CN114783547B - Accurate sewage classification treatment method - Google Patents

Abstract

The invention relates to the technical field of sewage classification treatment, and discloses a precise sewage classification treatment method, which comprises the following steps: constructing a sewage internal substance set, and constructing a chemical reaction relation rule set among internal substances according to the sewage internal substance set; constructing a classification decision tree according to the chemical reaction relation rule set; classifying the collected sewage according to the classification decision tree; constructing a sewage decision-making treatment model by using deep reinforcement learning, and training the model by using a Q-learning algorithm; the sewage after mixed treatment is input into a trained sewage decision treatment model, and the sewage decision treatment model outputs a mixed sewage treatment method which comprises a mixed sewage treatment flow, a mixed sewage treatment agent ratio and content. The method realizes the classification decision of the sewage classification treatment method by utilizing the decision tree, and realizes the accurate classification treatment of substances in different pollutants by utilizing the sewage treatment agent configuration method based on reinforcement learning, thereby avoiding secondary pollution.

Description

Accurate sewage classification treatment method

Technical Field

The invention relates to the technical field of sewage classification treatment, in particular to a precise sewage classification treatment method.

Background

The existing sewage classification treatment method does not consider the problem of secondary pollution in sewage treatment caused by chemical reaction among substances in the sewage, for example, metal salts can be formed by the chemical reaction of heavy metals and Polycyclic Aromatic Hydrocarbons (PAHs), secondary pollution is caused, toxic gases can be generated by the reaction of organic phosphorus and chlorine, and inorganic substances can be formed by the reaction of heavy metals (Cu) and organic chlorine. Meanwhile, in the anaerobic digestion process, organic pollutants are decomposed into CH ₄、CO₂、CO、H₂ O and H ₂ S by facultative bacteria and anaerobic bacteria under anaerobic conditions, but persistent toxic organic pollutants can be generated under anaerobic conditions, and the patent provides an accurate sewage classification treatment method aiming at the problem.

Disclosure of Invention

The invention provides a precise sewage classification treatment method, which aims to (1) realize the precise classification treatment of substances in different pollutants by utilizing sewage treatment agents, and avoid secondary pollution caused by environmental change of sewage treatment; and (2) realizing the classification decision of the sewage classification treatment method.

The invention provides a precise sewage classification treatment method, which comprises the following steps:

S1: constructing a sewage internal substance set, and constructing a chemical reaction relation rule set among internal substances according to the sewage internal substance set, wherein each rule in the chemical reaction relation rule set indicates that the two types of sewage cannot be mixed, or else, chemical reaction can occur to cause secondary pollution;

s2: constructing a classification decision tree according to the chemical reaction relation rule set, wherein the construction principle of the classification decision tree is that substances with more occurrence times in the chemical reaction relation rule set are closer to the root node of the classification decision tree;

s3: classifying the collected sewage according to the classification decision tree, and mixing the sewage on the same leaf node;

S4: constructing a sewage treatment decision optimization data set, constructing a sewage decision treatment model by using deep reinforcement learning, and training the model by using a Q-learning algorithm to obtain an optimal sewage decision treatment model;

S5: inputting the sewage after mixed treatment into an optimal sewage decision treatment model, and outputting a mixed sewage treatment method by the sewage decision treatment model, wherein the mixed sewage treatment method comprises a mixed sewage treatment flow, and the ratio and the content of a mixed sewage treatment agent.

As a further improvement of the present invention:

and in the step S1, constructing a sewage internal substance set, which comprises the following steps:

the constructed sewage internal substance set is as follows:

{x₁,x₂,x₃,x₄,x₅,x₆,x₇}

Wherein:

x ₁ represents heavy metal Zn pollutants in sewage;

x ₂ represents heavy metal Cu pollutants in sewage;

x ₃ represents heavy metal Ag pollutant in sewage;

x ₄ represents toxic organic pollutants in sewage, such as polychlorinated benzofuran and polychlorinated biphenyl;

x ₅ represents polycyclic aromatic hydrocarbon PAHs in the sewage;

x ₆ represents the organic chlorine pesticide OCPs in the sewage;

x ₇ represents organic phosphorus in sewage;

In one embodiment of the invention, the toxic organic pollutants are from microparticles produced after incomplete incineration of the organic pollutants, the polycyclic aromatic hydrocarbon PAHs are from combustion pyrolysis, natural pyrolysis and oil pollution, the organochlorine pesticide OCPs are from pesticide and industrial drug pollution, the organic phosphorus is from human body excretion, washing wastewater and food residues, and the heavy metal pollutants are from heavy metal pollutants discharged from automobile exhaust and factories.

