Prediction model for concentration change of dissolved oxygen in riverway water body and construction method and application of prediction model

Abstract

The invention discloses a river channel water dissolved oxygen concentration change prediction model and a construction method and application thereof, and is characterized in that a dissolved oxygen prediction method comprises the following steps: monitoring a predicted target water body to obtain a dissolved oxygen data sample of the target water body; performing PARAFAC analysis on the EEM data of the dissolved organic matters in the target water body to obtain result data of each fluorescent component; analyzing the result data of each component after the PARAFAC analysis so as to determine the tryptophan component of the high oxygen consumption substance; constructing a novel dissolved oxygen kinetic model based on the tryptophan-like component; the dissolved oxygen prediction model is used for predicting the dissolved oxygen of the water body, and has the advantages that the test and data processing method of the model is simple, rapid and efficient, the classic model is more practical and accurate, the model can be used for predicting the attenuation trend of the dissolved oxygen in the polluted river channel, and reliable technical support is provided for monitoring abnormal fluctuation of the dissolved oxygen in the river.

Description

Technical Field

[0001] The present invention relates to the technical field of water quality prediction, and in particular to a prediction model for changes in dissolved oxygen concentration in river water, a construction method and an application thereof. Background Art

[0002] Dissolved oxygen (DO) is an indispensable component of aquatic ecosystems and is closely related to the quality and health of the aquatic environment. The dissolved oxygen level in a river can reflect the respiration level, primary productivity and environmental quality of aquatic organisms in the water area. Low dissolved oxygen conditions are conducive to denitrification and promote the release of active phosphorus and toxic metals (such as arsenic and chromium) from contaminated sediments, leading to endogenous pollution. Low dissolved oxygen conditions caused by eutrophication of river water bodies can easily lead to dead zones in water bodies. On the contrary, high dissolved oxygen conditions can promote the circulation of various elements, which play an important role in maintaining water balance. Predicting changes in dissolved oxygen concentration in rivers is conducive to early warning and forecasting of sudden low oxygen in the water environment and risk optimization control, thereby improving the ability of river water quality to resist shocks.

[0003] At present, the methods used for prediction of dissolved oxygen concentration are mostly focused on deep learning model calculation and estimation, or predicting the back-end value of DO through the prior value of DO. After continuous exploration and improvement, these methods have certain advantages in the accuracy of dissolved oxygen prediction, but their huge data requirements, complicated model parameter adjustment process and the disadvantages of not being widely popularized still have great limitations in practical applications. The presence of high oxygen-consuming substances in water bodies is the main influencing factor of DO concentration changes in water bodies. By characterizing the effect of high oxygen-consuming substances on dissolved oxygen concentration and making effective predictions, it can not only reduce the amount of data required for the model, but also make the model output results more accurate. In the classic Streeter-Phelos (SP) dissolved oxygen prediction model, BOD is used as the aerobic substance indicator parameter, but the BOD test step has the defects of long cycle and low prediction accuracy. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide an efficient, accurate and convenient prediction model for changes in dissolved oxygen concentration in river water bodies, and a construction method and application thereof.

[0005] The technical solution adopted by the present invention to solve the above technical problems is: a prediction model for changes in dissolved oxygen concentration in river water, wherein the prediction model is a tryptophan-DO model, and the formula is as follows: Where D t is the oxygen deficit at a certain moment, in mg·L -1 ; k1 is the first-order decay reaction rate constant of tryptophan-like components in the river, in d-1 ; k2 is the reoxygenation rate constant, in d -1 ; a′ is the effective factor; t is the time, unit d.

[0006] Further, the method for determining the first-order decay reaction rate constant k1 of the tryptophan-like component in the river is as follows: In the formula, CN t is the fluorescence intensity of the tryptophan-like component CN in the water body at time t, in units of RU; CN0 is the initial fluorescence intensity of CN, in units of RU; t is time, d; within 0-20h, the fluorescence intensity of the component CN in the river is measured every 4 hours, and the 5 time series data are substituted into formula (2) for regression analysis to obtain the first-order decay reaction rate constant k1 of the component CN in the river.

[0007] Furthermore, the reoxygenation rate constant k2 and the effective factor a′ are determined as follows: within 0-20 hours, the oxygen deficiency of the river is measured every 4 hours, and the five time series data are substituted into formula (1) for quantitative regression analysis to obtain the reoxygenation coefficient constant k2 and the effective factor a′.

