A water quality parameter monitoring and intelligent dosing method and system based on the Internet of Things

Abstract

The present application relates to a water quality parameter monitoring and intelligent dosing method and system based on the Internet of Things, the method comprising: obtaining time series water quality parameter data through multi-point sensing equipment, performing data preprocessing on the water quality parameter data, and obtaining a first water quality data set; performing time series analysis on the first water quality data set to obtain a water quality trend feature set; inputting the water quality trend feature set into a preset water quality state prediction model to obtain predicted water quality parameter values ​​within a preset future time period; performing anomaly analysis on the predicted water quality parameter values ​​and the preset parameter threshold range to obtain early warning data, the early warning data including pollution risk areas, abnormal time intervals, and risk categories; optimizing and calculating the dosing parameters based on the early warning data and historical governance data, and using a tree model to determine the first dosing plan. The present application can effectively improve the accuracy and timeliness of water pollution monitoring and governance, and provide an intelligent solution for water environment management.

Description

Technical Field

[0001] The present application relates to the technical field of water quality monitoring, and in particular to a water quality parameter monitoring and intelligent dosing method and system based on the Internet of Things. Background Art

[0002] With the accelerated pace of industrialization and urbanization, water pollution is becoming increasingly serious, and ensuring water quality safety has become a pressing need for social development. Traditional methods for managing water quality in rivers, lakes, reservoirs, and coastal waters rely heavily on manual sampling and laboratory analysis. These methods are not only time-consuming and costly, but also struggle to achieve real-time monitoring and rapid response, often failing to meet the complex and ever-changing water environment.

[0003] Because traditional water quality management methods lack timely data support, managers find it difficult to promptly and accurately understand changes in water quality. Adjustments to water treatment plans often rely on empirical judgment and lack scientific basis, which can easily lead to waste of resources or poor treatment effects. Summary of the Invention

[0004] Based on the above statements, this application provides a water quality parameter monitoring and intelligent dosing method and system based on the Internet of Things to solve the problem of poor dosing treatment effect caused by untimely and inaccurate target water quality monitoring.

[0005] According to the first aspect of the present application, a water quality parameter monitoring and intelligent dosing method based on the Internet of Things is provided, comprising:

[0006] Acquiring time-series water quality parameter data through a multi-point sensing device, and performing data preprocessing on the water quality parameter data to obtain a first water quality data set;

[0007] Performing time series analysis on the first water quality data set to obtain a water quality trend feature set;

[0008] Inputting the water quality trend feature set into a preset water quality state prediction model to obtain predicted water quality parameter values ​​within a preset future time period;

[0009] Performing an anomaly analysis on the predicted water quality parameter values ​​and the preset parameter threshold range to obtain early warning data, the early warning data including pollution risk areas, abnormal time intervals and risk categories;

[0010] According to the early warning data and historical treatment data, the dosing parameters are optimized and calculated through a tree model to determine a first dosing plan.

[0011] In some embodiments, the water quality parameter data includes dissolved oxygen, turbidity, pH, nitrogen content, and phosphorus content.

[0012] In some embodiments, a time series analysis is performed on the first water quality dataset to obtain a water quality trend feature set, including:

[0013] Using a time series decomposition algorithm to perform stratification processing on the first water quality data set to obtain multiple data subsets with different change characteristics;

[0014] determining trend slope data of the water quality parameters in the plurality of data subsets within each sliding time window;

[0015] Performing correction processing according to the trend slope data and a preset slope threshold range;

[0016] The trend slope data of the corrected water quality parameters are subjected to multi-dimensional feature extraction to obtain a water quality trend feature set.

[0017] In some embodiments, before determining the trend slope data of the water quality parameters in each time window in the plurality of data subsets, the method further comprises:

[0018] The water quality parameters in the plurality of data subsets are weighted according to preset environmental factor weights.

