CN118173188A

CN118173188A - Method, medium and system for determining material amount for processing sea black and odorous silt clay

Info

Publication number: CN118173188A
Application number: CN202410591473.7A
Authority: CN
Inventors: 王保栋; 梁斌; 王立彬; 肖志伟; 王瑞; 刘超; 郑为升
Original assignee: China Construction Eighth Bureau Development and Construction Co Ltd
Current assignee: China Construction Eighth Bureau Development and Construction Co Ltd
Priority date: 2024-05-14
Filing date: 2024-05-14
Publication date: 2024-06-11
Anticipated expiration: 2044-05-14
Also published as: CN118173188B

Abstract

The invention provides a method, medium and system for determining the material amount for treating sea-facing black and odorous silt clay, which belong to the technical field of sea-facing black and odorous silt clay treatment and comprise the following steps: obtaining physical and chemical parameters of black and odorous silt clay to be treated; matching to obtain M history records with highest matching degree; obtaining results of experiments through a small scale, and collecting processed result effect data; fine tuning the pre-trained processing material amount calculation model to obtain an optimized calculation model; amplifying to engineering scale, inputting the preliminary material quantity parameters corresponding to small-scale experiments into the optimization calculation model to obtain material quantity parameters suitable for actual engineering; according to the material amount parameters adapting to the actual engineering, treating the black and odorous silt clay to be treated, and optimizing the material amount parameters in the treatment process to obtain optimized material amount parameters in the actual treatment process; and in the treatment process, the obtained optimized material quantity parameters are sent to an operator in real time.

Description

Method, medium and system for determining material amount for processing sea black and odorous silt clay

Technical Field

The invention belongs to the technical field of sea-facing black and odorous silt clay treatment, and particularly relates to a method, medium and system for determining the material amount for sea-facing black and odorous silt clay treatment.

Background

The black and odorous silt clay is an organic pollutant material deposited in a typical urban black and odorous water body and generally contains a large amount of pollutants such as organic matters, heavy metals, pathogenic bacteria and the like. If the waste water is not properly treated, the waste water is directly stacked or buried, and serious secondary pollution is caused to the surrounding environment. Therefore, how to efficiently and economically treat such black and odorous silt clay has been an important research topic in the field of environmental control.

At present, common black and odorous silt clay treatment processes comprise a physicochemical method, a biological method, a pyrolysis method and the like. The physicochemical method mainly removes or stabilizes pollutants by flocculation precipitation, chemical oxidation reduction and other modes, the biological method utilizes the catabolism of aerobic or anaerobic microorganisms to reduce the content of organic pollutants, and the pyrolysis method is to decompose the organic matters into gas and liquid products with smaller molecules by high-temperature pyrolysis. The processes can realize harmless and recycling of black and odorous silt clay to a certain extent.

However, these existing treatments still have some problems and limitations in practical applications:

1. There is a large uncertainty in the choice of process parameters. The components and characteristics of the black and odorous muddy clay are greatly different, and the optimal treatment process parameters (such as the dosage of the flocculating agent, the dosage of the pH regulator and the like) of different samples are also greatly different. However, most of the current treatment processes rely on experience to set parameters, so that accurate optimization is difficult to be performed aiming at the characteristics of specific samples, and the treatment effect is often not ideal.

2. There is a high trial and error cost and time loss in the process. Due to the complexity of black and odorous muddy clay, a great deal of tests are usually required in the actual treatment process, and the process parameters are repeatedly adjusted to find a proper treatment scheme. This not only increases the processing cost, but also greatly extends the processing cycle.

In summary, the prior art has the technical problems of high difficulty in selecting process parameters and high trial-and-error cost.

Disclosure of Invention

In view of the above, the invention provides a method, medium and system for determining the amount of material used for processing the marine black and odorous silt clay, which can solve the technical problems of high difficulty in selecting technological parameters and high trial-and-error cost in the prior art.

