Cooperative disinfection regulation and control method and system based on multi-modal water quality perception

Abstract

The invention discloses a collaborative disinfection regulation and control method and system based on multi-modal water quality perception, and relates to the technical field of collaborative disinfection regulation and control, and the method comprises the following steps: constructing a disinfectant diffusion model, optimizing the disinfectant diffusion model, correcting simulation deviation, and generating a compensation data set; the optimized disinfectant diffusion model is obtained by optimizing an alignment constraint curved surface of multi-modal data through a disinfectant concentration spatial and temporal distribution thermodynamic diagram generated by the disinfectant diffusion model and according to gradient change characteristics of the thermodynamic diagram; and calculating the microbial inactivation rate and the residual quantity of the disinfectant in real time based on the compensation data, and regulating and controlling the disinfectant adding scheme in real time according to the deviation. According to the method, the alignment constraint curved surface of the multi-modal data is constructed based on the gradient characteristics of the thermodynamic diagram, so that accurate alignment of different modal data is realized, and the accuracy and efficiency of disinfection regulation and control are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of collaborative disinfection control, and more specifically, to a collaborative disinfection control method and system based on multi-modal water quality perception. Background Art

[0002] Water quality monitoring and disinfection control are crucial technologies in the field of water treatment, especially in the treatment of public drinking water and industrial water. With the continuous development of water quality monitoring technologies, more and more sensors and detection means are applied to real-time monitor the physical, chemical, and biological characteristics of water bodies. Existing water quality monitoring methods usually rely on a single type of data, such as physicochemical sensor data, spectral data, or microscopic imaging data. However, due to the characteristic differences of different data sources, these single-modal data often cannot comprehensively reflect the actual situation of water bodies. Therefore, how to efficiently fuse multi-modal data and improve the accuracy of water quality monitoring and disinfection control has become a technical problem to be solved urgently.

[0003] In recent years, methods based on multi-modal data fusion have been widely studied and applied. By integrating data from different sensing means, the dynamic changes of water bodies can be more accurately reflected, and then the dosage and distribution of disinfectants can be optimized. However, when existing fusion methods implement cross-modal data docking, they often ignore the problem of non-linear transformation of the feature space. Traditional fusion methods usually assume that data of each modality can be directly aligned in the same feature space, but in practical applications, different modalities of water quality data (such as physicochemical sensors, spectra, and microscopic imaging data) often exhibit significant non-linear characteristics under different spatial scales, measurement methods, and data distributions. Therefore, ignoring these non-linear feature transformations will lead to insufficient accuracy of data fusion and further affect the performance of the disinfection control system.

[0004] In the process of water quality monitoring and disinfection control, the accuracy of data is crucial. Especially in a complex water body environment, the changes in water quality are dynamic and affected by multiple factors (such as water flow, pollutant diffusion, microbial community, etc.). Therefore, how to overcome the limitations of existing methods, accurately process and fuse cross-modal data, and improve the accuracy of water quality perception and disinfection control has become the key to current technical research.

[0005] For example, a method, system and medium for regulating the water quality of shrimp farming disclosed in the invention patent announcement with the publication number of CN115793585A belong to the technical field of regulating the water quality of shrimp farming. The present invention constructs a water quality change prediction model, predicts the relevant harmful substance data of the current shrimp farming area according to the bait plan feeding amount data information and the water quality change prediction model to obtain a prediction result; formulates a water quality regulation plan for each shrimp growth stage according to the prediction result, and obtains the real-time colony structure information and algae data information in the current shrimp farming area; optimizes and adjusts the water quality regulation plan based on the real-time colony structure information and algae data information in the current shrimp farming area. By this method, water quality data can be predicted according to feeding data, and then ammonia nitrogen and nitrite nitrogen can be effectively controlled based on the ecological function data and synergistic effects of microalgae, heterotrophic bacteria and nitrifying bacteria in the farming area.

