A method and system for controlling the addition of water treatment chemicals

Abstract

This invention discloses a method and system for controlling the addition of water treatment chemicals, relating to the field of chemical addition control technology. This invention replaces traditional water quality sensors with high-frequency characteristics of pump current, effectively overcoming the monitoring lag caused by water flow impact in the slide's landing area, and achieving second-level detection of crowd events. A dynamic game model integrates multiple objective constraints such as chemical cost, water quality risk, and equipment wear, breaking through the limitations of single-index optimization, and generating addition decisions that balance economy and safety under instantaneous pollution impact. Pulse-type addition and dynamic flow velocity binding, combined with a spatiotemporal decoupling compensation strategy, significantly improve the diffusion efficiency of chemicals in turbulent environments, avoiding local concentration accumulation or insufficiency.

Description

Technical Field

[0001] This invention relates to the field of chemical dosing control technology, and in particular to a method and system for controlling the dosing of water treatment chemicals. Background Technology

[0002] Currently, large water parks are generally equipped with clustered high-altitude water slides, which creates intermittent crowd surges, resulting in a large amount of human metabolites (sweat, sebum) and sunscreen chemicals being instantly introduced into local water areas. The pollutant input exhibits pulse-like burst characteristics, and conventional water quality monitoring has a lag of minutes. Furthermore, the hydraulic disturbance in the water landing area of ​​slides is severe, and the diffusion efficiency of chemicals fluctuates significantly due to turbulence. Existing solutions, such as the IoT-based multi-probe dynamic feedback control and adaptive fuzzy PID algorithm (Eco-Dosing system) published in CN117886381B, perform well under steady-state conditions, but are difficult to cope with dynamic loads that change on a second-by-second basis.

[0003] Newer intelligent dosing systems mostly rely on online sensor networks for closed-loop control, but they exhibit significant shortcomings in crowded water slide scenarios. Key indicators such as residual chlorine and ORP sensors need to be immersed at specific depths and maintain stable contact. However, the continuous water flow in the slide's landing area causes severe fluctuations in probe measurements, leading to misjudgments in the control algorithm, such as misreading excessively high residual chlorine levels and stopping dosing. Furthermore, mainstream machine learning prediction models like LSTM rely on historical water quality data for training, but the highly random timing of visitor groups makes it difficult to construct effective temporal features. Some solutions employ high-dose pre-dosing to mitigate instantaneous pollution, resulting in excessive consumption of chemicals during low-load periods. The accumulation of sodium hypochlorite decomposition products causes irritating odors, affecting comfort. Therefore, a new water treatment chemical dosing control scheme is urgently needed to address these issues. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] This invention provides a method and system for controlling the addition of water treatment chemicals, which solves the problem that existing solutions rely on hysteresis sensors and fixed control logic, making it difficult to cope with the impact of pulsed crowds in water parks, resulting in oscillating chemical addition and local water quality loss.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a method for controlling the addition of water treatment agents, comprising,

[0008] Step S1: Real-time acquisition of water pump operating current signal in the target water area, extraction of high-frequency current components in the preset frequency band of the current signal, and generation of current disturbance intensity time sequence.

[0009] Step S2: Input the time series sequence of the current disturbance intensity into the pre-trained dynamic game model and output the baseline coefficient of the dosage of the agent corresponding to the current disturbance characteristics. The dynamic game model is generated based on historical crowd event data and corresponding water quality parameter changes.

[0010] Step S3: Calculate the target dosage of the agent for the current time period based on the agent dosage benchmark coefficient, and control the dosing device to perform pulse dosing operation upstream of the preset water drop area. The duration of the pulse dosing is negatively correlated with the water flow velocity.

[0011] Step S4: Obtain water quality feedback data from sparse sensor nodes downstream of the dosing area. When water quality parameters are detected to deviate from the expected threshold, activate the spatiotemporal decoupling compensation strategy to adjust the spatiotemporal distribution parameters of subsequent dosing pulses.

[0012] In a preferred embodiment of the water treatment agent dosing control method of the present invention, the preset frequency band in step S1 is 20-30Hz, and the extraction of the high-frequency component of the current is achieved by discrete wavelet transform, with the wavelet basis function being Daubechies4.

[0013] As a preferred embodiment of the water treatment agent dosing control method of the present invention, the generation constraints of the Nash equilibrium solution during the training process of the dynamic game model include: agent dosing cost weight, water quality exceeding risk weight, and equipment operation frequency penalty term, wherein the water quality exceeding risk weight is dynamically adjusted according to the time period of the crowd event.