The invention is realized in the following way:

The first aspect of the invention provides a method for determining the treatment material amount of the black and odorous silt clay in the sea, which comprises the following steps of;

S10, obtaining physical and chemical parameters of black and odorous silt clay to be treated;

s20, based on the physicochemical parameters, matching is carried out in a preset black and odorous silt clay treatment process material history database, and M histories with highest matching degree are obtained;

s30, obtaining a result of a small-scale experiment, wherein the small-scale experiment is to treat the black and odorous silt clay to be treated according to the treatment process materials of the M histories, and collect the effect data of the treated result;

S40, fine tuning is carried out on the pre-trained processing material amount calculation model according to the M historical records and the corresponding small-scale experiment results, so that an optimized calculation model is obtained;

S50, selecting a history record with a corresponding result effect within a preset threshold range according to the small-scale experiment result, and clustering to obtain a preliminary material amount parameter;

s60, amplifying to an engineering scale, and inputting the preliminary material quantity parameters corresponding to a small-scale experiment into the optimization calculation model to obtain the material quantity parameters suitable for actual engineering;

S70, treating the black and odorous silt clay to be treated according to the material amount parameters adapting to the actual engineering, and optimizing the material amount parameters by utilizing a genetic algorithm according to the treatment effect of each process link in the treatment process to obtain optimized material amount parameters in the actual treatment process;

and S80, in the processing process, sending the obtained optimized material quantity parameters to an operator in real time.

Typically M is an integer greater than 10.

On the basis of the technical scheme, the method for determining the material amount for treating the marine black and odorous silt clay can be further improved as follows:

wherein the physicochemical parameters at least comprise the water content, organic matter content, pH value, cation exchange capacity and heavy metal content of the black and odorous silt clay.

The specific step of the S20 comprises the steps of firstly carrying out standardization processing on materialized parameter values in each history record to eliminate dimension differences, then calculating Euclidean distances between materialized parameters of a current sample to be processed and materialized parameters of each history record, sequencing according to the Euclidean distances from small to large, and selecting the first M records with the smallest distance as a matching result to be output.

The processing material amount calculation model is a machine learning model based on a neural network and comprises an input layer, a hidden layer and an output layer, wherein the number of nodes of the input layer is equal to the number of physicochemical parameters of black and odorous silt clay, and the number of nodes of the output layer is equal to the number of types of processing materials.

When the pre-trained processing material amount calculation model is subjected to fine tuning optimization, a transfer learning method is adopted, partial parameters related to the current sample in the pre-trained model are kept unchanged, and only specific parameters are subjected to fine tuning, so that the fitting capacity of the model to the current sample is improved.

The clustering adopts a k-means algorithm, the history record is divided into k clusters according to the processing material quantity parameters, and the clustering center point is the preliminary material quantity parameter.

K is generally 5 to 10.

The genetic algorithm comprises the following specific steps:

step 1, coding the material quantity parameters of each process link into individual gene strings;

step 2, initializing a random generation initial population;

step 3, calculating the fitness function value of each individual;

Step 4, selecting an individual with higher fitness by adopting a roulette selection operator as a parent;

Step 5, intersecting and mutating parent individuals to generate new offspring individuals;

Step 6, replacing individuals with lower adaptability in the parent population with newly generated offspring individuals to form a new population;

and 7, repeating the steps 3-6 until the termination condition is met, namely the maximum iteration times or the fitness function value is smaller than a preset threshold value.

Further, the fitness function is specifically:

Wherein R is the removal rate of harmful substances, eta is the dry weight increase rate, and alpha and beta are regulating factors; by minimizing this function, the removal rate and dry weight increase rate can be optimized simultaneously.

A second aspect of the present invention provides a computer readable storage medium, wherein the computer readable storage medium has stored therein program instructions, which when executed, are configured to perform a method for selecting a parameter of a black and odorous silt clay treatment process as described above.

A third aspect of the present invention provides a marine black and odorous muddy clay treatment material amount determination system, which comprises the above-mentioned computer-readable storage medium.

Compared with the prior art, the method, the medium and the system for determining the material amount for treating the marine black and odorous silt clay have the beneficial effects that:

1. can rapidly determine proper treatment process and material proportion according to the physicochemical characteristics of the black and odorous silt clay to be treated. According to the method, firstly, through physical and chemical parameter matching, records most similar to a current sample are searched in a historical processing database, and preliminary processing parameters are obtained. The applicability of these parameters on the current sample was then verified using a small-scale experiment and further optimization corrections were performed using a machine learning model. Thus, the time for parameter determination can be greatly shortened, and a large amount of trial-and-error cost is reduced.

2. Dynamic optimization parameters are realized in the actual processing process. In the method, when the actual engineering scale is processed, a genetic algorithm is adopted to continuously optimize the processing material parameters, and the optimal technological parameter configuration is found according to the fed-back real-time processing effect data. The closed-loop control mode ensures that the treatment process can be continuously and automatically optimized, and the treatment efficiency and stability are greatly improved.