In the step S1, a chemical reaction relation rule set between internal substances is constructed according to the internal substance set of the sewage, and the method comprises the following steps:

the constructed rule set of chemical reaction relation is as follows:

{(x₁,x₅),(x₂,x₅),(x₃,x₅),(x₄,x₇),(x₆,x₇),(x₂,x₆)}

Wherein:

Any group (x _i,x_j) in the chemical reaction relation rule set represents a chemical reaction relation rule, and each rule represents that the two types of sewage cannot be mixed, or else chemical reaction can occur to cause secondary pollution; in one embodiment of the invention, the heavy metal and polycyclic aromatic hydrocarbon PAHs react chemically to form metal salt, secondary pollution is caused, organic phosphorus and chlorine react to generate toxic gas, and heavy metal Cu and organic chlorine react to form inorganic matters.

In the step S2, a classification decision tree is constructed according to a chemical reaction relation rule set, and the method comprises the following steps:

Constructing a classification decision tree according to the chemical reaction relation rule set, wherein the construction principle of the classification decision tree is that substances with more occurrence times in the chemical reaction relation rule set are closer to the root node of the classification decision tree, and the root node of the decision tree is set as

The first level of the decision tree is a root node x ₀, the left node of the second level is denoted as x ₅₊, and the right node is denoted as x _5-, wherein x ₅₊ indicates that a substance x ₅,x_5- exists in sewage and no substance x ₅ exists in the sewage;

the left and right nodes of the second layer node are { x ₆₊,x_6- }, wherein x ₆₊ represents that the substance x ₆,x_6- exists in the sewage and x ₆ does not exist in the sewage;

the left node and the right node of the third layer node are respectively { x ₇₊,x_7- }; the left node and the right node of the fourth layer node are { x ₂₊,x_2- }; left and right nodes of the fifth layer node are { x ₄₊,x_4- }; the left node and the right node of the sixth layer of nodes are { x ₁₊,x_1- }; the left node and the right node of the seventh layer node are { x ₃₊,x_3- };

Traversing to a root node by taking a leaf node of a decision tree as an initial point to obtain ancestors of all the leaf nodes, wherein the ancestors of the nodes represent all the nodes branched from the root node to the node, judging whether any two nodes in the ancestors of the leaf nodes meet any chemical reaction relation rule, if so, deleting the ancestors of the leaf nodes to obtain a leaf node ancestor set, and any ith group of leaf node ancestors in the leaf node ancestor set are Where n represents the number of nodes in the ancestor of the ith group of leaf nodes,/>Representing the internal substances of sewage, wherein any two nodes in the ancestor of the leaf node do not meet the rule of chemical reaction relation.

And in the step S3, the collected sewage is classified according to a classification decision tree, and the method comprises the following steps:

Classifying the collected sewage according to the classification decision tree, and mixing the sewage on the same leaf node, namely, the internal substances of the sewage are ancestor of the same leaf node And mixing the sewage containing the substances in the sewage.

And in the step S4, a sewage decision treatment model is constructed by utilizing deep reinforcement learning, and the method comprises the following steps:

Constructing a sewage treatment decision optimization data set data, wherein the format of the sewage treatment decision optimization data set is as follows:

Wherein:

Represents the content of the internal substances x _i in the sewage, i.e./> The content of the kth x _i in the sewage treatment decision optimization data set is represented, and K represents the type of the content of the x _i in the sewage;

Represents that the content of p x _i is/> The content of the sewage treatment agent for carrying out the internal substance x _i treatment on the sewage, in one specific embodiment of the invention, the content of the sewage treatment agent is determined for the sewage with different internal substance contents according to the expert experience;

initializing a mixed sewage state space S ₀＝(n₀,L₀ of a sewage decision treatment model, wherein n ₀ represents the types of internal substances in mixed sewage in an initial state, and L ₀ represents the content of corresponding internal substances in the mixed sewage in the initial state; initializing the action space of a sewage decision treatment model as

The constructed sewage decision treatment model is as follows:

Wherein:

Represents the content of internal substances x _j in the mixed sewage, and comes from L ₀ in a state space; /(I) Belongs to the sewage internal substance content in sewage treatment decision optimization data set,/>Belongs to the action space;

w _j represents the content of sewage treatment agent If the mixed sewage does not contain internal substances x _c, c=1, 2,3,4,5,6,7, w _c is set to 0;

The input of the sewage decision-making treatment model is a mixed sewage state space S ₀＝(n₀,L₀), the model respectively comprises the types and the corresponding contents of internal substances in mixed sewage, and the model output is the content and the proportion of sewage treatment agents corresponding to different internal substances of sewage.