[0014] As a preferred embodiment of the water treatment agent dosing control method of the present invention, wherein: in step S2, based on historical crowd event amplitude and corresponding water quality parameter change data, a first... The state vector for a given time period is used as input to obtain the baseline coefficient for the optimal dosage during that time period. Specific steps include:

[0015] The combined representation of crowd disturbance and water quality status is:

[0016] ,

[0017] in, Indicates the first The intensity of the current disturbance during the time period. Indicates the first Water quality parameter vectors for different time periods Indicates the first The system state vector for a given time period;

[0018] Construct a utility function to find a balance between cost, risk, and frequency. The function formula is as follows:

[0019] ,

[0020] in, This represents the baseline coefficient variable for drug dosage. Indicates the weight of drug dosage costs. Indicates the first Risk weighting of water quality exceeding standards during different time periods This represents the water quality risk measurement function. Represents a vector of water quality parameters. This indicates the penalty weight based on the frequency of device actions. Indicates the frequency of equipment operation. Indicates the first The utility function for a given time period;

[0021] Under the Nash equilibrium condition, let

[0022] ,

[0023] ,

[0024] in, Indicates water quality threshold, Indicates the dose effect coefficient;

[0025] when When the time is right, an analytical equilibrium solution can be obtained:

[0026] ,

[0027] in, Indicates the first The optimal benchmark coefficient for the time period;

[0028] Training a nonlinear mapping model:

[0029] ,

[0030] in, Indicates the preceding The perturbation intensity sequence for each time period, Represents the parameter vector A defined mapping function Represents the baseline coefficients output by the model;

[0031] Obtained through loss minimization training The formula is:

[0032] ,

[0033] in, Indicates training loss, Represents the regularization coefficient. This represents the norm of the model parameters.

[0034] As a preferred embodiment of the water treatment agent dosing control method of the present invention, wherein: in the pulse dosing operation, the single dosing amount satisfy: ,in, Based on the baseline dosage, The instantaneous gradient value of the time series sequence of current disturbance intensity. It is a saturated nonlinear function.

[0035] In a preferred embodiment of the water treatment agent dosing control method of the present invention, step S3, setting the pulse dosing duration is as follows:

[0036] An inverse power law is used to ensure that the duration decreases proportionally as the flow velocity increases, while the sensitivity is adjusted exponentially.

[0037] ,

[0038] in, Indicates the first The duration of a single injection pulse during a given period. Indicates the duration of the reference pulse. Indicates the reference water flow velocity. Indicates the first Water flow rate over a period of time Indicates the influence of flow velocity index. ;

[0039] Using a semi-saturated function for upper and lower limit constraints:

[0040] ,

[0041] in, Indicates the maximum allowable pulse duration. Indicates the half-saturated flow rate. Indicates the saturation curve shape index. .

[0042] As a preferred embodiment of the water treatment agent dosing control method of the present invention, the spatiotemporal decoupling compensation strategy includes:

[0043] The target water area is divided into a hydraulic impact core zone, a transitional diffusion zone, and a marginal stability zone.

[0044] In the core area, a feedforward compensation mode is adopted to directly add a dose of medicine that is proportional to the historical deviation.

[0045] In the transition diffusion region, a feedback smoothing mode is enabled to attenuate and correct the amplitude of the unapplied pulse based on downstream sensor data.

[0046] In a preferred embodiment of the water treatment agent dosing control method of the present invention, in step S4, when the downstream sensing node detects that the water quality parameter deviates from the threshold, the agent dosage is added proportionally according to the historical deviation.

[0047] In a preferred embodiment of the water treatment agent dosing control method of the present invention, step S4, wherein the step of adding an additional agent dosage proportional to the historical deviation includes:

[0048] The formula for quantifying the degree of water quality deviation in the current period is as follows:

[0049] ,

[0050] in, Indicates the first Water quality parameter vectors for different time periods Represents the water quality threshold vector. Indicates the first The instantaneous deviation vector of the time period; when any component When the trigger threshold is exceeded, a compensation strategy is activated. Indicates the index of the water quality parameter vector components;

[0051] The historical offset is accumulated using an exponentially decaying formula, as follows:

[0052] ,

[0053] in, Indicates the first Historical deviation vector over a period of time This indicates a deviation from the attenuation coefficient, and , Indicates the first Historical deviation over a period of time Indicates the first Instantaneous deviation over a given period;

[0054] The historical deviation is linearly proportional to the compensation dose. The proportionality coefficient is obtained using offline regression calibration and is expressed as follows:

[0055] ,

[0056] in, Indicates the first The vector of additional drug doses required during the specified time period. Represents the compensation ratio coefficient matrix;

[0057] Through historical samples Perform regularized least squares estimation:

[0058] ,

[0059] in, Indicates the first The actual compensation dose for each sample This represents the regularization coefficient used to prevent overfitting. For sample index.

