A rapid prediction and precise control method for hydrogen sulfide in urban drainage networks based on artificial intelligence

Through the gradient enhancement decision tree algorithm combined with mechanism model, a prediction and control model of sulfide and hydrogen sulfide is constructed, which solves the problem of predicting and controlling hydrogen sulfide concentration in urban drainage pipelines, and achieves rapid and accurate prediction and control of hydrogen sulfide concentrations. , improves the safety of the drainage pipe network.

CN118444575BActive Publication Date: 2024-11-05SUN YAT SEN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410822063.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-25
Publication Date
2024-11-05
Estimated Expiration
2044-06-25

AI Technical Summary

Technical Problem

The existing technology cannot effectively predict and control the hydrogen sulfide concentration in urban drainage pipelines, resulting in frequent safety accidents and artificial intelligence lacks sufficient data for accurate predictions.

Method used

Using a gradient-enhancing decision tree algorithm, combined with mechanism model and artificial intelligence, the prediction and control model of sulfide and hydrogen sulfide are constructed by obtaining the water quality and water volume data of the drainage pipeline network, and the prediction and control model of sulfide and hydrogen sulfide are achieved quickly predicting and precise control of hydrogen sulfide concentration. .

Benefits of technology

It realizes rapid and accurate prediction and control of hydrogen sulfide concentration in the drainage pipeline network, overcomes the problems of long calculation time of mechanism model and insufficient artificial intelligence data, and improves prediction efficiency and control accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118444575B_ABST
    Figure CN118444575B_ABST
Patent Text Reader

Abstract

The present invention discloses a method, system, medium and equipment for rapid prediction and precise control of hydrogen sulfide in urban drainage pipe network based on artificial intelligence, and the method includes the following steps: obtaining a water quality and quantity sample data set; inputting the data set into a calibrated prediction model of hydrogen sulfide production, and outputting corresponding simulation concentration data under the conditions of adding and not adding drugs; constructing a sulfide and hydrogen sulfide prediction model according to the data set and the simulation concentration data without adding drugs; inputting real-time water quality and quantity data into the sulfide and hydrogen sulfide prediction model, and outputting predicted concentration; based on the sulfide and hydrogen sulfide prediction model, combining the data set, the conditions of adding drugs and the corresponding simulation concentration data, constructing a sulfide and hydrogen sulfide control model; using the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as the input of the sulfide and hydrogen sulfide control model, outputting control parameters and regulating. The present invention can realize rapid prediction and precise control of hydrogen sulfide concentration in drainage pipe network.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of sewage detection and treatment, and specifically, relates to a method for rapid prediction and precise control of hydrogen sulfide in an urban drainage network based on artificial intelligence. Background Art

[0002] Studies have shown that hydrogen sulfide concentrations of 100 ppm can cause damage to the human nervous system, while concentrations of 700 ppm or more can cause a serious risk of sudden death. During the operation and maintenance of urban drainage networks, safety accidents caused by hydrogen sulfide pollution occur frequently and result in the death of many people. Due to the complex production mechanism of hydrogen sulfide in urban drainage networks and the many influencing factors, the concentration of hydrogen sulfide fluctuates rapidly, and can even rise from 100 ppm to 1000 ppm in a very short time. It has the characteristics of acute toxicity. Therefore, conventional methods cannot accurately predict the concentration of hydrogen sulfide in urban drainage networks and control pollution.

[0003] Mechanism models can simulate the migration and transformation process of various substances, as well as the changes in time and space scales, to a certain extent, so as to achieve accurate prediction and control of hydrogen sulfide concentration in drainage pipe networks. For example, the prediction model of hydrogen sulfide generation disclosed in ZL 201610136984.5 can simulate the water quality, water quantity changes and hydrogen sulfide gas generation of the pipe network according to the historical water quality, water quantity and hydrogen sulfide gas conditions of each pipe in the pipe network; however, this operation method makes the mechanism model generally have the problem of long calculation time. When an emergency hydrogen sulfide toxicity event occurs, the mechanism model cannot meet the needs of rapid prediction of the rapidly changing hydrogen sulfide concentration in the pipe network and provide corresponding real-time control solutions, thus limiting its application in pipe network operation and maintenance management.

[0004] Although artificial intelligence can solve nonlinear relationships in complex systems and provide fast data processing and prediction capabilities, it has unique advantages in achieving fast and accurate prediction of hydrogen sulfide concentration in drainage networks. As a data-driven algorithm, artificial intelligence algorithms require a large amount of reliable data for training. However, the concentration of hydrogen sulfide in drainage networks is affected by a variety of factors, and the maintenance cost of network monitoring equipment is high and data collection is difficult, so it is difficult to collect data on a large scale, for a long time, and with multiple indicators. Therefore, artificial intelligence does not have enough relevant data to predict the generation and emission patterns of hydrogen sulfide in drainage networks. However, artificial intelligence currently does not have enough relevant data to predict the generation and emission patterns of hydrogen sulfide in drainage networks. Summary of the invention

[0005] The first purpose of the invention is to overcome the shortcomings and deficiencies in the prior art and to provide an artificial intelligence-based method for rapid prediction and precise control of hydrogen sulfide in urban drainage networks, thereby achieving rapid prediction and precise control of hydrogen sulfide concentration in drainage networks.

[0006] The second purpose of the present invention is to provide an artificial intelligence-based rapid prediction and precise control system for hydrogen sulfide in urban drainage networks.

[0007] The third object of the present invention is to provide a storage medium.

[0008] A fourth object of the present invention is to provide a computing device.

[0009] The purpose of the present invention is achieved through the following technical solution: a method for rapid prediction and precise control of hydrogen sulfide in urban drainage pipe networks based on artificial intelligence, comprising the steps of:

[0010] S1. Obtain a water quality and quantity sample data set of the drainage network;

[0011] S2, using the water quality and quantity sample data set as the input of the calibrated prediction model of hydrogen sulfide production, simulating under the conditions of no drug addition and drug addition, respectively, to obtain the simulated concentration data of sulfide and hydrogen sulfide corresponding to the target prediction position;

[0012] S3, constructing a sulfide and hydrogen sulfide prediction model based on a gradient boosting decision tree algorithm according to the water quality and quantity sample data set and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs;

[0013] S4, using the real-time water quality and water quantity data as the input of the sulfide and hydrogen sulfide prediction model, and outputting the predicted concentration data of sulfide and hydrogen sulfide at the target prediction position;

[0014] S5. Based on the sulfide and hydrogen sulfide prediction model, combined with the water quality and quantity sample data set, dosing conditions and simulated concentration data of sulfide and hydrogen sulfide during dosing, a sulfide and hydrogen sulfide control model is constructed;

[0015] S6. Using the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as inputs to the sulfide and hydrogen sulfide control model, outputting control parameters and performing regulation according to the control parameters.