In the step S4, training the model by using a Q-learning algorithm, including:

Training the constructed sewage decision treatment model by using a Q-learning algorithm to obtain the mixture ratio of sewage treatment agents of different sewage internal substances, wherein the Q-learning algorithm comprises the following steps:

Initializing a Q matrix of 1 row and 7 columns, wherein the initialized Q matrix is a null matrix; randomly generating a group of weight vectors (w _m1,w_m2,w_m3,w_m4,w_m5,w_m6,w_m7), wherein the initial value of m is 0, the update iteration times of the Q matrix are represented, and the weight vectors are filled into the Q matrix in sequence, so that matrix values from the 1 st column to the 7 th column in the Q matrix are [ w _m1,w_m2,w_m3,w_m4,w_m5,w_m6,w_m7 ];

Traversing a sewage treatment decision optimization data set data, and extracting sewage data U= { data ₁,data₂,…,data_u } which simultaneously contain all internal substances of sewage from the data, wherein U represents the number of the sewage data obtained by traversing, and data _u represents the content data of all the internal substances in the sewage;

constructing a weight vector value calculation function:

Wherein:

std (x _uj) represents the standard deviation of the sewage internal substance x _j in the sewage data U;

let m=m+1, update the value of Q matrix:

w_mj′＝w_mj+α[r(w_mj)+βmaxQ_m]

Wherein the method comprises the steps of

Alpha represents the update rate, which is set to 0.8;

Beta represents a decay coefficient, which is set to 0.6;

maxQ _m represents the maximum value in the Q matrix at the mth iteration, resulting in an updated Q matrix:

And taking a matrix value w _mj' in the Q matrix as an initial value of a weight vector w _j in the sewage decision treatment value to obtain an optimal sewage decision treatment model, setting w _c as 0 if no internal substances x _c, c=1, 2,3,4,5,6 and 7 exist in the mixed sewage, and carrying out normalization treatment on weights of other sewage internal substances.

In the step S5, the sewage after the mixed treatment is input into an optimal sewage decision treatment model, and the sewage decision treatment model outputs a mixed sewage treatment method, which comprises the following steps:

Inputting the mixed sewage into an optimal sewage decision treatment model, extracting a state space of the mixed sewage, namely the types and the corresponding contents of internal substances in the mixed sewage, taking the state space of the mixed sewage as the input of the sewage decision treatment model, outputting the model into the contents and the proportions of sewage treatment agents corresponding to different internal substances of the sewage, and adding the sewage treatment agents into the sewage according to the following sequence: a sewage treatment agent corresponding to heavy metal Ag, a sewage treatment agent corresponding to heavy metal Zn, a sewage treatment agent corresponding to toxic organic pollutants, a sewage treatment agent corresponding to heavy metal Cu, a sewage treatment agent corresponding to organic phosphorus, a sewage treatment agent corresponding to organochlorine pesticide and a sewage treatment agent corresponding to polycyclic aromatic hydrocarbon.

Compared with the prior art, the invention provides a precise sewage classification treatment method, which has the following advantages:

