Optimization method for ozone addition in ozone activated carbon process of water supply plant

Abstract

The invention provides a method for optimizing ozone addition in an ozone activated carbon process of a water supply plant, which comprises the following steps of: adopting a support vector regression algorithm, taking an ozone activated carbon process water inlet environment index and an ultraviolet absorption spectrum data removal rate as independent variables, and taking a specific organic index removal rate as a dependent variable; and training to obtain a specific organic index removal rate regression prediction model. The ozone dosage is changed within a certain range, the process organic index removal rate under each ozone dosage is calculated according to a removal rate regression prediction model, and the process cost per ton of water under each ozone dosage is calculated according to the liquid oxygen consumption rate and the ozone generation system power. And analyzing the grey correlation degree with the ideal scheme under each ozone dosage by adopting a grey correlation analysis method, and determining that the ozone dosage with the highest grey correlation degree is optimal. The method can assist the water supply plant in rapidly optimizing ozone adding of the ozone activated carbon process by using the ultraviolet absorption spectrum, and has important significance for realizing efficient decision-making of the ozone adding process of the water supply plant.

Description

Technical Field

[0001] The invention relates to the technical field of water treatment, and in particular to an optimization method for ozone addition in an ozone activated carbon process of a water supply plant. Background Art

[0002] With the continuous development of water treatment technology, the ozone-activated carbon process is being used more and more widely in drinking water treatment. Compared with traditional water supply processes, it has certain advantages in removing organic matter and reducing the generation of disinfection by-products. However, the optimization of ozone dosage remains one of the key challenges in the water treatment process. The ozone dosage not only directly affects the removal efficiency of organic matter, but also affects the operating cost of the water supply plant. At present, the determination of ozone dosage usually relies on experience, lacks a unified and scientific optimization method, and is difficult to dynamically adjust according to different water quality conditions, which may lead to insufficient or excessive ozone addition, thereby affecting water quality or increasing unnecessary operating costs.

[0003] Traditional water quality assessment usually relies on the permanganate index (COD Mn ), total organic carbon (TOC), dissolved organic carbon (DOC), trihalomethane generation potential (THMFP) and other conventional organic indicators often require complex experimental operations and expensive instruments and equipment, and it is difficult to achieve accurate online real-time monitoring, which cannot meet the increasing intelligent needs of water plants. In contrast, ultraviolet spectroscopy can quickly and easily reflect changes in organic matter concentration by detecting ultraviolet absorbance at a specific wavelength, and can serve as an important indicator of changes in organic matter concentration. Compared with traditional organic matter indicator detection methods, ultraviolet spectroscopy can not only achieve online monitoring, but also has low cost and easy operation. It can reflect the dynamic changes of water quality in real time and provides a new technical means for optimizing ozone dosage. Summary of the invention

[0004] The present invention provides a method for optimizing ozone dosage in an ozone activated carbon process of a water supply plant, so as to solve the technical problem in the prior art that it is difficult for a water supply plant to quickly and accurately obtain the operating effect of an ozone activated carbon process under different ozone dosages, and thus it is difficult to determine the optimal ozone dosage by balancing the organic matter removal rate and economic cost of the ozone activated carbon process.

[0005] In order to achieve the above object, the present invention proposes a method for optimizing ozone addition in an ozone activated carbon process of a water supply plant, comprising:

[0006] A method for optimizing ozone addition in an ozone activated carbon process of a water supply plant, characterized by comprising:

[0007] Establish a regression prediction model P-Model for the removal rate of organic indicators in the effluent relative to the influent of the ozone activated carbon process in a water supply plant and train the P-Model model;

[0008] The ozone dosage was changed, and the removal of organic matter in the ozone activated carbon process under different ozone dosages was analyzed by the P-Model model, and the optimal ozone dosage was determined in combination with the operation cost data.