[0019] In some embodiments, the training process of the water quality state prediction model includes:

[0020] A water quality status prediction model is constructed based on a long short-term memory neural network, where the input layer is the historical water quality trend characteristics and the output layer is the predicted values ​​of water quality parameters;

[0021] The mean square error between the predicted value of water quality parameters and the true value of water quality parameters is taken as the loss value;

[0022] Based on the loss value, executing the back propagation algorithm to update the water quality state prediction model parameters and optimize the water quality state prediction model;

[0023] When it is detected that the loss value of the model meets the conditions or the number of training times reaches the set upper limit, the training is considered to be completed and the trained model is obtained.

[0024] In some embodiments, performing an anomaly analysis on the predicted water quality parameter value and a preset parameter threshold range to obtain early warning data includes:

[0025] Determining abnormal data points and corresponding abnormal time intervals based on the predicted water quality parameter values ​​and preset parameter threshold ranges;

[0026] Determine whether the abnormal duration corresponding to the abnormal data point exceeds a preset time threshold range;

[0027] If yes, classify the abnormal time distribution and abnormal parameter type of the abnormal data point to obtain the risk category corresponding to the abnormal data point;

[0028] Acquiring historical water quality records, wherein the historical water quality records include abnormal areas and their corresponding abnormal categories and abnormal time periods;

[0029] The pollution risk area is determined based on the abnormal time interval and risk category of the abnormal data point in combination with the historical water quality records.

[0030] In some embodiments, based on the early warning data and historical treatment data, and by optimizing and calculating the dosing parameters through a tree model, a first dosing plan is obtained, including:

[0031] According to the abnormal time interval and risk category corresponding to the current pollution risk area, the relevant target governance records are matched in the historical governance data;

[0032] Determine the remediation type of the target remediation record based on the pollution area characteristics and time distribution in the target remediation record;

[0033] Determine whether the pollution parameter level corresponding to the current pollution risk area exceeds the preset pollution threshold range;

[0034] If so, a decision tree model is used to compare the target treatment record with the preset drug matching standard, and based on the treatment type, the initial dosing parameters are determined;

[0035] The initial dosing parameters are dynamically adjusted according to environmental variables and drug efficacy feedback information to determine a first dosing plan.

[0036] In some embodiments, it further includes:

[0037] Performing a simulation evaluation on the dosing effect of the first dosing scheme to obtain an evaluation result;

[0038] If the evaluation result does not meet the expected effect, iteratively optimize the first dosing plan to generate a second dosing plan;

[0039] Based on the second dosing plan, a dosing instruction is generated to control the dosing device to perform a dosing operation.

[0040] In some embodiments, the process of simulating and evaluating the dosing effect of the first dosing scheme includes: performing environmental simulation on the first dosing scheme through a simulation system to determine simulated water quality index data;

[0041] If the evaluation result does not meet the expected effect, the first dosing plan is iteratively optimized to generate a second dosing plan, including:

[0042] Preliminarily revising the dosing parameters of the first dosing scheme according to the water quality parameter values ​​in the simulated water quality index data and a preset simulation threshold range;

[0043] The dosing parameters after preliminary correction are simulated and optimized in combination with dynamic feedback data to determine the second dosing plan.

[0044] According to a second aspect of the present application, a water quality parameter monitoring and intelligent dosing system based on the Internet of Things is provided, comprising:

[0045] an acquisition and preprocessing module, configured to acquire time-series water quality parameter data through a multi-point sensing device, and perform data preprocessing on the water quality parameter data to obtain a first water quality data set;

[0046] a time analysis module, configured to perform time series analysis on the first water quality data set to obtain a water quality trend feature set;

[0047] A state prediction module is used to input the water quality trend feature set into a preset water quality state prediction model to obtain predicted water quality parameter values ​​within a preset future time period;

[0048] An anomaly analysis module, configured to perform an anomaly analysis on the predicted water quality parameter values ​​and a preset parameter threshold range to obtain early warning data, wherein the early warning data includes pollution risk areas, abnormal time intervals, and risk categories;

[0049] A scheme generation module is used to optimize and calculate the dosing parameters based on the early warning data and historical management data through a tree model to determine a first dosing scheme.