Firstly, the scheme provides a method for classifying and deciding substances in sewage, by classifying the substances in the sewage, the general types of pollutants in the sewage are toxic organic pollutants, polycyclic aromatic hydrocarbon, organic chlorine pesticides, organic phosphorus and heavy metals, a chemical reaction relation rule set among the substances in the sewage is constructed according to a sewage internal substance set, any group (x _i,x_j) of the chemical reaction relation rule set represents one chemical reaction relation rule, each rule represents that the two sewage cannot be mixed, or chemical reaction can occur, secondary pollution is caused, and the chemical reaction relation rule set is obtained. Constructing a classification decision tree according to the chemical reaction relation rule set, wherein the construction principle of the classification decision tree is that substances with more occurrence times in the chemical reaction relation rule set are closer to the root node of the classification decision tree, and the root node of the decision tree is set as The first level of the decision tree is a root node x ₀, the left node of the second level is denoted as x ₅₊, and the right node is denoted as x _5-, wherein x ₅₊ indicates that a substance x ₅,x_5- exists in sewage and no substance x ₅ exists in the sewage; the left and right nodes of the second layer node are { x ₆₊,x_6- }, wherein x ₆₊ represents that the substance x ₆,x_6- exists in the sewage and x ₆ does not exist in the sewage; the left node and the right node of the third layer node are respectively { x ₇₊,x_7- }; the left node and the right node of the fourth layer node are { x ₂₊,x_2- }; left and right nodes of the fifth layer node are { x ₄₊,x_4- }; the left node and the right node of the sixth layer of nodes are { x ₁₊,x_1- }; the left node and the right node of the seventh layer node are { x ₃₊,x_3- }; traversing to a root node by taking a leaf node of a decision tree as an initial point to obtain ancestors of all the leaf nodes, wherein the ancestors of the nodes represent all the nodes branched from the root node to the node, judging whether any two nodes in the ancestors of the leaf nodes meet any chemical reaction relation rule, if so, deleting the ancestors of the leaf nodes to obtain a leaf node ancestor set, and any ith group of leaf node ancestors in the leaf node ancestor set are/>Where n represents the number of nodes in the ancestor of the ith group of leaf nodes,/>Representing the internal substances of sewage, wherein any two nodes in the ancestor of the leaf node do not meet the rule of chemical reaction relation. Classifying the collected sewage according to the classification decision tree, and mixing the sewage on the same leaf node, namely that the internal substances of the sewage are ancestor of the same leaf node/>Mixing the sewage containing the substances in the sewage, and performing mixed treatment on the sewage according to a rule set of chemical reaction relations and ancestor/>, of the same leaf nodeThe internal substances of the sewage can not generate chemical reaction, so that secondary pollution is caused, and therefore, the part of sewage can be mixed, and pollutants in the mixed sewage can be treated by using the sewage treatment agent, so that more accurate sewage treatment is realized.

Meanwhile, the scheme proposes a sewage decision-making treatment model based on a reinforcement learning method, and a mixed sewage state space S ₀＝(n₀,L₀ of the sewage decision-making treatment model is initialized, wherein n ₀ represents the types of internal substances in mixed sewage in an initial state, and L ₀ represents the content of corresponding internal substances in the mixed sewage in the initial state; initializing the action space of a sewage decision treatment model asThe constructed sewage decision treatment model is as follows:

Wherein: Represents the content of internal substances x _j in the mixed sewage, and comes from L ₀ in a state space; /(I) Belongs to the sewage internal substance content in sewage treatment decision optimization data set,/>Belongs to the action space; w _j represents the content of sewage treatment agentIf the mixed sewage does not contain internal substances x _c, c=1, 2,3,4,5,6,7, w _c is set to 0; the input of the sewage decision-making treatment model is a mixed sewage state space S ₀＝(n₀,L₀), the model respectively comprises the types and the corresponding contents of internal substances in mixed sewage, and the model output is the content and the proportion of sewage treatment agents corresponding to different internal substances of sewage. Training the constructed sewage decision treatment model by utilizing a Q-learning algorithm to obtain the proportions of sewage treatment agents of different sewage internal substances, inputting the sewage after mixed treatment into the trained sewage decision treatment model, extracting the state space of the mixed sewage by the sewage decision treatment model, namely the types and the corresponding contents of the internal substances in the mixed sewage, taking the state space of the mixed sewage as the input of the sewage decision treatment model, and outputting the model as the contents and the proportions of the sewage treatment agents corresponding to the different sewage internal substances, thereby realizing the calculation of the contents and the proportions of the sewage treatment agents in the sewage treatment process, and adding the sewage treatment agents into the sewage according to the following sequence: a sewage treatment agent corresponding to heavy metal Ag, a sewage treatment agent corresponding to heavy metal Zn, a sewage treatment agent corresponding to toxic organic pollutants, a sewage treatment agent corresponding to heavy metal Cu, a sewage treatment agent corresponding to organic phosphorus, a sewage treatment agent corresponding to organochlorine pesticide and a sewage treatment agent corresponding to polycyclic aromatic hydrocarbon.

Drawings

FIG. 1 is a schematic flow chart of a method for classifying and treating sewage accurately according to an embodiment of the present invention;

the achievement of the objects, functional features and advantages of the present invention will be further described with reference to the accompanying drawings, in conjunction with the embodiments.

Detailed Description

It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

S1: and constructing a sewage internal substance set, and constructing a chemical reaction relation rule set among internal substances according to the sewage internal substance set.