[0016] Preferably, step S1 specifically includes: acquiring historical water quality and water quantity data of the drainage network, randomly changing the historical water quality and water quantity data within a set range, and generating a water quality and water quantity sample data set of the drainage network.

[0017] Preferably, step S2 specifically includes:

[0018] S21. Obtain pipeline parameters, historical water quality and quantity data, and historical sulfide and hydrogen sulfide concentration data of the drainage network, and calibrate the prediction model for hydrogen sulfide generation;

[0019] S22, using the water quality and quantity sample data set as the input of the prediction model of hydrogen sulfide production, and outputting the simulated concentration data of sulfide and hydrogen sulfide at the target prediction position when no drug is added;

[0020] S23, using the dosing conditions and the water quality and quantity sample data set as inputs of a prediction model for hydrogen sulfide production, and outputting simulated concentration data of sulfide and hydrogen sulfide when dosing at a target prediction position, wherein the dosing conditions include dosing amount and dosing position.

[0021] Preferably, step S3 specifically includes:

[0022] S31. Based on the gradient boosting decision tree algorithm, an initialized sulfide and hydrogen sulfide prediction model is constructed;

[0023] S32, taking the water quality and quantity sample data set as input, and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs as output, and dividing them into a training set and a validation set in proportion;

[0024] S33, inputting the training set into the initialized sulfide and hydrogen sulfide prediction model for training, and then using the verification set to further optimize the trained sulfide and hydrogen sulfide prediction model, so that the error between the prediction result of the sulfide and hydrogen sulfide prediction model and the verification data reaches the expected error range.

[0025] Preferably, step S5 specifically includes:

[0026] S51, taking the water quality and quantity sample data set and the dosing conditions as input, and the simulated concentration data of sulfide and hydrogen sulfide during dosing as output, and re-dividing the training set and the validation set according to the proportion, retraining the sulfide and hydrogen sulfide prediction model through the training set, and optimizing the trained sulfide and hydrogen sulfide prediction model through the validation set, so that the error between the sulfide and hydrogen sulfide concentrations in the drainage network predicted by the sulfide and hydrogen sulfide prediction model and the validation data reaches the expected error range;

[0027] S52. According to the retrained sulfide and hydrogen sulfide prediction model, the relationship between the dosing conditions and the simulated concentration data of sulfide and hydrogen sulfide during dosing is quickly determined to obtain formula (1) and formula (2):

[0028] DS α、β = model DS ( dos, place α , condition β ), formula (1),

[0029] H 2 S α、β = model H2S ( dos, place α , condition β ), formula (2),

[0030] in, model DS and model H2S They respectively represent the sulfide prediction relationship and hydrogen sulfide prediction relationship obtained according to the retrained sulfide and hydrogen sulfide prediction models, dos Indicates the dosage of the drug. place α Indicates α Dosing location, condition β Indicates β Water quality and quantity data, namely β Water quality and quantity conditions, DS α、β and H 2 S α、β Respectively expressed in α Dosing location and β Under the condition of water quality and quantity and the dosage is dos When , the predicted concentration of sulfide and hydrogen sulfide at the target prediction location;

[0031] S53. Establish a planning model with the minimum dosage as the objective function, the sulfide concentration and hydrogen sulfide concentration at the target predicted position after the dosage reach the corresponding preset standard concentration, and the optimal dosage sulfur control effect as the constraint condition, and obtain formula (3):

[0032] , formula (3),

[0033] in, a and b They represent the preset target concentration of sulfide and hydrogen sulfide at the target prediction location, min ( H 2 S β ) and min( DS β ) respectively represent that when the water quality, water quantity and dosage are the same and in the βWhen there are 3 dosing positions, the predicted concentrations of sulfide and hydrogen sulfide at the target predicted position are the smallest, that is, the dosing sulfur control effect is optimal;

[0034] S54. Combine formula (1), formula (2) and formula (3) to obtain the sulfide and hydrogen sulfide control model.

[0035] Preferably, step S6 specifically includes:

[0036] S61, determining the dosing conditions required to reduce the predicted concentration to a preset standard concentration through a sulfide and hydrogen sulfide control model, selecting the dosing conditions that meet the preset standards as a dosing strategy, and converting the dosing strategy into control parameters that can be recognized by a functional module in a dosing device;

[0037] S62. Intelligently and accurately control the functional modules of the current drainage network dosing equipment according to the control parameters.

[0038] Preferably, the water quality data of the water quality and quantity sample data set and the real-time water quality and quantity data both include pH, water temperature, alkalinity, dissolved oxygen, volatile fatty acids, sulfate, alkalinity, ammonia nitrogen, soluble sulfide, suspended solids and chemical oxygen demand.

[0039] An artificial intelligence-based rapid prediction and precise control system for hydrogen sulfide in urban drainage networks, including:

[0040] A first simulation data generation module is used to obtain a water quality and quantity sample data set of a drainage network;

[0041] The second simulation data generation module is used to use the water quality and quantity sample data set as the input of the prediction model of calibrated hydrogen sulfide production, simulate under the conditions of no drug addition and drug addition, and obtain the simulation concentration data of sulfide and hydrogen sulfide corresponding to the target prediction position;

[0042] A prediction model building module, used to build a sulfide and hydrogen sulfide prediction model based on a gradient boosting decision tree algorithm according to the water quality and quantity sample data set and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs;

[0043] A prediction module, used to use real-time water quality and water quantity data as input of the sulfide and hydrogen sulfide prediction model, and output the predicted concentration data of sulfide and hydrogen sulfide at the target prediction location;

[0044] A control model building module, for building a sulfide and hydrogen sulfide control model based on the sulfide and hydrogen sulfide prediction model, in combination with the water quality and quantity sample data set, dosing conditions and simulated concentration data of sulfide and hydrogen sulfide during dosing;

[0045] The control module is used to use the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as the input of the sulfide and hydrogen sulfide control model, output the control parameters and perform regulation according to the control parameters.