[0008] The present invention also provides a method for constructing the above-mentioned prediction model of dissolved oxygen concentration changes in river water, comprising the following steps: (1) Monitor the predicted target water body, use a multi-parameter water quality analyzer to monitor the dissolved oxygen concentration at each sampling point, and simultaneously obtain the three-dimensional fluorescence spectrum data corresponding to the target water body and the dissolved oxygen data; (2) parsing the three-dimensional fluorescence spectrum data and ultraviolet spectrum data obtained in step (1) using the parallel factor (PARAFAC) algorithm toolbox in Matlab software to obtain the fluorescence components and fluorescence intensity data of each sampling point; (3) performing quantitative regression analysis on the fluorescence intensity data of the fluorescent components at each sampling point obtained in step (2) and the corresponding dissolved oxygen data to determine that the key influencing factor component affecting the dissolved oxygen concentration level in the water body is a tryptophan-like component; (4) Construct a tryptophan-DO prediction model, as follows: Where D t is the oxygen deficit at a certain moment, in mg·L -1 ; k1 is the first-order decay reaction rate constant of tryptophan-like components in the river, in d -1 ; k2 is the reoxygenation rate constant, in d -1 ; a′ is the effective factor; t is the time, unit d.

[0009] Further, step (1) is specifically as follows: monitoring the urban river water body, measuring the dissolved oxygen concentration of the target water body and the three-dimensional fluorescence spectra of the target water body and the blank water sample, subtracting the blank sample fluorescence data from the target water body fluorescence data to remove the Raman scattering effect, setting the fluorescence intensity of the Rayleigh scattered line to missing, and setting the triangular area data on the fluorescence spectrum where the emission wavelength is less than the excitation wavelength to zero, to obtain the three-dimensional fluorescence spectrum data of the sample to be tested; the three-dimensional fluorescence measurement scanning conditions are as follows: Hitachi F-4600 fluorescence photometer, excitation wavelength 220-450nm, scanning interval 5nm, emission wavelength 260-600nm, scanning interval 1nm, slit width 5nm, scanning speed 2400nm·min -1 .

[0010] The present invention also provides an application of the tryptophan-like component-dissolved oxygen prediction model constructed by the above method in predicting dissolved oxygen in river water bodies, which is characterized by the following: the first-order decay reaction rate constant k1, the reoxygenation coefficient constant k2 and the effective factor a′ of the component CN in the river and the initial fluorescence intensity CN0 of the tryptophan-like component are input into the tryptophan-like component-dissolved oxygen model to calculate the oxygen deficit D at a certain moment. t , through oxygen deficiency D t The changes in dissolved oxygen concentration in the river on the second and third days are predicted.

[0011] Technical principle of the present invention: The method of realizing rapid prediction through the tryptophan-DO model of the present invention is mainly based on the influence of high oxygen-consuming substances on dissolved oxygen concentration. The main influencing factors of the change in dissolved oxygen concentration are analyzed and simulated, so as to realize the effective prediction of the change trend of dissolved oxygen. For different soluble organic matter in surface water, their specific chemical composition forms different fluorescence spectral characteristics. The use of PARAFAC fluorescent component (CN) instead of BOD will improve the characterization accuracy of the model independent variable for aerobic organic matter in the water body, making the analysis method more advanced, convenient and accurate. The present invention is based on this principle to establish a quantitative regression model between the fluorescence intensity of high oxygen-consuming substances and the change in dissolved oxygen concentration level, so as to achieve the purpose of rapid prediction of the change trend of dissolved oxygen in surface water.

[0012] Compared with the prior art, the advantages of the present invention are: a prediction model for changes in dissolved oxygen concentration in river water bodies, a construction method and an application thereof, based on the strong correlation between high oxygen-consuming organic matter and dissolved oxygen concentration levels in water bodies, and the powerful identification function of three-dimensional fluorescence spectroscopy for high oxygen-consuming substances, makes the operation simple, the detection cycle short and the cost low. After obtaining the three-dimensional fluorescence spectrum data of the sample, the PARAFAC calculation is performed and then inserted into the tryptophan-DO model, so that the current water body oxygen deficiency and the change trend of dissolved oxygen in the future period can be estimated, and the entire prediction process cycle is short and convenient. The prediction method of the present invention will provide a fast, efficient and reliable method for predicting dissolved oxygen concentration in surface water, and provide a theoretical basis and data support for surface water quality management and pollution early warning control.

