Water quality monitoring method based on turbidity compensation

Abstract

The invention discloses a water quality monitoring method based on turbidity compensation, which comprises the following steps: establishing a standard absorption spectrum curve of a plurality of turbidity substances, establishing a multivariate standard turbidity compensation database based on a polynomial fitting model, the data dimension of each multivariate standard turbidity compensation database being turbidity-wavelength-absorbance; acquiring the absorption spectrum and turbidity of the to-be-detected liquid; taking the wavelength of the visible light region as the reference wavelength of multi-element turbidity compensation, and carrying out regression analysis on the absorbance of the to-be-measured liquid with the current turbidity at the reference wavelength and the absorbance of the plurality of turbidity substances at the reference wavelength to obtain the contribution proportion of each turbidity substance; and calculating a total interference spectrum according to the contribution proportion of each turbidity substance, and subtracting the total interference spectrum from the absorption spectrum of the liquid to be detected to obtain a corrected spectrum. According to the method, spectral information of complex suspended matters in water is simulated through spectral information of various turbidity substances, the universality of a compensation model is improved, and the accuracy of water quality detection is improved.

Description

Technical Field

[0001] The present application relates to the field of water quality detection, and more specifically, to a water quality monitoring method based on turbidity compensation. Background Art

[0002] At present, COD water quality detection methods are divided into chemical methods and direct spectroscopic methods according to different detection principles. The chemical method uses reagents to react chemically with substances, and estimates the pollutant content by monitoring the reagent dosage and product concentration. For example, the CODCr-potassium dichromate method and the CODMn-potassium permanganate method have the advantages of high analytical accuracy and a wide detection range, but have a long measurement cycle and reagent contamination. The direct spectroscopic method calculates the content of different substances in the water sample by scanning the spectral information of the water sample according to the spectral properties, establishing a nonlinear solution model, and reflecting the complex information of water pollutants. It has the advantages of real-time, fast, and no secondary reagent pollution. It is a research focus in the field of domestic water quality detection.

[0003] However, the absorption spectrum of COD is easily interfered by turbidity. The suspended particles in the water interfere with the spectrum due to scattering, resulting in large errors in the subsequent COD model solution. Therefore, it is necessary to perform turbidity correction on the COD absorption spectrum.

[0004] Current full-spectrum turbidity correction methods fall into two main categories: direct solution algorithms, which directly predict the COD prediction error by establishing a turbidity-COD concentration prediction error model and remove the error from the COD concentration; and single turbidity-absorbance models, which subtract the interference spectrum from the original spectrum. For the second category, current mainstream methods only consider compensation correction for a single turbidity species, or select a single turbidity species from multiple turbidity species for compensation correction.

[0005] However, the turbidity substances in natural water bodies are not single, and the substances that constitute turbidity in different water bodies are different. Different turbidity substances have different interference mechanisms and interference contribution rates on the spectrum. Therefore, it is necessary to find a more suitable method for natural water spectral turbidity compensation to correct the deviation in the second type of method and improve the prediction accuracy of COD. Summary of the Invention

[0006] In view of this, this application proposes a water quality monitoring method based on turbidity compensation, which simulates the spectral information of complex suspended matter in water through the spectral information of multiple turbidity substances, improves the universality of the compensation model, and improves the accuracy of water quality detection.

[0007] According to the present application, a water quality monitoring method based on turbidity compensation is proposed, which includes the following steps:

[0008] S1 establishes standard absorption spectrum curves of various turbidity substances and establishes a multivariate standard turbidity compensation database based on a polynomial fitting model. The data dimensions of each multivariate standard turbidity compensation database are turbidity-wavelength-absorbance;

[0009] S2 obtains the absorption spectrum and turbidity of the liquid to be tested;

[0010] S3 uses the wavelength in the visible light region as the reference wavelength for multivariate turbidity compensation, and performs regression analysis on the absorbance of the current turbidity test liquid at the reference wavelength and the absorbance of multiple turbidity substances at the reference wavelength to obtain the contribution ratio of each turbidity substance;

[0011] S4 calculates the total interference spectrum according to the contribution ratio of each turbidity substance, and subtracts the total interference spectrum from the absorption spectrum of the liquid to be tested to obtain a correction spectrum;

[0012] The order of step S1 and step S2 can be swapped.

