A water washing filtrate treatment supervision system and method based on artificial intelligence

Abstract

The present invention discloses a water washing filtrate treatment supervision system and method based on artificial intelligence, which relates to the field of artificial intelligence technology. The method comprises the following steps: step 1, dividing the water washing filtrate blocks according to spatial position, arranging a sensor group to collect pollutant concentration data and auxiliary parameters and performing preprocessing; step 2, for water-soluble pollutants, constructing a concentration distribution map and calculating the overall concentration; step 3, for water-insoluble pollutants, setting weights in combination with particle dynamics, and weighted calculating the average concentration; step 4, setting the water washing filtrate discharge standard, and judging whether the pollutant concentration meets the discharge standard based on the interval; step 5, based on the discharge standard judgment result, dynamically adjusting the water washing filtrate treatment process through an artificial intelligence algorithm. The present invention can effectively improve the situation in the prior art where the concentration of the water washing filtrate is unevenly distributed in space, resulting in a lack of representativeness of the overall concentration data.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an artificial intelligence-based water wash filtrate processing monitoring system and method. Background Art

[0002] The supervision of water wash filtrate treatment is a key link in ensuring the coordinated development of industrial production and environmental protection. During the industrial production process, the water wash step produces a large amount of filtrate containing various pollutants. These pollutants include water-soluble and water-insoluble substances. If they are directly discharged without effective treatment, they will cause serious pollution to the soil, water bodies and other ecological environments. Therefore, accurately monitoring the concentration of pollutants in the water wash filtrate and treating it according to standards are important measures to avoid environmental pollution and achieve rational resource utilization. With the development of environmental monitoring technology and artificial intelligence technology, how to apply advanced technologies to the supervision of water wash filtrate treatment to improve treatment efficiency and supervision accuracy has become a hot topic in industry research.

[0003] However, the existing water wash filtrate treatment and supervision technology has many shortcomings. Since the concentration of water wash filtrate is often unevenly distributed in space, traditional methods often use single-point or limited-point sampling and monitoring, which makes it difficult to fully reflect the overall concentration situation, resulting in a lack of representative data. When calculating the concentration of pollutants, the differences in the nature of the pollutants are not fully considered, and a unified treatment method is used for water-soluble and non-water-soluble pollutants, which makes it impossible to accurately assess the concentration of pollutants. In addition, the existing treatment process lacks a dynamic adjustment mechanism. When there is a deviation between the actual pollutant concentration and the emission standard, it is difficult to quickly and effectively optimize the treatment plan. These problems result in poor water wash filtrate treatment, increased treatment costs, and potential risks to the environment. Summary of the Invention

[0004] The purpose of the present invention is to provide a water wash filtrate treatment supervision system and method based on artificial intelligence to solve the problems raised in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a water washing filtrate treatment supervision method based on artificial intelligence, the method comprising the following steps:

[0006] Step 1: Divide the water washing filtrate into blocks according to spatial location, arrange the sensor group to collect pollutant concentration data and auxiliary parameters and perform preprocessing;

[0007] Step 2: For water-soluble pollutants, construct a concentration distribution map and calculate the overall concentration;

[0008] Step 3: For water-insoluble pollutants, weights are set based on particle dynamics and the weighted average concentration is calculated;

[0009] Step 4: Set the discharge standard for the water wash filtrate and determine whether the pollutant concentration meets the discharge standard based on the interval;

[0010] Step 5: Based on the emission standard determination results, dynamically adjust the water wash filtrate treatment process through artificial intelligence algorithms.

[0011] In step 1, the water washing filtrate is divided into 1 to N blocks according to the spatial position; the block index is represented by n, n∈{1,2,…,N}; the block volume is represented by V1~V N ;

[0012] Deploy sensor groups in each block to directly or indirectly collect pollutant concentration data and auxiliary parameters from different parts; record the locations of corresponding sensor deployment points;

[0013] Pollutants include water-soluble pollutants and water-insoluble pollutants;

[0014] The concentration data of a water-soluble pollutant in different blocks is expressed as: C i,1 ~C i,N ; Where i is a positive integer, representing the sequence of water-soluble pollutant categories, i∈{1,2,…,I}, I is a positive integer, representing the number of water-soluble pollutant categories;

[0015] The concentration data of a certain non-water-soluble pollutant in different blocks is expressed as: c j,1 ~c j,N ; Wherein, j is a positive integer, representing the sequence of non-water-soluble pollutant categories, j∈{1,2,…,J}, J is a positive integer, representing the number of non-water-soluble pollutant categories;

[0016] Preprocessing of pollutant concentration data;

[0017] Pollutant data includes water-soluble pollutant concentration data and water-insoluble pollutant concentration data. For a certain pollutant data, if the pollutant data exceeds the preset range due to sensors or transmission, the average value of the pollutant concentration data of the remaining blocks is used to replace it.

[0018] For the replaced pollutant data, the mean and standard deviation are calculated to further standardize the pollutant data.