[0013] In summary, the present invention provides a surface water dissolved oxygen change trend forecasting technology that integrates sampling, analysis, and prediction, which can be effectively applied to actual water bodies, has the advantages of being cheap, fast, and reliable, and the forecast results are highly accurate. It can effectively solve the defects of large data sets and complex and cumbersome parameter adjustment processes when deep learning models predict dissolved oxygen in surface water, and provide a theoretical basis and data support for surface water quality management and pollution warning control. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A flow chart of the steps for constructing the prediction model of the present invention; Figure 2 The excitation-emission and loading diagrams of the five components (C1-C5) of EEM-PARAFAC are shown; Figure 3 The dynamic changes of dissolved oxygen (a), PARAFAC representative components C2 (b), C5 (c) and DOC (d) during the cultivation process; Figure 4 It is a linear regression diagram between each water quality parameter, fluorescent component and dissolved oxygen concentration in a specific implementation manner; Figure 5 The solid line is the prediction result of the tryptophan-DO model and the Streeter-Phelos model, and the dotted line is the prediction result of the Streeter-Phelos model. The dot is the true value of Dt; Figure 6 is the calculation result of the first-order decay reaction rate constant k1 of the tryptophan-like in the application example; Figure 7 is the calculation result of the reoxygenation coefficient k2 and the effective factor a′ in the application example; Figure 8This is the prediction result of dissolved oxygen content in urban rivers based on the tryptophan-DO model in the application example. DETAILED DESCRIPTION

[0015] The present invention is further described in detail below with reference to the accompanying drawings. 1. Specific embodiments A prediction model for changes in dissolved oxygen concentration in river water is provided. The prediction model is a tryptophan-DO model, and the formula is as follows: Where D t is the oxygen deficit at a certain moment, in mg·L -1 ; k1 is the first-order decay reaction rate constant of tryptophan-like components in the river, in d -1 ; k2 is the reoxygenation rate constant, in d -1 ; a' is the effective factor; t is the time, in units of d. The method for determining the first-order decay reaction rate constant k1 of the tryptophan-like components in the river is as follows: In the formula, CNt is the fluorescence intensity of the tryptophan-like component CN in the water body at time t, in units of RU; CN0 is the initial fluorescence intensity of CN, in units of RU; t is time, d; within 0-20h, the fluorescence intensity of the component CN in the river is measured every 4 hours, and the five time series data are substituted into formula (2) for regression analysis to obtain the first-order decay reaction rate constant k1 of the component CN in the river. The method for determining the reoxygenation rate constant k2 and the effective factor a′ is as follows: within 0-20h, the oxygen deficiency of the river is measured every 4 hours, and the five time series data are substituted into formula (1) for quantitative regression analysis to obtain the reoxygenation coefficient constant k2 and the effective factor a′.

[0017] Its construction method is as follows Figure 1 As shown, the following steps are included: 1. Monitor the predicted target water body, use a multi-parameter water quality analyzer to monitor the dissolved oxygen concentration of each sampling point, and obtain the three-dimensional fluorescence spectrum data of the target water body corresponding to the dissolved oxygen concentration data. Specifically, water samples were collected from urban rivers in Ningbo, a coastal city in southeastern China, and the influent of urban sewage treatment plants were collected for pollution load gradient experiments. In order to simulate the difference in the degree of influence of domestic sewage on urban rivers, we prepared four groups of samples, in which the ratios of filtered river water and sewage were 3:1, 1:1, 1:3 and 1:5 (v / v), respectively. These samples were recorded as R:S=3:1, R:S=1:1, R:S=1:3 and R:S=1:5, and three-stage cultivation experiments were carried out to obtain the dynamic change data of dissolved oxygen; the dissolved oxygen concentration of each river water sample was measured on site when collecting it. Each water sample was filtered through a filter membrane and subjected to three-dimensional fluorescence.

[0018] 2. The three-dimensional fluorescence spectrum data obtained in step 1 were analyzed using the parallel factor (PARAFAC) algorithm toolbox in Matlab software to obtain the fluorescence components and fluorescence intensity data of each sampling point; the three-dimensional fluorescence spectrum data were imported into the PARAFAC model, and finally 5 PARAFAC fluorescence components (such as Figure 2 and Table 1).