In summary, the technical problems of high difficulty in selecting process parameters and high trial-and-error cost in the prior art are solved by the scheme of the invention.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments of the present invention will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a method provided by the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

As shown in fig. 1, the invention provides a method for determining the treatment material amount of the black and odorous silt clay in the sea, which comprises the following steps:

S20, based on physicochemical parameters, matching is carried out in a preset black and odorous silt clay treatment process material history database, and M histories with highest matching degree are obtained;

S30, obtaining a result of a small-scale experiment, wherein the small-scale experiment is to treat black and odorous silt clay to be treated according to the treatment process materials of M historical records, and collect the effect data of the treated result;

S40, fine tuning the pre-trained processing material amount calculation model according to M historical records and corresponding small-scale experiment results to obtain an optimized calculation model;

S50, selecting a history record with a corresponding result effect within a preset threshold range according to a small-scale experiment result, and clustering to obtain a preliminary material amount parameter;

S60, amplifying to an engineering scale, and inputting the preliminary material quantity parameters corresponding to a small-scale experiment into an optimization calculation model to obtain the material quantity parameters suitable for actual engineering;

s70, treating the black and odorous silt clay to be treated according to the material amount parameters suitable for actual engineering, and optimizing the material amount parameters by using a genetic algorithm according to the treatment effect of each process link in the treatment process to obtain optimized material amount parameters in the actual treatment process;

The following describes in detail the specific embodiments of the above steps:

Step S10, obtaining physical and chemical parameters of black and odorous silt clay to be treated

The purpose of this step is to collect a series of physicochemical parameters of the sample of black and odorous muddy clay to be treated, providing a basis for the subsequent steps. Specifically, it is necessary to measure the following parameters of the black and odorous muddy clay, such as water content, organic matter content, pH value, cation exchange capacity, heavy metal content, etc. The physicochemical parameters can directly reflect the pollution characteristics and the treatment difficulty of the black and odorous silt clay. The parameters can be obtained by standard experimental analysis methods, such as a drying method for measuring the water content, a potassium dichromate capacity method for measuring the organic matter content, a pH electrode method for measuring the pH value, an ion exchange method for measuring the cation exchange capacity, an atomic absorption spectrometry for measuring the heavy metal content, and the like. These physicochemical parameter data will be used as input variables in subsequent steps.

Step S20, based on the physicochemical parameters, matching is carried out in a preset history database for the black and odorous silt clay treatment process materials, and M histories with highest matching degree are obtained

The aim of the step is to search M histories closest to the physicochemical parameters of the black and odorous silt clay to be treated in the existing black and odorous silt clay treatment process material history database. These histories include various process parameters during the previous treatment, such as flocculant usage, additive usage, pH adjuster usage, etc.

In particular, a distance metric based approach may be employed for matching. Firstly, standardized processing is carried out on the physicochemical parameter values in each history record, and dimension differences are eliminated. Then, the Euclidean distance between the physicochemical parameter of the current sample to be processed and each historical record physicochemical parameter is calculated:

；

Wherein, Is the i-th physicochemical parameter value of the current sample to be treated,/>The i-th materialized parameter value in the history record is n, and the n is the total materialized parameter number.

And selecting the first M records with the smallest distance as a matching result to output according to the sorting from small Euclidean distance to large Euclidean distance. The M records will provide reference basis for the subsequent steps. The value of M can be adjusted according to practical situations, and usually 5-10 values are suitable.

Step S30, obtaining a result of a small-scale experiment, wherein the small-scale experiment is to treat the black and odorous silt clay to be treated according to the treatment process materials recorded by the M histories, and collect the effect data of the treated result

The purpose of this step is to perform small-scale experimental treatment according to the M history matching records obtained in step S20 for the black and odorous muddy clay sample to be currently treated, and collect treatment effect data.

Specifically, first, laboratory-scale treatment is performed on the black and odorous silt clay sample to be treated according to treatment process parameters (including flocculant dosage, pH regulator dosage, reaction time, stirring intensity, and the like) in M histories. The experimental treatment process needs to strictly control various technological parameters, and collect index data of the treated samples, such as removal rate of harmful substances, dry weight increase rate, toxicity index and the like. These data will become training samples in subsequent steps.

The purpose of the small-scale experiment is to verify the applicability of the processing parameters in the history record on the current sample and provide basis for the subsequent steps. By comparing the experimental result with the history record, whether the characteristics of the current sample are obviously different from those of the history sample can be known.

Step S40, fine tuning the pre-trained processing material amount calculation model according to the M historical records and the corresponding small-scale experiment results to obtain an optimized calculation model

The purpose of this step is to optimize and fine tune the previously pre-trained calculation model of the amount of treatment material by using the small-scale experimental data obtained in step S30, so as to obtain a calculation model more suitable for the current black and odorous muddy clay to be treated.