[0019] In step 2, a concentration distribution map of water-soluble pollutants is constructed and the overall concentration is calculated;

[0020] For water-soluble pollutants C i , use Kriging interpolation method to perform spatial interpolation of concentration: C i (x,y,z)=μ i +Σ n=1 N λ n [Ci,n -μ i ];

[0021] Among them, λ n represents the weight calculated by the Kriging algorithm, Σ n=1 N λ n =1;C i (x,y,z) represents the water-soluble pollutant C at the point (x,y,z) to be estimated i Concentration data; C i,n Represents the water-soluble pollutant C at the sensor deployment point corresponding to the nth block i concentration data;

[0022] The three-dimensional space where the water washing filtrate is located is divided into volumes ΔV, and each volume ΔV is obtained by Kriging interpolation method. m The concentration C i,m (x,y,z); calculate the total concentration by integration; when ΔV m Approaching 0: ;

[0023] Where M is a positive integer, indicating the number of partitioned volumes; m∈{1,2,…,M}, indicating the sequence of partitioned volumes, ΔV m Indicates the size of the mth volume, C i,m (x,y,z) represents the water-soluble pollutant C at the mth volume corresponding to the estimated point (x,y,z) i Concentration data.

[0024] In step 3, for the water-insoluble pollutant c j : The concentration of the nth block is represented by c j,n ;

[0025] Obtaining water-insoluble pollutants c from auxiliary parameters j The average particle radius r j 、Insoluble pollutants c j Particle density ρ j , water density ρ w , water viscosity μ and gravitational acceleration g; calculate the sedimentation rate of the block: v j,n =2r j 2 (ρ j -ρ w )g / 9μ;

[0026] The sedimentation rate and the block volume V n Multiply them together to get the weight of the contribution to the overall concentration: w j,n =v j,n ×V n ;

[0027] w j,n As weight, the weighted average of all block concentrations is taken to get the overall concentration: c j,total =∑ n=1 N w j,n c j,n / ∑ n=1 N w j,n .

[0028] In step 4, based on the calculated concentrations of water-soluble and water-insoluble pollutants, it is determined whether the discharge standards are met;

[0029] Set emission standard ranges for water-soluble pollutant concentrations and non-water-soluble pollutant concentrations [C i,l1 ,C i,l2 ] and [c j,l1 ,c j,l2 ];

[0030] Comparison of water-soluble pollutant concentration C i,total and the concentration of water-insoluble pollutants c j,total and the corresponding emission standard range;

[0031] When the concentrations of all types of water-soluble and non-water-soluble pollutants are within the emission standard range, the wash filtrate is considered to meet the emission standards;

[0032] When there are water-soluble pollutants or insoluble pollutants whose concentrations are not within the emission standard range, the water-wash filtrate is considered to be not in compliance with the emission standard. A corresponding treatment plan is generated based on the difference between the concentrations of water-soluble pollutants or insoluble pollutants that are not within the emission standard range and the emission standard range, and feedback is provided.

[0033] In step 5, based on the emission standard determination results, the water wash filtrate treatment process is dynamically adjusted through artificial intelligence algorithms to achieve adaptive optimization of the treatment effect;

[0034] When the water wash filtrate does not meet the emission standards, identify the specific items that exceed the standards and the degree of excess (calculate the difference between the actual concentration and the standard range).

[0035] Establish a solution library containing multiple treatment strategies, each corresponding to a specific type of exceeded item, degree of exceeded item, and treatment equipment parameter combination; match similar cases from the solution library based on exceeded information and select candidate treatment solutions with higher priority;

[0036] A Deep Q-Network (DQN) was introduced to evaluate and optimize candidate solutions. The system's state was defined as multi-dimensional information such as the current pollutant concentration, equipment operating parameters, and treatment time. Actions were defined as different treatment solutions, and the reward function was set based on the degree to which the post-treatment pollutant concentration approaches the standard range (positive rewards were given if the standard was met, and negative rewards were given based on the degree of excess). Through continuous iterative training, the algorithm learned the optimal treatment strategy and generated an optimized treatment solution for the current excess.

[0037] The generated treatment plan is converted into specific equipment operating parameters, such as adjusting the water pump flow to control the water washing time, automatically calibrating the dosage of the dosing device, and adjusting the operating frequency of the filtration equipment. Through the Internet of Things technology, it connects with the water treatment equipment, transmits control instructions in real time, and realizes the automated configuration of treatment process parameters.

[0038] After the treatment process is adjusted, the sensor group continuously collects pollutant concentration and auxiliary parameter data, and transmits the data to the system at fixed intervals (e.g., 10 minutes). The system recalculates pollutant concentrations based on the new data and compares them with emission standards to determine whether the treatment effect meets expectations. If the expected effect is not achieved (e.g., pollutant concentrations decrease slowly or still exceed standards), the system returns to step 5 to regenerate the optimization plan.

[0039] After each treatment is completed, the pollutant concentration data before and after the treatment, the treatment plan adopted, the equipment operating parameters and the final treatment results and other information will be stored in the historical database as a reference experience for subsequent treatment.