[0009] Preferably, a regression prediction model P-Model of the removal rate of organic indicators in the effluent relative to the influent of the ozone activated carbon process of the water supply plant is established and the P-Model model is trained, which specifically includes the following steps:

[0010] According to the time difference Δt of water flow from the ozone activated carbon process water inlet collection point to the water outlet collection point i , analyze T i Environmental indicators, ultraviolet absorption spectrum data, organic indicators and T of the influent of the ozone activated carbon process at all times i +Δt i Ultraviolet absorption spectrum data and organic indexes of effluent from ozone activated carbon process at all times;

[0011] Calculate T i +Δt i The UV absorption spectrum data and organic indexes of the effluent from the ozone activated carbon process at the time T i The removal rate of organic matter in the influent of the ozone activated carbon process at all times;

[0012] A support vector regression algorithm was adopted, with the ozone activated carbon process influent environmental indicators and ultraviolet absorption spectrum data removal rate as independent variables, and the organic indicator removal rate as the dependent variable (in this application, the "organic indicator" means that for different water treatment standard requirements, the organic indicators are also different. For example, for standard A, the organic indicators are permanganate index and total organic carbon, and for standard B, the organic indicators are permanganate index, total organic carbon, dissolved organic carbon and trihalomethane generation potential, that is, different water treatment standards require different organic indicators for output removal rate), and a regression prediction model P-Model of the organic indicators of the effluent relative to the influent removal rate of the ozone activated carbon process of the water supply plant was obtained through training.

[0013] Preferably, a support vector regression algorithm is used, with the ozone activated carbon process influent environmental indicators and ultraviolet absorption spectrum data removal rate as independent variables and the organic indicator removal rate as the dependent variable, to train a regression prediction model P-Model for the relative influent removal rate of the ozone activated carbon process effluent organic indicators of the water supply plant, specifically including the following steps:

[0014] Obtain the independent variable data set X = {X1, X2, …, X i ,…,X n} and its correspondence (in this application, the "correspondence" means for the independent variable X a, whose dependent variable is y a , a is a positive integer) the dependent variable data set Y = {y1,y2,…,y i ,…,y n}; where y i is the specific organic index of the i-th sample, i is a positive integer; X i for:

[0015] X i =(x i1 ,x i2 ,x i3 ,x i4 ,x i5 ,x i6 ,x i7 ,x i8 ) T

[0016] Among them, x i1 、x i2 、x i3 、x i4 、x i5 、x i6 、x i7 、x i8 They respectively represent the influent flow rate, water temperature, pH, ammonia nitrogen of the i-th sample and the removal rate of the absorbance of the effluent at wavelengths of 220nm, 230nm, 254nm, and 272nm relative to the influent water.

[0017] Normalize all sample data sets X and Y to obtain the independent variable input matrix X′=[X1′,X2′,…,X i ′,…,X n ′] and its corresponding dependent variable input matrix Y′=[y1′,y2′,…,y i ′,…,y n ′]; where X i ′=(x i ′1,x i ′2,…,x i ′8) T ;

[0018] The basic form of the specific organic index removal rate regression prediction model trained by support vector regression is:

[0019] f(X ★ )=w T X ★ +b

[0020] Among them, X ★ is the input feature vector of the unknown sample independent variable data, w is the weight vector, and b is the bias term.

[0021] For the i-th sample, the objective function form and constraints of support vector regression training are:

[0022] Objective function:

[0023] Constraints: y i -(w T X i +b)≤ε+ξ i (ξ i ≥0)

[0024]

[0025] Among them, ||w|| 2 is the model complexity, ξ i and is a slack variable that allows the training error to exceed the tolerance ε, and C is a hyperparameter that balances the training error and model complexity.

[0026] The radial basis function is selected as the kernel function to map the input data of n samples into a high-dimensional space; the kernel function of the i-th sample is:

[0027]

[0028] Among them, σ is the width of the kernel function.

[0029] The Lagrange multiplier is introduced, and the objective function optimization is transformed into a dual problem through Lagrange duality; wherein the dual problem expression is:

[0030]

[0031]

[0032] Among them, α i , α j The Lagrange multiplier for handling positive errors exceeding the tolerance ε is, is the Lagrange multiplier for handling negative errors exceeding the tolerance ε.