[0013] Preferably, the plurality of turbidity substances include formazine, PS microparticles, kaolin and diatomaceous earth.

[0014] Preferably, in step S1 and step S2, the wavelength range of the absorption spectrum is 200-700 nm.

[0015] Preferably, step S3 includes the following steps:

[0016] S301 obtains multiple turbidity standard absorption spectra from a multivariate standard turbidity compensation database according to the turbidity of the liquid to be tested, and obtains the absorbance of the turbidity substance at the reference wavelength;

[0017] S302 establishes the turbidity substance contribution ratio solution equation:

[0018]

[0019] Among them, x i is the contribution ratio of turbidity substances to total turbidity, is the absorbance of the original absorption spectrum of the liquid to be tested at the reference wavelength, is the absorbance of the turbidity substance at the reference wavelength, stur is the turbidity substance, and vis is the reference wavelength of multivariate turbidity compensation;

[0020] S303 calculates the contribution ratio of turbidity substances using a regression model;

[0021] S304 calculates the regression error of the absorbance of the liquid to be tested at a reference wavelength in the visible light region based on the turbidity substance contribution ratio results calculated by each model, and selects the turbidity substance contribution ratio that minimizes the regression error;

[0022] S305 calculates the total interference spectrum based on the contribution ratio of the selected turbidity substances:

[0023]

[0024] Among them, x i is the contribution ratio of turbidity substances to total turbidity, is the total turbidity interference spectrum of the liquid to be measured, is the turbidity standard absorption spectrum of the turbidity substance under the current turbidity, stur is the turbidity substance, and any is all wavelengths of the full spectrum.

[0025] Preferably, the wavelength range of the reference wavelength in step S3 is 400-700 nm.

[0026] Preferably, the regression model of step S303 includes logistic regression LR, ordinary least squares regression OLS and support vector machine regression SVR.

[0027] Preferably, the calculation formula for the correction spectrum of the liquid to be tested in step S4 is:

[0028]

[0029] in, is the calibration spectrum of the liquid to be measured, is the original light of the liquid to be measured, is the total turbidity interference spectrum of the liquid to be measured.

[0030] According to the technical solution of the present application, a multivariate turbidity compensation model is established through the spectral and turbidity information in the visible light region, and the contribution ratio of multiple turbidity substances to the total interference spectrum / total turbidity is calculated. The spectral information of complex suspended matter in water is simulated by the spectral information of multiple turbidity substances, thereby improving the universality of the compensation model and the accuracy of water quality detection.

[0031] Other features and advantages of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings, which constitute part of this application, are used to provide a further understanding of the application, and the exemplary embodiments of the application and their descriptions are used to explain the application. In the accompanying drawings:

[0033] Figure 1 Schematic diagram of a process according to a preferred embodiment of the present application;

[0034] Figure 2 is the calibration spectrum of the liquid to be measured. DETAILED DESCRIPTION

[0035] The technical solution of the present application will be described in detail below with reference to the accompanying drawings and in combination with the implementation methods.

[0036] In natural water environments, different turbidity substances have different interference mechanisms and interference contribution rates on the spectrum. In order to compensate for the impact of suspended matter in water on the full spectrum, the existing turbidity correction methods usually use the following methods for turbidity correction: a single turbidity substance solution such as formazine suspension (turbidity standard substance) is used to establish a standard absorption spectrum curve, or the turbidity is calculated in the visible light region and then compensated, or the most suitable turbidity substance is found by using spectral similarity. However, the turbidity substance compensated in these methods is single, and the type of substance causing turbidity in water is not necessarily only one.

[0037] Therefore, the present application provides a water quality monitoring method based on turbidity compensation, which establishes a multivariate turbidity compensation model to adapt to the full-spectrum turbidity correction of complex water bodies in natural environments.