[0040] Kriging interpolation models and reinforcement learning models are retrained regularly (e.g., weekly or monthly) using accumulated historical data. For example, new data can be used to optimize the weight calculation parameters of kriging interpolation to improve concentration estimation accuracy; neural network parameters of reinforcement learning models can be updated to better adapt to treatment requirements under different operating conditions. Simultaneously, the solution library is optimized based on actual treatment results, inefficient solutions are removed, and new effective treatment strategies are added to continuously improve treatment capacity and regulatory efficiency.

[0041] An artificial intelligence-based water washing filtrate processing supervision system, the system includes a data acquisition module, a concentration calculation module, a standard judgment module and a processing optimization module;

[0042] The data acquisition module is used to divide the water washing filtrate blocks according to spatial positions, arrange sensor groups to collect pollutant concentration data and auxiliary parameters and perform preprocessing;

[0043] The concentration calculation module is used to construct a concentration distribution map and calculate the overall concentration for water-soluble pollutants; for insoluble pollutants, the weight is set in combination with particle dynamics to calculate the weighted average concentration.

[0044] The standard judgment module is used to set the discharge standard of the water washing filtrate and judge whether the pollutant concentration meets the discharge standard according to the interval;

[0045] The processing optimization module is used to dynamically adjust the water washing filtrate processing process through an artificial intelligence algorithm based on the emission standard determination results.

[0046] The data acquisition module includes a block division unit, a data acquisition unit and a data preprocessing unit;

[0047] The block division unit is used to divide the water washing filtrate blocks according to spatial positions;

[0048] The data acquisition unit is used to arrange the sensor group to collect pollutant concentration data and auxiliary parameters;

[0049] The data preprocessing unit is used to preprocess the collected pollutant concentration data.

[0050] The concentration calculation module includes an interpolation unit, a concentration calculation 1 unit, a weight calculation unit and a concentration calculation 2 unit;

[0051] The interpolation unit is used to perform spatial interpolation of concentration of water-soluble pollutants using the Kriging interpolation method;

[0052] The concentration calculation unit 1 is used to calculate the overall concentration of water-soluble pollutants by integration;

[0053] The weight calculation unit is used to set weights for water-insoluble pollutants in combination with particle dynamics;

[0054] The concentration calculation unit 2 is used for weighted calculation of the average concentration of water-insoluble pollutants.

[0055] The standard judgment module includes a standard setting unit and a matching comparison unit;

[0056] The standard setting unit is used to set the discharge standard of the water washing filtrate;

[0057] The matching and comparison unit is used to determine whether the pollutant concentration meets the emission standards according to the interval;

[0058] The processing optimization module includes an identification unit, a solution generation unit, an execution unit and an optimization unit;

[0059] The identification unit is used to identify the specific pollutant type and the degree of exceeding the standard when the water washing filtrate does not meet the emission standard;

[0060] The solution generation unit is used to match similar cases from the solution library according to the specific pollutant type and the degree of exceeding the standard, screen candidate solutions and evaluate and optimize the solutions;

[0061] The execution unit is used to convert the optimal solution into equipment control instructions to achieve automatic adjustment;

[0062] The optimization unit is used to optimize the model and solution library according to the actual processing effect.

[0063] Compared with the prior art, the beneficial effects of the present invention are: the partitioned collection and pretreatment method of the present invention can more accurately obtain the distribution information of pollutants in the water washing filtrate, provide a more accurate data basis for subsequent concentration calculation and processing, reduce processing errors caused by data errors, and improve the pertinence and effectiveness of treatment; the present invention adopts different concentration calculation methods for pollutants of different properties, fully considers the characteristics of the pollutants, and can more accurately reflect the actual situation of pollutants in the water washing filtrate, providing a more accurate basis for formulating reasonable treatment plans, and improving the treatment effects of different types of pollutants. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 This is a schematic diagram of the steps of an artificial intelligence-based water wash filtrate treatment supervision method of the present invention;

[0065] Figure 2 This is a flow chart of an artificial intelligence-based water wash filtrate treatment monitoring system of the present invention. DETAILED DESCRIPTION

[0066] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0067] Example: Figure 1-Figure 2 As shown, the present invention provides a technical solution, a water washing filtrate treatment supervision method based on artificial intelligence, which includes the following steps:

[0068] Step 1: Divide the water washing filtrate into blocks according to spatial location, arrange the sensor group to collect pollutant concentration data and auxiliary parameters and perform preprocessing;

[0069] Step 2: For water-soluble pollutants, construct a concentration distribution map and calculate the overall concentration;

[0070] Step 3: For water-insoluble pollutants, weights are set based on particle dynamics and the weighted average concentration is calculated;

[0071] Step 4: Set the discharge standard for the water wash filtrate and determine whether the pollutant concentration meets the discharge standard based on the interval;

[0072] Step 5: Based on the emission standard determination results, dynamically adjust the water wash filtrate treatment process through artificial intelligence algorithms.