[0033] The P-Model model is obtained by solving the dual problem, and the P-Model model is specifically expressed as follows:

[0034]

[0035] The output result y of the P-Model model ★ ′ is denormalized to obtain the prediction result y ★ .

[0036] Preferably, all sample data sets X and Y are normalized, and the specific calculation is as follows:

[0037]

[0038]

[0039]

[0040] max y =max(y1,y2,…,y i ,…,y n )

[0041] min y =min(y1,y2,…,y i ,…,y n )

[0042]

[0043] Preferably, the output result y of the P-Model model ★ ′ is denormalized to obtain the prediction result y ★ , specifically expressed as follows:

[0044]

[0045] Preferably, the water supply plant should use the latest independent variable data set and its corresponding dependent variable data set to retrain the P-Model at least once a year.

[0046] Preferably, the ozone dosage is changed, and the removal of organic matter in the ozone activated carbon process under different ozone dosages is analyzed by the P-Model model, and the optimal ozone dosage is determined in combination with the operating cost data, which specifically includes the following steps:

[0047] Change the ozone dosage. After the operation is stable under each ozone dosage, the time difference Δt from the water inlet collection point to the water outlet collection point of the ozone activated carbon process is calculated. j , analyze T j Environmental indicators, ultraviolet absorption spectrum data and T of process water at all times j +Δt j Ultraviolet absorption spectrum data of process effluent at all times;

[0048] According to the model P-Model, the removal rate of organic indexes in the effluent of the ozone activated carbon process relative to the influent water was calculated at each ozone dosage;

[0049] According to the liquid oxygen consumption rate and the power of the ozone generation system, calculate the cost per ton of water for the ozone activated carbon process at each ozone dosage;

[0050] The grey correlation analysis method is used to analyze the grey correlation between each ozone dosage and the ideal solution, and the ozone dosage with the highest grey correlation is determined to be the optimal one.

[0051] Preferably, the ozone dosage should be changed to ensure that the residual ozone concentration in the effluent from the ozone contact tank is not less than 0.02 mg / L and not more than 0.10 mg / L, and that the bromate concentration in the effluent from the ozone activated carbon process is lower than the standard limit of the water supply plant.

[0052] Preferably, the grey correlation analysis method is used to analyze the grey correlation degree between each ozone dosage and the ideal solution, including the following steps:

[0053] Different ozone dosage O1, O2, ...O j ,…,O m (where j is a positive integer), the removal rates of organic indicators in process effluent relative to influent are R1, R2, ...R j , …, R m The cost per ton of water is C1, C2, ...C j ,…,C m ;

[0054] The effluent organic index relative to the influent removal rate and the ton water cost data are standardized, and the standardized data under the jth ozone dosage is:

[0055]

[0056]

[0057] Grey correlation degree Π between the jth ozone dosage and the ideal solution j The calculation is as follows:

[0058]

[0059] Among them, α is the weight of the effluent organic index relative to the influent removal rate in evaluating the effect of ozone activated carbon process, It is the resolution coefficient of the effluent organic index relative to the influent removal rate and the cost per ton of water data.

[0060] Preferably, the ozone dosage O is calculated j Corresponding process effluent organic index relative to influent removal rate R j The following steps are included:

[0061] The r organic indicators including the concentration of permanganate index, total organic carbon and specific organic substances of concern to the water supply plant (in this application, the "specific organic substances" refer to other organic indicators selected by the water supply plant according to the water supply standard in addition to the permanganate index and total organic carbon. For example, for water supply plant C, its water supply is for drinking water for residents, and for water supply plant D, its water supply is for industrial water supply, then the specific organic indicators of water supply plant C and water supply plant D are different) are y1, y2, ..., y m , …, y r ;

[0062] Calculation of specific organic indicators m The average proportion of W in the water discharged from the factory in the past year m and the maximum occupancy rate M m ; The specific expressions are as follows:

[0063]

[0064] M m =max[max(y m,max / L m ),0.01]

[0065] Among them, y m,max , L m are specific indicators y m The measured maximum value and the factory water standard limit, N is the specific indicator y m Number of tests in factory water in at least the past year.

