Intelligent dosing method and system for sewage treatment based on artificial intelligence

Abstract

The invention belongs to the technical field of sewage treatment, and discloses an intelligent dosing method and system for sewage treatment based on artificial intelligence. The method comprises the following steps: arranging a sewage detection device in a sewage treatment pipe network of a sewage treatment plant; constructing an artificial intelligence model in the cloud data center; collecting real-time sewage quality direct measurement data by using the sewage quality direct measurement equipment of each sewage detection device; generating real-time sewage quality soft measurement data corresponding to the real-time sewage quality direct measurement data in the cloud data center; sewage detection data analysis is carried out in the cloud data center; in the cloud data center, a sewage treatment dosing control strategy is generated; and executing a real-time sewage treatment dosing control strategy by using the sewage treatment dosing equipment of each sewage detection device. According to the invention, the problems of inaccurate dosage, low processing efficiency, insufficient data utilization, high cost input and low informatization level in the prior art are solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of sewage treatment, and in particular relates to an intelligent dosing method and system for sewage treatment based on artificial intelligence. Background Art

[0002] The primary function of a sewage treatment plant is to treat wastewater to meet discharge standards or water quality requirements for reuse. Furthermore, sewage treatment plants are a vital component of urban infrastructure, playing a crucial role in improving environmental quality and enhancing people's quality of life. Sewage treatment plants treat wastewater by adding different chemicals based on its specific composition and treatment objectives. These chemicals primarily remove pollutants such as suspended solids, colloids, heavy metals, organic matter, nitrogen, and phosphorus.

[0003] The existing sewage treatment dosing technology has the following defects: 1) Inaccurate dosage: Traditional sewage treatment dosing methods often rely on manual experience and preset procedures, and are unable to adjust the dosage in real time and accurately according to changes in sewage water quality, resulting in excessive or insufficient dosage, affecting the treatment effect; 2) Low treatment efficiency: Due to the lack of intelligent control, existing technologies are slow to respond to complex and changing sewage water quality, resulting in low treatment efficiency and failure to meet the demand for efficient treatment; 3) Insufficient data utilization: Existing technologies do not fully utilize sewage detection data, lack effective data analysis and mining methods, and are unable to fully utilize historical data and real-time data to optimize the treatment process; 4) High cost: To ensure comprehensive sewage quality testing, a large amount of money is required to configure a full set of water quality sensors. This also results in waste of chemicals and excessive energy consumption, increasing the operating costs of sewage treatment plants. 5) Low level of informatization: The existing technology has a low level of informatization and lacks a unified management platform, making it impossible to achieve remote monitoring, data analysis, and intelligent decision-making, which affects management efficiency. Summary of the Invention

[0004] In order to solve the problems of inaccurate dosing, low treatment efficiency, insufficient data utilization, high cost investment and low level of informatization in the existing technology, the purpose of the present invention is to provide an intelligent dosing method and system for sewage treatment based on artificial intelligence.

[0005] The technical solution adopted in the present invention is: An artificial intelligence-based intelligent dosing method for sewage treatment includes the following steps: Install corresponding sewage detection devices at several detection locations in the sewage treatment network of the sewage treatment plant, and connect all sewage detection devices to the cloud data center; In the cloud data center, artificial intelligence algorithms are used to build sewage water quality soft measurement models, sewage detection data analysis models, and sewage treatment dosing control models; Use the sewage water quality direct measurement equipment of each sewage detection device to collect real-time sewage water quality direct measurement data at the corresponding detection location and upload it to the cloud data center; In the cloud data center, based on each real-time sewage water quality direct measurement data, the sewage water quality soft measurement model is used to generate corresponding real-time sewage water quality soft measurement data; In the cloud data center, based on a number of real-time sewage water quality direct measurement data and corresponding real-time sewage water quality soft measurement data, a sewage detection data analysis model is used to generate real-time sewage detection data analysis results; In the cloud data center, based on the real-time sewage detection data analysis results, the sewage treatment dosing control model is used to generate the corresponding real-time sewage treatment dosing control strategy and send it to each sewage detection device; Use the sewage treatment dosing equipment of each sewage detection device to implement real-time sewage treatment dosing control strategy and perform intelligent dosing on the sewage treatment network of the sewage treatment plant.

[0006] Furthermore, the detection positions of the sewage detection device include the outlet of the sewage discharge pipe of the sewage treatment plant, and the inlet, outlet and intersection of the sewage treatment pipe in the sewage treatment network of the sewage treatment plant. Furthermore, in the cloud data center, artificial intelligence algorithms are used to build a sewage water quality soft measurement model, a sewage detection data analysis model, and a sewage treatment dosing control module, including the following steps: In the cloud data center, a number of historical sewage water quality direct measurement data are collected and pre-processed to obtain a number of pre-processed historical sewage water quality direct measurement data; Based on some direct measurement data of historical sewage water quality after pretreatment, a sewage water quality soft measurement model is constructed using deep learning and decision tree fusion algorithm, and some predicted sewage water quality soft measurement data are generated; Based on several direct measurement data of historical sewage water quality after pretreatment and the corresponding soft measurement data of predicted sewage water quality, a sewage detection data analysis model was constructed using a deep learning and swarm intelligence optimization fusion algorithm, and several historical sewage detection data analysis results were generated; Based on the analysis results of several historical sewage detection data, a sewage treatment dosing control model was constructed using a deep learning and adversarial training fusion algorithm, and several historical sewage treatment dosing control experiences were generated.