[0060] Secondly, the present invention provides a water treatment agent dosing control system, comprising,

[0061] The current characteristic analysis module is used to perform current signal processing;

[0062] The game decision engine module, which incorporates the dynamic game model, is used to calculate the baseline coefficient for the amount of feed added.

[0063] The spatiotemporal dosing execution module includes multiple independently controllable dosing nozzle arrays for pulse dosing and compensation strategies.

[0064] The sparse sensor network module consists of at least three miniature water quality sensor nodes deployed downstream of the water flow, and the sensor nodes transmit data between each other via an ad hoc network protocol.

[0065] The beneficial effects of this invention are as follows: This invention replaces traditional water quality sensors with the high-frequency characteristics of water pump current, effectively overcoming the monitoring lag caused by the impact of water flow in the slide's landing area, and achieving second-level detection of crowd events; the dynamic game model integrates multiple objective constraints such as reagent cost, water quality risk, and equipment wear and tear, breaking through the limitations of single-index optimization, and generating dosing decisions that take into account both economy and safety under instantaneous pollution impact; pulsed dosing and dynamic binding of flow velocity, combined with a spatiotemporal decoupling compensation strategy, significantly improve the diffusion efficiency of reagents in turbulent environments, avoiding local concentration accumulation or insufficiency; the system uses sparse sensor network feedback to drive cross-regional collaborative control, reducing the repeated consumption caused by independent dosing in multiple pools, while relying on the historical deviation accumulation mechanism to enhance the response sensitivity to hidden pollution events. Attached Figure Description

[0066] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0067] Figure 1 This is a schematic flowchart of the dosing control method for water treatment agents in Example 1.

[0068] Figure 2 This is a schematic diagram of the framework of the water treatment agent dosing control system in Example 1. Detailed Implementation

[0069] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0070] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0071] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0072] Example 1, referring to Figure 1 and Figure 2 This embodiment provides a method for controlling the addition of water treatment chemicals, including the following steps:

[0073] Step S1: Real-time acquisition of water pump operating current signal in the target water area, extraction of high-frequency current components in the preset frequency band of the current signal, and generation of current disturbance intensity time series.

[0074] In step S1, the preset frequency band is 20-30Hz, and the extraction of the high-frequency components of the current is achieved by discrete wavelet transform, with the wavelet basis function being Daubechies4.

[0075] Step S2: Input the time series sequence of current disturbance intensity into the pre-trained dynamic game model and output the baseline coefficient of the dosage of the agent corresponding to the current disturbance characteristics. The dynamic game model is generated based on historical human tide event data and corresponding water quality parameter changes.

[0076] During the training process of the dynamic game model, the constraints for generating the Nash equilibrium solution include: the weight of the cost of adding the reagent, the weight of the risk of exceeding the water quality standard, and the penalty term for the frequency of equipment operation. Among them, the weight of the risk of exceeding the water quality standard is dynamically adjusted according to the time period of the crowd event.

[0077] In step S2, based on historical data on the magnitude of crowd events and corresponding changes in water quality parameters, the first... The state vector for a given time period is used as input to obtain the baseline coefficient for the optimal dosage during that time period. Specific steps include:

[0078] The combined representation of crowd disturbance and water quality status is:

[0079] ,

[0080] in, Indicates the first The intensity of the current disturbance during the time period. Indicates the first Water quality parameter vectors for different time periods Indicates the first The system state vector for a given time period;

[0081] Construct a utility function to find a balance between cost, risk, and frequency. The function formula is as follows:

[0082] ,

[0083] in, This represents the baseline coefficient variable for drug dosage. Indicates the weight of drug dosage costs. Indicates the first Risk weighting of water quality exceeding standards during different time periods This represents the water quality risk measurement function. Represents a vector of water quality parameters. This indicates the penalty weight based on the frequency of device actions. Indicates the frequency of equipment operation. Indicates the first The utility function for a given time period;