The pre-trained processing material amount calculation model can be constructed by adopting a machine learning method based on a neural network, wherein input variables comprise physicochemical parameters of black and odorous silt clay, and output variables are the optimal amount of various processing materials. A large amount of historical sample data is adopted for training, so that the model can capture the complex nonlinear relation between the properties of black and odorous silt clay and the treatment materials.

The new data obtained in the small-scale experiment of step S30 can be used to fine tune the pre-trained model. Specifically, a migration learning method can be adopted to fix part of parameters of the pre-training model, and fine tuning optimization is performed on only part of parameters related to the current sample. The optimization objective function may be set to:

；

Wherein, Actual material amount of small-scale experimental result,/>For the predicted amount of material of the model,/>The number of samples for small-scale experiments. By minimizing this objective function, the predicted outcome of the model can be made as close as possible to the actual experimental data.

Through the optimization of the step, the original processing material amount calculation model is more fit with the characteristics of the black and odorous silt clay to be processed at present, and more accurate material amount prediction is provided for the subsequent step.

Step S50, according to the small-scale experimental result, selecting a history record with a corresponding result effect within a preset threshold range, and clustering to obtain a preliminary material amount parameter

The purpose of this step is to further analyze the small-scale experimental result of step S30, select the history records with better treatment effect, and perform cluster analysis on these records to obtain the preliminary treatment dosage parameters.

First, it is necessary to set some desired treatment effect index such as a removal rate of harmful substances, a dry weight increase rate, etc., and to set a corresponding threshold range. Then, the small-scale experimental result of step S30 is screened, and only the history records of which the treatment effect indexes meet the preset threshold range are reserved.

Next, cluster analysis is performed on these screened histories. The k-means method may be used to cluster according to the process dosage parameters (flocculant dosage, pH adjuster dosage, etc.). The purpose of clustering is to categorize together histories with similar material ratios to provide preliminary material amount parameters for subsequent steps.

The cluster center of the cluster can be considered as the preliminary dosage parameter. For example, if k-means clustering is employed, k cluster centers can be obtained as preliminary usage parameters. These parameters will be further optimized and adjusted in subsequent steps.

The setting of the threshold value needs to be determined in combination with actual process requirements and experience. For example, the harmful substance removal rate threshold may be set to 90% or more, the dry weight increase rate threshold may be set to 50% or more, and the like.

Step S60, amplifying to engineering scale, adopting preliminary material quantity parameters corresponding to small-scale experiments, and inputting the preliminary material quantity parameters into the optimization calculation model to obtain material quantity parameters suitable for actual engineering

The purpose of this step is to apply the preliminary material amount parameter obtained in step S50 to the actual engineering scale treatment process of the black and odorous silt clay, and obtain the final material amount parameter suitable for engineering practice by using the optimized treatment material amount calculation model (step S40).

Specifically, the preliminary dosage parameters obtained in step S50, such as the flocculant dosage, the pH adjuster dosage, etc., are first scaled up to the actual engineering scale for processing. At the same time, these parameters are input into the processing material amount calculation model optimized in step S40, and prediction and optimization are performed by using the model.

The optimization objective function of the model may be set to:

；

Wherein, Is the removal rate of harmful substances,/>For the dry weight increase rate, α and β are the adjustment factors, with a value ranging from [0,1], for balancing the relative importance of the pest removal rate and the dry weight increase rate. By minimizing this objective function, the final material quantity parameter that meets the actual requirements of the project can be obtained.

The adjustment factors α and β here may be set according to actual processing targets. For example, if the removal rate of harmful substances is the primary objective, it is possible to set α to be large and β to be small, and if the increase in dry weight is also important, the value of β can be appropriately increased.

Through the optimization calculation of the step, the final material quantity parameter suitable for the actual condition of the current engineering can be obtained, and a basis is provided for the subsequent actual treatment.

Step S70, treating the black and odorous silt clay to be treated according to the material amount parameters adapting to the actual engineering, and optimizing the material amount parameters by utilizing a genetic algorithm according to the treatment effect of each process link in the treatment process to obtain optimized material amount parameters in the actual treatment process

The purpose of this step is to apply the final material quantity parameter obtained in step S60 to the actual engineering treatment process, monitor the treatment effect of each process link in real time in the treatment process, and dynamically optimize the material quantity parameter by using a genetic algorithm to obtain the final optimized material quantity parameter.