[0007] Furthermore, the sewage water quality soft measurement model is constructed based on the RF-MLP algorithm, and the sewage water quality soft measurement model includes a key feature extraction module constructed based on the RF algorithm and a sewage water quality soft measurement module constructed based on the MLP algorithm, which are connected in sequence.

[0008] Furthermore, the sewage detection data analysis model is constructed based on the CNN-DBN-IWOA algorithm, and the sewage detection data analysis model includes a matrix feature extraction module constructed based on the CNN algorithm, a sewage detection data analysis module constructed based on the DBN algorithm, and an analysis result optimization module constructed based on the IWOA algorithm, which are connected in sequence.

[0009] Furthermore, the sewage treatment dosing control model is constructed based on the MOPPO-cGAN algorithm, and the sewage treatment dosing control model includes a sewage treatment dosing control module constructed based on the MOPPO algorithm and an adversarial training module constructed based on the cGAN algorithm, which are connected in sequence. The sewage treatment dosing control module is provided with an objective function set, an experience replay pool, an Actor network, a Critic network and an intelligent agent, and the adversarial training module is provided with a generator and a discriminator.

[0010] Furthermore, in the cloud data center, based on each real-time sewage water quality direct measurement data, a sewage water quality soft measurement model is used to generate corresponding real-time sewage water quality soft measurement data, including the following steps: In the cloud data center, the key feature extraction module of the sewage water quality soft measurement model is used to extract several real-time key features with high correlation between the real-time sewage water quality direct measurement data and the sewage water quality soft measurement indicators; According to several real-time key features, the sewage water quality soft measurement module of the sewage water quality soft measurement model is used to perform sewage water quality soft measurement prediction and obtain corresponding real-time sewage water quality soft measurement data; All real-time sewage water quality direct measurement data received by the cloud data center are traversed to obtain some real-time sewage water quality soft measurement data.

[0011] Furthermore, in the cloud data center, based on a number of real-time sewage water quality direct measurement data and corresponding real-time sewage water quality soft measurement data, a sewage detection data analysis model is used to generate real-time sewage detection data analysis results, including the following steps: In the cloud data center, the real-time sewage water quality soft measurement data corresponding to each real-time sewage water quality direct measurement data are combined to obtain a number of real-time sewage water quality comprehensive measurement data; Converting a plurality of real-time sewage water quality comprehensive measurement data into a real-time sewage water quality comprehensive measurement data matrix, and inputting the real-time sewage water quality comprehensive measurement data matrix into a sewage detection data analysis model; Use the matrix feature extraction module of the sewage detection data analysis model to extract the real-time matrix features of the real-time sewage water quality comprehensive measurement data matrix; According to the real-time matrix characteristics, the sewage detection data analysis module of the sewage detection data analysis model is used to perform sewage detection data analysis and prediction to obtain the real-time sewage detection data analysis probability distribution; The analysis result optimization module of the sewage detection data analysis model is used to optimize the analysis results of the real-time sewage detection data analysis probability distribution to obtain the real-time sewage detection data analysis results.

[0012] Furthermore, in the cloud data center, based on the real-time sewage detection data analysis results, a sewage treatment dosing control model is used to generate a corresponding real-time sewage treatment dosing control strategy, and the strategy is sent to each sewage detection device, including the following steps: In the cloud data center, the real-time sewage detection data analysis results are analyzed to obtain several real-time sewage detection data analysis states; According to a number of real-time sewage detection data analysis states, the state space of the intelligent agent of the sewage treatment dosing control module in the sewage treatment dosing control model is updated to obtain an updated state space; Randomly extract a number of historical sewage treatment dosing control experiences from the experience playback pool of the sewage treatment dosing control module, and generate a number of possible sewage treatment dosing control actions based on the number of historical sewage treatment dosing control experiences; According to several possible sewage treatment dosing control actions, the action space of the intelligent agent of the sewage treatment dosing control module is updated to obtain an updated action space; A real-time objective function is selected from the objective function set of the sewage treatment dosing control module. Based on the real-time objective function, an intelligent agent is used to control the critic network to generate the real-time value of all possible sewage treatment dosing control actions in the updated action space for each real-time sewage detection data analysis state in the updated state space. Based on several real-time values, use intelligent agents to control the Actor network and generate the probability distribution of all possible sewage treatment dosing control actions corresponding to each real-time sewage detection data analysis state; The possible sewage treatment dosing control action with the highest probability distribution in the updated action space is used as the execution sewage treatment dosing control action for the real-time sewage detection data analysis state; Integrate all real-time sewage detection data in the updated state space to analyze the state of the sewage treatment dosing control action, obtain the real-time sewage treatment dosing control strategy, and send it to each sewage detection device.

[0013] An artificial intelligence-based intelligent dosing system for sewage treatment is used to implement an intelligent dosing method for sewage treatment. The system includes a cloud data center and several sewage detection devices. The several sewage detection devices are all communicatively connected to the cloud data center, and each sewage detection device is provided with a direct sewage quality measurement device and a sewage treatment dosing device. The cloud data center includes a model construction unit, a sewage quality soft measurement unit, a sewage detection data analysis unit, and a sewage treatment dosing control unit, which are connected in sequence.