[0084] Under the Nash equilibrium condition, let

[0085] ,

[0086] ,

[0087] in, Indicates water quality threshold, Indicates the dose effect coefficient;

[0088] when When the time is right, an analytical equilibrium solution can be obtained:

[0089] ,

[0090] in, Indicates the first The optimal benchmark coefficient for the time period;

[0091] Training a nonlinear mapping model:

[0092] ,

[0093] in, Indicates the preceding The perturbation intensity sequence for each time period, Represents the parameter vector A defined mapping function Represents the baseline coefficients output by the model;

[0094] Obtained through loss minimization training The formula is:

[0095] ,

[0096] in, Indicates training loss, Represents the regularization coefficient. Represents the norm of the model parameters;

[0097] Specifically, a state vector is introduced to effectively integrate current disturbances with water quality parameters, providing accurate input for dynamic game theory. The utility function design takes into account economic costs, water quality risks, and equipment usage frequency. The Nash equilibrium solution ensures that the injection decision achieves a reasonable trade-off among multiple objectives. The combination of linearization of the risk measure function and analytical equilibrium solution greatly reduces the online computational complexity. Subsequently, the nonlinear mapping model based on the analytical solution achieves rapid response and robust generalization to unknown disturbance sequences through a least squares plus regularization training strategy.

[0098] Step S3: Calculate the target dosage of the agent for the current time period based on the agent dosage benchmark coefficient, and control the dosing device to perform pulse dosing operation upstream of the preset water drop area. The duration of pulse dosing is negatively correlated with the water flow velocity.

[0099] In pulse dosing operations, the amount of substance added in a single operation is... satisfy: ,in, Based on the baseline dosage, The instantaneous gradient value of the time series sequence of current disturbance intensity. It is a saturated nonlinear function;

[0100] Step S3, the method for setting the pulse dosing duration is as follows:

[0101] An inverse power law is used to ensure that the duration decreases proportionally as the flow velocity increases, while the sensitivity is adjusted exponentially.

[0102] ,

[0103] in, Indicates the first The duration of a single injection pulse during a given period. Indicates the duration of the reference pulse. Indicates the reference water flow velocity. Indicates the first Water flow rate over a period of time Indicates the influence of flow velocity index. ;

[0104] Using a semi-saturated function for upper and lower limit constraints:

[0105] ,

[0106] in, Indicates the maximum allowable pulse duration. Indicates the half-saturated flow rate. Indicates the saturation curve shape index. .

[0107] Specifically, by combining inverse power law and saturated nonlinearity, the duration of the applied pulses can be precisely controlled at different flow velocities. The inverse power law relationship predominates, ensuring a simple and adjustable negative correlation between pulse duration and flow velocity in the low-to-medium velocity range. The model can be flexibly configured according to the requirements for uniformity of dosing. The saturation model provides upper and lower limit protection against extreme flow velocities, preventing excessively long or short dosing times when the flow rate is abnormal, thus improving safety and stability. Both models are described with a small number of parameters, making them easy to calibrate and identify online. Combined with the aforementioned benchmark coefficients... and nonlinear functions A complete flow rate adaptive dosing strategy is formed to effectively balance the efficiency of reagent utilization and the requirements for water quality compliance, while reducing wear and energy consumption caused by frequent start-up and shutdown of equipment.

[0108] Step S4: Obtain water quality feedback data from sparse sensor nodes downstream of the dosing area. When water quality parameters are detected to deviate from the expected threshold, activate the spatiotemporal decoupling compensation strategy to adjust the spatiotemporal distribution parameters of subsequent dosing pulses.

[0109] Spatiotemporal decoupling compensation strategies include:

[0110] The target water area is divided into a hydraulic impact core zone, a transitional diffusion zone, and a marginal stability zone.

[0111] In the core area, a feedforward compensation mode is adopted to directly add a dose of medicine that is proportional to the historical deviation.

[0112] In the transition diffusion region, a feedback smoothing mode is activated to attenuate and correct the amplitude of the unadded pulse based on downstream sensor data.

[0113] In step S4, when the downstream sensor node detects that the water quality parameter deviates from the threshold, the dosage of the agent is increased proportionally according to the historical deviation.