Firstly, inputting the material amount parameters obtained in the step S60 into an actual engineering treatment process, and treating a black and odorous silt clay sample to be treated. In various links of the treatment process, such as flocculation, precipitation, dehydration and the like, the treatment effect indexes such as the removal rate of harmful substances, the dry weight increase rate and the like are monitored and collected in real time.

These treatment effect indices are then optimized using genetic algorithms. The genetic algorithm is a random search optimization algorithm based on natural selection and genetic mechanism, and can effectively find a globally optimal solution. Specifically, the following steps may be employed:

1. coding, namely coding the material quantity parameters of each process link into an individual gene string.

2. Initial population-an initial population comprising a plurality of individuals is randomly generated.

3. Fitness evaluation-the fitness function value of each individual is calculated, where the objective function in step S60 may be used.

4. Selecting, namely selecting an individual with higher fitness as a father by adopting methods such as roulette selection and the like.

5. Crossover and mutation, namely crossover and mutation operation is carried out on the selected parent individuals, and new offspring individuals are generated.

6. And replacing, namely replacing individuals with lower adaptability in the parent population by newly generated offspring individuals to form a new population.

7. And (3) iterating, namely repeating the steps 3-6 until a termination condition is met (such as the maximum iteration number or the objective function value is smaller than a preset threshold value).

Through multiple rounds of iterative optimization, the optimal material quantity parameters in the treatment process can be obtained. These parameters can be fed back in real time to the operator, guiding the actual process.

The method has the advantages that feedback information in the actual treatment process can be fully utilized, the material quantity parameters can be dynamically adjusted, the treatment effect is improved, and the operation cost is reduced.

Step S80, in the process of treatment, the obtained optimized material quantity parameters are sent to an operator in real time

The purpose of this step is to feed back the optimized material quantity parameters obtained in step S70 in real time to the personnel responsible for the actual processing operation, providing decision basis for them.

Specifically, various optimized process dosage parameters (such as flocculant dosage, pH regulator dosage, etc.) can be displayed to an operator in a chart or digital form and updated in real time. Operators can adjust the actual adding amount accordingly, and ensure that the treatment effect reaches the standard.

Meanwhile, various index data (such as the removal rate of harmful substances, the dry weight increasing rate and the like) collected in the treatment process can be fed back to operators in real time, so that the operators can know the treatment effect in time, and a basis is provided for the next operation decision.

In general, this step aims to establish a real-time feedback mechanism between the process and the operator, so that the operation is more intelligent and accurate, thereby improving the processing effect and reducing the cost and time consumption.

Specifically, the principle of the invention is as follows: the historical processing data is fully utilized, and the quick parameter determination and dynamic optimization aiming at specific sample characteristics are realized by combining machine learning and intelligent optimization technology. Specifically, the method mainly comprises the following key technical principles:

1. Historical data matching based on physicochemical parameters

The composition and characteristics of the black and odorous muddy clay are greatly different, which determines that the optimal treatment process parameters are also greatly different. The method of the invention firstly establishes a characteristic vector of a sample to be treated by measuring a series of physicochemical parameters such as water content, organic matter content, pH value, cation exchange capacity, heavy metal content and the like of the sample. And searching a record most similar to the physicochemical parameters of the current sample in a pre-established historical processing database to be used as a primary processing technological parameter reference. The matching method based on the similarity can quickly lock the applicable process of the current sample and provide a reasonable starting point for subsequent optimization.

2. Parameter correction based on small-scale experiments

Although preliminary processing parameters can be obtained through historical data matching, due to the complexity of the black and odorous muddy clay, certain uncertainty still exists in the applicability of these parameters to actual samples. Therefore, the method of the present invention also requires small-scale process experimental verification of the current sample on a laboratory scale and collection of process effect data. The experimental data can be directly used as a training sample for subsequent optimization, can also help judge the applicability of the historical parameters on the current sample, and provides a basis for further adjustment of the parameters.

3. Parameter optimization based on machine learning model

In order to further improve the accuracy of the treatment parameters, the invention adopts a machine learning method based on a neural network to construct a prediction model, takes the physicochemical parameters of the black and odorous silt clay as input variables, and takes the treatment material amount (such as the dosage of a flocculating agent, the dosage of a pH regulator and the like) as output variables. The model can effectively learn and capture the complex relationship between the characteristics of the black and odorous silt clay and the treatment materials. On the basis of small-scale experimental data, the pre-training model is optimized and corrected by adopting a transfer learning method, so that the model can be better adapted to the characteristics of the current sample, and more accurate material quantity prediction is obtained.