[0073] In step 1, the water washing filtrate is divided into 1 to N blocks according to the spatial position; the block index is represented by n, n∈{1,2,…,N}; the block volume is represented by V1~V N ;

[0074] Deploy sensor groups in each block to directly or indirectly collect pollutant concentration data and auxiliary parameters from different parts; record the locations of corresponding sensor deployment points;

[0075] Pollutants include water-soluble pollutants and water-insoluble pollutants;

[0076] The concentration data of a water-soluble pollutant in different blocks is expressed as: C i,1 ~C i,N ; Where i is a positive integer, representing the sequence of water-soluble pollutant categories, i∈{1,2,…,I}, I is a positive integer, representing the number of water-soluble pollutant categories;

[0077] The concentration data of a certain non-water-soluble pollutant in different blocks is expressed as: c j,1 ~c j,N ; Wherein, j is a positive integer, representing the sequence of non-water-soluble pollutant categories, j∈{1,2,…,J}, J is a positive integer, representing the number of non-water-soluble pollutant categories;

[0078] Preprocessing of pollutant concentration data;

[0079] Pollutant data includes water-soluble pollutant concentration data and water-insoluble pollutant concentration data. For a certain pollutant data, if the pollutant data exceeds the preset range due to sensors or transmission, the average value of the pollutant concentration data of the remaining blocks is used to replace it.

[0080] For the replaced pollutant data, the mean and standard deviation are calculated to further standardize the pollutant data.

[0081] In step 2, a concentration distribution map of water-soluble pollutants is constructed and the overall concentration is calculated;

[0082] For water-soluble pollutants C i , use Kriging interpolation method to perform spatial interpolation of concentration: C i (x,y,z)=μ i +Σ n=1 N λ n [Ci,n -μ i ];

[0083] Among them, λ n represents the weight calculated by the Kriging algorithm, Σ n=1 N λ n =1;C i (x,y,z) represents the water-soluble pollutant C at the point (x,y,z) to be estimated i Concentration data; C i,n Represents the water-soluble pollutant C at the sensor deployment point corresponding to the nth block i concentration data;

[0084] The three-dimensional space where the water washing filtrate is located is divided into volumes ΔV, and each volume ΔV is obtained by Kriging interpolation method. m The concentration C i,m (x,y,z); calculate the total concentration by integration; when ΔV m Approaching 0: ;

[0085] Where M is a positive integer, indicating the number of partitioned volumes; m∈{1,2,…,M}, indicating the sequence of partitioned volumes, ΔV m Indicates the size of the mth volume, C i,m (x,y,z) represents the water-soluble pollutant C at the mth volume corresponding to the estimated point (x,y,z) i Concentration data.

[0086] In step 3, for the water-insoluble pollutant c j : The concentration of the nth block is represented by c j,n ;

[0087] Obtaining water-insoluble pollutants c from auxiliary parameters j The average particle radius r j 、Insoluble pollutants c j Particle density ρ j , water density ρ w , water viscosity μ and gravitational acceleration g; calculate the sedimentation rate of the block: v j,n =2r j 2 (ρ j -ρ w )g / 9μ;

[0088] The sedimentation rate and the block volume V n Multiply them together to get the weight of the contribution to the overall concentration: w j,n =v j,n ×V n ;

[0089] w j,n As weight, the weighted average of all block concentrations is taken to get the overall concentration: c j,total =∑ n=1 N w j,n c j,n / ∑ n=1 N w j,n .

[0090] In step 4, based on the calculated concentrations of water-soluble and water-insoluble pollutants, it is determined whether the discharge standards are met;

[0091] Set emission standard ranges for water-soluble pollutant concentrations and non-water-soluble pollutant concentrations [C i,l1 ,C i,l2 ] and [c j,l1 ,c j,l2 ];

[0092] Comparison of water-soluble pollutant concentration C i,total and the concentration of water-insoluble pollutants c j,total and the corresponding emission standard range;

[0093] When the concentrations of all types of water-soluble and non-water-soluble pollutants are within the emission standard range, the wash filtrate is considered to meet the emission standards;

[0094] When there are water-soluble pollutants or insoluble pollutants whose concentrations are not within the emission standard range, the water-wash filtrate is considered to be not in compliance with the emission standard. A corresponding treatment plan is generated based on the difference between the concentrations of water-soluble pollutants or insoluble pollutants that are not within the emission standard range and the emission standard range for feedback;

[0095] In step 5, based on the emission standard determination results, the water wash filtrate treatment process is dynamically adjusted through artificial intelligence algorithms to achieve adaptive optimization of the treatment effect;

[0096] When the water wash filtrate does not meet the emission standards, identify the specific items that exceed the standards and the degree of excess (calculate the difference between the actual concentration and the standard range).

[0097] Establish a solution library containing multiple treatment strategies, each corresponding to a specific type of exceeded item, degree of exceeded item, and treatment equipment parameter combination; match similar cases from the solution library based on exceeded information and select candidate treatment solutions with higher priority;

[0098] A Deep Q-Network (DQN) was introduced to evaluate and optimize candidate solutions. The system's state was defined as multi-dimensional information such as the current pollutant concentration, equipment operating parameters, and treatment time. Actions were defined as different treatment solutions, and the reward function was set based on the degree to which the post-treatment pollutant concentration approaches the standard range (positive rewards were given if the standard was met, and negative rewards were given based on the degree of excess). Through continuous iterative training, the algorithm learned the optimal treatment strategy and generated an optimized treatment solution for the current excess.