[0014] The beneficial effects of the present invention are: The present invention provides an artificial intelligence-based sewage treatment intelligent dosing method and system, which can mine deep information in sewage water quality data by monitoring sewage water quality in real time and using artificial intelligence algorithms to build sewage water quality soft measurement models, sewage detection data analysis models and sewage treatment dosing control models, and achieve more accurate dosing control through intelligent sewage treatment dosing control, thereby avoiding excessive or insufficient dosing, and significantly improving sewage treatment effects; the provided integrated sewage water quality direct measurement, sewage water quality soft measurement, sewage detection data analysis and sewage treatment dosing control process, and intelligent control of dosing operations can quickly respond to changes in sewage water quality, adjust treatment strategies, effectively improve sewage treatment efficiency, and meet the requirements of efficient treatment demand; through advanced data analysis and mining technology, make full use of historical data and real-time data, continuously optimize the sewage treatment process, and realize data-driven intelligent decision-making; the sewage water quality soft measurement model can predict water quality indicators based on direct measurement data, avoiding the configuration of expensive equipment, reducing hardware cost investment, and through intelligent dosing control, reduce chemical waste and excessive energy consumption, while optimizing sensor configuration and reducing the operating costs of sewage treatment plants; establish a unified information management platform in the cloud data center to realize remote monitoring, data analysis and intelligent decision-making, improve management efficiency, and promote the modern operation of sewage treatment plants; through precise control of the dosing process, improve the quality of recycled water and increase the recycling rate of water resources.

[0015] Other beneficial effects of the present invention will be further described in the specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of the intelligent dosing method for sewage treatment based on artificial intelligence in the present invention.

[0017] Figure 2 This is a structural block diagram of the artificial intelligence-based intelligent dosing system for sewage treatment in the present invention. DETAILED DESCRIPTION

[0018] The present invention will be further explained below with reference to the accompanying drawings and specific embodiments.