[0114] Step S4, which involves adding a dose of medication proportional to the historical deviation, includes:

[0115] The formula for quantifying the degree of water quality deviation in the current period is as follows:

[0116] ,

[0117] in, Indicates the first Water quality parameter vectors for different time periods Represents the water quality threshold vector. Indicates the first The instantaneous deviation vector of the time period; when any component When the trigger threshold is exceeded, a compensation strategy is activated. Indicates the index of the water quality parameter vector components;

[0118] The historical offset is accumulated using an exponentially decaying formula, as follows:

[0119] ,

[0120] in, Indicates the first Historical deviation vector over a period of time This indicates a deviation from the attenuation coefficient, and , Indicates the first Historical deviation over a period of time Indicates the first Instantaneous deviation over a given period;

[0121] The historical deviation is linearly proportional to the compensation dose. The proportionality coefficient is obtained using offline regression calibration and is expressed as follows:

[0122] ,

[0123] in, Indicates the first The vector of additional drug doses required during the specified time period. Represents the compensation ratio coefficient matrix;

[0124] Through historical samples Perform regularized least squares estimation:

[0125] ,

[0126] in, Indicates the first The actual compensation dose for each sample This represents the regularization coefficient used to prevent overfitting. For sample index;

[0127] Specifically, by exponentially smoothing and accumulating the instantaneous deviation, the impact of recent large deviation events can be preserved while gradually reducing the weight of long-term data, thus realizing a mechanism for dynamic memory of water quality deviations. When the deviation exceeds the threshold, the linear proportional coefficient is used to quickly compensate for the historical cumulative deviation, ensuring the system's timely response to sudden water quality fluctuations. The offline regression calibration of the proportional coefficient, while satisfying the data fit, provides robust and reliable parameters for online real-time compensation by controlling the size and stability of the coefficient through regularization.

[0128] This embodiment also provides a water treatment agent dosing control system, including:

[0129] The current characteristic analysis module is used to perform current signal processing;

[0130] The game decision engine module has a built-in dynamic game model for calculating the baseline coefficient of the injection amount.

[0131] The spatiotemporal dosing execution module includes multiple independently controllable dosing nozzle arrays for pulse dosing and compensation strategies.

[0132] The sparse sensor network module consists of at least three miniature water quality sensor nodes deployed downstream of the water flow, with data transmitted between the sensor nodes via an ad hoc network protocol.

[0133] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for controlling the addition of water treatment chemicals, characterized in that, include, Step S1: Real-time acquisition of water pump operating current signal in the target water area, extraction of high-frequency current components in the preset frequency band of the current signal, and generation of current disturbance intensity time sequence. Step S2: Input the time series sequence of the current disturbance intensity into the pre-trained dynamic game model and output the baseline coefficient of the dosage of the agent corresponding to the current disturbance characteristics. The dynamic game model is generated based on historical crowd event data and corresponding water quality parameter changes. Step S3: Calculate the target dosage of the agent for the current time period based on the agent dosage benchmark coefficient, and control the dosing device to perform pulse dosing operation upstream of the preset water drop area. The duration of the pulse dosing is negatively correlated with the water flow velocity. Step S4: Obtain water quality feedback data from sparse sensor nodes downstream of the dosing area. When water quality parameters are detected to deviate from the expected threshold, activate the spatiotemporal decoupling compensation strategy to adjust the spatiotemporal distribution parameters of subsequent dosing pulses. The spatiotemporal decoupling compensation strategy includes: The target water area is divided into a hydraulic impact core zone, a transitional diffusion zone, and a marginal stability zone. In the core area, a feedforward compensation mode is adopted to directly add a drug dosage that is proportional to the historical deviation. In the transition diffusion region, a feedback smoothing mode is activated to attenuate and correct the amplitude of the unadded pulse based on downstream sensor data. In step S4, when the downstream sensor node detects that the water quality parameter deviates from the threshold, the dosage of the agent is increased proportionally according to the historical deviation. In step S4, the step of adding an additional dose of medicine proportional to the historical deviation includes: The formula for quantifying the degree of water quality deviation in the current period is as follows: ， in, Indicates the first Water quality parameter vectors for different time periods Represents the water quality threshold vector. Indicates the first The instantaneous deviation vector of the time period; when any component When the trigger threshold is exceeded, a compensation strategy is activated. Indicates the index of the water quality parameter vector components; The historical offset is accumulated using an exponentially decaying formula, as follows: ， in, Indicates the first Historical deviation vector over a period of time This indicates a deviation from the attenuation coefficient, and , Indicates the first Historical deviation over a period of time Indicates the first Instantaneous deviation over a given period; The historical deviation is linearly proportional to the compensation dose. The proportionality coefficient is obtained using offline regression calibration and is expressed as follows: ， in, Indicates the first The vector of additional drug doses required during the specified time period. Represents the compensation ratio coefficient matrix; Through historical samples Perform regularized least squares estimation: ， in, Indicates the first The actual compensation dose for each sample This represents the regularization coefficient used to prevent overfitting. For sample index.