[0046] A storage medium stores a program, which, when executed by a processor, implements the above-mentioned method for rapid prediction and precise control of hydrogen sulfide in an urban drainage network based on artificial intelligence.

[0047] A computing device includes a processor and a memory for storing programs executable by the processor. When the processor executes the program stored in the memory, the above-mentioned artificial intelligence-based method for rapid prediction and precise control of hydrogen sulfide in urban drainage networks is implemented.

[0048] Compared with the prior art, the present invention has the following advantages and effects:

[0049] (1) The present invention provides a method for rapid prediction and precise control of hydrogen sulfide in an urban drainage network based on artificial intelligence. The method mainly adopts a method of combining a mechanism model with artificial intelligence (gradient boosting decision tree algorithm). The method uses a calibrated prediction model of hydrogen sulfide production to generate a large amount of reliable simulation data for learning and training artificial intelligence models (including sulfide and hydrogen sulfide prediction models and control models), thereby resolving the conflict that artificial intelligence requires a large amount of training data but drainage networks are difficult to collect data on a large scale, for a long time, and with multiple indicators; a sulfide and hydrogen sulfide prediction model is constructed using a gradient boosting decision tree algorithm to achieve rapid prediction of hydrogen sulfide concentration in drainage networks, thereby overcoming the problem of long calculation time commonly found in existing mechanism models; the sulfide and hydrogen sulfide prediction model is then retrained and constraints are added to establish a sulfide and hydrogen sulfide control model, thereby quickly obtaining control parameters to achieve real-time and precise control of sulfide and hydrogen sulfide concentrations, thereby overcoming the defect that the mechanism model is difficult to accurately control the real-time changing sulfide and hydrogen sulfide concentrations in real time due to the lag in calculation.

[0050] (2) The present invention predicts concentration through the sulfide and hydrogen sulfide prediction model. The fitting degree between the prediction result and the simulation result of the prediction model of hydrogen sulfide production is 0.99, and the fitting degree between the prediction result and the actual measurement result of hydrogen sulfide concentration change in the drainage network is 0.72. The change of hydrogen sulfide concentration in the network can be predicted more accurately. On this basis, the prediction time required by the present invention is only 0.14 seconds. Compared with the 70 minutes required for the simulation of the prediction model of hydrogen sulfide production in the prior art, the calculation efficiency is improved by nearly 12,000 times, which greatly improves the prediction efficiency and realizes real-time rapid prediction.

[0051] (3) The present invention controls the concentration through a sulfide and hydrogen sulfide control model, with 20 ppm as the preset target concentration of hydrogen sulfide at the target prediction location. The model can calculate the dosing conditions for reducing the hydrogen sulfide concentration corresponding to each set of water quality and quantity data to the preset target concentration within 1 second and screen out the dosing strategy that meets the requirements, and control the concentration of hydrogen sulfide at the target prediction location in real time and accurately below 20 ppm. The method for rapid prediction and precise control of hydrogen sulfide in urban drainage networks based on artificial intelligence of the present invention can achieve efficient and precise control of hydrogen sulfide.

[0052] (4) The present invention provides an artificial intelligence-based system for rapid prediction and precise control of hydrogen sulfide concentration in urban drainage networks. By combining the mechanism model with the artificial intelligence model, the present invention realizes rapid prediction and precise control of hydrogen sulfide concentration in urban drainage networks, providing strong technical support for sewage treatment in drainage networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 The present invention is a flow chart of a method for rapid prediction and precise control of hydrogen sulfide in an urban drainage network based on artificial intelligence.

[0054] Figure 2 This is a diagram showing the effect of hydrogen sulfide gas concentration prediction using an artificial intelligence-based method for rapid prediction and precise control of hydrogen sulfide in urban drainage networks in an embodiment of the present invention.

[0055] Figure 3 This is a diagram showing the effect of controlling the concentration of hydrogen sulfide gas using an artificial intelligence-based method for rapidly predicting and accurately controlling hydrogen sulfide in urban drainage networks in an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The present invention will be further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0057] Example 1

[0058] like Figure 1 The figure shows a flow chart of a method for rapid prediction and precise control of hydrogen sulfide in a city drainage network based on artificial intelligence according to the present invention, which comprises the following steps:

[0059] S1. Obtain a water quality and quantity sample data set of the drainage network;

[0060] S2, using the water quality and quantity sample data set as the input of the calibrated prediction model of hydrogen sulfide production, simulating under the conditions of no drug addition and drug addition, respectively, to obtain the simulated concentration data of sulfide and hydrogen sulfide corresponding to the target prediction position;

[0061] S3, constructing a sulfide and hydrogen sulfide prediction model based on a gradient boosting decision tree algorithm according to the water quality and quantity sample data set and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs;

[0062] S4, using the real-time water quality and water quantity data as the input of the sulfide and hydrogen sulfide prediction model, and outputting the predicted concentration data of sulfide and hydrogen sulfide at the target prediction position;

[0063] S5. Based on the sulfide and hydrogen sulfide prediction model, combined with the water quality and quantity sample data set, dosing conditions and simulated concentration data of sulfide and hydrogen sulfide during dosing, a sulfide and hydrogen sulfide control model is constructed;

[0064] S6. Using the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as inputs to the sulfide and hydrogen sulfide control model, outputting control parameters and performing regulation according to the control parameters.

[0065] Specifically, the present invention provides a method for rapid prediction and precise control of hydrogen sulfide in urban drainage pipe networks based on artificial intelligence. The method mainly combines a mathematical mechanism model with artificial intelligence (gradient boosting decision tree algorithm), and compared with the prior art, it has achieved significant technical effects: it can realize rapid prediction and precise control of hydrogen sulfide concentration in drainage pipe networks. The method uses a calibrated prediction model of hydrogen sulfide production to generate a large amount of reliable simulation data for learning and training artificial intelligence models (including sulfide and hydrogen sulfide prediction models and control models), solving the conflict between artificial intelligence requiring a large amount of training data and the difficulty of drainage pipe networks to collect data on a large scale, for a long time, and with multiple indicators. Through the trained sulfide and hydrogen sulfide prediction model, the concentration of hydrogen sulfide gas in the drainage pipe network is quickly predicted, overcoming the problem of long calculation time commonly existing in the existing mechanism model; then, a sulfide and hydrogen sulfide control model is established based on the sulfide and hydrogen sulfide prediction model, and the control parameters are output to realize real-time and precise control of the concentration of sulfide and hydrogen sulfide, breaking the limitation that the mechanism model is limited to the hysteresis of calculation and it is difficult to perform real-time and precise control of the concentration of sulfide and hydrogen sulfide that changes in real time.