[0019] Example 1: like Figure 1 As shown, this embodiment provides an artificial intelligence-based intelligent dosing method for sewage treatment, comprising the following steps: S1: Install corresponding sewage detection devices at several detection locations in the sewage treatment network of the sewage treatment plant, and connect all sewage detection devices to the cloud data center; The detection locations of the sewage detection device include the outlet of the sewage discharge pipe of the sewage treatment plant, and the inlet, outlet and intersection of the sewage treatment pipes in the sewage treatment network of the sewage treatment plant; It is used to collect water quality data of sewage in sewage treatment pipes within the sewage treatment network, and to perform dosing operations conveniently and quickly, so that chemical agents can quickly enter the sewage treatment network; S2: In the cloud data center, artificial intelligence algorithms are used to build a sewage water quality soft measurement model, a sewage detection data analysis model, and a sewage treatment dosing control model, including the following steps: S2-1: In the cloud data center, collect some historical sewage water quality direct measurement data and pre-process them to obtain some pre-processed historical sewage water quality direct measurement data; Direct wastewater quality measurement data include easily measurable variables (such as flow rate, temperature, pH value, dissolved oxygen, etc.). Direct wastewater quality measurement data are used as a data basis to predict key water quality indicators (such as chemical oxygen demand, biochemical oxygen demand, total nitrogen, total phosphorus, etc.) that may require complex, time-consuming or expensive analytical equipment or cannot be directly measured; S2-2: Based on some direct measurement data of historical sewage water quality after pretreatment, a sewage water quality soft measurement model is constructed using deep learning and decision tree fusion algorithm, and some predicted sewage water quality soft measurement data are generated; The sewage water quality soft measurement model is constructed based on the Random Forest (RF)-Multilayer Perceptron (MLP) algorithm, and the sewage water quality soft measurement model includes a key feature extraction module constructed based on the RF algorithm and a sewage water quality soft measurement module constructed based on the MLP algorithm, which are connected in sequence; The key feature extraction module uses ensemble learning of multiple decision trees to screen several important features of direct sewage quality measurement data. Among these important features, the key features that are relevant to the prediction of key water quality indicators are selected as the feature basis for subsequent sewage quality soft measurement prediction. The sewage quality soft measurement module is used to perform sewage quality soft measurement prediction based on several key features. S2-3: Based on several direct measurement data of historical sewage water quality after pretreatment and the corresponding soft measurement data of predicted sewage water quality, a sewage detection data analysis model is constructed using a deep learning and swarm intelligence optimization fusion algorithm, and several historical sewage detection data analysis results are generated; The sewage detection data analysis model is built based on the Convolutional Neural Network (CNN)-Deep Belief Network (DBN)-Improved Whale Optimization Algorithm (IWOA) algorithm. The model includes a matrix feature extraction module based on the CNN algorithm, a sewage detection data analysis module based on the DBN algorithm, and an analysis result optimization module based on the IWOA algorithm. The matrix feature extraction module, based on a deep network structure, mines the deep features of the sewage water quality comprehensive measurement data matrix to improve the accuracy of data analysis. The multi-layer structure of the sewage detection data analysis module learns the complex patterns and inherent relationships in the matrix features. Based on the matrix features, it conducts sewage detection data analysis and prediction, and obtains the probability distribution of sewage detection data analysis. The analysis result optimization module optimizes the recognition results, effectively improving the accuracy and stability of the analysis results, enhancing the optimization ability of the model, and making the analysis results more reliable. S2-4: Based on the analysis results of several historical sewage detection data, a sewage treatment dosing control model is constructed using a deep learning and adversarial training fusion algorithm, and several historical sewage treatment dosing control experiences are generated; The sewage treatment dosing control model is constructed based on the Multi-Objective Proximal Policy Optimization (MOPPO)-Conditional Generative Adversarial Network (cGAN) algorithm. The sewage treatment dosing control model includes a sewage treatment dosing control module constructed based on the MOPPO algorithm and an adversarial training module constructed based on the cGAN algorithm, which are connected in sequence. The sewage treatment dosing control module is equipped with an objective function set, an experience replay pool, an actor network, a critic network, and an intelligent agent. The adversarial training module is equipped with a generator and a discriminator. The Actor network of the sewage treatment dosing control module is responsible for outputting the probability distribution of the actions that should be taken in a given state. The goal is to learn an optimal strategy, that is, to maximize the long-term cumulative reward. In the continuous action space, the Actor network usually outputs a mean and an optional variance parameter to describe the probability distribution of the action. The Critic network is responsible for evaluating the value of a given state, that is, predicting the expected return that can be obtained by starting from this state and following the current strategy. It usually outputs a scalar value representing the value of the state or the state-action value. The experience replay pool is used to store historical experience for reuse during the training process. The objective function set includes functions of multiple defined data processing objectives, including minimizing the drug loss rate, minimizing the sewage pollutant content, maximizing the drug utilization rate, maximizing the dosing response efficiency, etc. The generator of the adversarial training module is used to generate strategies, and the discriminator is used to distinguish between the generated strategies and the optimal strategies. Through this adversarial training process, the sewage treatment dosing control module can learn more effective strategies. At the same time, the adversarial training module helps ensure the diversity and quality of strategies, thereby finding better optimal solutions in multi-objective optimization problems. Based on the analysis results of several historical sewage detection data, a sewage treatment dosing control model was constructed using a deep learning and adversarial training fusion algorithm. This model also generated several historical sewage treatment dosing control experiences, including the following steps: S2-4-1: Use the MOPPO-GAN algorithm to build an initial sewage treatment dosing control model; the initial sewage treatment dosing control model includes an initial sewage treatment dosing control strategy generation module and an initial adversarial training module; S2-4-2: Set the objective function set, experience replay pool, actor network, critic network and intelligent agent for the initial sewage treatment dosing control strategy generation module; S2-4-3: Use the sewage treatment dosing control strategy generation problem as the initial simulation environment for the sewage treatment dosing control strategy generation module, and set the action space and state space for the intelligent agent; S2-4-4: Based on any objective function in the objective function set and according to the analysis results of a number of historical sewage detection data, the initial sewage treatment dosing control strategy generation module is pre-trained to obtain a pre-trained sewage treatment dosing control strategy generation module, and a number of historical sewage treatment dosing control strategies and corresponding historical sewage treatment dosing control experiences are generated; S2-4-5: Based on the analysis results of several historical sewage detection data and the corresponding historical sewage treatment dosing control strategies, the initial generator of the initial adversarial training module is optimized and trained to obtain an optimized generator, and several sewage treatment dosing control strategies are generated; S2-4-6: Based on several historical sewage treatment dosing control strategies and corresponding generated sewage treatment dosing control strategies, the initial discriminator of the initial adversarial training module is optimized and trained to obtain an optimized discriminator, and several historical discrimination results are generated; S2-4-7: Use the pre-trained Critic network of the sewage treatment dosing control strategy generation module to obtain several rewards for generating sewage treatment dosing control strategies, and optimize the Actor network of the pre-trained sewage treatment dosing control strategy generation module based on the several rewards to obtain an optimized Actor network; S2-4-8: Optimize the critic network of the pre-trained sewage treatment dosing control strategy generation module based on the rewards for generating several sewage treatment dosing control strategies and the corresponding historical discrimination results to obtain an optimized critic network; S2-4-9: Traverse all objective functions in the objective function set and repeat the above adversarial training steps to obtain an optimized sewage treatment dosing control strategy generation module provided with an optimized Actor network and an optimized Critic network and an optimized adversarial training module provided with an optimized discriminator and an optimized discriminator; S2-4-10: Integrate the optimized sewage treatment dosing control strategy generation module and the optimized adversarial training module to obtain the final sewage treatment dosing control model, and store several historical sewage treatment dosing control experiences in the experience replay pool; S3: Use the sewage water quality direct measurement equipment of each sewage detection device to collect real-time sewage water quality direct measurement data at the corresponding detection location and upload it to the cloud data center; S4: In the cloud data center, based on each real-time sewage water quality direct measurement data, the sewage water quality soft measurement model is used to generate corresponding real-time sewage water quality soft measurement data, including the following steps: S4-1: In the cloud data center, the key feature extraction module of the sewage water quality soft measurement model is used to extract several real-time key features with high correlation between the real-time sewage water quality direct measurement data and the sewage water quality soft measurement indicators; S4-2: Based on several real-time key features, the sewage water quality soft measurement module of the sewage water quality soft measurement model is used to perform sewage water quality soft measurement prediction and obtain corresponding real-time sewage water quality soft measurement data; S4-3: Traverse all real-time sewage water quality direct measurement data received by the cloud data center to obtain some real-time sewage water quality soft measurement data; S5: In the cloud data center, based on a number of real-time sewage water quality direct measurement data and corresponding real-time sewage water quality soft measurement data, a sewage detection data analysis model is used to generate real-time sewage detection data analysis results, including the following steps: S5-1: In the cloud data center, the real-time sewage water quality soft measurement data corresponding to each real-time sewage water quality direct measurement data are combined to obtain a number of real-time sewage water quality comprehensive measurement data; S5-2: converting a number of real-time sewage water quality comprehensive measurement data into a real-time sewage water quality comprehensive measurement data matrix, and inputting the real-time sewage water quality comprehensive measurement data matrix into a sewage detection data analysis model; S5-3: Use the matrix feature extraction module of the sewage detection data analysis model to extract the real-time matrix features of the real-time sewage water quality comprehensive measurement data matrix; S5-4: Based on the real-time matrix characteristics, the sewage detection data analysis module of the sewage detection data analysis model is used to perform sewage detection data analysis and prediction to obtain a real-time sewage detection data analysis probability distribution; S5-5: Using the analysis result optimization module of the sewage detection data analysis model, the analysis result of the real-time sewage detection data analysis probability distribution is optimized to obtain the real-time sewage detection data analysis result, including the following steps: S5-5-1: Analyze the probability distribution of real-time sewage detection data, set the real-time adjustment parameters of the probability distribution of real-time sewage detection data analysis as the solution vector of the analysis result optimization module, and initialize it according to the solution vector to obtain several initial solutions; Specifically, the Tent-Logistic-Cosine chaotic mapping sequence is used for initialization to generate several initial solutions of the analysis result optimization module, and an initial IWOA population consisting of several initial IWOA individuals (initial solutions) is obtained; The formula is:

[0020] Where, IWOA individuals (initial solutions) generated by the Tent-Logistic-Cosine chaotic mapping sequence; is a randomly generated IWOA individual; are preset parameters; i is the individual indicator of IWOA; the initial population is generated by the Tent-Logistic-Cosine chaotic mapping sequence. Compared with the randomly distributed population, the improved initial position distribution of the whale population is more uniform, which expands the search range of the whale population in space and increases the diversity of the group position. To a certain extent, it improves the defect of the algorithm that it is easy to fall into local extreme values, thereby improving the optimization efficiency of the algorithm; S5-5-2: Minimizing the prediction error is the optimization goal, the optimization goal is used as the fitness function, and the IWOA population parameters and maximum number of iterations of the IWOA optimization algorithm are set; The formula is:

[0021] Where, For IWOA individuals The fitness function of is the prediction error value; is the IWOA individual variable; is the IWOA individual indicator; S5-5-3: Random Generation p ,like p <0. 5 and| A |<1, execute the convergence factor improved encirclement prey behavior, update the IWOA population position, if p <0. 5 and| A | ≥1, perform the search behavior with improved convergence factor and update the IWOA population position. If p ≥0.5, execute the bubble net attack behavior with improved convergence factor, update the IWOA population, and obtain an updated IWOA population including several updated IWOA individuals; p To update the parameters, A The step size coefficient for introducing convergence factor optimization; The formula for surrounding prey behavior is:

[0022] Where, IWOA individuals updated for the prey encirclement behavior; is the optimal IWOA individual; is the distance between the current IWOA individual and the optimal IWOA individual; , is the convergence factor that decreases from 2 to 0; For the n The guidance coefficient of the search space node for the IWOA individual; The formula for the convergence factor is:

[0023] Where, a is the convergence factor; tanh(.) is the hyperbolic tangent function; t 、 t max are the current number of iterations and the maximum number of iterations respectively; a max 、 a min are the maximum and minimum values of the convergence factor respectively; λ is the deceleration rate parameter, k is the decrement period parameter, λ =-2π , k = π ; In the early stage of iteration, a The value of is larger, the newer A Also larger, | A | ≥1, which makes the IWOA algorithm search for prey for a long time in the early stage of iteration, enhancing the global search ability of the algorithm. a The value of is smaller, the newer A Also smaller, | A |<1, the IWOA algorithm will be in the behavior of surrounding the prey for a long time in the early iteration, which will enhance the algorithm's local surrounding ability and improve the local hunting ability; The formula for searching for prey is:

[0024] Where, IWOA individuals updated for prey search behavior; is a random IWOA individual selected from the IWOA population; For the n The guidance coefficient of the search space node for the IWOA individual; are the current iteration number respectively; The formula for the attack behavior of Bubble Network is:

[0025] Where, IWOA individuals updated for the Bubble Network attack behavior; is the distance between the current IWOA individual and the prey; b is the constant defining the screw equation, ; l is a random number between [−1, 1]; For the n The guidance coefficient of the search space node for the IWOA individual; S5-5-4: Use the dynamic reverse learning algorithm to perform dynamic reverse learning on the updated IWOA population to obtain the reverse IWOA population. The formula of dynamic reverse learning strategy is:

[0026] Where, For the i IWOA Individual j The reverse solution position of dimension; For the i IWOA Individual j The forward solution position of the dimension; The updated IWOA populations are j Upper and lower bounds of the dimension; is the decreasing inertia factor, ; are the current number of iterations and the maximum number of iterations respectively; S5-5-5: Use the Gaussian mutation algorithm to perform Gaussian mutation on the updated IWOA population to generate a Gaussian mutated IWOA population including several Gaussian mutated IWOA individuals; The formula is:

[0027] Where, It is a dynamically reversed IWOA individual; For the updated IWOA individual; is the standard deviation; The mean is 0 and the variance is A random number drawn from the normal distribution is used to simulate the Gaussian variation process; S5-5-6: Calculate the fitness value of each IWOA individual in the updated IWOA population, the reverse IWOA population, and the Gaussian-mutated IWOA population, and take the IWOA individual with the minimum fitness value as the optimal individual; S5-5-7: If the number of iterations of iterative optimization reaches the maximum number of iterations or the fitness value of the optimal individual meets the requirements, the optimal individual will be output; S5-5-8: Decode the solution vector of the optimal individual to obtain the optimal real-time adjustment parameter, and adjust the real-time sewage detection data analysis probability distribution based on the optimal real-time adjustment parameter to obtain the adjusted real-time sewage detection data analysis probability distribution; S5-5-9: taking the real-time sewage detection data analysis prediction label with the highest probability in the adjusted real-time sewage detection data analysis probability distribution as the real-time sewage detection data analysis result; Sewage testing data analysis results include: water quality classification results (based on predicted indicator values, determining the sewage water quality category, such as Class V, Class V, Class IV, etc., to meet national or local emission standards), anomaly detection results (detecting whether there are abnormal data, such as abnormal water quality caused by sudden pollutant leaks or equipment failures), trend analysis (analyzing the changing trends of water quality indicators, such as rising, falling, or stable, to predict future water quality changes), etc. S6: In the cloud data center, based on the real-time sewage detection data analysis results, a sewage treatment dosing control model is used to generate a corresponding real-time sewage treatment dosing control strategy, and the strategy is sent to each sewage detection device, including the following steps: S6-1: In the cloud data center, the real-time sewage detection data analysis results are analyzed to obtain several real-time sewage detection data analysis states; S6-2: Based on the state analysis of a number of real-time sewage detection data, the state space of the intelligent agent of the sewage treatment dosing control module in the sewage treatment dosing control model is updated to obtain an updated state space; S6-3: randomly extracting a number of historical sewage treatment dosing control experiences from the experience playback pool of the sewage treatment dosing control module, and generating a number of possible sewage treatment dosing control actions based on the number of historical sewage treatment dosing control experiences; S6-4: updating the action space of the intelligent agent of the sewage treatment dosing control module according to several possible sewage treatment dosing control actions to obtain an updated action space; S6-5: Select a real-time objective function from the objective function set of the sewage treatment dosing control module, and based on the real-time objective function, use the intelligent agent to control the critic network to generate the real-time value of all possible sewage treatment dosing control actions in the updated action space for each real-time sewage detection data analysis state in the updated state space; S6-6: Based on several real-time values, use intelligent agents to control the Actor network to generate the probability distribution of all possible sewage treatment dosing control actions corresponding to each real-time sewage detection data analysis state; S6-7: taking the possible sewage treatment dosing control action with the highest probability distribution in the updated action space as the executed sewage treatment dosing control action for the real-time sewage detection data analysis state; S6-8: Integrate all real-time sewage detection data in the updated state space to analyze the execution of sewage treatment dosing control actions, obtain a real-time sewage treatment dosing control strategy, and send it to each sewage detection device; The dosing control strategy for sewage treatment includes: dosing amount setting action (setting the dosage of specific chemical agents, such as coagulants, flocculants, oxidants, reducing agents, disinfectants, etc.), dosing frequency setting action (setting the dosing time interval, such as once an hour, every half hour, once every two hours), dosing point setting action (setting the dosing location to better mix the agent and sewage), agent type setting action (setting the use of different types of chemical agents according to changes in water quality), mixing intensity setting action (starting the intensity of the agitator or mixing equipment to improve the mixing effect of the agents), pH value setting action (adjusting the pH value of the sewage by adding acid or alkali to facilitate certain chemical processes), dissolved oxygen setting action (adjusting the aeration volume to control the dissolved oxygen content in the sewage and affect the biological treatment effect), etc. S7: Using the sewage treatment dosing equipment of each sewage detection device, executing a real-time sewage treatment dosing control strategy, and performing intelligent dosing on the sewage treatment pipe network of the sewage treatment plant, including the following steps: S7-1: using the sewage treatment dosing equipment of each sewage detection device, analyzing the received real-time sewage treatment dosing control strategy to confirm the content of the real-time sewage treatment dosing control strategy related to the current sewage detection device; S7-2: Generate real-time control instructions for the sewage treatment dosing equipment based on the real-time sewage treatment dosing control strategy. The real-time control instructions include a real-time start instruction for the target chemical type to be added, a real-time duration instruction for adding the target chemical dosage, and a real-time addition time instruction for the time of adding the chemical. S7-3: According to the real-time control instructions, the sewage treatment dosing equipment is used to add the target chemical agent at the target chemical agent dosage to the corresponding position of the sewage treatment pipeline network at the target time.

[0028] Example 2: like Figure 2 As shown, this embodiment provides an artificial intelligence-based intelligent dosing system for sewage treatment, which is used to implement an intelligent dosing method for sewage treatment. The system includes a cloud data center and several sewage detection devices. The several sewage detection devices are all communicatively connected to the cloud data center, and each sewage detection device is provided with a direct sewage quality measurement device and a sewage treatment dosing device. The cloud data center includes a model building unit, a sewage quality soft measurement unit, a sewage detection data analysis unit, and a sewage treatment dosing control unit connected in sequence; The sewage detection device is used to use sewage water quality direct measurement equipment to collect real-time sewage water quality direct measurement data at the corresponding detection location and upload it to the cloud data center; the sewage treatment dosing equipment is used to implement real-time sewage treatment dosing control strategies and perform intelligent dosing on the sewage treatment pipe network of the sewage treatment plant; A model building unit is used to use artificial intelligence algorithms to build a sewage water quality soft measurement model, a sewage detection data analysis model, and a sewage treatment dosing control model; The sewage water quality soft measurement unit is used to generate corresponding real-time sewage water quality soft measurement data based on each real-time sewage water quality direct measurement data using the sewage water quality soft measurement model; A sewage detection data analysis unit is used to generate real-time sewage detection data analysis results using a sewage detection data analysis model based on a number of real-time sewage water quality direct measurement data and corresponding real-time sewage water quality soft measurement data; The sewage treatment dosing control unit is used to generate a corresponding real-time sewage treatment dosing control strategy based on the real-time sewage detection data analysis results and use the sewage treatment dosing control model, and send it to each sewage detection device.