[0006] For example, a method for joint regulation and control of the biological safety of pipe network water quality based on real-time ArcGis disclosed in the invention patent announcement with the publication number of CN102306021A has the following steps: (1) Import the residual chlorine data of each water quality monitoring point in the water supply pipe network into the geographic information system; (2) Analyze the spatial variation law of the attenuation of disinfectants in the current water in the transmission and distribution pipe network with reference to the distribution of water plants, secondary pressurization pump stations and transmission and distribution pipe networks, and conduct regulation of the residual chlorine in the pipe network; (3) Conduct a secondary analysis of the distribution of residual chlorine in the pipe network, and observe whether the most unfavorable point in the transmission and distribution of the pipe network meets the minimum value of residual chlorine required for model calculation to control the microbial index to be qualified; (4) If the most unfavorable point in the transmission and distribution of the pipe network still does not meet the minimum value required by the model, repeat step (2) until the distribution of residual chlorine in the pipe network basically meets the control conditions of the microbial index at the most unfavorable point in the transmission and distribution of the pipe network. The present invention realizes the compliance of the microbial index of the pipe network and makes the spatial and temporal distribution of the residual chlorine in the pipe network more uniform, reducing the dosage of disinfectants and the generation amount of disinfection by-products.

[0007] In the above disclosed technical solutions, there are at least the following technical problems: Most of the existing water quality monitoring and disinfection control systems are based on single-modal data, such as traditional sensor data or spectral data. However, these methods have certain limitations. Especially due to the complexity of water quality, single-modal data is difficult to comprehensively and accurately reflect all the quality indicators of water bodies. And the existing fusion methods do not fully consider the non-linear transformation of the feature space, resulting in insufficient fusion accuracy of cross-modal data, which affects the effect of disinfection regulation.

[0008] In view of the above problems, the present invention proposes a solution. Summary of the Invention

[0009] To overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a collaborative disinfection regulation method and system based on multimodal water quality perception. By constructing an alignment constraint surface for multimodal data based on the gradient features of the heat map, it solves the problem that the existing fusion methods do not fully consider the non-linear transformation of the feature space, resulting in insufficient fusion accuracy of cross-modal data.

[0010] To achieve the above object, the present invention provides the following technical solutions: A collaborative disinfection regulation method based on multimodal water quality perception, comprising the following steps: constructing a disinfectant diffusion model, optimizing the disinfectant diffusion model, correcting the simulation deviation, and generating a compensation data set; the optimized disinfectant diffusion model is obtained by generating a heat map of the spatio-temporal distribution of the disinfectant concentration by the disinfectant diffusion model, and constructing an alignment constraint surface for multimodal data according to the gradient change characteristics of the heat map; calculating the microbial inactivation rate and the disinfectant residue in real time based on the compensation data, and regulating the disinfectant dosing scheme in real time according to the deviation.

[0011] In a preferred embodiment, the heat map of the spatio-temporal distribution of the disinfectant concentration is specifically: dividing the water body into several grid units, setting the size of each grid unit according to the simulation accuracy requirement, and setting initial conditions and boundary conditions for each grid unit; solving the diffusion of the disinfectant in the water body based on a fluid dynamics simulation software to obtain the disinfectant concentration values of each grid unit at different time steps; according to the calculation results of each time step, deriving the disinfectant concentration values of each grid unit at the corresponding time points to form a complete spatio-temporal distribution data set; based on the data visualization tool, converting the spatial distribution result of the disinfectant concentration in the spatio-temporal distribution data set into a heat map.

[0012] In a preferred embodiment, constructing the alignment constraint surface for multimodal data according to the gradient change characteristics of the heat map is specifically: calculating the temporal gradient and spatial gradient of the concentration for each spatio-temporal point, and performing normalization processing to generate a gradient weight matrix; mapping the multimodal data to the spatio-temporal points and spatial coordinates respectively, and calculating the time difference and spatial Euclidean distance between each modal data point and the nearest spatio-temporal grid point of the heat map; defining alignment constraint conditions according to the gradient weight matrix, time difference and spatial Euclidean distance, and performing local alignment on the multimodal data; based on the global optimization method of gradient descent, adjusting the spatial distribution of each modal data to minimize the deviation between the spatial distribution and the constraint surface to obtain the final aligned data set.