[0066] In this embodiment, the prediction model of hydrogen sulfide production used is an existing mathematical mechanism model, which is used to predict the concentration changes of sulfide and hydrogen sulfide in the drainage network according to the water quality changes of the drainage network, such as the prediction model of hydrogen sulfide production in Chinese invention patent ZL201610136984.5, which consists of a water quality module, a dynamic biofilm module and a liquid-gas diffusion module. The development of this model adopts a segmented plug flow simulation process, which divides the entire pipeline into n sections, and each section is regarded as a biofilm reactor. In each sewage section, the heterogeneous biofilm process and its effect on sulfur and nitrogen conversion are simulated, so as to achieve the prediction of sulfide concentration changes in sewage, and the hydrogen sulfide gas concentration is calculated by gas release model simulation. According to the setting of plug flow segmented simulation, the effluent water quality of the previous section can be simulated as the inlet water quality of the next section, and the simulated water quality and water quantity data are input into the model, and finally the simulated sulfide data, simulated hydrogen sulfide gas data and simulated dosing data of the target prediction position are simulated. Other similar mechanism models that predict changes in the concentrations of sulfide and hydrogen sulfide in the drainage network based on changes in the water quality of the drainage network may also be used.

[0067] Specifically, by inputting real-time water quality and quantity data into the sulfide and hydrogen sulfide prediction model, the real-time predicted concentration of sulfide and hydrogen sulfide in the current drainage network can be quickly obtained according to the changes in various water quality and quantity data. For example, when the concentration of sulfate in the drainage network increases, the concentration of sulfide and hydrogen sulfide at the target prediction position in the drainage network will increase simultaneously; when the flow rate in the drainage network increases, the concentration of sulfide and hydrogen sulfide will decrease. In this embodiment, the preset standard concentration meets the concentration range of relevant laws and regulations, and the specific standard concentration is set according to the actual conditions of different cities and different drainage networks.

[0068] The water quality data of the water quality and quantity sample data set and the real-time water quality and quantity data include pH, water temperature, alkalinity, dissolved oxygen, volatile fatty acids, sulfate, alkalinity, ammonia nitrogen, dissolved sulfide, suspended solids and chemical oxygen demand.

[0069] Step S1 specifically includes: obtaining historical water quality and water quantity data of the drainage network, randomly changing the historical water quality and water quantity data within a set range, and generating a water quality and water quantity sample data set of the drainage network.

[0070] Specifically, in this embodiment, within a certain range, through the random function randomGenerate a number of random numbers, which are used as random fluctuation factors to multiply the historical water quality and water quantity data of the current drainage network, thereby generating a variety of simulated water quality and water quantity data. Specifically, first use a random function to generate a set of random number sequences that obey a specified distribution. Then, multiply these random numbers by the historical water quality and water quantity data of the current drainage network one by one to obtain a new set of data. For example, the water volume changes to 0.7 times the historical water volume, and the sulfate concentration changes to 1.2 times the historical sulfate quality concentration. This method effectively simulates the changes that may occur in the historical water quality and water quantity data of the current drainage network under different conditions, and generates a large amount of simulated water quality and water quantity data to form a water quality and water quantity sample data set.

[0071] Step S2 specifically includes:

[0072] S21, obtaining pipeline parameters, historical water quality and quantity data, and historical sulfide and hydrogen sulfide concentration data of the drainage network, and calibrating a prediction model for hydrogen sulfide production, wherein the prediction model for hydrogen sulfide production is used to predict changes in sulfide and hydrogen sulfide concentrations in the drainage network according to changes in water quality of the drainage network;

[0073] S22, using the water quality and quantity sample data set as the input of the prediction model of hydrogen sulfide production, and outputting the simulated concentration data of sulfide and hydrogen sulfide at the target prediction position when no drug is added;

[0074] S23, using the dosing conditions and the water quality and quantity sample data set as inputs of a prediction model for hydrogen sulfide production, and outputting simulated concentration data of sulfide and hydrogen sulfide when dosing at a target prediction position, wherein the dosing conditions include dosing amount and dosing position.

[0075] Specifically, when the mechanism model used is the prediction model of hydrogen sulfide production in Chinese invention patent ZL201610136984.5, the pipe parameters of the drainage network include pipe diameter, pipe length, design slope, roughness coefficient and total designed sewage volume. If other mechanism models that can predict the concentration changes of sulfides and hydrogen sulfide in the drainage network according to the water quality changes of the drainage network are used, the pipe parameters are selected according to the corresponding model requirements.

[0076] Step S3 specifically includes:

[0077] S31. Based on the gradient boosting decision tree algorithm, an initialized sulfide and hydrogen sulfide prediction model is constructed;

[0078] S32, taking the water quality and quantity sample data set as input, and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs as output, and dividing them into a training set and a validation set in proportion;

[0079] S33, input the training set into the initialized sulfide and hydrogen sulfide prediction model for training, and then use the verification set to further optimize the trained sulfide and hydrogen sulfide prediction model, so that the error between the prediction result of the sulfide and hydrogen sulfide prediction model and the verification data reaches the expected error range. Specifically, the present invention uses the gradient boosting decision tree machine learning algorithm to construct a prediction model. The algorithm is an iterative decision tree algorithm, which generates several weak learners through multiple rounds of iterations, and continuously fits the residuals of the previous round of weak classifiers through the negative gradient of the loss function during the iteration process, so that the residual estimate of each round is gradually reduced and finally tends to 0; finally, the results of all learners are added up to obtain the final prediction result. Among them, the weak learner selected by the present invention is a classification regression tree (cart tree).