And in the step S1, constructing a sewage internal substance set, which comprises the following steps:

the constructed sewage internal substance set is as follows:

{x₁,x₂,x₃,x₄,x₅,x₆,x₇}

Wherein:

x ₁ represents heavy metal Zn pollutants in sewage;

x ₂ represents heavy metal Cu pollutants in sewage;

x ₃ represents heavy metal Ag pollutant in sewage;

x ₄ represents toxic organic pollutants in sewage, such as polychlorinated benzofuran and polychlorinated biphenyl;

x ₅ represents polycyclic aromatic hydrocarbon PAHs in the sewage;

x ₆ represents the organic chlorine pesticide OCPs in the sewage;

x ₇ represents organic phosphorus in sewage;

In one embodiment of the invention, the toxic organic pollutants are from microparticles produced after incomplete incineration of the organic pollutants, the polycyclic aromatic hydrocarbon PAHs are from combustion pyrolysis, natural pyrolysis and oil pollution, the organochlorine pesticide OCPs are from pesticide and industrial drug pollution, the organic phosphorus is from human body excretion, washing wastewater and food residues, and the heavy metal pollutants are from heavy metal pollutants discharged from automobile exhaust and factories.

In the step S1, a chemical reaction relation rule set between internal substances is constructed according to the internal substance set of the sewage, and the method comprises the following steps:

the constructed rule set of chemical reaction relation is as follows:

{(x₁,x₅),(x₂,x₅),(x₃,x₅),(x₄,x₇),(x₆,x₇),(x₂,x₆)}

Wherein:

Any group (x _i,x_j) in the chemical reaction relation rule set represents a chemical reaction relation rule, and each rule represents that the two types of sewage cannot be mixed, or else chemical reaction can occur to cause secondary pollution; in one embodiment of the invention, the heavy metal and polycyclic aromatic hydrocarbon PAHs react chemically to form metal salt, secondary pollution is caused, organic phosphorus and chlorine react to generate toxic gas, and heavy metal Cu and organic chlorine react to form inorganic matters.

S2: and constructing a classification decision tree according to the chemical reaction relation rule set.

In the step S2, a classification decision tree is constructed according to a chemical reaction relation rule set, and the method comprises the following steps:

Constructing a classification decision tree according to the chemical reaction relation rule set, wherein the construction principle of the classification decision tree is that substances with more occurrence times in the chemical reaction relation rule set are closer to the root node of the classification decision tree, and the root node of the decision tree is set as

The first level of the decision tree is a root node x ₀, the left node of the second level is denoted as x ₅₊, and the right node is denoted as x _5-, wherein x ₅₊ indicates that a substance x ₅,x_5- exists in sewage and no substance x ₅ exists in the sewage;

the left and right nodes of the second layer node are { x ₆₊,x_6- }, wherein x ₆₊ represents that the substance x ₆,x_6- exists in the sewage and x ₆ does not exist in the sewage;

the left node and the right node of the third layer node are respectively { x ₇₊,x_7- }; the left node and the right node of the fourth layer node are { x ₂₊,x_2- }; left and right nodes of the fifth layer node are { x ₄₊,x_4- }; the left node and the right node of the sixth layer of nodes are { x ₁₊,x_1- }; the left node and the right node of the seventh layer node are { x ₃₊,x_3- };

Traversing to a root node by taking a leaf node of a decision tree as an initial point to obtain ancestors of all the leaf nodes, wherein the ancestors of the nodes represent all the nodes branched from the root node to the node, judging whether any two nodes in the ancestors of the leaf nodes meet any chemical reaction relation rule, if so, deleting the ancestors of the leaf nodes to obtain a leaf node ancestor set, and any ith group of leaf node ancestors in the leaf node ancestor set are Where n represents the number of nodes in the ancestor of the ith group of leaf nodes,/>Representing the internal substances of sewage, wherein any two nodes in the ancestor of the leaf node do not meet the rule of chemical reaction relation.

S3: and classifying the collected sewage according to the classification decision tree, and carrying out mixed treatment on the sewage on the same leaf node.

And in the step S3, the collected sewage is classified according to a classification decision tree, and the method comprises the following steps:

Classifying the collected sewage according to the classification decision tree, and mixing the sewage on the same leaf node, namely, the internal substances of the sewage are ancestor of the same leaf node And mixing the sewage containing the substances in the sewage.

S4: constructing a sewage treatment decision optimization data set, constructing a sewage decision treatment model by using deep reinforcement learning, and training the model by using a Q-learning algorithm to obtain an optimal sewage decision treatment model.