[0013] In a preferred embodiment, the multimodal data includes physicochemical sensor data, spectral data, and microscopic imaging data. The local alignment of the multimodal data specifically includes: aligning the physicochemical sensor data with the concentration gradient region in the heat map by optimizing the coordinates of the sensor measurement points; matching the spectral signals in spatial positions according to the absorption characteristics of the disinfectant at different wavelengths to ensure their consistency with the concentration gradient region; and aligning the microbial aggregation area in the microscopic image with the high-gradient region of the disinfectant concentration through an image registration algorithm.

[0014] In a preferred embodiment, optimizing the disinfectant diffusion model, correcting the simulation deviation, and generating an adjusted data set specifically involves: performing forward correction on the disinfectant diffusion model based on the aligned data set to obtain an optimized disinfectant diffusion model; reversely compensating and correcting the attenuation function of the spectral data and constructing a turbidity compensation matrix according to the optimized disinfectant diffusion model to obtain compensated spectral data; calculating the disinfectant concentration at each moment and each spatial position in the water body based on the optimized diffusion model and the compensated spectral data, and generating a new spatio-temporal distribution data set of the disinfectant concentration.

[0015] In a preferred embodiment, the forward correction of the disinfectant diffusion model specifically involves: constructing a deviation function based on the aligned multimodal data and the prediction results of the preliminary disinfectant diffusion model; adjusting the parameters of the diffusion model with the goal of minimizing the deviation function based on the genetic algorithm; and updating the disinfectant diffusion model according to the optimized model parameters to generate new spatio-temporal distribution data of the disinfectant concentration.

[0016] In a preferred embodiment, reversely compensating and correcting the attenuation function of the spectral data and constructing a turbidity compensation matrix according to the optimized disinfectant diffusion model specifically involves: establishing a preliminary spectral attenuation model and updating the spectral attenuation model, where the updated spectral attenuation model is updated by identifying the deviation through comparing the optimized disinfectant concentration distribution with the actually observed spectral data; analyzing the influence of the water turbidity in different regions on the spectral and microscopic imaging data, and constructing a turbidity compensation matrix in combination with the disinfectant concentration; applying the compensation matrix to the spectral and microscopic imaging data for reverse compensation to obtain corrected spectral data and generating a new compensated data set.

[0017] In a preferred embodiment, the method calculates the microbial inactivation rate and the disinfectant residue in real time based on compensation data, and adjusts the disinfectant dosing scheme in real time according to the deviation. Specifically, it calculates the disinfectant residue in the water body and the microbial inactivation rate based on the compensation data and a preset microbial inactivation model. The compensation data includes disinfectant concentration data. If there is a deviation between the microbial inactivation rate and the preset target inactivation rate, the control system automatically adjusts the dosing amount and position of the disinfectant. If the disinfectant residue exceeds the preset safety range, the multi-objective particle swarm optimization algorithm is used to optimize the two objectives of the microbial inactivation rate and the disinfectant residue simultaneously, and dynamically adjust the dosing amount and position of the disinfectant under different control constraints.

[0018] Technical effects and advantages of the collaborative disinfection control method and system based on multi-modal water quality perception of the present invention: 1. By monitoring the disinfectant dosing amount and water body flow parameters in real time and combining with hydrodynamic simulation, the present invention generates a spatio-temporal distribution heat map of the disinfectant concentration. This method can accurately simulate the diffusion behavior of the disinfectant in water and dynamically adjust the disinfectant dosing amount according to real-time data to ensure the best disinfection effect. This precise control can effectively reduce resource waste and improve the efficiency of water quality treatment.

[0019] 2. By constructing an alignment constraint surface for multi-modal data based on the gradient characteristics of the heat map, the present invention realizes the precise alignment of different modal data. This method solves the problem of insufficient cross-modal data fusion accuracy in the prior art. Through global optimization and local alignment, it ensures a high degree of consistency of different types of data and improves the accuracy and efficiency of disinfection control. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a schematic flow chart of the collaborative disinfection control method based on multi-modal water quality perception of the present invention; Figure 2 is a schematic structural diagram of the collaborative disinfection control system based on multi-modal water quality perception of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0022] Embodiment 1 Figure 1 The collaborative disinfection control method based on multi-modal water quality perception of the present invention is given, including the following steps: S1. Construct a disinfectant diffusion model, optimize the disinfectant diffusion model, correct the simulation deviation, and generate a compensation data set. The optimized disinfectant diffusion model is obtained by generating a heat map of the spatio-temporal distribution of disinfectant concentration through the disinfectant diffusion model and constructing an alignment constraint surface for multi-modal data based on the gradient change characteristics of the heat map. The disinfectant diffusion model is constructed based on the real-time disinfectant dosage and water body flow parameters.