[0019] Table 1 Emission and excitation maxima and characteristics of each PARAFAC component 1 The excitation wavelengths in brackets represent secondary peaks.

[0020] The above PARAFAC algorithm is based on the trilinear decomposition theory, that is, assuming that at a certain excitation emission wavelength, the fluorescence intensity of a component is a trilinear function of the concentration of the component and the specific absorption / emission spectral properties. The model is as follows: Where F refers to the number of components in the model, which can be determined by the kernel consistency function; x ijk is the fluorescence intensity data of sample i at emission wavelength j and excitation wavelength k, i.e., the constituent elements of the cubic matrix X(I×J×K) of F; a if is the factor score, reflecting the concentration percentage of component f in sample i, i.e., the constituent element of the component matrix A(I×F); b jf is the load, which is linearly related to the fluorescence quantum yield of the component at the emission wavelength j, i.e., the constituent element of the component matrix E(J×F); c kf is the load, which is proportional to the specific absorption coefficient of component f at the excitation wavelength k, i.e., the constituent element of the component matrix C(K×F). ijkrepresents the residual, including noise and unmodeled data signals. It is solved by using the alternating least squares algorithm to reduce the residual sum of squares (SSR). When SSR < 10 -6 When , the model is considered to have reached convergence.

[0021] 3. Perform quantitative regression analysis on the fluorescence intensity data of the fluorescent components at each sampling point obtained in step 2 and the corresponding dissolved oxygen data to determine that the key influencing factor component affecting the dissolved oxygen concentration level in the water body is the tryptophan-like component. The process analysis is as follows: The dynamic changes of dissolved oxygen during the culture process can be analyzed according to the biodegradation kinetics. The changes of dissolved oxygen in water are based on the balance between the oxygen consumption of microorganisms and the reoxygenation of air. By dividing the instantaneous dissolved oxygen concentration by the initial dissolved oxygen concentration of that stage and normalizing the dissolved oxygen concentration at each stage, the dynamic changes of dissolved oxygen concentration can be more intuitively revealed (such as Figure 3 The dynamic changes of PARAFAC components during the culture process can be used to reveal the nature of the biodegradation process. Figure 3 b and 3c show the dynamic changes of two representative components of PARAFAC (i.e., C2 and C5) during the cultivation period. Among the five components, the tryptophan-like component C2 showed a significant downward trend, indicating that the tryptophan-like component is the main contributor to readily biodegradable organic carbon (BDOC) and an important driving factor for biodegradation activity. Figure 3 d shows the dynamic changes of chemical indicators during the entire cultivation process. Dissolved organic carbon (DOC) showed an overall downward trend during the cultivation process, with the degradation rate of DOC in stage I accounting for 50.2%-80.3% of the total degradation rate, indicating that microorganisms in this stage preferentially utilized soluble DOC. In stages II and III, since the residual DOC was more difficult to degrade, the DOC concentration slowly decreased and tended to stabilize, which corresponded to the dynamic changes in the content of the tryptophan-like component C2.

[0022] Through the cultivation experiment, an important driving factor of biodegradation was identified, namely the tryptophan-like component C2 in the sewage. If the content of the tryptophan-like component C2 can be used to simulate the dynamic changes of dissolved oxygen content, it will be an important breakthrough in predicting the dissolved oxygen consumption mechanism during biodegradation. On this basis, the water quality parameters and fluorescence intensity of the water samples were linearly regressed with dissolved oxygen, such as Figure 4 As shown in the figure, dissolved oxygen is strongly negatively correlated with the tryptophan-like component C2 (0.719), but has a weak negative correlation with the terrestrial humus-like component C5 (0.104). This shows that the tryptophan-like component C2 from domestic sewage is the main factor determining the trend of dissolved oxygen concentration in rivers, while humus-like substances have no significant relationship with the change of dissolved oxygen concentration due to their own biological intractability.