4. Dynamic optimization based on genetic algorithm

In the actual engineering scale treatment process, the method also adopts a genetic algorithm to continuously optimize the treatment material parameters. Specifically, the material quantity parameters of each process link are encoded into a gene string of an individual to construct an initial population. And then calculating the fitness of each individual according to the treatment effect indexes (such as the removal rate, the dry weight increase rate and the like) monitored in real time, and finally obtaining the optimal material quantity parameter through continuous iterative optimization by genetic operations such as selection, crossover, variation and the like. The dynamic optimization mode can ensure that parameters are continuously adjusted in actual operation, improve the treatment effect and reduce the operation cost.

The following is a specific embodiment of the present invention:

In the embodiment, the black and odorous silt clay deposited in a certain sea black and odorous water body is taken as an object, and the high-efficiency harmless treatment of the black and odorous silt clay is successfully realized by adopting the treatment process parameter selection method provided by the invention. The specific process is as follows:

1. obtaining physical and chemical parameters of black and odorous silt clay to be treated

Firstly, a series of physicochemical property tests are carried out on the collected black and odorous silt clay samples, and the main physicochemical parameter data are obtained as follows:

The water content is 78.2%;

32.5% of organic matter;

The pH value is 6.8;

cation exchange capacity 18.6 cmol/kg;

copper (Cu) content 210 mg/kg;

Cadmium (Cd) content of 3.2 mg/kg;

the lead (Pb) content was 102 mg/kg.

These physicochemical parameter data will be used as input variables for the subsequent process parameter determination.

2. Matching in a history database

And according to the obtained physicochemical parameter data, matching is carried out in a pre-established black and odorous silt clay treatment process material history database. The database contains 800 historical processing records.

Firstly, standardized processing is carried out on the physicochemical parameter data in each history record, and dimension differences are eliminated. And then calculating Euclidean distance between the physicochemical parameter of the current sample and each historical record, and sorting according to the distance from small to large. The first 10 records with the smallest distance are selected as the matching result to be output, and the specific table is as follows:

Sequence number	Water content	Organic matter	PH value of	Cation exchange capacity cmol/kg	Cu content mg/kg	Cd content mg/kg	Pb content mg/kg	Dosage of flocculant kg/t	The dosage of the pH regulator is kg/t	Harmful substance removal rate	Rate of dry weight increase
												1	75.1%	29.8%	6.9	17.3	195	2.8	95	180	35	91.2%	54.3%
2	77.4%	31.2%	6.7	18.1	208	3.0	98	185	38	89.7%	52.1%
												3	73.5%	28.6%	7.0	16.9	191	2.6	92	175	33	92.4%	56.5%
4	79.1%	33.1%	6.6	19.2	215	3.3	105	190	41	88.4%	50.7%
												5	74.8%	30.3%	6.8	17.8	201	2.9	96	182	36	90.6%	53.4%
6	76.9%	30.9%	6.7	18.4	212	3.1	102	188	39	89.2%	51.6%
												7	72.3%	27.8%	7.1	16.5	187	2.5	89	172	32	93.1%	57.8%
8	78.6%	32.7%	6.5	19.6	220	3.4	108	193	42	87.6%	49.8%
												9	74.2%	29.5%	6.9	17.4	197	2.7	93	179	35	91.0%	54.0%
10	76.3%	30.6%	6.8	18.0	205	3.0	99	185	37	89.8%	52.3%

From the table, the physicochemical parameters of the 10 histories are closer to those of the current sample, so that reference basis can be provided for subsequent experimental verification and parameter optimization.

3. Verification of small-scale experiments

According to the processing technological parameters in the 10 historical records, small-scale experimental processing is carried out on the current black and odorous silt clay sample in a laboratory. The specific operation is as follows:

(1) Flocculation reaction, namely adding a sample to be treated into a 10L reaction kettle, adding a flocculating agent, and stirring for reaction for 30 minutes. The flocculant dosage was adjusted in the range of 180-193 kg/t by reference to the history in the above table. The stirring intensity is controlled to be 100-120 r/min.

(2) And (3) precipitation separation, namely standing for 30 minutes after flocculation reaction is finished, and carrying out solid-liquid separation. The precipitate (i.e. treated black and odorous muddy clay) is collected and the supernatant is discharged.

(3) Dewatering, namely adding the precipitate into a filter cake machine, and dewatering and drying until the water content is less than or equal to 40%.