[0023] Disinfectant dosage: Collect real-time disinfectant dosage data. Usually, the flow rate and concentration of disinfectant are monitored by sensors, or the input amount of disinfectant each time is recorded through an automated system.

[0024] Water body flow parameters: Collect relevant parameters of water body flow through flow velocity sensors, fluid mechanics simulation, or on-site measurement data, including flow velocity, flow direction, temperature, turbulence intensity, etc. These parameters are crucial for accurately simulating the diffusion of disinfectant in water.

[0025] The generated heat map of the spatio-temporal distribution of disinfectant concentration is specifically as follows: Collect real-time disinfectant dosage and water body flow parameter data, including flow velocity, flow direction, temperature, turbulence intensity, disinfectant injection rate, and injection concentration parameters in the water body. Establish a fluid mechanics model, simulate the diffusion behavior of disinfectant in the water body through computational fluid dynamics, and discretize and solve the disinfectant diffusion equation based on the finite difference method to generate spatio-temporal distribution data of disinfectant concentration. The model includes a disinfectant diffusion equation, a turbulence model, and boundary condition settings. Based on the spatio-temporal distribution data, using a data visualization tool, convert the spatial distribution result of disinfectant concentration into a heat map, where different concentration values are represented by different colors for intuitive observation of the distribution of disinfectant in the water body.

[0026] Boundary conditions: Include the inlet and outlet conditions of the water body, as well as the reflection or absorption conditions of the surface boundary and bottom boundary of the water body.

[0027] The generation of spatio-temporal distribution data of disinfectant concentration is specifically as follows: Divide the water body into several grid cells, set the size of each grid cell according to the simulation accuracy requirements, and set initial conditions and boundary conditions for each grid cell. The initial conditions and boundary conditions include the initial concentration of disinfectant, flow velocity, flow direction, and other environmental parameters. Based on fluid mechanics simulation software, solve the diffusion of disinfectant in the water body to obtain the disinfectant concentration values of each grid cell at different time steps. According to the calculation results at each time step, the disinfectant concentration values of each grid cell at the corresponding time points are derived to form a complete spatio-temporal distribution data set.

[0028] The disinfectant diffusion model is specifically as follows:

[0029] Among them, is the disinfectant concentration, is the diffusion time, is the water velocity vector, is the diffusion coefficient, is the disinfectant dosing source.

[0030] The disinfectant dosing source represents all factors that cause changes in the disinfectant concentration except for convection (water body flow) and diffusion (molecular diffusion or turbulent diffusion).

[0031] The multi-modal data includes physicochemical sensor data, spectral data, and microscopic imaging data; Constructing an alignment constraint surface for multi-modal data according to the gradient change characteristics of the heat map is specifically as follows: Calculate the temporal gradient and spatial gradient of the concentration at each spatio-temporal point and perform normalization processing to generate a gradient weight matrix; Map the multi-modal data to spatio-temporal points and spatial coordinates respectively, and calculate the time difference and spatial Euclidean distance between each modal data point and the nearest spatio-temporal grid point of the heat map; Define alignment constraint conditions according to the gradient weight matrix, time difference, and spatial Euclidean distance, and perform local alignment on the multi-modal data; Based on the global optimization method of gradient descent, adjust the spatial distribution of each modal data to minimize the deviation between the spatial distribution and the constraint surface, and obtain the final aligned data set.

[0032] The global optimization method is quantified by the following loss function, specifically as follows:

[0033] Among them, is the alignment result of the j-th modal data, is the ideal result of the target constraint surface, is the adjustment parameter, is the constraint surface deviation.