[0029] The present invention provides an artificial intelligence-based sewage treatment intelligent dosing method and system, which can mine deep information in sewage water quality data by monitoring sewage water quality in real time and using artificial intelligence algorithms to build sewage water quality soft measurement models, sewage detection data analysis models and sewage treatment dosing control models, and achieve more accurate dosing control through intelligent sewage treatment dosing control, thereby avoiding excessive or insufficient dosing, and significantly improving sewage treatment effects; the provided integrated sewage water quality direct measurement, sewage water quality soft measurement, sewage detection data analysis and sewage treatment dosing control process, and intelligent control of dosing operations can quickly respond to changes in sewage water quality, adjust treatment strategies, effectively improve sewage treatment efficiency, and meet the requirements of efficient treatment demand; through advanced data analysis and mining technology, make full use of historical data and real-time data, continuously optimize the sewage treatment process, and realize data-driven intelligent decision-making; the sewage water quality soft measurement model can predict water quality indicators based on direct measurement data, avoiding the configuration of expensive equipment, reducing hardware cost investment, and through intelligent dosing control, reduce chemical waste and excessive energy consumption, while optimizing sensor configuration and reducing the operating costs of sewage treatment plants; establish a unified information management platform in the cloud data center to realize remote monitoring, data analysis and intelligent decision-making, improve management efficiency, and promote the modern operation of sewage treatment plants; through precise control of the dosing process, improve the quality of recycled water and increase the recycling rate of water resources.

[0030] The present invention is not limited to the above optional embodiments. Anyone can derive various other forms of products based on the teachings of the present invention. The above specific embodiments should not be construed as limiting the scope of protection of the present invention. The scope of protection of the present invention shall be based on the scope defined in the claims, and the description can be used to interpret the claims.

Claims

1. An artificial intelligence-based intelligent dosing method for sewage treatment, characterized by: The steps include: Install corresponding sewage detection devices at several detection locations in the sewage treatment network of the sewage treatment plant, and connect all sewage detection devices to the cloud data center; In the cloud data center, artificial intelligence algorithms are used to build sewage water quality soft measurement models, sewage detection data analysis models, and sewage treatment dosing control models; Use the sewage water quality direct measurement equipment of each sewage detection device to collect real-time sewage water quality direct measurement data at the corresponding detection location and upload it to the cloud data center; In the cloud data center, based on each real-time sewage water quality direct measurement data, the sewage water quality soft measurement model is used to generate corresponding real-time sewage water quality soft measurement data; In the cloud data center, based on a number of real-time sewage water quality direct measurement data and corresponding real-time sewage water quality soft measurement data, a sewage detection data analysis model is used to generate real-time sewage detection data analysis results; In the cloud data center, based on the real-time sewage detection data analysis results, the sewage treatment dosing control model is used to generate the corresponding real-time sewage treatment dosing control strategy and send it to each sewage detection device; Use the sewage treatment dosing equipment of each sewage detection device to implement real-time sewage treatment dosing control strategy and perform intelligent dosing on the sewage treatment network of the sewage treatment plant.

2. The artificial intelligence-based intelligent dosing method for sewage treatment according to claim 1, characterized in that: The detection positions of the sewage detection device include the outlet of the sewage discharge pipe of the sewage treatment plant, and the inlet, outlet and intersection of the sewage treatment pipe in the sewage treatment network of the sewage treatment plant.

3. The artificial intelligence-based intelligent dosing method for sewage treatment according to claim 2, characterized in that: In the cloud data center, artificial intelligence algorithms are used to build a sewage water quality soft measurement model, a sewage detection data analysis model, and a sewage treatment dosing control module, including the following steps: In the cloud data center, a number of historical sewage water quality direct measurement data are collected and pre-processed to obtain a number of pre-processed historical sewage water quality direct measurement data; Based on some direct measurement data of historical sewage water quality after pretreatment, a sewage water quality soft measurement model is constructed using deep learning and decision tree fusion algorithm, and some predicted sewage water quality soft measurement data are generated; Based on several direct measurement data of historical sewage water quality after pretreatment and the corresponding soft measurement data of predicted sewage water quality, a sewage detection data analysis model was constructed using a deep learning and swarm intelligence optimization fusion algorithm, and several historical sewage detection data analysis results were generated; Based on the analysis results of several historical sewage detection data, a sewage treatment dosing control model was constructed using a deep learning and adversarial training fusion algorithm, and several historical sewage treatment dosing control experiences were generated.

4. The artificial intelligence-based intelligent dosing method for sewage treatment according to claim 2, characterized in that: The sewage water quality soft measurement model is constructed based on the RF-MLP algorithm, and the sewage water quality soft measurement model includes a key feature extraction module constructed based on the RF algorithm and a sewage water quality soft measurement module constructed based on the MLP algorithm, which are connected in sequence.