(4) And (5) measuring indexes such as harmful substance removal rate, dry weight increase rate, heavy metal content and the like of the samples before and after treatment.

Through small-scale experiments, the following result data were obtained:

the removal rate of harmful substances is 91.3 percent;

The dry weight increase rate is 53.2%;

Copper (Cu) content of 22 mg/kg (harmful substance removal rate 89.5%);

cadmium (Cd) content of 0.4 mg/kg (harmful substance removal rate of 87.5%);

lead (Pb) content of 12 mg/kg (harmful substance removal rate 88.2%).

Compared with the data in the 10 historical records, the processing effect of the small-scale experiment is relatively similar, and the applicability of the historical process parameters on the current sample is verified. And basic data is provided for subsequent parameter optimization.

4. Optimizing a process material amount calculation model

Based on the small-scale experimental data, the previously pre-trained process material amount calculation model is further optimized. The model is constructed by adopting a machine learning method based on a neural network, wherein the input variable is the physicochemical parameter of the black and odorous silt clay, and the output variable is the flocculant dosage and the pH regulator dosage.

Firstly, physicochemical parameter data and processing result data obtained by a small-scale experiment are used as new training samples to be input into a pre-training model. Then, a transfer learning method is adopted, fine tuning optimization is carried out on only part of parameters related to the current sample, and other parameters are kept unchanged. The optimization objective function is:

；

Wherein, For practical amounts of flocculant and pH regulator in small-scale experiments,/>For the amount of the model prediction,Is the number of samples. By minimizing this objective function, the predicted outcome of the model is made as close as possible to the actual usage.

And obtaining a processing material amount calculation model which is more fit with the characteristics of the current black and odorous silt clay sample through 200 rounds of iterative optimization. The prediction precision of the model is obviously improved, and the average relative error is controlled within 5%.

5. Determining preliminary dosage parameters

Next, based on the results of the small-scale experiment in step 3, further analysis was performed on the above 10 histories.

Firstly, setting threshold requirements of some treatment effect indexes, wherein the removal rate of harmful substances is more than or equal to 90%, and the dry weight increase rate is more than or equal to 50%. The 10 histories were screened and 7 records were found to meet these requirements.

Then, the 7 histories meeting the requirements are subjected to cluster analysis, and the 7 histories are divided into 3 clusters by adopting a k-means algorithm. The obtained 3 clustering center points are the preliminarily determined material quantity parameters:

The dosage of the flocculant is 183 kg/t;

the dosage of the pH regulator is 36 kg/t.

These parameters will serve as initial input values for the subsequent actual engineering process.

6. Scaling up to engineering scale for treatment

And amplifying to an actual engineering scale according to the preliminarily determined material amount parameters, and treating a 100 m-wave black and odorous silt clay sample. At the same time, these parameters are also input into the optimized processing material amount calculation model to further optimize.

The optimization objective function is:

；

Wherein R is the removal rate of harmful substances, and eta is the dry weight increase rate. By minimizing this weighted objective function, the optimal material quantity parameters for engineering scale processing are obtained as:

the dosage of the flocculant is 188 kg/t;

the dosage of the pH regulator is 39 kg/t.

7. Dynamic optimization in actual processing

The determined optimal material amount parameter is input into the actual engineering treatment process. In each link of treatment, such as flocculation, precipitation, dehydration and the like, the treatment effect index data including harmful substance removal rate, dry weight increase rate, heavy metal harmful substance removal rate and the like are collected in real time.

These real-time data are then dynamically optimized using genetic algorithms. The method comprises the following specific steps:

(1) Coding, namely coding two parameters of flocculant dosage and pH regulator dosage into an 8-bit binary gene string.

(2) Initial population-initial population comprising 50 individuals was randomly generated.

(3) Fitness evaluation by objective functionAs a fitness function, a fitness value for each individual is calculated.

(4) Selecting, namely selecting an individual with higher fitness as a parent by adopting a roulette selection operator.

(5) Crossover and mutation by applying crossover probability of 0.7 and mutation probability of 0.03 to the selected parent individuals to generate new offspring individuals.

(6) And replacing, namely replacing individuals with lower adaptability in the parent population by newly generated offspring individuals to form a new population.

(7) And (5) iterating, namely repeating the steps (3) - (6), and iterating 150 times.

After 150 rounds of iterative optimization, the optimal material quantity parameters in the treatment process are obtained:

the dosage of the flocculant is 192 kg/t;

the dosage of the pH regulator is 41 kg/t.