[0080] The specific algorithm principle is as follows:

[0081] Algorithm input: training set samples [( x 1 , y 1 ), x 2 , y 2 ),…,( x m , y m )],in x i Indicates input, y i Indicates the output, i represents the sample size, and i = 1, 2, 3, …, m ;

[0082] Algorithm output: classification and regression tree f ( x ), that is, the predicted concentrations of sulfide and hydrogen sulfide at the target prediction location;

[0083] Loss function: L ( y , f ( x )),in, y represents the simulated concentration of sulfide and hydrogen sulfide at the target prediction location, f ( x ) represents the predicted concentration of sulfide and hydrogen sulfide at the target prediction location;

[0084] Initial classification and regression tree:

[0085] , formula (4),

[0086] in, c for y i The average value of

[0087] For each iteration (i.e. each classification and regression tree) t (t = 1, 2, 3, …, T):

[0088] For each sample i , calculate the residuals of the simulated concentrations of sulfide and hydrogen sulfide at the target prediction location and the predicted concentrations of sulfide and hydrogen sulfide:

[0089] , formula (5),

[0090] use( x i ,r ti ) fits the classification and regression tree to obtain the tth classification and regression tree, and its corresponding leaf node area is R tj ,in j is the number of leaf nodes, and the best fitting value is calculated for the leaf node area:

[0091] , formula (6),

[0092] Update the classification and regression tree:

[0093] , formula (7),

[0094] in, j =1, 2, 3, ..., J , I represents the indicator function;

[0095] The final classification and regression tree expression is:

[0096] , formula (8),

[0097] In this embodiment, the parameter values ​​used in the quick prediction model are shown in Table 1.

[0098] Table 1 Hyperparameter types and selection ranges

[0099]

[0100] Among them, the learning rate refers to the contribution of each tree to the final classification and regression tree, the maximum depth of the tree refers to the longest path length from the root node of the tree to the deepest leaf node, and the maximum number of features of the tree refers to the maximum number of features that the algorithm can select each time a node splits.

[0101] Specifically, the selection and proportion of the training set and the validation set are selected according to the actual situation. In this embodiment, for example, 10,000 groups of simulated concentration data of sulfide and hydrogen sulfide without adding drugs can be generated according to 10,000 groups of water quality and water quantity sample data, and 8,000 groups of them are extracted as training data at any time according to the classification method of 4:1, and the remaining 2,000 groups are used as validation data. The training data is input into the initialized sulfide and hydrogen sulfide prediction model for training and learning, and the parameter setting of the prediction model is adjusted based on the training results to improve the prediction accuracy and prediction efficiency of the prediction model; the validation data is then input into the trained sulfide and hydrogen sulfide prediction model, and the generalization of the model is obtained according to the prediction performance of the prediction model for the validation data, and the prediction model is further optimized according to the performance, and the prediction model with the best optimization result is selected to obtain the final sulfide and hydrogen sulfide prediction model. Through the above training and verification process, the sulfide and hydrogen sulfide prediction model autonomously learns the intrinsic relationship in the water quality quantity data in the current drainage network, and realizes the rapid prediction of the future changes in sulfide and hydrogen sulfide in the drainage network.

[0102] like Figure 2 The figure shows the effect of hydrogen sulfide gas concentration prediction using an artificial intelligence-based urban drainage network hydrogen sulfide rapid prediction and precise control method in the embodiment of the present invention. The fitting degree R between the prediction results of the sulfide and hydrogen sulfide prediction model of the present invention and the simulation results of the existing prediction model for hydrogen sulfide generation 2 The fitting degree R is 0.99, which is consistent with the measured results of the change of hydrogen sulfide concentration in the drainage network. 2 is 0.72, which shows that the method of the present invention can more accurately predict the change of hydrogen sulfide gas concentration in the drainage network. In terms of prediction time, the sulfide and hydrogen sulfide prediction model of the present invention only takes 0.14 seconds to predict in this specific embodiment, which is nearly 12,000 times more efficient than the 70 minutes required for the existing prediction model of hydrogen sulfide production, greatly improving the prediction efficiency.

[0103] Step S5 specifically includes:

[0104] S51, taking the water quality and quantity sample data set and the dosing conditions as input, and the simulated concentration data of sulfide and hydrogen sulfide during dosing as output, and re-dividing the training set and the validation set according to the proportion, retraining the sulfide and hydrogen sulfide prediction model through the training set, and optimizing the trained sulfide and hydrogen sulfide prediction model through the validation set, so that the error between the sulfide and hydrogen sulfide concentrations in the drainage network predicted by the sulfide and hydrogen sulfide prediction model and the validation data reaches the expected error range;

[0105] S52. According to the retrained sulfide and hydrogen sulfide prediction model, the relationship between the dosing conditions and the simulated concentration data of sulfide and hydrogen sulfide during dosing is quickly determined to obtain formula (1) and formula (2):

[0106] DS α、β = model DS ( dos, place α , condition β ), formula (1),

[0107] H 2 S α、β = model H2S ( dos, place α , condition β ), formula (2),

[0108] in, model DS and model H2S They respectively represent the sulfide prediction relationship and hydrogen sulfide prediction relationship obtained according to the retrained sulfide and hydrogen sulfide prediction models, dos Indicates the dosage of the drug. place α Indicates α Dosing location, condition β Indicates β Water quality and quantity data, namely β Water quality and quantity conditions, DS α、β and H 2 S α、β Respectively expressed in α Dosing location and β Under the condition of water quality and quantity and the dosage is dos When , the predicted concentration of sulfide and hydrogen sulfide at the target prediction location;

[0109] S53. Establish a planning model with the minimum dosage as the objective function, the sulfide concentration and hydrogen sulfide concentration at the target predicted position after the dosage reach the corresponding preset standard concentration, and the optimal dosage sulfur control effect as the constraint condition, and obtain formula (3):

[0110] , formula (3),

[0111] in,a and b They represent the preset target concentration of sulfide and hydrogen sulfide at the target prediction location, min ( H 2 S β ) and min( DS β ) respectively represent that when the water quality, water quantity and dosage are the same and in the β When there are 3 dosing positions, the predicted concentrations of sulfide and hydrogen sulfide at the target predicted position are the smallest, that is, the dosing sulfur control effect is optimal;

[0112] S54. Combine formula (1), formula (2) and formula (3) to obtain the sulfide and hydrogen sulfide control model.

