvent to be measured, a light source for irradiating the solvent to be measured with light, and a spectroscope for separating light transmitted and scattered through the solvent to be measured. A neural network consisting of an input layer, an intermediate layer, and an output layer, in which a vessel, a first computing unit that computes absorbance from the obtained spectral spectrum or light intensity, and a weight coefficient determined by learning by a back propagation method are stored; A predetermined operation is performed on the output from the output layer of the neural network and the conversion unit for normalizing the absorbance obtained from the first arithmetic unit to 0 to 1 and all organic carbon and total nitrogen An apparatus for calculating concentrations of ammonia nitrogen and total phosphorus, and an output, display, and storage device. 9. An apparatus according to claim 7, wherein said means for dispersing light transmitted and scattered through the solvent to be measured includes a first photodetector for measuring the intensity of light transmitted through the solvent to be measured. The apparatus further comprises at least one second photodetector for measuring the intensity of light scattered from the solvent, and calculates a light absorbance from the light intensity. An apparatus for quantifying a plurality of components in sewage, comprising: a third calculation unit that converts the absorption cross section (absorption coefficient) of the sample into a scattering cross section (scattering coefficient) per unit volume of the medium to be measured. 10. The apparatus according to claim 9, wherein the third arithmetic unit is provided with a conversion table for calculating the transmitted and scattered light intensities for various absorption and scattering coefficients by solving a transport equation. Equipment for multiple components in sewage.