[0034] Defining alignment constraint conditions according to the time difference and spatial Euclidean distance is specifically as follows: Temporal alignment constraint:

[0035] Spatial alignment constraint: Gradient consistency constraint: Among them, is the time deviation, is the basic time tolerance, is the gradient weight coefficient, is the gradient amplitude, is the spatial distance, is the basic spatial tolerance, is the gradient sensitivity coefficient, is the gradient of modal data features, is the preset direction consistency threshold.

[0036] The local alignment of the multi-modal data is specifically as follows: For physicochemical sensor data, by optimizing the coordinates of the sensor measurement points, align them with the concentration gradient region in the heat map; For spectral data, according to the absorption characteristics of the disinfectant at different wavelengths, match the spectral signals in the spatial position to ensure their consistency with the concentration gradient region; For microscopic imaging data, align the microbial aggregation area in the microscopic image with the high gradient area of the disinfectant concentration through an image registration algorithm.

[0037] The optimization of the disinfectant diffusion model, correction of the simulation deviation, and generation of a compensation data set are specifically as follows: Based on the aligned data set, perform forward correction on the disinfectant diffusion model to obtain an optimized disinfectant diffusion model; According to the optimized disinfectant diffusion model, reverse-compensate and correct the attenuation function of the spectral data and construct a turbidity compensation matrix to obtain compensated spectral data; Based on the optimized diffusion model and the compensated spectral data, calculate the disinfectant concentration at each moment and each spatial position in the water body to generate a new spatio-temporal distribution data set of the disinfectant concentration.

[0038] The forward correction of the disinfectant diffusion model is specifically as follows: According to the prediction results of the aligned multi-modal data and the preliminary disinfectant diffusion model, construct a deviation function; Based on the genetic algorithm, with the goal of minimizing the deviation function, adjust the parameters of the diffusion model; According to the optimized model parameters, update the disinfectant diffusion model to generate new spatio-temporal distribution data of the disinfectant concentration.

[0039] The adjustment of the parameters of the diffusion model includes: Diffusion coefficient: Adjust the diffusion rate of the disinfectant in the water body according to the comparison of multi-modal data; Reaction rate: Optimize the reaction process rate of the disinfectant in the water body; Turbulence effect parameter: To correct the non-uniform diffusion of disinfectant caused by turbulence, a turbulence diffusion model is usually used to capture the impact of turbulence on the disinfectant.

[0040] The deviation function is specifically:

[0041] Where is the deviation function, is the predicted concentration of the diffusion model at spatial point and time step and is the observed concentration of the multimodal data.

[0042] According to the optimized disinfectant diffusion model, the attenuation function of the spectral data is compensated backward and a turbidity compensation matrix is constructed, specifically: Analyze the spectral attenuation effect in the water body and establish a preliminary spectral attenuation model based on the water body characteristics; the water body characteristics include turbidity, particulate matter, and dissolved substance concentration; By comparing the optimized disinfectant concentration distribution and the actually observed spectral data, identify the deviation in the spectral data, and update the attenuation model according to the optimized disinfectant concentration; According to the optimized disinfectant diffusion model, analyze the influence of turbidity in different regions of the water body on the spectral and microscopic imaging data; Combining the turbidity of the water body and the optimized disinfectant concentration, construct a turbidity compensation matrix; Apply the compensation matrix to the spectral data and microscopic imaging data for backward compensation to obtain the compensated spectral data; Based on the corrected attenuation function and the compensated spectral data, generate a new compensated data set.

[0043] The preliminary spectral attenuation model is specifically:

[0044] The updated attenuation model is specifically:

[0045] Where is the preliminary spectral attenuation model, is the spectral wavelength, is the attenuation coefficient, indicating the attenuation intensity of light in water, is the disinfectant concentration, is a preset correction coefficient, is the difference between the optimized disinfectant concentration and the measured value, is the corrected attenuation function.

[0046] S2. Calculate the microbial inactivation rate and the disinfectant residue in real time based on the compensation data, and adjust the disinfectant dosing plan in real time according to the deviation.