And in the step S4, a sewage decision treatment model is constructed by utilizing deep reinforcement learning, and the method comprises the following steps:

Constructing a sewage treatment decision optimization data set data, wherein the format of the sewage treatment decision optimization data set is as follows:

Wherein:

Represents the content of the internal substances x _i in the sewage, i.e./> The content of the kth x _i in the sewage treatment decision optimization data set is represented, and K represents the type of the content of the x _i in the sewage;

Represents that the content of p x _i is/> The content of the sewage treatment agent for carrying out the internal substance x _i treatment on the sewage, in one specific embodiment of the invention, the content of the sewage treatment agent is determined for the sewage with different internal substance contents according to the expert experience;

initializing a mixed sewage state space S ₀＝(n₀,L₀ of a sewage decision treatment model, wherein n ₀ represents the types of internal substances in mixed sewage in an initial state, and L ₀ represents the content of corresponding internal substances in the mixed sewage in the initial state; initializing the action space of a sewage decision treatment model as

The constructed sewage decision treatment model is as follows:

Wherein:

Represents the content of internal substances x _j in the mixed sewage, and comes from L ₀ in a state space; /(I) Belongs to the sewage internal substance content in sewage treatment decision optimization data set,/>Belongs to the action space;

w _j represents the content of sewage treatment agent If the mixed sewage does not contain internal substances x _c, c=1, 2,3,4,5,6,7, w _c is set to 0;

The input of the sewage decision-making treatment model is a mixed sewage state space S ₀＝(n₀,L₀), the model respectively comprises the types and the corresponding contents of internal substances in mixed sewage, and the model output is the content and the proportion of sewage treatment agents corresponding to different internal substances of sewage.

In the step S4, training the model by using a Q-learning algorithm, including:

Training the constructed sewage decision treatment model by using a Q-learning algorithm to obtain the mixture ratio of sewage treatment agents of different sewage internal substances, wherein the Q-learning algorithm comprises the following steps:

Initializing a Q matrix of 1 row and 7 columns, wherein the initialized Q matrix is a null matrix; randomly generating a group of weight vectors (w _m1,w_m2,w_m3,w_m4,w_m5,w_m6,w_m7), wherein the initial value of m is 0, the update iteration times of the Q matrix are represented, and the weight vectors are filled into the Q matrix in sequence, so that matrix values from the 1 st column to the 7 th column in the Q matrix are [ w _m1,w_m2,w_m3,w_m4,w_m5,w_m6,w_m7 ];

Traversing a sewage treatment decision optimization data set data, and extracting sewage data U= { data ₁,data₂,…,data_u } which simultaneously contain all internal substances of sewage from the data, wherein U represents the number of the sewage data obtained by traversing, and data _u represents the content data of all the internal substances in the sewage;

constructing a weight vector value calculation function:

Wherein:

std (x _uj) represents the standard deviation of the sewage internal substance x _j in the sewage data U;

let m=m+1, update the value of Q matrix:

w_mj′＝w_mj+α[r(w_mj)+βmaxQ_m]

Wherein the method comprises the steps of

Alpha represents the update rate, which is set to 0.8;

Beta represents a decay coefficient, which is set to 0.6;

maxQ _m represents the maximum value in the Q matrix at the mth iteration, resulting in an updated Q matrix:

And taking a matrix value w _mj' in the Q matrix as an initial value of a weight vector w _j in the sewage decision treatment value to obtain an optimal sewage decision treatment model, setting w _c as 0 if no internal substances x _c, c=1, 2,3,4,5,6 and 7 exist in the mixed sewage, and carrying out normalization treatment on weights of other sewage internal substances.

S5: inputting the sewage after mixed treatment into an optimal sewage decision treatment model, and outputting a mixed sewage treatment method by the sewage decision treatment model, wherein the mixed sewage treatment method comprises a mixed sewage treatment flow, and the ratio and the content of a mixed sewage treatment agent.

In the step S5, the sewage after the mixed treatment is input into an optimal sewage decision treatment model, and the sewage decision treatment model outputs a mixed sewage treatment method, which comprises the following steps:

Inputting the mixed sewage into an optimal sewage decision treatment model, extracting a state space of the mixed sewage, namely the types and the corresponding contents of internal substances in the mixed sewage, taking the state space of the mixed sewage as the input of the sewage decision treatment model, outputting the model into the contents and the proportions of sewage treatment agents corresponding to different internal substances of the sewage, and adding the sewage treatment agents into the sewage according to the following sequence: a sewage treatment agent corresponding to heavy metal Ag, a sewage treatment agent corresponding to heavy metal Zn, a sewage treatment agent corresponding to toxic organic pollutants, a sewage treatment agent corresponding to heavy metal Cu, a sewage treatment agent corresponding to organic phosphorus, a sewage treatment agent corresponding to organochlorine pesticide and a sewage treatment agent corresponding to polycyclic aromatic hydrocarbon.

It should be noted that, the foregoing reference numerals of the embodiments of the present invention are merely for describing the embodiments, and do not represent the advantages and disadvantages of the embodiments. And the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, apparatus, article, or method. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, apparatus, article, or method that comprises the element.