The parameter values are improved compared with the initially determined optimal parameters, mainly because feedback data can be obtained in real time in the actual processing process, and the processing effect is further optimized through dynamic adjustment of a genetic algorithm.

8. Verification of treatment effect

And carrying out actual engineering treatment on the black and odorous silt clay sample of which the quantity is 100m by adopting the optimized material quantity parameters. The treatment results were as follows:

The removal rate of harmful substances is 92.1 percent;

56.3% of dry weight increase rate;

copper (Cu) content 18 mg/kg (harmful substance removal rate 91.9%);

cadmium (Cd) content of 0.3 mg/kg (harmful substance removal rate 90.6%);

lead (Pb) content 9 mg/kg (harmful substance removal rate 91.2%).

Compared with the small-scale experimental result, the method has the advantages that various indexes of actual engineering treatment are improved, key performance indexes such as harmful substance removal rate and dry weight increase rate reach or exceed 90%, the heavy metal content is greatly reduced, and the method meets the requirements of relevant environmental protection standards. This fully demonstrates the effectiveness and superiority of the proposed process parameter selection method.

Compared with the traditional processing mode of parameter determination by experience, the method has the following remarkable advantages:

1. Reducing a lot of trial and error costs and time. The conventional method requires a large amount of experimental trials to determine appropriate processing parameters, and is inefficient. The method can quickly lock the optimal parameters suitable for the current sample characteristics through historical data matching, small-scale experiment verification and machine learning model optimization.

2. Dynamic optimization of the processing process is realized. In actual engineering operation, the invention adopts the genetic algorithm to continuously optimize the material consumption parameters, automatically adjusts according to the real-time feedback data, ensures continuous improvement of the treatment effect and greatly improves the operation stability.

3. The treatment effect is greatly improved. Compared with the traditional method, the method has the advantages that the core indexes such as the harmful substance removal rate and the dry weight increase rate are improved by about 10%, the heavy metal removal effect is more ideal, and the overall treatment performance is more excellent.

The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention.

Claims

1. The method for determining the material amount for treating the sea-facing black and odorous silt clay is characterized by comprising the following steps of;

2. The method for determining the amount of black and odorous sludge clay treatment material in the sea according to claim 1, wherein the physicochemical parameters at least include water content, organic matter content, pH value, cation exchange capacity, and heavy metal content of the black and odorous sludge clay.

3. The method for determining the amount of the used material for treating the marine black and odorous silt clay according to claim 1 is characterized by comprising the specific steps of firstly carrying out standardization treatment on materialized parameter values in each history record to eliminate dimension differences, then calculating Euclidean distances between materialized parameters of a current sample to be treated and materialized parameters of each history record, sequencing according to the Euclidean distances from small to large, and selecting the first M records with the smallest distances as matching results to be output.

4. The method for determining the treatment material amount of the black and odorous muddy clay in the sea according to claim 1, wherein the treatment material amount calculation model is a neural network-based machine learning model and comprises an input layer, a hidden layer and an output layer, the number of nodes of the input layer is equal to the number of physicochemical parameters of the black and odorous muddy clay, and the number of nodes of the output layer is equal to the number of types of the treatment material.

5. The method for determining the treatment material amount of the marine black and odorous muddy clay according to claim 1, wherein when the pre-trained treatment material amount calculation model is subjected to fine adjustment optimization, a transfer learning method is adopted, partial parameters related to the current sample in the pre-trained model are kept unchanged, and only specific parameters are subjected to fine adjustment so as to improve the fitting capacity of the model to the current sample.

6. The method for determining the treatment material amount of the black and odorous muddy clay in the sea according to claim 1, wherein the clustering adopts a k-means algorithm, the history record is divided into k clusters according to the treatment material amount parameter, and the clustering center point is the preliminary material amount parameter.

7. The method for determining the amount of the sludge clay treatment material for the sea-going black and odorous in the sea according to claim 1, wherein the specific steps of the genetic algorithm include:

step 2, initializing a random generation initial population;

step 3, calculating the fitness function value of each individual;

8. The method for determining the amount of material used for treating marine black and odorous silt clay according to claim 7, wherein the fitness function value is specifically:

wherein R is the removal rate of harmful substances, eta is the dry weight increase rate, and alpha and beta are regulating factors.

9. A computer readable storage medium, wherein program instructions are stored in the computer readable storage medium, which program instructions, when executed, are adapted to carry out a method for selecting a process parameter for the treatment of black and odorous silty clay according to any one of claims 1-8.

10. A marine black and odorous muddy clay treatment material amount determination system, characterized by comprising the computer readable storage medium of claim 9.