[0047] The calculation of the microbial inactivation rate and the disinfectant residue in real time based on the compensation data, and the real-time adjustment of the disinfectant dosing plan according to the deviation are specifically as follows: Calculate the disinfectant residue in the water body and the inactivation rate of microorganisms based on the compensation data and a preset microbial inactivation model. The compensation data includes disinfectant concentration data. If there is a deviation between the microbial inactivation rate and the preset target inactivation rate, the control system automatically adjusts the dosing amount and position of the disinfectant. If the disinfectant residue exceeds the preset safety range, then optimize the two objectives of the microbial inactivation rate and the disinfectant residue simultaneously based on the multi-objective particle swarm optimization algorithm, and dynamically adjust the dosing amount and position of the disinfectant under different control constraints.

[0048] The control system automatically adjusts the dosing amount and position of the disinfectant, specifically as follows: The control system combines PID control and regional optimization strategies for linkage adjustment. Linkage control algorithm: Use an integrated optimization algorithm (such as a multi-objective optimization algorithm) to optimize the dosing position while adjusting the dosing amount, so that the dosing amount and position of the disinfectant complement each other to achieve the best inactivation effect and disinfectant residue control.

[0049] Furthermore, the target area (where the microbial density is relatively high) may require more disinfectant, so the dosing amount can be increased while ensuring that the disinfectant concentration in this area is high enough. For the area with excessive reaction, the dosing amount can be reduced while adjusting its dosing position to avoid over-disinfection.

[0050] Example 2. A collaborative disinfection control system based on multi-modal water quality perception includes the following modules: Disinfectant model construction and data alignment module: Used to construct a disinfectant diffusion model, optimize the disinfectant diffusion model, correct the simulation deviation, and generate a compensation data set. The optimized disinfectant diffusion model is obtained by generating a heat map of the spatio-temporal distribution of the disinfectant concentration through the disinfectant diffusion model and constructing an alignment constraint surface of multi-modal data according to the gradient change characteristics of the heat map. Feedback control adjustment module: Used to calculate the microbial inactivation rate and the disinfectant residue in real time based on the compensation data, and adjust the disinfectant dosing plan in real time according to the deviation.

[0051] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0052] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product.

[0053] Those of ordinary skill in the art can realize that the modules and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional technician can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0054] In addition, in each embodiment of this application, the functional modules can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0055] As described above, this is only a specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0056] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A collaborative disinfection control method based on multimodal water quality perception, characterized in that: The following steps are involved: Constructing a disinfectant diffusion model, optimizing the disinfectant diffusion model, correcting simulation deviations, and generating a compensation data set; the optimized disinfectant diffusion model generates a disinfectant concentration spatiotemporal distribution heat map through the disinfectant diffusion model, and constructing an alignment constraint surface of multimodal data for optimization according to the gradient change characteristics of the heat map; The microbial inactivation rate and disinfectant residual amount are calculated in real time based on the compensation data, and the disinfectant dosing plan is adjusted in real time according to the deviation.

2. The collaborative disinfection control method based on multimodal water quality perception according to claim 1 is characterized in that: The generated spatiotemporal distribution heat map of disinfectant concentration is specifically: Divide the water body into several grid cells, set the size of each grid cell according to the simulation accuracy requirements, and set initial conditions and boundary conditions for each grid cell; The diffusion of disinfectants in water bodies is solved based on fluid mechanics simulation software to obtain the disinfectant concentration value of each grid cell at different time steps; According to the calculation results of each time step, the disinfectant concentration value of each grid unit at the corresponding time point is derived to form a complete spatiotemporal distribution data set; The spatiotemporal distribution data set is used to convert the spatial distribution results of disinfectant concentration into a heat map based on a data visualization tool.

3. The collaborative disinfection control method based on multimodal water quality perception according to claim 2 is characterized in that: The method of constructing the alignment constraint surface of multimodal data according to the gradient change characteristics of the heat map is as follows: The time gradient and space gradient of concentration are calculated for each time-space point, and normalized to generate a gradient weight matrix; Map the multimodal data to spatiotemporal points and spatial coordinates respectively, and calculate the time difference and spatial Euclidean distance between each modal data point and the nearest spatiotemporal grid point of the heat map; Define alignment constraints based on gradient weight matrix, time difference and spatial Euclidean distance, and locally align multimodal data; Based on the global optimization method of gradient descent, the spatial distribution of each modal data is adjusted to minimize the deviation between the spatial distribution and the constraint surface to obtain the final aligned data set.