[0038] like Figure 1 As shown, the water quality monitoring method based on turbidity compensation includes the following steps:

[0039] S1 establishes standard absorption spectrum curves of various turbidity substances, and establishes a multivariate standard turbidity compensation database based on a polynomial fitting model. The data dimensions of each multivariate standard turbidity compensation database are turbidity-wavelength-absorbance; preferably, the wavelength range of the absorption spectrum is 200-700nm.

[0040] S2 obtains the absorption spectrum and turbidity of the liquid to be tested; preferably, the wavelength range of the absorption spectrum is 200 to 700 nm.

[0041] It is further explained that the order of step S1 and step S2 can be swapped.

[0042] S3 uses the wavelength in the visible light region as the reference wavelength for multivariate turbidity compensation, and performs regression analysis on the absorbance of the current turbidity liquid at the reference wavelength and the absorbance at the reference wavelengths of multiple turbidity substances to obtain the contribution ratio of each turbidity substance; preferably, the wavelength range of the reference wavelength is 400-700nm.

[0043] To further illustrate, step S3 may include the following steps:

[0044] S301 obtains multiple turbidity standard absorption spectra from a multivariate standard turbidity compensation database according to the turbidity of the liquid to be tested, and obtains the absorbance of the turbidity substance at the reference wavelength;

[0045] S302 establishes the turbidity substance contribution ratio solution equation:

[0046]

[0047] Among them, x iis the contribution ratio of turbidity substances to total turbidity, is the absorbance of the original absorption spectrum of the liquid to be tested at the reference wavelength, is the absorbance of the turbidity substance at the reference wavelength, stur is the turbidity substance, and vis is the reference wavelength of multivariate turbidity compensation;

[0048] S303 uses a regression model to calculate the contribution ratio of turbidity substances; preferably, the regression model includes logistic regression LR, ordinary least squares regression OLS and support vector machine regression SVR.

[0049] S304 calculates the regression error of the absorbance of the liquid to be tested at a reference wavelength in the visible light region based on the turbidity substance contribution ratio results calculated by each model, and selects the turbidity substance contribution ratio that minimizes the regression error;

[0050] S305 calculates the total interference spectrum based on the contribution ratio of the selected turbidity substances:

[0051]

[0052] Among them, x i is the contribution ratio of turbidity substances to total turbidity, is the total turbidity interference spectrum of the liquid to be measured, is the turbidity standard absorption spectrum of the turbidity substance under the current turbidity, stur is the turbidity substance, and any is all wavelengths of the full spectrum.

[0053] S4 calculates the total interference spectrum based on the contribution ratio of each turbidity substance, and subtracts the total interference spectrum from the absorption spectrum of the liquid to be tested to obtain a correction spectrum. Preferably, the correction spectrum calculation formula of the liquid to be tested is:

[0054]

[0055] in, is the calibration spectrum of the liquid to be measured, is the original light of the liquid to be measured, is the total turbidity interference spectrum of the liquid to be measured.

[0056] Example 1

[0057] According to an embodiment of the water quality monitoring method based on turbidity compensation of the present application, natural water bodies can be tested. Considering that natural water bodies may contain sediment and microplastic pollution, a variety of turbidity substances include formazin, PS particles, kaolin and diatomaceous earth.

[0058] First, prepare solutions of varying turbidity, scan absorption spectra, and establish standard absorption spectral curves for various turbidity substances. Prepare a series of gradient formazine solutions to establish standard absorption spectral curves for formazine. Prepare a kaolin suspension and a series of gradient kaolin solutions to establish standard absorption spectral curves for kaolin. Prepare a diatomaceous earth suspension and a series of gradient diatomaceous earth solutions to establish standard absorption spectral curves for diatomaceous earth. Prepare a series of gradient PS microparticle solutions using PS microparticle solutions of varying particle size specifications as mother solutions to establish standard absorption spectral curves for PS microparticles.