[0050] According to a third aspect of the present application, a computer device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method described in any one of the aforementioned embodiments when executing the computer program.

[0051] According to a fourth aspect of the present application, a storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described in any one of the aforementioned embodiments are performed.

[0052] Compared with the prior art, the technical solution of this application has the following beneficial technical effects:

[0053] (1) This application forms a sensor network by deploying multi-point sensing devices in the water body. In this way, time series water quality parameter data can be obtained in real time based on the Internet of Things, and key water quality parameter data can be provided in a timely manner.

[0054] (2) This application can reduce noise interference signals by performing pre-processing steps such as denoising, data standardization and data cleaning on water quality parameter data, ensure the integrity, consistency and accuracy of the data, facilitate the system's subsequent classification, analysis and other operations, and improve the accuracy of the analysis.

[0055] (3) This application obtains a water quality trend feature set by performing time series analysis on the first water quality data set, determines multi-dimensional characteristics such as long-term trend characteristics and short-term fluctuation characteristics of water quality parameters, reveals the changing patterns of water quality parameter data, and provides necessary data support for model prediction.

[0056] (4) This application obtains the predicted water quality parameter values ​​within a preset future time period by inputting the water quality trend feature set into a pre-built water quality status prediction model; the water quality trend feature set is processed by the water quality status prediction model to obtain the predicted water quality parameter values ​​within the future time period, thereby realizing trend prediction and knowing the future water quality status in advance, which facilitates the system to formulate a treatment plan in advance.

[0057] (5) This application can achieve early warning monitoring by conducting anomaly analysis on the predicted water quality parameter values ​​and the preset parameter threshold range; through anomaly analysis, early warning data such as pollution risk areas, abnormal time intervals and risk categories can be obtained, thereby improving the overall efficiency from anomaly identification to regional positioning; it helps to discover potential problems in advance and provide strong support for water body protection.

[0058] (6) This application optimizes and calculates the dosing parameters based on early warning data and historical treatment data through a tree model to determine the first dosing plan; through the close connection of matching records, reagent comparison, parameter adjustment and scheme integration, it can significantly improve the overall efficiency from pollution identification to treatment plan generation; subsequently, the dosing effect of the first dosing plan can be evaluated through simulation, and iterative optimization can be performed when necessary, ultimately driving the automated equipment to perform precise dosing operations; effectively improving the accuracy and timeliness of water pollution control, and providing an intelligent solution for water environment management. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a flow chart of a water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to an embodiment of the present application;

[0060] Figure 2 This is a flow chart of a water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to another embodiment of the present application;

[0061] Figure 3 This is a schematic diagram of the principle of a water quality parameter monitoring and intelligent dosing system based on the Internet of Things according to an embodiment of the present application.

[0062] Figure 4 A schematic diagram of the principles of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0063] To facilitate understanding of the present application, the present application will be described more fully below with reference to the accompanying drawings. The accompanying drawings provide embodiments of the present application. However, the present application may be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to make the disclosure of the present application more thorough and comprehensive.

[0064] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application pertains. The terms used herein in the specification of this application are for the purpose of describing specific embodiments only and are not intended to limit this application.

[0065] Water quality monitoring and treatment currently face numerous challenges. Traditional methods are unable to capture dynamic information such as changes in water appearance and color, or algae growth, resulting in delayed water quality assessments. The lack of timely data makes it extremely difficult to predict water quality trends, making it difficult for managers to formulate proactive responses. Inaccurate predictions directly impact the accuracy of treatment plans. For example, adjustments to dosage and type of dosing often rely on empirical judgment without scientific basis, which can lead to wasted resources and poor treatment results.