[0023] In order to further analyze the correlation between dissolved oxygen and chemical indices, this study conducted the following correlation studies, such as Figure 4 As shown in Figure 2, the relationship between dissolved oxygen and conductivity, TDN, TDP and DOC is weak, R 2 Adj. The values ​​are 0.369, 0.376, 0.262 and 0.128 respectively. The conductivity is mainly affected by the Na + Mg 2+ , Ca 2+ , Cl- and SO4 2- The effect of plasma strength is not entirely dominated by microbial activity. Nutrients, such as TDN and TDP, may not be the main consumption and limiting factor of the biodegradation process driven by protein substances. As for DOC, although DOC showed a rapid decline in the early stage of cultivation, DOC also contained some organic carbon that could not be biodegraded. Dissolved oxygen and ammonia nitrogen had a moderate relationship, R 2 Adj. =0.533, which may be due to the fact that ammonia nitrogen can reflect the input of domestic sewage to a certain extent. In addition, dissolved oxygen concentration has no direct relationship with turbidity, R 2 Adj =-0.042. This is because under natural water conditions, turbidity is affected by many factors, such as soil particles, residues, organisms and colloids, so there is no direct intrinsic relationship between dissolved oxygen and turbidity in rivers. Therefore, tryptophan-like components are determined to be the key influencing factors of dissolved oxygen.

[0024] 4. Based on the time series data of tryptophan-like components, a tryptophan-like component-dissolved oxygen (tryptophan-DO) prediction model is constructed. The formula is as follows: Where D t is the oxygen deficit at a certain moment, in mg·L -1 ; k1 is the first-order decay reaction rate constant of tryptophan-like components in the river, in d -1 ; k2 is the reoxygenation rate constant, in d -1 ; a′ is the effective factor; t is the time, unit d. The calculation process of the prediction model is as follows: (1) The incubation experiment was divided into three stages. The target water body of the river was aerated to the maximum dissolved oxygen concentration and then placed in an incubator for observation. The changes in dissolved oxygen concentration were observed. This was the first stage, which lasted from 0 to 88 hours. After 88 hours, the river water was aerated again to the maximum dissolved oxygen concentration, and then the changes in dissolved oxygen concentration were observed. This was the second stage, which lasted for 72 hours. After the end of the second stage, the river water was aerated again to the maximum dissolved oxygen concentration for the third time, and the changes in dissolved oxygen concentration were observed. This was the third stage, which lasted for 64 hours.

[0025] Since the microbial activity in the first stage of cultivation is strong, the changes in dissolved oxygen concentration are representative. Therefore, the dissolved oxygen change data of stage I was used to establish a dynamic dissolved oxygen model. The Streeter-Phelos (SP) model is a classic river model for hundreds of years. It is based on the following three assumptions: (a) The oxygen consumption in the river is caused by BOD decay, and the dissolved oxygen in the river comes from atmospheric reoxygenation; (b) The BOD decay and reoxygenation process in the river is a first-order reaction; (c) The reaction rate is constant. Therefore, the attenuation of BOD can be calculated as follows: Where: L t Refers to the BOD concentration in a river at a certain moment, in mg·L -1 ; L0 refers to the initial concentration of BOD in the river, in mg·L -1 ; k1 refers to the attenuation rate coefficient of BOD in the river, unit d -1 ; t is time, unit is d.

[0026] (2) The fluorescence intensity of the tryptophan-like component C2 is proportional to the amount of organic matter. Since the oxygen consumption for complete oxidation of organic matter is proportional to the amount of organic matter, BOD can be calculated as follows: L t =aC2 t (4) Where C2 t Refers to the fluorescence intensity of C2 within a certain period of time, unit RU.

[0027] Substituting equation (3) and L0=aC20 into equation (4), we get the following equation: Where C20 refers to the initial fluorescence intensity of C2 (same meaning as CN in formula (2)), unit: RU.

[0028] Formula (5) was used to quantitatively regress the C2 fluorescence intensity change data of the four groups in stage I to determine the speed coefficient k1. The k1 values ​​obtained by regression were between 0.10 and 0.12, which were very close, so k1 was taken as 0.1 in this study.