[0113] Specifically, in this embodiment, the dosing conditions are input into the sulfide and hydrogen sulfide prediction model, so that the model can autonomously learn the variable of the dosing conditions. Through model training and parameter tuning, the error between the prediction results of the sulfide and hydrogen sulfide prediction model and the verification data reaches the expected error range, and the change trend of the sulfide and hydrogen sulfide gas concentrations can be accurately predicted according to the water quality and quantity sample data set and the dosing conditions. For example, after adding the dosing amount, the sulfide and hydrogen sulfide concentrations at the target prediction position of the drainage network will decrease to varying degrees; when agents are added at different positions, the degree of reduction in the sulfide and hydrogen sulfide concentrations at the target prediction position will be different.

[0114] According to the trained sulfide and hydrogen sulfide prediction model, the internal relationship between different dosing conditions and sulfide and hydrogen sulfide concentrations can be quickly obtained. In this embodiment, formula (1) and formula (2) are used to express their internal relationship. Then min (x) is used as the objective function and constraints are added: including that the sulfide and hydrogen sulfide concentrations must meet the standards, and the sulfur control effect is the best among all dosing positions (that is, the sulfide and hydrogen sulfide concentrations are the lowest after dosing). That is, a planning model is established according to formula (3) to obtain the minimum dosing amount that meets the constraints.

[0115] The method for rapid prediction and precise control of hydrogen sulfide in urban drainage pipe networks based on artificial intelligence of the present invention is used to control the concentration of hydrogen sulfide gas. The control effect diagram is shown in FIG. Figure 3As shown. In this embodiment, 20ppm is used as the preset standard concentration of hydrogen sulfide gas at the target prediction position. The sulfide and hydrogen sulfide control model of the present invention can calculate the dosing conditions for reducing the hydrogen sulfide gas concentration corresponding to each set of input water quality and quantity data to the preset standard concentration within 1s, and screen out the dosing strategy that meets the requirements, and accurately control the concentration of hydrogen sulfide gas at the target prediction position in real time to below 20ppm. It can be seen that the method for rapid prediction and precise control of hydrogen sulfide in urban drainage pipe networks based on artificial intelligence of the present invention realizes rapid and precise control of hydrogen sulfide gas concentration.

[0116] Step S6 specifically includes:

[0117] S61, determining the dosing conditions required to reduce the predicted concentration to a preset standard concentration at the target predicted position through a sulfide and hydrogen sulfide control model, selecting the dosing conditions that meet the preset standards as a dosing strategy, and converting the dosing strategy into control parameters that can be recognized by a functional module in the dosing device;

[0118] S62. Intelligently and accurately control the functional modules of the current drainage network dosing equipment according to the control parameters.

[0119] Specifically, in this embodiment, the predicted concentration data of sulfide and hydrogen sulfide are input, and the sulfide and hydrogen sulfide control model is used to calculate that a variety of dosing conditions can make the concentrations of sulfide and hydrogen sulfide gases at the target predicted position meet the standards. For example, if the preset standard is the lowest cost, the dosing condition that meets the lowest cost is selected as the dosing strategy, and the dosing amount and dosing position that can minimize the cost are obtained based on the dosing strategy, and the control parameters are output as the basis for regulation, so as to adjust the dosing amount and dosing position of the dosing equipment in the drainage network in real time.

[0120] After completing the regulation of sulfide and hydrogen sulfide gases in the drainage network, the sulfide and hydrogen sulfide prediction model and control model will save the prediction data and control parameters. Among them, the sulfide and hydrogen sulfide control model can be verified and corrected according to the actual sulfide and hydrogen sulfide gas concentration data recorded by the current drainage network at the preset time point.

[0121] Example 2

[0122] An artificial intelligence-based rapid prediction and precise control system for hydrogen sulfide in urban drainage networks, including:

[0123] A first simulation data generation module is used to obtain a water quality and quantity sample data set of a drainage network;

[0124] The second simulation data generation module is used to use the water quality and quantity sample data set as the input of the prediction model of calibrated hydrogen sulfide production, simulate under the conditions of no drug addition and drug addition, and obtain the simulation concentration data of sulfide and hydrogen sulfide corresponding to the target prediction position;

[0125] A prediction model building module, used to build a sulfide and hydrogen sulfide prediction model based on a gradient boosting decision tree algorithm according to the water quality and quantity sample data set and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs;

[0126] A prediction module, used to use real-time water quality and water quantity data as input of the sulfide and hydrogen sulfide prediction model, and output the predicted concentration data of sulfide and hydrogen sulfide at the target prediction location;

[0127] A control model building module, for building a sulfide and hydrogen sulfide control model based on the sulfide and hydrogen sulfide prediction model, in combination with the water quality and quantity sample data set, dosing conditions and simulated concentration data of sulfide and hydrogen sulfide during dosing;

[0128] The control module is used to use the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as the input of the sulfide and hydrogen sulfide control model, output the control parameters and perform regulation according to the control parameters.

[0129] Specifically, the present invention provides an artificial intelligence-based rapid prediction and precise control system for hydrogen sulfide in urban drainage networks. By combining a mechanism model with an artificial intelligence model, it achieves rapid prediction and precise control of hydrogen sulfide concentration in urban drainage networks, providing strong technical support for sewage treatment in drainage networks.

[0130] Example 3

[0131] A storage medium stores a program, and when the program is executed by a processor, the method for rapid prediction and precise control of hydrogen sulfide in an urban drainage network based on artificial intelligence described in Example 1 is implemented as follows:

[0132] S1. Obtain a water quality and quantity sample data set of the drainage network;

[0133] S2, using the water quality and quantity sample data set as the input of the calibrated prediction model of hydrogen sulfide production, simulating under the conditions of no drug addition and drug addition, respectively, to obtain the simulated concentration data of sulfide and hydrogen sulfide corresponding to the target prediction position;

[0134] S3, constructing a sulfide and hydrogen sulfide prediction model based on a gradient boosting decision tree algorithm according to the water quality and quantity sample data set and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs;

[0135] S4, using the real-time water quality and water quantity data as the input of the sulfide and hydrogen sulfide prediction model, and outputting the predicted concentration data of sulfide and hydrogen sulfide at the target prediction position;

[0136] S5. Based on the sulfide and hydrogen sulfide prediction model, combined with the water quality and quantity sample data set, dosing conditions and simulated concentration data of sulfide and hydrogen sulfide during dosing, a sulfide and hydrogen sulfide control model is constructed;

[0137] S6, using the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as inputs of the sulfide and hydrogen sulfide control model, outputting control parameters and regulating according to the control parameters. In the above process, the specific processing process is as described in Example 1, which will not be repeated here.