[0066] Based on this, the embodiment of the present application proposes a water quality parameter monitoring and intelligent dosing method and system based on the Internet of Things to solve the problem of poor dosing treatment effect caused by untimely and inaccurate target water quality monitoring.

[0067] See Figure 1 , Figure 1 A schematic diagram of a process for monitoring water quality parameters and intelligent dosing based on the Internet of Things according to an embodiment of the present application is shown. The process of monitoring water quality parameters and intelligent dosing based on the Internet of Things according to an embodiment of the present application includes the following steps:

[0068] S101, acquiring time-series water quality parameter data through a multi-point sensing device, and performing data preprocessing on the water quality parameter data to obtain a first water quality data set;

[0069] S102, performing time series analysis on the first water quality data set to obtain a water quality trend feature set;

[0070] S103, inputting the water quality trend feature set into a preset water quality state prediction model to obtain predicted water quality parameter values ​​within a preset future time period;

[0071] S104: Perform an anomaly analysis on the predicted water quality parameter values ​​and the preset parameter threshold ranges to obtain early warning data, which includes pollution risk areas, abnormal time intervals, and risk categories;

[0072] S105: Optimize and calculate dosing parameters based on the early warning data and historical treatment data through a tree model to determine a first dosing plan.

[0073] In this embodiment, water quality parameters are collected in real time by multi-point sensing equipment to form time series water quality parameter data. The first water quality data set is obtained by preprocessing the water quality parameter data such as data cleaning, which can improve the accuracy and consistency of the basic data; the water quality change trend is determined by performing time series analysis on the first water quality data set, and a water quality trend feature set is obtained to provide key data support for water quality prediction; the water quality trend feature set is processed by a water quality state prediction model to obtain predicted water quality parameter values ​​in the future time period, realize trend prediction, and facilitate the formulation of treatment plans in advance; the predicted water quality parameter values ​​and the preset parameter threshold range are analyzed for abnormalities to realize early warning monitoring; based on the early warning data and historical treatment data, the dosing parameters can be quickly optimized and calculated using a tree model to determine the first dosing plan, which can effectively improve the accuracy and timeliness of water pollution monitoring and treatment, and provide an intelligent solution for water environment management.

[0074] In step S101, time series water quality parameter data is acquired through a multi-point sensing device, and data preprocessing is performed on the water quality parameter data to obtain a first water quality data set.

[0075] Specifically, by deploying multiple sensing devices in the water body to form a sensor network, the Internet of Things (IoT) is used to acquire real-time water quality parameter data. This water quality parameter data includes, but is not limited to, key indicators such as dissolved oxygen, turbidity, pH, nitrogen content, and phosphorus content. This time-series water quality parameter data is used as the raw data set and transmitted to a cloud storage platform.

[0076] For example, ten sensors were deployed in a lake under monitoring, collecting data every five minutes. The data showed dissolved oxygen values ​​between 6.5 and 8.2 mg / L, turbidity between 3.0 and 10.0 NTU, and pH between 6.8 and 7.5. These water quality parameter data were transmitted to a cloud-based storage platform, creating a preliminary, organized water quality monitoring record for subsequent analysis.

[0077] Data preprocessing is performed on water quality parameter data, including denoising, data standardization, and data cleaning. Specifically, a filtering algorithm can be used to denoise the water quality parameter data in the water quality monitoring records to reduce noise interference signals. For the denoised water quality parameter data, the parameter ranges are adjusted and the scale is unified to achieve data standardization. The standardized water quality parameter data can be cleaned to identify and correct outliers and missing values. For example, outliers and missing values ​​in the water quality parameter data can be interpolated and supplemented. If there are data points in the water quality parameter data that exceed the preset parameter threshold range, the abnormal data segments can be marked.