[0029] (3) According to the SP model, the rate of change of river hypoxia is the sum of oxygen consumption rate and reoxygenation rate. The dissolved oxygen dynamic model can be expressed by the formula: Where k2 refers to the reoxygenation rate coefficient, unit d -1 ;D t is the oxygen deficit at a certain moment, in mg·L -1 Since the water samples were fully aerated at the beginning of each stage, we assumed that the dissolved oxygen concentration was saturated after each aeration, so the hypoxia value was calculated as follows: D t =DO0-DO t (7), where DO0 refers to the initial dissolved oxygen concentration after aeration, in mg·L -1 ;DO t Refers to the dissolved oxygen concentration at a certain moment, in mg·L -1 Integrating equation (6), we get the following equation: Where D0 refers to the initial hypoxia value, in mg·L -1 According to the above assumption, D0 = 0. Therefore, formula (8) can be simplified as follows: The regression analysis of the four groups of tryptophan-like component C2 data was performed using formula (1), as shown in Figure 5 The k2 values ​​of each group obtained by regression are 0.75, 0.98, 0.86 and 0.82, respectively. 2 Adj. The values ​​are between 0.798 and 0.989, and the values ​​of the effective factor a′ are 2.481, 3.296, 2.616, and 2.426, respectively. Figure 5 The solid lines show that the R 2 Adj. The values ​​were 0.968, 0.988, 0.882 and 0.840, respectively, indicating that tryptophan-like components can be used to effectively predict the changes in dissolved oxygen during the entire biodegradation process. The test and analysis methods of the tryptophan-like component-dissolved oxygen kinetic model are simple, fast and efficient, making the model more practical and accurate. The model can be used to predict the lack of dissolved oxygen in rivers after sewage discharge and provide technical support for abnormal fluctuations in dissolved oxygen in rivers.

[0030] In order to test the superiority of the tryptophan-like component-dissolved oxygen kinetic model in characterizing the aerobic organic matter in water, the BOD index was used as the input variable and input into the SP model for prediction calculation. Four groups of Dt prediction curves under different domestic sewage loads were obtained, such as Figure 5 The dotted line shows that by comparing the model prediction value with the actual DO value, it can be found that there is a significant gap between the theoretical concentration of DO and the actual concentration, and it does not have a good prediction effect on the dissolved oxygen content of the water body. This also proves the scientificity, rationality and irreplaceability of selecting tryptophan-like components to predict DO concentration.

[0031] 2. Application Examples To verify the model, random sampling was conducted in the rivers of Ningbo. The river sections with regional characteristics and without devices that interfere with dissolved oxygen in the water (aeration equipment and water trucks, etc.) were located and sampled. The water samples of the river sections were sampled every 4 hours. The filtered samples of the first day (a total of 5 samples) were added to a quartz cuvette and placed in a Hitachi F-4600 fluorescence photometer. The excitation wavelength was set to 220-450nm, the scanning interval was 5nm, the emission wavelength was set to 260-600nm, the scanning interval was 1nm, the slit width was 5nm, and the scanning speed was 2400nm·min -1 , ultrapure water was treated in the same way, and the samples and blanks were scanned to obtain the fluorescence data of the samples and blanks. The spectral data of the calibration sample set and the samples were analyzed using the PARAFAC algorithm toolbox in Matlab, and the tryptophan-like components were selected, and the corresponding 5 groups of fluorescence data intensity time series data were exported and substituted into The first-order decay reaction rate constant k1 of tryptophan-like is 0.45, such as Figure 6 shown.

[0032] Substituting the first-order decay reaction rate constant k1 of the tryptophan-like component and the five sets of oxygen deficiency time series data on the first day into the tryptophan-like component-dissolved oxygen (tryptophan-DO) model (1), the river water reoxygenation rate coefficient k2 was calculated to be 2.15 and the effective factor a′ was 8.8, as shown in Figure 7 As shown, the tryptophan-DO model is established as

[0033] The tryptophan-like component-dissolved oxygen model was further used to verify the dissolved oxygen concentration prediction performance of the present invention, such as Figure 8 The figure shows the comparison between the predicted dissolved oxygen concentration and the actual value on the second and third days. The predicted dissolved oxygen concentration graphs on the second and third days have the following characteristics: (1) the slope is close to 1; (2) the intercept is close to 0; (3) R 2>0.85. This indicates that the tryptophan-DO model has a high prediction accuracy in actual rivers.

[0034] It can be seen that the surface water dissolved oxygen change trend forecasting technology integrated with sampling, analysis and prediction developed by the present invention can be effectively applied to actual water bodies, and has the advantages of being cheap, fast and reliable. The forecast results are highly accurate and can effectively solve the defects of large data sets and complex and cumbersome parameter adjustment process when deep learning models predict dissolved oxygen in surface water, providing a theoretical basis and data support for surface water quality management and pollution warning control.

[0035] The above description is not intended to limit the present invention, and the present invention is not limited to the above examples. Any changes, modifications, additions or substitutions made by a person skilled in the art within the spirit and scope of the present invention shall also fall within the protection scope of the present invention.