4. The collaborative disinfection control method based on multimodal water quality perception according to claim 3 is characterized in that: The multimodal data includes physical and chemical sensor data, spectral data and microscopic imaging data. Local alignment of the multimodal data specifically includes: The physicochemical sensor data is aligned with the concentration gradient area in the thermodynamic map by optimizing the coordinates of the sensor measurement points; The spectral data matches the spectral signal in spatial position according to the absorption characteristics of the disinfectant at different wavelengths to ensure its consistency with the concentration gradient area; The microscopic imaging data is aligned with the microorganism aggregation area in the microscopic image and the high gradient area of ​​disinfectant concentration through an image registration algorithm.

5. The collaborative disinfection control method based on multimodal water quality perception according to claim 4 is characterized in that: The optimization of the disinfectant diffusion model, correction of simulation deviation, and generation of a compensation data set are specifically as follows: Based on the aligned data set, the disinfectant diffusion model is forward calibrated to obtain an optimized disinfectant diffusion model; According to the optimized disinfectant diffusion model, the attenuation function of the spectral data is corrected by reverse compensation and a turbidity compensation matrix is ​​constructed to obtain compensated spectral data; Based on the optimized diffusion model and compensated spectral data, the disinfectant concentration at each time and spatial position in the water body is calculated to generate a new spatiotemporal distribution data set of disinfectant concentration.

6. The collaborative disinfection control method based on multimodal water quality perception according to claim 5 is characterized in that: The forward correction of the disinfectant diffusion model is specifically as follows: Constructing a deviation function based on the aligned multimodal data and the predictions of the preliminary disinfectant diffusion model; Based on genetic algorithm, the parameters of the diffusion model are adjusted with the goal of minimizing the deviation function; According to the optimized model parameters, the disinfectant diffusion model is updated to generate new spatiotemporal distribution data of disinfectant concentration.

7. The collaborative disinfection control method based on multimodal water quality perception according to claim 6 is characterized in that: According to the optimized disinfectant diffusion model, the attenuation function of the spectral data is corrected by reverse compensation and the turbidity compensation matrix is ​​constructed, which is specifically: Establishing a preliminary spectral attenuation model and updating the spectral attenuation model, wherein the updated spectral attenuation model is updated by comparing the optimized disinfectant concentration distribution with the actually observed spectral data and identifying deviations; Analyze the impact of water turbidity in different areas on spectral and microscopic imaging data, and construct a turbidity compensation matrix based on disinfectant concentration; The compensation matrix is ​​applied to the spectral and microscopic imaging data for reverse compensation to obtain the corrected spectral data and generate a new compensation data set.

8. The collaborative disinfection control method based on multimodal water quality perception according to claim 7 is characterized in that: The real-time calculation of microbial inactivation rate and disinfectant residual amount based on compensation data, and real-time regulation of disinfectant dosing scheme according to deviation, are specifically as follows: Calculating the residual amount of disinfectant in the water and the inactivation rate of microorganisms based on compensation data and a preset microbial inactivation model, wherein the compensation data includes disinfectant concentration data; If there is a deviation between the microbial inactivation rate and the preset target inactivation rate, the control system automatically adjusts the dosage and position of the disinfectant; If the residual amount of disinfectant exceeds the preset safety range, the two objectives of microbial inactivation rate and disinfectant residual amount are optimized simultaneously based on the multi-objective particle swarm optimization algorithm, and the dosage and position of the disinfectant are dynamically adjusted under different control constraints.

9. A system for implementing the collaborative disinfection control method based on multimodal water quality perception as described in any one of claims 1 to 8, characterized in that: Includes the following modules: Disinfectant model construction and data alignment module: used to construct a disinfectant diffusion model, optimize the disinfectant diffusion model, correct simulation deviations, and generate a compensation data set; the optimized disinfectant diffusion model is obtained by constructing an alignment constraint surface of multimodal data based on the temporal and spatial distribution heat map of the disinfectant concentration generated by the disinfectant diffusion model, and according to the gradient change characteristics of the heat map; Feedback control adjustment module: used to calculate the microbial inactivation rate and disinfectant residual amount in real time based on the compensation data, and adjust the disinfectant dosing plan in real time according to the deviation.

Images