[0078] For example, in the actual application of water body monitoring, obtaining raw data records from the collection source is the starting point of the entire data processing process. Raw data often contain noise interference, such as unstable sensor readings due to environmental fluctuations. To address this problem, a filtering algorithm can be used for preliminary processing. For example, a smoothing filtering algorithm can be used to weaken short-term abnormal fluctuations to obtain a more stable data combination. In another lake monitoring project, the dissolved oxygen data of a certain sensor frequently jumped in a short period of time, ranging from 5.0 to 9.0 mg / L. After filtering, the data fluctuations were smoothed to a reasonable range of 6.2 to 7.8 mg / L. This processing method helps to reduce the interference of noise on subsequent analysis.

[0079] For example, it is particularly important to adjust the range of each parameter for the denoised water quality parameter data. The dimensions and numerical ranges of different parameters vary greatly. For example, turbidity may be between 2.0 and 15.0 NTU, while pH is between 6.5 and 8.0. Direct analysis may lead to deviations. Therefore, by adjusting the data to a unified scale, such as mapping all parameter values ​​to the range of 0 to 1, subsequent comparison and processing can be facilitated. In a specific lake monitoring case, standardized dissolved oxygen and turbidity data were used for comparative analysis, and it was found that the parameter distribution characteristics in certain areas were more consistent. In this way, data standardization can facilitate subsequent feature extraction.

[0080] For example, for the standardized water quality parameter data, if there are data points that deviate from the preset threshold range, data cleaning is essential. Assuming that the preset threshold range of dissolved oxygen is 5.5 to 9.0 mg / L, the data at a certain time point is suddenly missing or shows an abnormal value of 12.0 mg / L, then through data cleaning, and combined with the data of the previous and next time points to fill in the gaps, it can be corrected to a reasonable value of around 7.0 mg / L. For example, in the records of a monitoring point, turbidity data for three consecutive hours is missing. By analyzing adjacent points and historical trends, it is supplemented to 8.5 NTU. In this way, the integrity and regularity of the data sequence can be ensured through data cleaning methods, laying the foundation for subsequent processing.

[0081] In step S102, a time series analysis is performed on the first water quality data set to obtain a water quality trend feature set.

[0082] In some embodiments, step S102 of the present application includes:

[0083] Using a time series decomposition algorithm to perform stratification processing on the first water quality data set, a plurality of data subsets with different change characteristics are obtained;

[0084] Determine trend slope data of water quality parameters within each sliding time window in multiple data subsets;

[0085] Perform correction processing based on trend slope data and preset slope threshold range;

[0086] The trend slope data of the corrected water quality parameters are subjected to multi-dimensional feature extraction to obtain a water quality trend feature set.

[0087] Specifically, a time series decomposition algorithm is used to stratify the water quality parameter data of the first water quality data set, aiming to separate long-term trend data, periodic fluctuation data, and random noise data to obtain multiple data subsets. For multiple data subsets, a sliding time window is used to extract the change characteristics of the water quality parameter data in different time periods, and the trend slope data in each time window is determined. If the trend slope data exceeds the preset slope threshold range, the abnormal interval is corrected to obtain the corrected trend slope data. For the corrected trend slope data, the change law of the water quality parameter data is quantified by feature extraction technology to generate a water quality trend feature set containing multi-dimensional features.

[0088] Time series decomposition algorithms play a central role in processing the first water quality dataset. Their essence lies in deconstructing complex water quality parameter data (i.e., water quality parameter sequences) and separating them into multiple layers, including long-term trend data, cyclical fluctuations, and random noise. In this embodiment, the time series decomposition algorithm can employ the STL (Seasonal Trend Decomposition Using Loess) algorithm, a technique for time series decomposition that decomposes time series data into three components: trend, seasonal, and residual. The STL method offers flexibility and robustness, making it applicable to a wide variety of time series data, particularly when processing data with complex seasonal components. For example, lake dissolved oxygen monitoring, whose time series are subject to both macro-seasonal variations and micro-meteorological disturbances, can be applied. Applying this decomposition technique, it is possible to accurately isolate long-term trends that reveal the evolution of the water quality system and identify cyclical fluctuations driven by circadian rhythms. This multi-level analytical mechanism effectively decouples and traces the sources of variability in water quality parameters, laying a clear, structured foundation for subsequent in-depth analysis.