[0138] In this embodiment, the storage medium may be a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, a USB flash drive, a mobile hard disk, or the like.

[0139] Example 4

[0140] A computing device includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the method for rapid prediction and precise control of hydrogen sulfide in an urban drainage network based on artificial intelligence described in Example 1 is implemented as follows:

[0141] S1. Obtain a water quality and quantity sample data set of the drainage network;

[0142] S2, using the water quality and quantity sample data set as the input of the calibrated prediction model of hydrogen sulfide production, simulating under the conditions of no drug addition and drug addition, respectively, to obtain the simulated concentration data of sulfide and hydrogen sulfide corresponding to the target prediction position;

[0143] S3, constructing a sulfide and hydrogen sulfide prediction model based on a gradient boosting decision tree algorithm according to the water quality and quantity sample data set and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs;

[0144] S4, using the real-time water quality and water quantity data as the input of the sulfide and hydrogen sulfide prediction model, and outputting the predicted concentration data of sulfide and hydrogen sulfide at the target prediction position;

[0145] S5. Based on the sulfide and hydrogen sulfide prediction model, combined with the water quality and quantity sample data set, dosing conditions and simulated concentration data of sulfide and hydrogen sulfide during dosing, a sulfide and hydrogen sulfide control model is constructed;

[0146] S6. Using the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as inputs to the sulfide and hydrogen sulfide control model, outputting control parameters and performing regulation according to the control parameters.

[0147] In the above process, the specific processing process is as described in Example 1 and will not be repeated here.

[0148] In this embodiment, the computing device may be a terminal device such as a desktop computer, a laptop computer, a PDA handheld terminal, a tablet computer, etc.

[0149] The above embodiments are preferred implementations of the present invention and are not intended to limit the present invention. Any other changes or other equivalent replacement methods that do not deviate from the technical solutions of the present invention are included in the protection scope of the present invention.

Claims

1. A rapid prediction and precise control method for hydrogen sulfide in urban drainage networks based on artificial intelligence, It is characterized in that Includes steps: S1. Obtain a water quality and quantity sample data set of the drainage network; S2, using the water quality and quantity sample data set as the input of the calibrated prediction model of hydrogen sulfide production, simulating under the conditions of no drug addition and drug addition, respectively, to obtain the simulated concentration data of sulfide and hydrogen sulfide corresponding to the target prediction position; S3, constructing a sulfide and hydrogen sulfide prediction model based on a gradient boosting decision tree algorithm according to the water quality and quantity sample data set and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs; S4, using the real-time water quality and water quantity data as the input of the sulfide and hydrogen sulfide prediction model, and outputting the predicted concentration data of sulfide and hydrogen sulfide at the target prediction position; S5. Based on the sulfide and hydrogen sulfide prediction model, combined with the water quality and quantity sample data set, dosing conditions and simulated concentration data of sulfide and hydrogen sulfide during dosing, a sulfide and hydrogen sulfide control model is constructed; S6. Using the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as inputs to the sulfide and hydrogen sulfide control model, outputting control parameters and performing regulation according to the control parameters; Step S5 specifically includes: S51, taking the water quality and quantity sample data set and the dosing conditions as input, and the simulated concentration data of sulfide and hydrogen sulfide during dosing as output, and re-dividing the training set and the validation set according to the proportion, retraining the sulfide and hydrogen sulfide prediction model through the training set, and optimizing the trained sulfide and hydrogen sulfide prediction model through the validation set, so that the error between the sulfide and hydrogen sulfide concentrations in the drainage network predicted by the sulfide and hydrogen sulfide prediction model and the validation data reaches the expected error range; S52. According to the retrained sulfide and hydrogen sulfide prediction model, the relationship between the dosing conditions and the simulated concentration data of sulfide and hydrogen sulfide during dosing is quickly determined to obtain formula (1) and formula (2): DS α、β = model DS ( dos, place α , condition β ), Formula (1), H 2 S α、β = model H2S ( dos, place α , condition β ), Formula (2), in, model DS and model H2S They respectively represent the sulfide prediction relationship and hydrogen sulfide prediction relationship obtained according to the retrained sulfide and hydrogen sulfide prediction models, dos Indicates the dosage of the drug. place α Indicates α Dosing location, condition β Indicates β Water quality and quantity data, namely β Water quality and quantity conditions, DS α、β and H 2 S α、β Respectively expressed in α Dosing location and β Under the condition of water quality and quantity and the dosage is dos When , the predicted concentration of sulfide and hydrogen sulfide at the target prediction location; S53. Establish a planning model with the minimum dosage as the objective function, the sulfide concentration and hydrogen sulfide concentration at the target predicted position after the dosage reach the corresponding preset standard concentration, and the optimal dosage sulfur control effect as the constraint condition, and obtain formula (3): , Formula (3), in, a and b They represent the preset standard concentration of sulfide and hydrogen sulfide at the target prediction location, respectively, min ( H 2 S β ) and min( DS β ) respectively represent that when the water quality, water quantity and dosage are the same and in the β When there are 3 dosing positions, the predicted concentrations of sulfide and hydrogen sulfide at the target predicted position are the smallest, that is, the dosing sulfur control effect is optimal; S54. Combine formula (1), formula (2) and formula (3) to obtain the sulfide and hydrogen sulfide control model.

2. According to claim 1, a method for rapid prediction and precise control of hydrogen sulfide in urban drainage pipe networks based on artificial intelligence, It is characterized in that Step S1 specifically includes: obtaining historical water quality and water quantity data of the drainage network, randomly changing the historical water quality and water quantity data within a set range, and generating a water quality and water quantity sample data set of the drainage network.