From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) as described above, comprising instructions for causing a terminal device (which may be a mobile phone, a computer, a server, or a network device, etc.) to perform the method according to the embodiments of the present invention.

The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.

Claims (2)

1. The accurate sewage classification treatment method is characterized by comprising the following steps of:

S1: constructing a sewage internal substance set, and constructing a chemical reaction relation rule set among internal substances according to the sewage internal substance set, wherein each rule in the chemical reaction relation rule set indicates that the two types of sewage cannot be mixed, or else, chemical reaction can occur to cause secondary pollution;

constructing a collection of wastewater internals, comprising: the constructed sewage internal substance set is as follows:

{x₁,x₂,x₃,x₄,x₅,x₆,x₇}

Wherein:

x ₁ represents heavy metal Zn pollutants in sewage;

x ₂ represents heavy metal Cu pollutants in sewage;

x ₃ represents heavy metal Ag pollutant in sewage;

x ₄ represents toxic organic pollutants in sewage;

x ₅ represents polycyclic aromatic hydrocarbon PAHs in the sewage;

x ₆ represents the organic chlorine pesticide OCPs in the sewage;

x ₇ represents organic phosphorus in sewage;

constructing a rule set of chemical reaction relations among internal substances according to the sewage internal substance collection, wherein the rule set comprises the following steps:

the constructed rule set of chemical reaction relation is as follows:

{(x₁,x₅),(x₂,x₅),(x₃,x₅),(x₄,x₇),(x₆,x₇),(x₂,x₆)}

Wherein:

Any group (x _i,x_j) in the chemical reaction relation rule set represents a chemical reaction relation rule, and each rule represents that the two types of sewage cannot be mixed, or else chemical reaction can occur to cause secondary pollution;

s2: constructing a classification decision tree according to the chemical reaction relation rule set, wherein the construction principle of the classification decision tree is that substances with more occurrence times in the chemical reaction relation rule set are closer to the root node of the classification decision tree;

Constructing a classification decision tree according to the chemical reaction relation rule set, wherein the construction principle of the classification decision tree is that substances with more occurrence times in the chemical reaction relation rule set are closer to the root node of the classification decision tree, and the root node of the decision tree is set as

The first level of the decision tree is a root node x ₀, the left node of the second level is denoted as x ₅₊, and the right node is denoted as x _5-, wherein x ₅₊ indicates that a substance x ₅,x_5- exists in sewage and no substance x ₅ exists in the sewage;

the left and right nodes of the second layer node are { x ₆₊,x_6- }, wherein x ₆₊ represents that the substance x ₆,x_6- exists in the sewage and x ₆ does not exist in the sewage;

the left node and the right node of the third layer node are respectively { x ₇₊,x_7- }; the left node and the right node of the fourth layer node are { x ₂₊,x_2- }; left and right nodes of the fifth layer node are { x ₄₊,x_4- }; the left node and the right node of the sixth layer of nodes are { x ₁₊,x_1- }; the left node and the right node of the seventh layer node are { x ₃₊,x_3- };

Traversing to a root node by taking a leaf node of a decision tree as an initial point to obtain ancestors of all the leaf nodes, wherein the ancestors of the nodes represent all the nodes branched from the root node to the node, judging whether any two nodes in the ancestors of the leaf nodes meet any chemical reaction relation rule, if so, deleting the ancestors of the leaf nodes to obtain a leaf node ancestor set, and any ith group of leaf node ancestors in the leaf node ancestor set are Where n represents the number of nodes in the ancestor of the ith group of leaf nodes,/>Representing the internal substances of sewage, wherein any two nodes in leaf node ancestors do not meet the rule of chemical reaction relation;

s3: classifying the collected sewage according to the classification decision tree, and mixing the sewage on the same leaf node;

and in the step S3, the collected sewage is classified according to a classification decision tree, and the method comprises the following steps:

Classifying the collected sewage according to the classification decision tree, and mixing the sewage on the same leaf node, namely, the internal substances of the sewage are ancestor of the same leaf node Mixing the sewage containing the substances in the sewage;

S4: constructing a sewage treatment decision optimization data set, constructing a sewage decision treatment model by using deep reinforcement learning, and training the model by using a Q-learning algorithm to obtain an optimal sewage decision treatment model;

Constructing a sewage treatment decision optimization data set data, wherein the format of the sewage treatment decision optimization data set is as follows:

Wherein:

Represents the content of the internal substances x _i in the sewage, i.e./> The content of the kth x _i in the sewage treatment decision optimization data set is represented, and K represents the type of the content of the x _i in the sewage;