[0066] According to the average occupancy rate W m and the maximum occupancy rate M m , calculate the specific indicator y m The weight w m , specifically expressed as follows:

[0067]

[0068] Input data X j =(x j1 ,x j2 ,…,x j8 ) T Normalization is performed to obtain the input matrix X j ′=(x j ′1,x j ′2,…,x j ′8) T ; The specific expressions are as follows:

[0069]

[0070] The input matrix Xj Substitute the r organic index removal rate regression prediction model including permanganate index, total organic carbon and the concentration of specific organic matter concerned by the water supply plant into the prediction value Y of r organic index j ={y j1 ,y j2 ,…,y jm ,…,y jr}, then R j for:

[0071]

[0072] Preferably, the water supply plant should determine the optimal ozone dosage at least once every quarter.

[0073] The present invention provides an optimization method for ozone addition in an ozone activated carbon process of a water supply plant. The method adopts a support vector regression algorithm, takes the ozone activated carbon process influent environmental index and ultraviolet absorption spectrum data removal rate as independent variables, takes the specific organic index removal rate as a dependent variable, and trains to obtain a regression prediction model for the specific organic index removal rate; the ozone dosage is changed within a certain range, and the process organic index removal rate under each ozone dosage is calculated according to the regression prediction model for the specific organic index removal rate; the process water cost per ton under each ozone dosage is calculated according to the liquid oxygen consumption rate and the ozone generation system power; the gray correlation analysis method is used to analyze the gray correlation degree with the ideal solution under each ozone dosage, and the ozone dosage with the highest gray correlation degree is determined to be the optimal. The present invention can assist water supply plants in rapidly optimizing the ozone addition in the ozone activated carbon process using ultraviolet absorption spectrum data, and is of great significance for realizing efficient decision-making in the ozone addition process in water supply plants. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0075] Figure 1 It is a flow chart of steps of an optimization method for adding ozone in an ozone activated carbon process of a water supply plant disclosed by the present invention;

[0076] Figure 2 is the regression prediction model of the permanganate index in the embodiment;

[0077] Figure 3 It is the regression prediction model of total organic carbon in the embodiment. DETAILED DESCRIPTION

[0078] In order to make the above and other features and advantages of the present invention more clear, the present invention is further described below in conjunction with the accompanying drawings. It should be understood that the specific embodiments given herein are for the purpose of explaining to those skilled in the art and are only exemplary and not restrictive.

[0079] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise", "axial", "radial", "circumferential" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0080] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0081] The implementation object of this embodiment is a large-scale water plant in East China, with a daily water supply scale of 700,000 tons. There are four water production lines in the plant, and the embodiment of the present invention is implemented in one of the production lines with a water supply scale of 250,000 tons / day and using ozone activated carbon as a deep treatment process. Figure 1 As shown, this embodiment provides an optimization method for ozone addition in an ozone activated carbon process of a water supply plant, comprising the following steps:

[0082] S1, establish a regression prediction model P-Model for the removal rate of organic indicators in the effluent of the water supply plant ozone activated carbon process relative to the influent and train the P-Model model, which specifically includes the following steps:

[0083] S11, based on the time difference Δt of the water flow from the ozone activated carbon process water inlet collection point to the water outlet collection point i , analyze T i Environmental indicators, UV absorption spectrum data, organic indicators and T of process water at all times i +Δt iUltraviolet absorption spectrum data and organic indicators of process effluent at all times, including environmental indicators such as flow rate, water temperature, pH, and ammonia nitrogen; ultraviolet absorption spectrum data including absorbance at wavelengths of 220nm, 230nm, 254nm, and 272nm; and organic indicators including permanganate index, total organic carbon, and 2-MIB concentration;

[0084] S12, calculate T i +Δt i The UV absorption spectrum data and organic indexes of the effluent from the ozone activated carbon process at the time T i The removal rate of the influent of the ozone activated carbon process at all times;