[0099] The generated treatment plan is converted into specific equipment operating parameters, such as adjusting the water pump flow to control the water washing time, automatically calibrating the dosage of the dosing device, and adjusting the operating frequency of the filtration equipment. Through the Internet of Things technology, it connects with the water treatment equipment, transmits control instructions in real time, and realizes the automated configuration of treatment process parameters.

[0100] After the treatment process is adjusted, the sensor group continuously collects pollutant concentration and auxiliary parameter data, and transmits the data to the system at fixed intervals (e.g., 10 minutes). The system recalculates pollutant concentrations based on the new data and compares them with emission standards to determine whether the treatment effect meets expectations. If the expected effect is not achieved (e.g., pollutant concentrations decrease slowly or still exceed standards), the system returns to step 5 to regenerate the optimization plan.

[0101] After each treatment is completed, the pollutant concentration data before and after the treatment, the treatment plan adopted, the equipment operating parameters and the final treatment results and other information will be stored in the historical database as a reference experience for subsequent treatment.

[0102] Kriging interpolation models and reinforcement learning models are retrained regularly (e.g., weekly or monthly) using accumulated historical data. For example, new data can be used to optimize the weight calculation parameters of kriging interpolation to improve concentration estimation accuracy; neural network parameters of reinforcement learning models can be updated to better adapt to treatment requirements under different operating conditions. Simultaneously, the solution library is optimized based on actual treatment results, inefficient solutions are removed, and new effective treatment strategies are added to continuously improve treatment capacity and regulatory efficiency.

[0103] An artificial intelligence-based water washing filtrate processing supervision system, the system includes a data acquisition module, a concentration calculation module, a standard judgment module and a processing optimization module;

[0104] The data acquisition module is used to divide the water washing filtrate blocks according to spatial positions, arrange sensor groups to collect pollutant concentration data and auxiliary parameters and perform preprocessing;

[0105] The concentration calculation module is used to construct a concentration distribution map and calculate the overall concentration for water-soluble pollutants; for insoluble pollutants, the weight is set in combination with particle dynamics to calculate the weighted average concentration.

[0106] The standard judgment module is used to set the discharge standard of the water washing filtrate and judge whether the pollutant concentration meets the discharge standard according to the interval;

[0107] The processing optimization module is used to dynamically adjust the water washing filtrate processing process through an artificial intelligence algorithm based on the emission standard determination results.

[0108] The data acquisition module includes a block division unit, a data acquisition unit and a data preprocessing unit;

[0109] The block division unit is used to divide the water washing filtrate blocks according to spatial positions;

[0110] The data acquisition unit is used to arrange the sensor group to collect pollutant concentration data and auxiliary parameters;

[0111] The data preprocessing unit is used to preprocess the collected pollutant concentration data.

[0112] The concentration calculation module includes an interpolation unit, a concentration calculation 1 unit, a weight calculation unit and a concentration calculation 2 unit;

[0113] The interpolation unit is used to perform spatial interpolation of concentration of water-soluble pollutants using the Kriging interpolation method;

[0114] The concentration calculation unit 1 is used to calculate the overall concentration of water-soluble pollutants by integration;

[0115] The weight calculation unit is used to set weights for water-insoluble pollutants in combination with particle dynamics;

[0116] The concentration calculation unit 2 is used for weighted calculation of the average concentration of water-insoluble pollutants.

[0117] The standard judgment module includes a standard setting unit and a matching comparison unit;

[0118] The standard setting unit is used to set the discharge standard of the water washing filtrate;

[0119] The matching and comparison unit is used to determine whether the pollutant concentration meets the emission standards according to the interval;

[0120] The processing optimization module includes an identification unit, a solution generation unit, an execution unit and an optimization unit;

[0121] The identification unit is used to identify the specific pollutant type and the degree of exceeding the standard when the water washing filtrate does not meet the emission standard;

[0122] The solution generation unit is used to match similar cases from the solution library according to the specific pollutant type and the degree of exceeding the standard, screen candidate solutions and evaluate and optimize the solutions;

[0123] The execution unit is used to convert the optimal solution into equipment control instructions to achieve automatic adjustment;

[0124] The optimization unit is used to optimize the model and solution library according to the actual processing effect.

[0125] In this example, a printing and dyeing company has multiple washing workshops, generating large amounts of wash filtrate daily. This wash filtrate contains water-soluble dyes (such as reactive dyes and direct dyes) and insoluble pollutants (such as suspended fiber particles and grease). To ensure that the wash filtrate meets discharge standards and avoids environmental pollution, the company implements an AI-based filtrate treatment and monitoring method.

[0126] Step 1: water washing filtrate block division and data acquisition preprocessing;

[0127] The company divided the wash filtrate treatment pool into eight zones (N=8) based on spatial location. The zones have volumes of V1 = 10m³, V2 = 12m³, V3 = 8m³, V4 = 15m³, V5 = 10m³, V6 = 13m³, V7 = 9m³, and V8 = 11m³. Sensors were deployed in each zone to collect pollutant concentration data and auxiliary parameters (water temperature, pH value, etc.).