[0089] For multiple data subsets after decomposition, the use of sliding windows can capture the dynamic characteristics of water quality parameters in different time periods. Taking dissolved oxygen as an example, by setting a 7-day sliding time window, the changing trend of the water quality parameter in each window can be carefully analyzed, and the trend slope of each period can be accurately calculated. For example, if the trend slope of dissolved oxygen in a certain time window is -0.5 mg / L / day, this may suggest that the water quality during this period is gradually deteriorating; on the contrary, if the trend slope in another time window is 0.2 mg / L / day, it may indicate that the water quality is stabilizing or even improving. This analysis method based on sliding windows can effectively reveal the subtle changes in water quality parameters in the short term, and provide strong support for the dynamic monitoring and management of lake ecosystems.

[0090] Anomaly analysis is performed on the trend slope data of water quality parameters within each time window in multiple data subsets. If the trend slope data of any water quality parameter within a time period exceeds the preset slope threshold range, the abnormal interval (i.e., the abnormal time window) is corrected to obtain the corrected trend slope data. For example, taking lake dissolved oxygen as an example, the slope threshold range can be set to -0.3 to 0.3 mg / L / day. If the trend slope reaches -0.8 mg / L / day within a certain week, the moving average method can be used to smooth the trend slope of the abnormal interval to bring it closer to the average level of the previous few weeks. This correction process can prevent short-term anomalies from interfering with overall trend judgment, providing more stable input for subsequent analysis.

[0091] In a possible embodiment, before determining the trend slope data of the water quality parameters in the multiple data subsets within each time window, the method of this embodiment further includes: performing weighted processing on the water quality parameters in the multiple data subsets according to preset environmental factor weights.

[0092] Specifically, different seasons have different impacts on water quality parameters. For example, in lake monitoring, dissolved oxygen may be more affected by temperature in the summer. Therefore, the weight of its summer water quality parameter data can be appropriately increased. For example, the environmental factor weight of dissolved oxygen in the summer is set to 1.2, while the environmental factor weights of other seasons are set to 1.0. The environmental factor weights for each water quality parameter can be set according to actual conditions, and this embodiment does not impose any restrictions here. In this way, the weighted processing can more realistically reflect the impact of environmental factors on water quality parameters, which helps to improve the pertinence of subsequent feature extraction.

[0093] After obtaining the corrected trend data, the changing patterns of water quality parameters can be further quantified through feature extraction, and a water quality trend feature set with multi-dimensional features can be generated. In this embodiment, the Pandas library or the NumPy library, or a machine learning model or a deep learning model can be used to implement the feature extraction task. For example, taking the dissolved oxygen monitored in lakes as an example, the monthly average value, fluctuation amplitude, and change frequency of the dissolved oxygen trend slope can be extracted. Assuming that the average dissolved oxygen value in a certain month is 6.5 mg / L and the fluctuation amplitude is 1.2 mg / L, combined with the characteristics of the turbidity data, a comprehensive feature set can be constructed, that is, the water quality trend feature set of dissolved oxygen. This feature set provides a multi-dimensional perspective for subsequent water quality assessment, which helps to more comprehensively understand the inherent laws of water quality changes.

[0094] In some embodiments, a multi-dimensional water quality status chart can be generated based on a water quality trend feature set, achieving visualization. This allows for clear observation of the trend changes in various water quality parameters. Furthermore, comparative analysis can be performed based on changes in point parameters within the water quality trend feature set to determine whether there are localized areas of abnormal water quality, providing a reference for subsequent monitoring.