Claims

1. A prediction model for changes in dissolved oxygen concentration in river water. Features The prediction model is a tryptophan-like component-dissolved oxygen model, and the formula is as follows: Where D t is the oxygen deficit at a certain moment, in mg·L -1 ;k 1 is the first-order decay reaction rate constant of tryptophan-like components in rivers, in d -1 ;k 2 is the reoxygenation rate constant, in d -1 ; a' is the effective factor; t is the time, unit d.

2. A prediction model for changes in dissolved oxygen concentration in river water according to claim 1, Features: The first-order decay reaction rate constant k of the tryptophan-like component in the river is 1 The determination method is as follows: Where, CNt is the fluorescence intensity of the tryptophan-like component CN in water at time t, in units of RU; CN 0 is the initial fluorescence intensity of CN, in units of RU; t is time, in d; within 0-20 h, the fluorescence intensity of the component CN in the river was measured every 4 hours, and the five time series data were substituted into formula (2) for regression analysis to obtain the first-order decay reaction rate constant k of the component CN in the river 1 .

3. A prediction model for changes in dissolved oxygen concentration in river water according to claim 2, Features: The reoxygenation rate constant k 2 The method for determining the effective factor a' is as follows: within 0-20 hours, measure the oxygen deficiency of the river every 4 hours, substitute the five time series data into formula (1) for quantitative regression analysis, and obtain the reoxygenation coefficient constant k 2 and the effective factor a'.

4. A method for constructing a prediction model for changes in dissolved oxygen concentration in river water according to any one of claims 1 to 3, Features The following steps are involved: (1) Monitor the predicted target water body, use a multi-parameter water quality analyzer to monitor the dissolved oxygen concentration at each sampling point, and simultaneously obtain the three-dimensional fluorescence spectrum data corresponding to the target water body and the dissolved oxygen data; (2) parsing the three-dimensional fluorescence spectrum data and ultraviolet spectrum data obtained in step (1) using the parallel factor (PARAFAC) algorithm toolbox in Matlab software to obtain the fluorescence components and fluorescence intensity data of each sampling point; (3) performing quantitative regression analysis on the fluorescence intensity data of the fluorescent components at each sampling point obtained in step (2) and the corresponding dissolved oxygen data to determine that the key influencing factor component affecting the dissolved oxygen concentration level in the water body is a tryptophan-like component; (4) Construct a tryptophan-like component-dissolved oxygen prediction model, as follows: Where D t is the oxygen deficit at a certain moment, in mg·L -1 ;k 1 is the first-order decay reaction rate constant of tryptophan-like components in rivers, in d -1 ;k 2 is the reoxygenation rate constant, in d -1 ; a' is the effective factor; t is the time, unit d.

5. The method for constructing a prediction model for changes in dissolved oxygen concentration in river water according to claim 4, Features Step (1) specifically comprises: monitoring the water body of an urban river, determining the dissolved oxygen concentration of the target water body and the three-dimensional fluorescence spectra of the target water body and a blank water sample, subtracting the fluorescence data of the blank sample from the fluorescence data of the target water body to remove the Raman scattering effect, setting the fluorescence intensity of the Rayleigh scattered line to be missing, and setting the data of the triangular area on the fluorescence spectrum where the emission wavelength is less than the excitation wavelength to zero, thereby obtaining the three-dimensional fluorescence spectrum data of the sample to be tested; the scanning conditions for the three-dimensional fluorescence measurement are as follows: Hitachi F-4600 fluorescence photometer, excitation wavelength 220-450nm, scanning interval 5nm, emission wavelength 260-600nm, scanning interval 1nm, slit width 5nm, scanning speed 2400nm·min -1 .

6. An application of the tryptophan-like component-dissolved oxygen prediction model constructed by the method according to claim 4 to predict dissolved oxygen in river water bodies, Features The details are as follows: The first-order decay reaction rate constant k of the CN component in the river is input into the tryptophan-like component-dissolved oxygen model. 1 , reoxygenation coefficient constant k 2 and effective factor a' and tryptophan-like initial fluorescence intensity CN 0 , calculate the oxygen deficit D at a certain moment t , through oxygen deficiency D t The changes in dissolved oxygen concentration in the river on the second and third days are predicted.

Images