[0128] It is known that water-soluble pollutants include reactive dyes (i=1) and direct dyes (i=2), and non-water-soluble pollutants include fiber particles (j=1) and grease (j=2). The reactive dye concentration data (mg / L) collected in each block is C 1,1 =50, C 1,2 =48, C 1,3 =200, C 1,4 =60, C 1,5 =45, C 1,6 =52, C 1,7 =58, C 1,8 =47; direct dye concentration data (mg / L) is C 2,1 =35, C 2,2 =32, C 2,3 =38, C 2,4 =40, C 2,5 =30, C 2,6 =36, C 2,7 =39, C 2,8 =33.

[0129] Fiber particle concentration data (mg / L) is c 1,1 =20, c 1,2 =22, c 1,3 =18, c 1,4 =25, c 1,5=19, c 1,6 =23, c 1,7 =21, c 1,8 =24; oil concentration data (mg / L) is c 2,1 =15, c 2,2 =13, c 2,3 =17, c 2,4 =19, c 2,5 =12, c 2,6 =16, c 2,7 =18, c 2,8 =14.

[0130] During the data collection process, it was found that C 1,3 The concentration data for the 7 blocks was 200 (clearly exceeding the preset range [0,100]), so we replaced it with the average value of the reactive dye concentration data from the remaining 7 blocks. Similarly, after processing the other abnormal data, we calculated the mean and standard deviation to standardize the pollutant data.

[0131] Step 2: Water-soluble pollutant concentration treatment

[0132] For the active dye (i=1), the concentration is spatially interpolated using the Kriging interpolation method. The weights λ1-λ8 are calculated by the Kriging algorithm to satisfy Σ n=1 8 λ n =1. Taking a point to be estimated (x, y, z) as an example, according to formula C i (x,y,z)=μ i +Σ n=1 N λ n [C i,n -μ i ], combined with the reactive dye concentration data of each block after treatment, the reactive dye concentration at that point was calculated.

[0133] The three-dimensional space where the water filtrate is located is divided into 100 small volumes ΔV (M=100), and the Kriging interpolation method is used to obtain the value of each volume ΔV m The concentration C 1,m (x,y,z). When ΔV m When it approaches 0, the total concentration C of the reactive dye is calculated by integration. 1,total Similarly, calculate the total concentration C of direct dyes 2,total .

[0134] Step 3: Treatment of non-water-soluble pollutant concentration

[0135] Taking fiber particles (j=1) as an example, the average particle radius r1=0.05mm, particle density ρ1=1.2g / cm³, and water density ρw =1g / cm³, water viscosity μ=0.001Pa・s and acceleration due to gravity g=9.8m / s².

[0136] Calculate the sedimentation rate v of the first block 1,1 ; Multiply the sedimentation rate by the block volume V1 to obtain the weight of the contribution to the overall concentration: w 1,1 .

[0137] Similarly, calculate the weights of other blocks, using w j,n As the weight, the weighted average of all block concentrations is taken to obtain the overall concentration of fiber particles: c 1,total =∑ n=1 8 w 1,n c 1,n / ∑ n=1 8 w 1,n Similarly, calculate the overall concentration of oil c 2,total .

[0138] Step 4: Emission standard determination;

[0139] The emission standards for reactive dyes stipulated by the local environmental protection department are in the range of [0,50] mg / L, direct dyes in the range of [0,35] mg / L, fiber particles in the range of [0,20] mg / L, and grease in the range of [0,15] mg / L.

[0140] Compare the calculated total concentration of reactive dyes C 1,total =53mg / L, total concentration of direct dye C 2,total =36mg / L, fiber particle overall concentration c 1,total =22mg / L, total oil concentration c 2,total =16mg / L and the corresponding emission standard range, it was found that the concentrations of various pollutants exceeded the standards, and it was considered that the water washing filtrate did not meet the emission standards.

[0141] Based on the difference between the concentration of each pollutant and the emission standard range, a corresponding treatment plan is generated for feedback: reactive dyes exceed the standard by 3 mg / L, direct dyes exceed the standard by 1 mg / L, fiber particles exceed the standard by 2 mg / L, and oils and fats exceed the standard by 1 mg / L.

[0142] Step 5: Dynamic adjustment of processing flow;

[0143] The company's established solution library contains multiple treatment strategies. Based on the exceedance information, similar cases are matched from the solution library to select high-priority candidate solutions. A Deep Q-Network (DQN) is introduced to evaluate and optimize the candidate solutions. The state of the water wash filtrate treatment system is defined as a multi-dimensional information set including the current pollutant concentration, equipment operating parameters (pump flow rate, dosing device dosage, filtration equipment operating frequency), and treatment time. Actions are defined as different treatment solutions, and the reward function is set as the degree to which the post-treatment pollutant concentration approaches the standard range.

[0144] Through continuous iterative training, the algorithm generates an optimized treatment plan for the current exceeding standard situation: increasing the water pump flow from the original 50m³ / h to 60m³ / h, and extending the water washing time; increasing the dosage of active dye treatment agent in the dosing device from the original 10kg / h to 12kg / h, and the dosage of direct dye treatment agent from 8kg / h to 9kg / h; and increasing the operating frequency of the filtration equipment from the original 20 times / hour to 25 times / hour.