5. The artificial intelligence-based intelligent dosing method for sewage treatment according to claim 2, characterized in that: The sewage detection data analysis model is constructed based on the CNN-DBN-IWOA algorithm, and the sewage detection data analysis model includes a matrix feature extraction module constructed based on the CNN algorithm, a sewage detection data analysis module constructed based on the DBN algorithm, and an analysis result optimization module constructed based on the IWOA algorithm, which are connected in sequence.

6. The artificial intelligence-based intelligent dosing method for sewage treatment according to claim 2, characterized in that: The sewage treatment dosing control model is constructed based on the MOPPO-cGAN algorithm, and the sewage treatment dosing control model includes a sewage treatment dosing control module constructed based on the MOPPO algorithm and an adversarial training module constructed based on the cGAN algorithm, which are connected in sequence. The sewage treatment dosing control module is provided with an objective function set, an experience replay pool, an Actor network, a Critic network and an intelligent agent, and the adversarial training module is provided with a generator and a discriminator.

7. The artificial intelligence-based intelligent dosing method for sewage treatment according to claim 4, characterized in that: In the cloud data center, based on each real-time sewage water quality direct measurement data, the sewage water quality soft measurement model is used to generate the corresponding real-time sewage water quality soft measurement data, including the following steps: In the cloud data center, the key feature extraction module of the sewage water quality soft measurement model is used to extract several real-time key features with high correlation between the real-time sewage water quality direct measurement data and the sewage water quality soft measurement indicators; According to several real-time key features, the sewage water quality soft measurement module of the sewage water quality soft measurement model is used to perform sewage water quality soft measurement prediction and obtain corresponding real-time sewage water quality soft measurement data; All real-time sewage water quality direct measurement data received by the cloud data center are traversed to obtain some real-time sewage water quality soft measurement data.

8. The artificial intelligence-based intelligent dosing method for sewage treatment according to claim 5, characterized in that: In the cloud data center, based on a number of real-time sewage water quality direct measurement data and corresponding real-time sewage water quality soft measurement data, a sewage detection data analysis model is used to generate real-time sewage detection data analysis results, including the following steps: In the cloud data center, the real-time sewage water quality soft measurement data corresponding to each real-time sewage water quality direct measurement data are combined to obtain a number of real-time sewage water quality comprehensive measurement data; Converting a plurality of real-time sewage water quality comprehensive measurement data into a real-time sewage water quality comprehensive measurement data matrix, and inputting the real-time sewage water quality comprehensive measurement data matrix into a sewage detection data analysis model; Use the matrix feature extraction module of the sewage detection data analysis model to extract the real-time matrix features of the real-time sewage water quality comprehensive measurement data matrix; According to the real-time matrix characteristics, the sewage detection data analysis module of the sewage detection data analysis model is used to perform sewage detection data analysis and prediction to obtain the real-time sewage detection data analysis probability distribution; The analysis result optimization module of the sewage detection data analysis model is used to optimize the analysis results of the real-time sewage detection data analysis probability distribution to obtain the real-time sewage detection data analysis results.

9. The artificial intelligence-based intelligent dosing method for sewage treatment according to claim 6, characterized in that: In the cloud data center, based on the real-time sewage detection data analysis results, the sewage treatment dosing control model is used to generate a corresponding real-time sewage treatment dosing control strategy and send it to each sewage detection device, including the following steps: In the cloud data center, the real-time sewage detection data analysis results are analyzed to obtain several real-time sewage detection data analysis states; According to a number of real-time sewage detection data analysis states, the state space of the intelligent agent of the sewage treatment dosing control module in the sewage treatment dosing control model is updated to obtain an updated state space; Randomly extract a number of historical sewage treatment dosing control experiences from the experience playback pool of the sewage treatment dosing control module, and generate a number of possible sewage treatment dosing control actions based on the number of historical sewage treatment dosing control experiences; According to several possible sewage treatment dosing control actions, the action space of the intelligent agent of the sewage treatment dosing control module is updated to obtain an updated action space; A real-time objective function is selected from the objective function set of the sewage treatment dosing control module. Based on the real-time objective function, an intelligent agent is used to control the critic network to generate the real-time value of all possible sewage treatment dosing control actions in the updated action space for each real-time sewage detection data analysis state in the updated state space. Based on several real-time values, use intelligent agents to control the Actor network and generate the probability distribution of all possible sewage treatment dosing control actions corresponding to each real-time sewage detection data analysis state; The possible sewage treatment dosing control action with the highest probability distribution in the updated action space is used as the execution sewage treatment dosing control action for the real-time sewage detection data analysis state; Integrate all real-time sewage detection data in the updated state space to analyze the state of the sewage treatment dosing control action, obtain the real-time sewage treatment dosing control strategy, and send it to each sewage detection device.

10. An artificial intelligence-based intelligent dosing system for sewage treatment, used to implement the intelligent dosing method for sewage treatment according to any one of claims 1 to 9, characterized in that: The system includes a cloud data center and several sewage detection devices. The several sewage detection devices are all communicatively connected to the cloud data center, and each sewage detection device is equipped with a sewage water quality direct measurement device and a sewage treatment dosing device. The cloud data center includes a model construction unit, a sewage water quality soft measurement unit, a sewage detection data analysis unit and a sewage treatment dosing control unit connected in sequence.

Images