[0095] Through the processing of the above links, this embodiment forms a tight logical chain for the entire process from time series decomposition to feature extraction, ensuring the in-depth mining and effective utilization of water quality monitoring data, and laying a solid foundation for subsequent analysis.

[0096] In step S103, the water quality trend feature set is input into a preset water quality state prediction model to obtain predicted water quality parameter values ​​within a preset future time period.

[0097] In some embodiments, the training process of the water quality state prediction model includes:

[0098] A water quality status prediction model is constructed based on a long short-term memory neural network, where the input layer is the historical water quality trend characteristics and the output layer is the predicted values ​​of water quality parameters;

[0099] The mean square error between the predicted value of water quality parameters and the true value of water quality parameters is taken as the loss value;

[0100] Based on the loss value, the back propagation algorithm is executed to update the parameters of the water quality state prediction model and optimize the water quality state prediction model;

[0101] When it is detected that the loss value of the model meets the conditions or the number of training times reaches the set upper limit, the training is considered to be completed and the trained model is obtained.

[0102] Specifically, in this embodiment, a water quality state prediction model can be constructed based on a long-short-term memory neural network. A training set is constructed, which includes historical water quality trend features and corresponding true values ​​of water quality parameters. This historical water quality trend feature set can be obtained through preprocessing and time series analysis of historical water quality parameter data. The true values ​​of water quality parameters refer to the true values ​​of each water quality parameter in the historical water quality data. The historical water quality trend features are used as input to the model. The predicted water quality parameter values ​​output by the model are compared with the true values ​​to determine a loss value. Based on the loss value, a backpropagation algorithm is executed to update the parameters of the water quality state prediction model and optimize the water quality state prediction model. When the model loss value is detected to be less than a preset loss threshold range (e.g., a loss value between 0.08 and 0.1) or the model has been trained a preset number of times (e.g., 200 times), training is determined to be complete. The model can then be verified using a test set and a validation set to obtain a trained water quality state prediction model.

[0103] During the model's formal use, the water quality trend feature set is input into a preset water quality state prediction model to obtain predicted water quality parameter values ​​for a preset future time period, thereby obtaining a time series of predicted water quality parameter values. This preset future time period can be set to the next three months, one month, or fifteen days, depending on actual circumstances and is not a limitation in this embodiment. For example, in lake monitoring, the water quality state prediction model estimates that dissolved oxygen may drop to 5.8 mg / L and turbidity may rise to 15 NTU in the next month.

[0104] Furthermore, in some embodiments, the water quality prediction model can generate a prediction result set containing remediation decision-making information based on predicted water quality parameter values, water quality trend characteristics, and current environmental conditions, such as recommendations for adding aeration equipment or reducing pollution sources. This multi-dimensional processing can provide management departments with more instructive information.

[0105] It should be noted that in lake water quality monitoring, the water quality status prediction model can achieve accurate prediction based on the water quality trend feature set, help identify potential water quality risks, facilitate the formulation of response measures in advance, and provide strong support for water body protection.

[0106] In step S104, an abnormality analysis is performed on the predicted water quality parameter value and the preset parameter threshold range to obtain early warning data, which includes pollution risk areas, abnormal time intervals and risk categories.

[0107] In some embodiments, step S104 of the present application includes:

[0108] Determine abnormal data points and corresponding abnormal time intervals based on the predicted water quality parameter values ​​and the preset parameter threshold ranges;

[0109] Determine whether the abnormal duration corresponding to the abnormal data point exceeds the preset time threshold range;

[0110] If yes, classify the abnormal time distribution and abnormal parameter type of the abnormal data point to obtain the risk category corresponding to the abnormal data point;

[0111] Obtain historical water quality records, including abnormal areas and their corresponding abnormal categories and abnormal time periods;

[0112] Based on the abnormal time interval and risk category of the abnormal data point and combined with historical water quality records, the pollution risk area is determined.