[0145] By connecting to water treatment equipment through IoT technology, control commands are transmitted in real time, enabling automated configuration of treatment process parameters. After adjusting the treatment process, the sensor group collects pollutant concentration and auxiliary parameter data every 10 minutes and transmits it to the system. The system recalculates pollutant concentrations based on the new data and determines whether the treatment effect is meeting expectations. If not, the system returns to step 5 to regenerate an optimization plan.

[0146] After each treatment is completed, information such as pollutant concentration data before and after treatment, the treatment plan used, equipment operating parameters, and final treatment results is stored in a historical database. Kriging interpolation models and reinforcement learning models are retrained weekly using accumulated historical data to optimize the weight calculation parameters of Kriging interpolation and update the neural network parameters of the reinforcement learning model. The solution library is also optimized, inefficient solutions are removed, and new effective treatment strategies are added.

[0147] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for supervising the treatment of wash filtrate based on artificial intelligence, characterized by: The method comprises the following steps: Step 1: Divide the water washing filtrate into blocks according to spatial location, arrange the sensor group to collect pollutant concentration data and auxiliary parameters and perform preprocessing; Step 2: For water-soluble pollutants, construct a concentration distribution map and calculate the overall concentration; Step 3: For water-insoluble pollutants, weights are set based on particle dynamics and the weighted average concentration is calculated; Step 4: Set the discharge standard for the water wash filtrate and determine whether the pollutant concentration meets the discharge standard based on the interval; Step 5: Based on the emission standard determination results, the water wash filtrate treatment process is dynamically adjusted through an artificial intelligence algorithm; In step 1, the water washing filtrate is divided into 1 to N blocks according to the spatial position; the block index is represented by n, n∈{1,2,…,N}; the block volume is represented by V1~V N ; Deploy sensor groups in each block to directly or indirectly collect pollutant concentration data and auxiliary parameters from different parts; record the locations of corresponding sensor deployment points; Pollutants include water-soluble pollutants and water-insoluble pollutants; The concentration data of a water-soluble pollutant in different blocks is expressed as: C i,1 ~C i,N ; Where i is a positive integer, representing the sequence of water-soluble pollutant categories, i∈{1,2,…,I}, I is a positive integer, representing the number of water-soluble pollutant categories; The concentration data of a certain non-water-soluble pollutant in different blocks is expressed as: c j,1 ~c j,N ; Wherein, j is a positive integer, representing the sequence of non-water-soluble pollutant categories, j∈{1,2,…,J}, J is a positive integer, representing the number of non-water-soluble pollutant categories; Pollutant data include water-soluble pollutant concentration data and water-insoluble pollutant concentration data; For a certain pollutant data, if the pollutant data exceeds the preset range due to sensor or transmission, it will be replaced with the average value of the pollutant concentration data of the remaining blocks; For the replaced pollutant data, the mean and standard deviation are calculated to further standardize the pollutant data; In step 2, for water-soluble pollutants Ci, the kriging interpolation method is used to perform spatial interpolation of concentration: C i (x,y,z)=μ i +Σ n=1 N λ n [C i,n -μ i ]; Among them, λ n represents the weight calculated by the Kriging algorithm, Σ n=1 N λ n =1;C i (x,y,z) represents the water-soluble pollutant C at the point (x,y,z) to be estimated i Concentration data; C i,n Represents the water-soluble pollutant C at the sensor deployment point corresponding to the nth block i Concentration data; μ i Indicates water-soluble pollutants C i The mean of the concentration data in all blocks The three-dimensional space where the water washing filtrate is located is divided into volumes ΔV, and each volume ΔV is obtained by Kriging interpolation method. m The concentration C i,m (x,y,z); calculate the total concentration by integration; when ΔV m Approaching 0: ; Where M is a positive integer, indicating the number of partitioned volumes; m∈{1,2,…,M}, indicating the sequence of partitioned volumes, ΔV m represents the size of the mth volume, C i,m (x,y,z) represents the water-soluble pollutant C at the mth volume corresponding to the estimated point (x,y,z) i concentration data; In step 3, for the water-insoluble pollutant c j : The concentration of the nth block is represented by c j,n ; Obtaining water-insoluble pollutants c from auxiliary parameters j The average particle radius r j 、Insoluble pollutants c j Particle density ρ j , water density ρ w , water viscosity μ1 and gravitational acceleration g; calculate the sedimentation rate of the block: v j,n =2r j 2 (ρ j -ρ w )g / 9μ1; The sedimentation rate and the block volume V n Multiply them together to get the weight of the contribution to the overall concentration: w j,n =v j,n ×V n ; w j,n As weight, the weighted average of all block concentrations is taken to get the overall concentration: c j,total =∑ n=1 N w j,n c j,n / ∑ n=1 N w j,n .