Represents that the content of p x _i is/> The content of sewage treatment agent for carrying out internal substance x _i treatment on sewage;

initializing a mixed sewage state space S ₀＝(n₀,L₀ of a sewage decision treatment model, wherein n ₀ represents the types of internal substances in mixed sewage in an initial state, and L ₀ represents the content of corresponding internal substances in the mixed sewage in the initial state; initializing the action space of a sewage decision treatment model as

The constructed sewage decision treatment model is as follows:

Wherein:

Represents the content of internal substances x _j in the mixed sewage, and comes from L ₀ in a state space; /(I) Belongs to the sewage internal substance content in sewage treatment decision optimization data set,/>Belongs to the action space;

w _j represents the content of sewage treatment agent If the mixed sewage does not contain internal substances x _c, c=1, 2,3,4,5,6,7, w _c is set to 0;

The input of the sewage decision-making treatment model is a mixed sewage state space S ₀＝(n₀,L₀), the model respectively comprises the types and the corresponding contents of internal substances in mixed sewage, and the model output is the content and the proportion of sewage treatment agents corresponding to different internal substances of sewage;

S5: inputting the sewage after mixed treatment into an optimal sewage decision treatment model, and outputting a mixed sewage treatment method by the sewage decision treatment model, wherein the mixed sewage treatment method comprises a mixed sewage treatment flow, a mixed sewage treatment agent ratio and content;

Training the constructed sewage decision treatment model by using a Q-learning algorithm to obtain the mixture ratio of sewage treatment agents of different sewage internal substances, wherein the Q-learning algorithm comprises the following steps:

Initializing a Q matrix of 1 row and 7 columns, wherein the initialized Q matrix is a null matrix; randomly generating a group of weight vectors (w _m1,w_m2,w_m3,w_m4,w_m5,w_m6,w_m7), wherein the initial value of m is 0, the update iteration times of the Q matrix are represented, and the weight vectors are filled into the Q matrix in sequence, so that matrix values from the 1 st column to the 7 th column in the Q matrix are [ w _m1,w_m2,w_m3,w_m4,w_m5,w_m6,w_m7 ];

Traversing a sewage treatment decision optimization data set data, and extracting sewage data U= { data ₁,data₂,…,data_u } which simultaneously contain all internal substances of sewage from the data, wherein U represents the number of the sewage data obtained by traversing, and data _u represents the content data of all the internal substances in the sewage;

constructing a weight vector value calculation function:

Wherein:

std (x _uj) represents the standard deviation of the sewage internal substance x _j in the sewage data U;

let m=m+1, update the value of Q matrix:

w_mj ^′＝w_mj+α[r(w_mj)+βmaxQ_m]

Wherein the method comprises the steps of

Alpha represents the update rate, which is set to 0.8;

Beta represents a decay coefficient, which is set to 0.6;

maxQ _m represents the maximum value in the Q matrix at the mth iteration, resulting in an updated Q matrix:

[w_m1 ^′,w_m2 ^′,w_m3 ^′,w_m4 ^′,w_m5 ^′,w_m6 ^′,w_m7 ^′]

And taking a matrix value w _mj ^′ in the Q matrix as an initial value of a weight vector w _j in the sewage decision treatment value to obtain an optimal sewage decision treatment model, setting w _c as 0 if no internal substances x _c, c=1, 2,3,4,5,6 and 7 exist in the mixed sewage, and carrying out normalization treatment on weights of other sewage internal substances.

2. The precise sewage classification processing method according to claim 1, wherein in the step S5, the sewage after the mixed processing is input into an optimal sewage decision processing model, and the sewage decision processing model outputs the mixed sewage processing method, comprising:

Inputting the mixed sewage into an optimal sewage decision treatment model, extracting a state space of the mixed sewage, namely the types and the corresponding contents of internal substances in the mixed sewage, taking the state space of the mixed sewage as the input of the sewage decision treatment model, outputting the model into the contents and the proportions of sewage treatment agents corresponding to different internal substances of the sewage, and adding the sewage treatment agents into the sewage according to the following sequence: a sewage treatment agent corresponding to heavy metal Ag, a sewage treatment agent corresponding to heavy metal Zn, a sewage treatment agent corresponding to toxic organic pollutants, a sewage treatment agent corresponding to heavy metal Cu, a sewage treatment agent corresponding to organic phosphorus, a sewage treatment agent corresponding to organochlorine pesticide and a sewage treatment agent corresponding to polycyclic aromatic hydrocarbon.