3. According to claim 1, a method for rapid prediction and precise control of hydrogen sulfide in urban drainage pipe networks based on artificial intelligence, It is characterized in that Step S2 specifically includes: S21. Obtain pipeline parameters, historical water quality and quantity data, and historical sulfide and hydrogen sulfide concentration data of the drainage network, and calibrate the prediction model for hydrogen sulfide generation; S22, using the water quality and quantity sample data set as the input of the prediction model of hydrogen sulfide production, and outputting the simulated concentration data of sulfide and hydrogen sulfide at the target prediction position when no drug is added; S23, using the dosing conditions and the water quality and quantity sample data set as inputs of a prediction model for hydrogen sulfide production, and outputting simulated concentration data of sulfide and hydrogen sulfide when dosing at a target prediction position, wherein the dosing conditions include dosing amount and dosing position.

4. According to claim 1, a method for rapid prediction and precise control of hydrogen sulfide in urban drainage network based on artificial intelligence, It is characterized in that Step S3 specifically includes: S31. Based on the gradient boosting decision tree algorithm, an initialized sulfide and hydrogen sulfide prediction model is constructed; S32, taking the water quality and quantity sample data set as input, and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs as output, and dividing them into a training set and a validation set in proportion; S33, inputting the training set into the initialized sulfide and hydrogen sulfide prediction model for training, and then using the verification set to further optimize the trained sulfide and hydrogen sulfide prediction model, so that the error between the prediction result of the sulfide and hydrogen sulfide prediction model and the verification data reaches the expected error range.

5. According to claim 1, a method for rapid prediction and precise control of hydrogen sulfide in urban drainage network based on artificial intelligence, It is characterized in that Step S6 specifically includes: S61, determining the dosing conditions required to reduce the predicted concentration to a preset standard concentration through a sulfide and hydrogen sulfide control model, selecting the dosing conditions that meet the preset standards as a dosing strategy, and converting the dosing strategy into control parameters that can be recognized by a functional module in a dosing device; S62. Intelligently and accurately control the functional modules of the current drainage network dosing equipment according to the control parameters.

6. According to claim 1, a method for rapid prediction and precise control of hydrogen sulfide in urban drainage network based on artificial intelligence, It is characterized in that The water quality data of the water quality and quantity sample data set and the real-time water quality and quantity data include pH, water temperature, alkalinity, dissolved oxygen, volatile fatty acids, sulfate, alkalinity, ammonia nitrogen, dissolved sulfide, suspended solids and chemical oxygen demand.

7. A rapid prediction and precise control system for hydrogen sulfide in urban drainage networks based on artificial intelligence. It is characterized in that include: A first simulation data generation module is used to obtain a water quality and quantity sample data set of a drainage network; The second simulation data generation module is used to use the water quality and quantity sample data set as the input of the prediction model of calibrated hydrogen sulfide production, simulate under the conditions of no drug addition and drug addition, and obtain the simulation concentration data of sulfide and hydrogen sulfide corresponding to the target prediction position; A prediction model building module, used to build a sulfide and hydrogen sulfide prediction model based on a gradient boosting decision tree algorithm according to the water quality and quantity sample data set and the simulated concentration data of sulfide and hydrogen sulfide without adding drugs; A prediction module, used to use real-time water quality and water quantity data as input of the sulfide and hydrogen sulfide prediction model, and output the predicted concentration data of sulfide and hydrogen sulfide at the target prediction location; A control model building module, for building a sulfide and hydrogen sulfide control model based on the sulfide and hydrogen sulfide prediction model, in combination with the water quality and quantity sample data set, dosing conditions and simulated concentration data of sulfide and hydrogen sulfide during dosing; A control module, used to use the real-time water quality and quantity data, the predicted concentration data and the preset standard concentration as inputs of the sulfide and hydrogen sulfide control model, output control parameters and perform regulation according to the control parameters; Control model building modules, including: The first construction module is used to take the water quality and quantity sample data set and the dosing conditions as input, and the simulated concentration data of sulfide and hydrogen sulfide during dosing as output, and re-divide the training set and the validation set according to the proportion, retrain the sulfide and hydrogen sulfide prediction model through the training set, and optimize the trained sulfide and hydrogen sulfide prediction model through the validation set, so that the error between the sulfide and hydrogen sulfide prediction model predicting the sulfide and hydrogen sulfide concentration in the drainage network and the validation data reaches the expected error range; The second building block is used to quickly determine the relationship between the dosing conditions and the simulated concentration data of sulfide and hydrogen sulfide during dosing based on the retrained sulfide and hydrogen sulfide prediction model, and obtain formula (1) and formula (2): DS α、β = model DS ( dos, place α , condition β ), Formula (1), H 2 S α、β = model H2S ( dos, place α , condition β ), Formula (2), in, model DS and model H2S They respectively represent the sulfide prediction relationship and hydrogen sulfide prediction relationship obtained according to the retrained sulfide and hydrogen sulfide prediction models, dos Indicates the dosage of the drug. place α Indicates α Dosing location, condition β Indicates β Water quality and quantity data, namely β Water quality and quantity conditions, DS α、β and H 2 S α、β Respectively expressed in α Dosing location and β Under the condition of water quality and quantity and the dosage is dos When , the predicted concentration of sulfide and hydrogen sulfide at the target prediction location; The third construction module is used to establish a planning model with the minimum dosage as the objective function, the sulfide concentration and hydrogen sulfide concentration at the target predicted position after the dosage reach the corresponding preset standard concentration, and the optimal dosage sulfur control effect as the constraint condition, and obtain formula (3): , Formula (3), in, a and b They represent the preset standard concentration of sulfide and hydrogen sulfide at the target prediction location, respectively, min ( H 2 S β ) and min( DS β ) respectively represent that when the water quality, water quantity and dosage are the same and in the β When there are 3 dosing positions, the predicted concentrations of sulfide and hydrogen sulfide at the target predicted position are the smallest, that is, the dosing sulfur control effect is optimal; The fourth building block is used to combine formula (1), formula (2) and formula (3) to obtain the sulfide and hydrogen sulfide control model.

8. A storage medium, It is characterized in that A program is stored, and when the program is executed by a processor, an artificial intelligence-based method for rapid prediction and precise control of hydrogen sulfide in an urban drainage network as described in any one of claims 1 to 6 is implemented.

9. A computing device, It is characterized in that It includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, it implements the method for rapid prediction and precise control of hydrogen sulfide in an urban drainage network based on artificial intelligence as described in any one of claims 1 to 6.