2. The artificial intelligence-based water wash filtrate treatment supervision method according to claim 1, characterized in that: In step 4, the emission standard range intervals [C i,l1 ,C i,l2 ] and [c j,l1 ,c j,l2 ]; Comparison of water-soluble pollutant concentration C i,total and the concentration of water-insoluble pollutants c j,total and the corresponding emission standard range; When the concentrations of all types of water-soluble and non-water-soluble pollutants are within the emission standard range, the wash filtrate is considered to meet the emission standards; When there are water-soluble pollutants or insoluble pollutants whose concentration is not within the emission standard range, the water-wash filtrate is considered to be not in compliance with the emission standard. A corresponding treatment plan is generated based on the difference between the concentration of water-soluble pollutants or insoluble pollutants that are not within the emission standard range and the emission standard range for feedback.

3. The artificial intelligence-based water wash filtrate treatment supervision method according to claim 2, characterized in that: In step 5, when the water wash filtrate does not meet the discharge standards, the specific pollutant type and the degree of excess are identified; Establish a solution library containing multiple treatment strategies, each corresponding to different pollutant types, exceedance levels, and treatment equipment parameter combinations; match similar cases from the solution library based on specific pollutant types and exceedance levels, and screen candidate solutions; A deep Q-network is used to evaluate and optimize candidate solutions. The filtrate state is defined as the multi-dimensional information of the current pollutant concentration, equipment operating parameters, and treatment time. Actions are defined as different treatment solutions, and the reward function is set as the degree to which the post-treatment pollutant concentration approaches the standard range. Continuous iterative training enables the algorithm to learn the optimal treatment strategy and generate an optimized treatment solution tailored to the specific pollutant type and level of exceedance. The generated treatment plan is converted into specific equipment operating parameters, connected to the water treatment equipment through Internet of Things technology, and control instructions are transmitted in real time to achieve automatic configuration of treatment process parameters; After the treatment process is adjusted, the sensor group continuously collects pollutant concentration and auxiliary parameter data, recalculates pollutant concentration at preset fixed intervals, and compares it with the emission standards to determine whether the treatment effect meets the expectations. If the expected effect is not achieved, return to step 5 to regenerate the optimization plan; The pollutant concentration data before and after treatment, the treatment scheme adopted, the equipment operating parameters and the final treatment results are stored in the historical database, and the accumulated historical data are used to retrain the Kriging interpolation model and the reinforcement learning model regularly; at the same time, the scheme library is optimized according to the actual treatment effect, including deleting old schemes and adding new treatment strategies.

4. An artificial intelligence-based water wash filtrate treatment monitoring system, applying the artificial intelligence-based water wash filtrate treatment monitoring method according to any one of claims 1 to 3, characterized in that: The system includes a data acquisition module, a concentration calculation module, a standard judgment module and a processing optimization module; The data acquisition module is used to divide the water washing filtrate blocks according to spatial positions, arrange sensor groups to collect pollutant concentration data and auxiliary parameters and perform preprocessing; The concentration calculation module is used to construct a concentration distribution map and calculate the overall concentration for water-soluble pollutants; for insoluble pollutants, the weight is set in combination with particle dynamics to calculate the weighted average concentration. The standard judgment module is used to set the discharge standard of the water washing filtrate and judge whether the pollutant concentration meets the discharge standard according to the interval; The processing optimization module is used to dynamically adjust the water washing filtrate processing process through an artificial intelligence algorithm based on the emission standard determination results.

5. The artificial intelligence-based water washing filtrate processing monitoring system according to claim 4 is characterized in that: The data acquisition module includes a block division unit, a data acquisition unit and a data preprocessing unit; The block division unit is used to divide the water washing filtrate blocks according to spatial positions; The data acquisition unit is used to arrange the sensor group to collect pollutant concentration data and auxiliary parameters; The data preprocessing unit is used to preprocess the collected pollutant concentration data.

6. The artificial intelligence-based water wash filtrate processing monitoring system according to claim 5, characterized in that: The concentration calculation module includes an interpolation unit, a concentration calculation 1 unit, a weight calculation unit and a concentration calculation 2 unit; The interpolation unit is used to perform spatial interpolation of concentration of water-soluble pollutants using the Kriging interpolation method; The concentration calculation unit 1 is used to calculate the overall concentration of water-soluble pollutants by integration; The weight calculation unit is used to set weights for water-insoluble pollutants in combination with particle dynamics; The concentration calculation unit 2 is used for weighted calculation of the average concentration of water-insoluble pollutants.

7. The artificial intelligence-based water wash filtrate processing monitoring system according to claim 6, characterized in that: The standard judgment module includes a standard setting unit and a matching comparison unit; The standard setting unit is used to set the discharge standard of the water washing filtrate; The matching and comparison unit is used to determine whether the pollutant concentration meets the emission standards according to the interval; The processing optimization module includes an identification unit, a solution generation unit, an execution unit and an optimization unit; The identification unit is used to identify the specific pollutant type and the degree of excess when the water washing filtrate does not meet the emission standards; the solution generation unit is used to match similar cases from the solution library according to the specific pollutant type and the degree of excess, screen candidate solutions and evaluate and optimize the solutions; the execution unit is used to convert the optimal solution into equipment control instructions to achieve automatic adjustment; the optimization unit is used to optimize the model and solution library according to the actual treatment effect.

Images