2. The method for controlling the addition of water treatment agents as described in claim 1, characterized in that, In step S1, the preset frequency band is 20-30Hz, and the extraction of the high-frequency components of the current is achieved by discrete wavelet transform, with the wavelet basis function being Daubechies4.

3. The method for controlling the addition of water treatment agents as described in claim 1, characterized in that, During the training process of the dynamic game model, the constraints for generating the Nash equilibrium solution include: the weight of the cost of adding the reagent, the weight of the risk of exceeding the water quality standard, and the penalty term for the frequency of equipment operation. The weight of the risk of exceeding the water quality standard is dynamically adjusted according to the time period of the crowd event.

4. The method for controlling the addition of water treatment agents as described in claim 3, characterized in that, In step S2, based on historical data on the magnitude of crowd events and corresponding changes in water quality parameters, the first... The state vector for a given time period is used as input to obtain the baseline coefficient for the optimal dosage during that time period. Specific steps include: The combined representation of crowd disturbance and water quality status is: ， in, Indicates the first The intensity of the current disturbance during the time period. Indicates the first Water quality parameter vectors for different time periods Indicates the first The system state vector for a given time period; Construct a utility function to find a balance between cost, risk, and frequency. The function formula is as follows: ， in, This represents the baseline coefficient variable for drug dosage. Indicates the weight of drug dosage costs. Indicates the first Risk weighting of water quality exceeding standards during different time periods This represents the water quality risk measurement function. Represents a vector of water quality parameters. This indicates the penalty weight based on the frequency of device actions. Indicates the frequency of equipment operation. Indicates the first The utility function for a given time period; Under the Nash equilibrium condition, let ， ， in, Indicates water quality threshold, Indicates the dose effect coefficient; when When the time is right, an analytical equilibrium solution can be obtained: ， in, Indicates the first The optimal benchmark coefficient for the time period; Training a nonlinear mapping model: ， in, Indicates the preceding The perturbation intensity sequence for each time period, Represents the parameter vector A defined mapping function Represents the baseline coefficients output by the model; Obtained through loss minimization training The formula is: ， in, Indicates training loss, Represents the regularization coefficient. This represents the norm of the model parameters.

5. The method for controlling the addition of water treatment agents as described in claim 1, characterized in that, In the pulsed dosing operation, the amount of substance added in a single dose... satisfy: ,in, Based on the baseline dosage, The instantaneous gradient value of the time series sequence of current disturbance intensity. It is a saturated nonlinear function.

6. The method for controlling the addition of water treatment agents as described in claim 5, characterized in that, Step S3, the method for setting the pulse dosing duration is as follows: An inverse power law is used to ensure that the duration decreases proportionally as the flow velocity increases, while the sensitivity is adjusted exponentially. ， in, Indicates the first The duration of a single injection pulse during a given period. Indicates the duration of the reference pulse. Indicates the reference water flow velocity. Indicates the first Water flow rate over a period of time Indicates the influence of flow velocity index. ; Using a semi-saturated function for upper and lower limit constraints: ， in, Indicates the maximum allowable pulse duration. Indicates the half-saturated flow rate. Indicates the saturation curve shape index. .

7. A dosing control system for a water treatment agent, based on the dosing control method for a water treatment agent according to any one of claims 1 to 6, characterized in that, include: The current characteristic analysis module is used to perform current signal processing; The game decision engine module, which incorporates the dynamic game model, is used to calculate the baseline coefficient for the amount of feed added. The spatiotemporal dosing execution module includes multiple independently controllable dosing nozzle arrays for pulse dosing and compensation strategies. The sparse sensor network module consists of at least three miniature water quality sensor nodes deployed downstream of the water flow, and the sensor nodes transmit data between each other via an ad hoc network protocol.

Images