[0085] S13, using support vector regression algorithm, taking ozone activated carbon process influent environmental indicators and ultraviolet absorption spectrum data removal rate as independent variables, taking specific organic index removal rate as dependent variable, training to obtain a specific organic index removal rate regression prediction model, including the following steps:

[0086] Collect the independent variable data set X={X1,X2,…,X i ,…,X n} and its corresponding dependent variable data set Y = {y1,y2,…,y i ,…,y n}, where n = 60, y i is the specific organic index of the i-th sample, X i for:

[0087] X i =(x i1 ,x i2 ,x i3 ,x i4 ,x i5 ,x i6 ,x i7 ,x i8 ) T

[0088] Where x i1 、x i2 、x i3 、x i4 、x i5 、x i6 、x i7 、x i8 They respectively represent the influent flow rate, water temperature, pH, ammonia nitrogen of the i-th sample and the removal rate of the absorbance of the effluent at wavelengths of 220nm, 230nm, 254nm, and 272nm relative to the influent water.

[0089] After normalizing all sample data sets X and Y, we can get the independent variable input matrix X′=[X1′,X2′,…,X i ′,…,X n ′] and its corresponding dependent variable input matrix Y′=[y1′,y2′,…,y i ′,…,y n ′], where X i ′=(x i ′1,x i ′2,…,x i ′8) T . Normalization is done using the following method:

[0090]

[0091]

[0092]

[0093] max y =max(y1,y2,…,y i ,…,y n )

[0094] min y =min(y1,y2,…,y i ,…,y n )

[0095]

[0096] The basic form of the specific organic index removal rate regression prediction model trained by support vector regression is:

[0097] f(X ★ )=w T X ★ +b

[0098] In the formula, X ★ is the input feature vector of the unknown sample independent variable data, w is the weight vector, and b is the bias term.

[0099] For the i-th sample, the objective function form and constraints of support vector regression training are:

[0100] Objective function:

[0101] Constraints: y i -(w T X i +b)≤ε+ξ i (ξ i ≥0)

[0102]

[0103] In the formula, ||w|| 2 is the model complexity, ξ i and is a slack variable that allows the training error to exceed the tolerance ε, and C is a hyperparameter that balances the training error and model complexity.

[0104] The radial basis function is selected as the kernel function to map the input data of n samples to a high-dimensional space, where the kernel function of the i-th sample is:

[0105]

[0106] Where σ is the width of the kernel function.

[0107] The Lagrange multiplier is introduced, and the objective function optimization is transformed into a dual problem through Lagrange duality. The dual problem expression is:

[0108]

[0109]

[0110] In the formula, α i , α j The Lagrange multiplier for handling positive errors exceeding the tolerance ε is, is the Lagrange multiplier for handling negative errors exceeding the tolerance ε.

[0111] The regression prediction model is obtained by solving the dual problem:

[0112]

[0113] The output results of the regression prediction model are denormalized to obtain the prediction result y ★ :

[0114]

[0115] The model training process is carried out using the built-in function of the SVR (Support Vector Regression) class in the svm (Support Vector Machines) module in Python's scikit-learn machine learning library. Since the 2-MIB concentration in the influent of the ozone activated carbon process is basically below the detection limit, the regression prediction model of the permanganate index and total organic carbon is mainly established, such as Figure 2 and Figure 3 shown.

[0116] For the permanganate index y1 and total organic carbon y2, calculate their average proportion W in the water discharged from the factory in the past year. m and the maximum occupancy rate M m :

[0117]

[0118] M m =max[max(y m,max / L m ),0.01]

[0119] Among them, y m,max , L m are specific indicators y m The measured maximum value and the factory water standard limit, N is the specific indicator y m Number of tests in factory water in at least the past year.

[0120] Then the specific indicator y m The weight w m for:

[0121]

[0122] From October 2023 to October 2024, the average and maximum percentages of permanganate index in the outlet water were 0.45 and 0.53 respectively, and the average and maximum percentages of total organic carbon in the outlet water were 0.60 and 0.68 respectively. The weights of permanganate index and total organic carbon were 0.43 and 0.57 respectively.