[0059] A polynomial fit is performed on the absorbance and turbidity at each wavelength, with the order set to 3, to establish a multivariate standard turbidity compensation database. This database includes turbidity compensation databases for formazin, diatomaceous earth, kaolin, and PS particles. Simply measuring the turbidity of the test liquid directly yields standard absorption spectra for multiple turbidity substances at that turbidity.

[0060] Prepare a standard sample, based on COD solutions of varying concentrations, and add multiple turbidity solutions of varying volumes in a certain proportion to form a multi-mixed solution. Measure the turbidity of the prepared multi-mixed solution and scan to obtain the original absorption spectrum of the liquid to be tested.

[0061] Select the reference wavelength as 550nm, and the absorbance of the absorption spectrum of the liquid to be tested at 550nm The absorbance of the four turbidity substances at 550nm is used to establish the solution equation:

[0062]

[0063] Among them, x i is the contribution ratio of turbidity substances to total turbidity, is the absorbance of the original absorption spectrum of the liquid to be tested at 550nm, is the absorbance of turbidity substance at 550nm;

[0064] The regression models logistic regression (LR), ordinary least squares regression (OLS) and support vector machine regression (SVR) were used to calculate the contribution ratio of turbidity substances and the fitting error at 550 nm.

[0065] The fitting errors of each model at 550 nm are as follows:

[0066]

[0067] The SVR fitting error is the smallest. The contribution ratio of turbidity substances under this model is selected to calculate the total interference spectrum of the liquid to be tested:

[0068]

[0069] in, is the total turbidity interference spectrum of the liquid to be measured, is the turbidity standard absorption spectrum of the turbidity substance under the current turbidity, and stur is the turbidity substance.

[0070] The total interference spectrum is calculated based on the contribution ratio of each turbidity substance. The total interference spectrum is subtracted from the absorption spectrum of the liquid to be tested, and the correction spectrum of the liquid to be tested is calculated as:

[0071]

[0072] from Figure 2 It can be seen that the spectrum after compensation is almost consistent with the sample with zero turbidity and the same COD concentration, and the absorbance error is below 0.06, indicating a good compensation effect.

[0073] Example 2

[0074] A large number of water samples were collected from a pond in Shenzhen. The turbidity of the water samples was measured with a turbidity analyzer, the absorption spectrum was obtained with a spectrometer, and the true value of the COD concentration was obtained using the chemical method according to the national standard.

[0075] Taking 550nm as the reference wavelength for multivariate turbidity compensation, the absorbance of the current turbidity test liquid at a wavelength of 550nm is regressed and analyzed with the absorbance of multiple turbidity substances at a wavelength of 550nm to obtain the contribution ratio of each turbidity substance; the total interference spectrum is calculated based on the contribution ratio of the turbidity substances, and the total interference spectrum is subtracted from the absorption spectrum of the test liquid to obtain the correction spectrum of the pond water sample.

[0076] The partial least squares method is used as the prediction model, and the absorbance between 200 and 400 nm is used as the input of the prediction model to predict the COD concentration of the water sample. The error results are shown in the following table:

[0077]

[0078] Therefore, the determination coefficient of the training set under the multivariate turbidity compensation method is better, and the root mean square error of the test set prediction is lower.

[0079] For the absorption spectrum of pond water samples, the absorbance after formazin compensation is less than 0, which does not conform to the Lambert-Beer law. The training set determination coefficient of compensation for other single turbidity substances (kaolin, diatomaceous earth, PS particles) is lower, and the test set prediction root mean square error is higher. The reason is that the types of turbidity substances in natural water bodies are not single.

[0080] Compared with the prior art, the present invention has the following beneficial effects:

[0081] 1. A multivariate turbidity compensation model is established based on the spectrum and turbidity information in the visible light region. The contribution ratio of multiple turbidity substances to the total interference spectrum / total turbidity is calculated. The spectral information of complex suspended matter in water is simulated with the spectral information of multiple turbidity substances, which improves the universality of the compensation model and the accuracy of water quality detection.