[0123] S2, changing the ozone dosage, analyzing the removal of organic matter in the ozone activated carbon process under different ozone dosages through the P-Model model, and determining the optimal ozone dosage in combination with the operating cost data, specifically including the following steps:

[0124] S21, in November 2024, the ozone dosage O1, O2, O3 were changed to 0.3mg / L, 0.45mg / L and 0.6mg / L respectively. The residual ozone concentration in the effluent of the ozone contact tank under the three dosages was not less than 0.02mg / L and not more than 0.10mg / L. At the same time, the bromate concentration of the effluent of the ozone activated carbon was lower than the standard limit of 0.005mg / L of the water supply plant. After each ozone dosage condition was running stably, according to the time difference Δt of the water flow from the water inlet collection point to the water outlet collection point of the ozone activated carbon process j , analyze T j Environmental indicators, ultraviolet absorption spectrum data and T of process water at all times j +Δt j Ultraviolet absorption spectrum data of process effluent at all times;

[0125] S22, input data X j =(x j1 ,x j2 ,…,x j8 ) T (j=1,2,3) The input matrix X is obtained by normalizing it using the following method j ′=(x j ′1,x j ′2,…,x j ′8) T :

[0126]

[0127] The input matrix X j Substituting the permanganate index and total organic carbon into the regression prediction model, the predicted values ​​of the two organic indicators under three ozone dosages were obtained as Y1={1.4%, 21.6%}, Y2={3.3%, 24.9%}, and Y1={6.5%, 22.7%}, respectively. After actual testing, it was found that the predicted relative errors of the permanganate index and total organic carbon were no more than 5%. The relative removal rates of organic indicators in the process effluent under different ozone dosages were R1, R2, and R3, respectively, 12.9%, 15.6%, and 15.7%, respectively.

[0128] S23, according to the liquid oxygen consumption rate and the power of the ozone generation system, the cost of ozone activated carbon process per ton of water C1, C2, C3 at each ozone dosage is calculated to be 0.12, 0.14, 0.17 yuan respectively;

[0129] S24, using grey correlation analysis to analyze the grey correlation between each ozone dosage and the ideal solution, and determining the ozone dosage with the highest grey correlation as the optimal one. The steps include:

[0130] For different ozone dosages O1, O2, and O3, the removal rates of organic indicators in process effluent relative to influent are R1, R2, and R3, and the cost per ton of water is C1, C2, and C3, respectively. First, the data of the removal rate of organic indicators in effluent relative to influent and the cost per ton of water are standardized. The standardized data under the jth ozone dosage are:

[0131]

[0132]

[0133] Grey correlation degree ∏ between the jth ozone dosage and the ideal solution j The calculation is as follows:

[0134]

[0135] In the formula, α is the weight of the effluent organic index relative to the influent removal rate in evaluating the effect of ozone activated carbon process, It is the resolution coefficient of the effluent organic index relative to the influent removal rate and the cost per ton of water data.

[0136] In the embodiment, α is taken as 0.4 according to the practice of the water plant. All of them are 0.5, then the grey correlation degree of different ozone dosages O1, O2, O3 with the ideal solution is ∏ j They are 0.73, 0.71 and 0.60 respectively, and it is determined that the ozone dosage of 0.30 mg / L is the optimal dosage under the current conditions.

[0137] The above description is only a preferred embodiment of the present invention and does not limit the present invention in other forms. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A method for optimizing ozone addition in an ozone activated carbon process in a water supply plant, characterized in that: include: Establish a regression prediction model P-Model for the removal rate of organic indicators in the effluent relative to the influent of the ozone activated carbon process in a water supply plant and train the P-Model model; The ozone dosage was changed, and the removal of organic matter in the ozone activated carbon process under different ozone dosages was analyzed by the P-Model model, and the optimal ozone dosage was determined in combination with the operation cost data.

2. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 1, characterized in that: The P-Model is established to predict the relative removal rate of organic indexes in the effluent of the ozone activated carbon process in the water supply plant and the P-Model is trained, which specifically includes the following steps: According to the time difference Δt of water flow from the ozone activated carbon process water inlet collection point to the water outlet collection point i , analyze T i Environmental indicators, ultraviolet absorption spectrum data, organic indicators and T of the influent of the ozone activated carbon process at all times i +Δt i Ultraviolet absorption spectrum data and organic indexes of effluent from ozone activated carbon process at all times; Calculate T i +Δt i The UV absorption spectrum data and organic indexes of the effluent from the ozone activated carbon process at the moment T i The removal rate of organic matter in the influent of the ozone activated carbon process at all times; Support vector regression algorithm was used to train the regression prediction model P-Model of organic index removal rate relative to influent of ozone activated carbon process in water supply plant, taking the removal rate of influent environmental index and ultraviolet absorption spectrum data of ozone activated carbon process as independent variables and the removal rate of organic index as dependent variable.

3. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 2, characterized in that: Environmental indicators include flow rate, water temperature, pH, ammonia nitrogen, ultraviolet absorption spectrum data include absorbance at wavelengths of 220nm, 230nm, 254nm, and 272nm, and organic indicators include permanganate index, total organic carbon, and the concentration of specific organic substances that water supply plants are concerned about.

4. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 2, characterized in that: The support vector regression algorithm was used, with the ozone activated carbon process influent environmental indicators and ultraviolet absorption spectrum data removal rate as independent variables and the organic indicator removal rate as the dependent variable, and the training was used to obtain the regression prediction model P-Model of the relative influent removal rate of the ozone activated carbon process effluent organic indicators of the water supply plant, which specifically includes the following steps: Obtain the independent variable data set X = {X1, X2, …, X i ,…,X n } and its corresponding dependent variable data set Y = {y1,y2,…,y i ,…,y n }; where y i is the specific organic index of the i-th sample, i is a positive integer; X i for: X i =(x i1 ,x i2 ,x i3 ,x i4 ,x i5 ,x i6 ,x i7 ,x i8 ) T ; Among them, x i1 、x i2 、x i3 、x i4 、x i5 、x i6 、x i7 、x i8 They represent the flow rate, water temperature, pH, ammonia nitrogen of the influent of the i-th sample, and the removal rate of the absorbance of the effluent at wavelengths of 220 nm, 230 nm, 254 nm, and 272 nm relative to the influent; Normalize all sample data sets X and Y to obtain the independent variable input matrix X ′ =[X1 ′ ,X2 ′ ,…,X i ′ ,…,X n ′ ] and its corresponding dependent variable input matrix Y ′ =[y1 ′ ,y2 ′ ,…,y i ′ ,…,y n ′ ]; where X i ′ =(x i ′ 1,x i ′ 2,…,x i ′ 8) T ; The basic form of the specific organic index removal rate regression prediction model trained by support vector regression is: f(X ★ )=w T X ★ +b; Among them, X ★ is the input feature vector of the unknown sample independent variable data, w is the weight vector, and b is the bias term; For the i-th sample, the objective function form and constraints of support vector regression training are: Objective function: Constraints: y i -(w T X i +b)≤ε+ξ i (ξ i ≥0); Among them, ||w|| 2 is the model complexity, ξ i and is a slack variable that allows the training error to exceed the tolerance ε, and C is a hyperparameter that balances the training error and model complexity; The radial basis function is selected as the kernel function to map the input data of n samples into a high-dimensional space; the kernel function of the i-th sample is: Among them, σ is the width of the kernel function; The Lagrange multiplier is introduced, and the objective function optimization is transformed into a dual problem through Lagrange duality; wherein the dual problem expression is: Among them, α i , α j The Lagrange multiplier for handling positive errors exceeding the tolerance ε is, Lagrange multiplier for handling negative errors exceeding the tolerance ε; The P-Model model is obtained by solving the dual problem, and the P-Model model is specifically expressed as follows: The output result y of the P-Model model ★ ′ Perform denormalization to obtain the predicted result y ★ .

5. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 4, characterized in that: All sample data sets X and Y are normalized. The specific calculation is as follows: max y =max(y1,y2,…,y i ,…,and n ); min y =min(y1,y2,…,y i ,…,y n ); 6. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 4, characterized in that: The output result y′ of the P-Model model ＊ Perform denormalization to obtain the predicted result y ＊ , specifically expressed as follows:

7. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 2, characterized in that: The water supply plant should use the latest independent variable data set and its corresponding dependent variable data set to retrain the P-Model at least once a year.

8. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 1, characterized in that: When changing the ozone dosage, the residual ozone concentration in the effluent from the ozone contact tank should be no less than 0.02 mg / L and no more than 0.10 mg / L. At the same time, the bromate concentration in the effluent from the ozone activated carbon process should be lower than the standard limit of the water supply plant.

9. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 1, characterized in that: The ozone dosage is changed, and the removal of organic matter in the ozone activated carbon process under different ozone dosages is analyzed by the P-Model model, and the optimal ozone dosage is determined in combination with the operating cost data, which specifically includes the following steps: Change the ozone dosage. After the operation is stable under each ozone dosage, the time difference Δt from the water inlet collection point to the water outlet collection point of the ozone activated carbon process is calculated. j , analyze T j Environmental indicators, ultraviolet absorption spectrum data and T of process water at all times j +Δt j Ultraviolet absorption spectrum data of process effluent at all times; According to the model P-Model, the removal rate of organic indexes in the effluent of the ozone activated carbon process relative to the influent water was calculated at each ozone dosage; According to the liquid oxygen consumption rate and the power of the ozone generation system, calculate the cost per ton of water for the ozone activated carbon process at each ozone dosage; The grey correlation analysis method is used to analyze the grey correlation between each ozone dosage and the ideal solution, and the ozone dosage with the highest grey correlation is determined to be the optimal one.

10. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 9, characterized in that: The grey correlation analysis method is used to analyze the grey correlation degree between each ozone dosage and the ideal solution, including the following steps: Different ozone dosage O1, O2, ...O j ,…,O m (where j is a positive integer), the removal rates of organic indicators in process effluent relative to influent are R1, R2, ...R j , …, R m The cost per ton of water is C1, C2, ...C j ,…,C m ; The effluent organic index relative to the influent removal rate and the ton water cost data are standardized, and the standardized data under the jth ozone dosage is: Grey correlation degree Π between the jth ozone dosage and the ideal solution j The calculation is as follows: Among them, α is the weight of the effluent organic index relative to the influent removal rate in evaluating the effect of ozone activated carbon process, It is the resolution coefficient of the effluent organic index relative to the influent removal rate and the cost per ton of water data.

11. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 10, characterized in that: Calculate the ozone dosage j Corresponding process effluent organic index relative to influent removal rate R j The following steps are included: The r organic indicators, including the permanganate index, total organic carbon, and the concentration of specific organic substances of concern to water supply plants, are y1, y2, …, y m , …, y r ; Calculation of specific organic indicators m The average proportion of W in the water discharged from the factory in the past year m and the maximum occupancy rate M m ; The specific expressions are as follows: M m =max[max(y m,max / L m ),0.01]; Among them, y m,max , L m are specific indicators y m The measured maximum value and the factory water standard limit, N is the specific indicator y m The number of tests in the factory water in the past year or more; According to the average occupancy rate W m and the maximum occupancy rate M m , calculate the specific indicator y m The weight w m , specifically expressed as follows: Input data X j =(x j1 ,x j2 ,…,x j8 ) T Normalization is performed to obtain the input matrix X j ′ =(x j ′ 1,x j ′ 2,…,x j ′ 8) T ; The specific expressions are as follows: The input matrix X j ′ Substitute the r organic index removal rate regression prediction model including permanganate index, total organic carbon and specific organic matter concentration of concern to the water supply plant to obtain the predicted value Y of r organic index j ={y j1 ,y j2 ,…,y jm ,…,y jr }, then R j for:

12. The method for optimizing ozone addition in an ozone activated carbon process in a water supply plant according to claim 9, characterized in that: Water supply plants should conduct the above determination of the optimal ozone dosage at least once every quarter.

Images