[0082] 2. By establishing a nonlinear solution model, the content of specific substances in water samples is calculated to reflect the complex information of water pollutants. It has the advantages of real-time, fast, and no secondary reagent pollution.

[0083] The preferred embodiments of the present application are described in detail above. However, the present application is not limited to the specific details of the above embodiments. Within the technical concept of the present application, various simple modifications can be made to the technical solution of the present application, and these simple modifications all fall within the scope of protection of the present application.

[0084] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any appropriate manner unless there is any contradiction. In order to avoid unnecessary repetition, this application will not further describe various possible combinations.

[0085] In addition, the various embodiments of the present application may be arbitrarily combined, and as long as they do not violate the concept of the present application, they should also be regarded as the contents disclosed by the present invention.

Claims

1. A water quality monitoring method based on turbidity compensation, characterized in that: The method comprises the following steps: S1 establishes standard absorption spectrum curves of various turbidity substances and establishes a multivariate standard turbidity compensation database based on a polynomial fitting model. The data dimensions of each multivariate standard turbidity compensation database are turbidity-wavelength-absorbance; S2 obtains the absorption spectrum and turbidity of the liquid to be tested; S3 uses the wavelength in the visible light region as the reference wavelength for multivariate turbidity compensation, and performs regression analysis on the absorbance of the current turbidity test liquid at the reference wavelength and the absorbance of multiple turbidity substances at the reference wavelength to obtain the contribution ratio of each turbidity substance; S4 calculates the total interference spectrum according to the contribution ratio of each turbidity substance, and subtracts the total interference spectrum from the absorption spectrum of the liquid to be tested to obtain a correction spectrum; The order of step S1 and step S2 can be swapped.

2. The water quality monitoring method based on turbidity compensation according to claim 1, characterized in that: Various turbidity substances include formazine, PS particles, kaolin and diatomaceous earth.

3. The water quality monitoring method based on turbidity compensation according to claim 1, characterized in that: In step S1 and step S2, the wavelength range of the absorption spectrum is 200 to 700 nm.

4. The water quality monitoring method based on turbidity compensation according to claim 1, characterized in that: Step S3 includes the following steps: S301 obtains multiple turbidity standard absorption spectra from a multivariate standard turbidity compensation database according to the turbidity of the liquid to be tested, and obtains the absorbance of the turbidity substance at the reference wavelength; S302 establishes the turbidity substance contribution ratio solution equation: Among them, x i is the contribution ratio of turbidity substances to total turbidity, is the absorbance of the original absorption spectrum of the liquid to be tested at the reference wavelength, is the absorbance of the turbidity substance at the reference wavelength, stur is the turbidity substance, and vis is the reference wavelength of multivariate turbidity compensation; S303 calculates the contribution ratio of turbidity substances using a regression model; S304 calculates the regression error of the absorbance of the liquid to be tested at a reference wavelength in the visible light region based on the turbidity substance contribution ratio results calculated by each model, and selects the turbidity substance contribution ratio that minimizes the regression error; S305 calculates the total interference spectrum based on the contribution ratio of the selected turbidity substances: Among them, x i is the contribution ratio of turbidity substances to total turbidity, is the total turbidity interference spectrum of the liquid to be measured, is the turbidity standard absorption spectrum of the turbidity substance under the current turbidity, stur is the turbidity substance, and any is all wavelengths of the full spectrum.

5. The water quality monitoring method based on turbidity compensation according to claim 3, characterized in that: The wavelength range of the reference wavelength in step S3 is 400 to 700 nm.

6. The water quality monitoring method based on turbidity compensation according to claim 3, characterized in that: The regression models in step S303 include logistic regression LR, ordinary least squares regression OLS and support vector machine regression SVR.

7. The water quality monitoring method based on turbidity compensation according to claim 1, characterized in that: The calculation formula for the correction spectrum of the liquid to be tested in step S4 is: in, is the calibration spectrum of the liquid to be measured, is the original light of the liquid to be measured, is the total turbidity interference spectrum of the liquid to be measured.

Images