[0113] Specifically, the predicted water quality parameter values ​​of each water quality parameter are compared one by one with the corresponding parameter threshold range, and the data of each time period is processed separately to obtain abnormal data points that exceed the preset parameter threshold range, and determine the time interval corresponding to the abnormal data point to ensure that the identification of abnormal data points has time accuracy. For example, suppose in a lake water quality monitoring project, the prediction results show that the dissolved oxygen value may drop to 4.5 mg / L on certain days in the next month, while the preset parameter threshold range is 5.0 to 10 mg / L. In this way, abnormal data points can be quickly screened out and their corresponding abnormal time intervals can be recorded, such as the 10th to 15th day of the next month.

[0114] If an abnormal data point persists for longer than a preset time threshold (e.g., 3 to 8 days) within a certain time interval, an early warning mechanism is automatically triggered. The abnormal time distribution and abnormal parameter type of the abnormal data point are classified to determine the corresponding risk category. This risk category indicates the risk of the abnormal data point, such as low oxygen risk, high turbidity risk, high acidity risk, etc.

[0115] For example, using the above-mentioned dissolved oxygen prediction as an example, if the dissolved oxygen value remains below 5.0 mg / L for five consecutive days within the aforementioned abnormal time period, a potential risk is identified. At this point, an early warning mechanism can be triggered to categorize the abnormal data points based on their time distribution and parameter type. For example, the abnormal dissolved oxygen signal can be classified as "hypoxia risk" and the time period of occurrence (i.e., the abnormal time period) can be noted. This provides a clear basis for subsequent risk identification.

[0116] Obtain historical water quality records, which include abnormal areas and corresponding abnormal categories. The purpose of obtaining historical water quality records is to locate the current abnormal data points with reference to historical water quality abnormalities. The abnormal area refers to the spatial location where water quality abnormalities occurred during the historical monitoring of water sources. The abnormal category refers to the water quality abnormality that occurred in the abnormal area, such as low oxygen risk, high turbidity risk, high acidity risk, etc. In this embodiment, it is necessary to classify the water quality abnormal data of the historical water quality records in advance according to the abnormal time distribution and abnormal parameter type, determine the abnormal category corresponding to the abnormal area, and mark the abnormal time period. Furthermore, the abnormal category and abnormal time period corresponding to each abnormal area in the historical water quality record are matched with the risk category and abnormal time interval corresponding to the continuous abnormal data point to locate the potential pollution risk area.

[0117] For example, in a lake monitoring project, if historical water quality records indicate frequent hypoxia risk in the northwest region of the lake during past summers, and current warning data also indicates similar abnormal time periods and risk categories, the pollution risk area can be located in the northwest. This positioning method can transform abstract signals into specific spatial scopes, providing targeted reference for governance.

[0118] In some embodiments, targeted suggestions for adjusting the governance plan are generated based on the pollution risk areas, and the distribution characteristics of the pollution risk areas are prioritized to determine the final content of the adjustment plan.

[0119] For example, adjustments to remediation plans for pollution risk areas will be made, with recommendations generated and prioritized based on regional distribution characteristics. For example, in a lake monitoring project, if the northwest region is close to industrial outfalls and has a higher risk of hypoxia, priority will be given to strengthening pollution source control in that area, such as limiting emissions or adding water circulation equipment. For lower-risk areas, routine monitoring can be arranged. The final remediation plan will clearly identify priority remediation areas and specific measures to ensure appropriate resource allocation.

[0120] It's important to note that the close integration of these steps improves overall efficiency, from anomaly identification to regional location. From data comparison and alert triggering to regional location and adjustment of treatment plans, each step is centered around the dynamics of lake water quality, ensuring the practicality of treatment recommendations. This systematic approach helps identify potential problems early and provides strong support for water protection.

[0121] In step S105, the dosing parameters are optimized and calculated based on the early warning data and the historical treatment data through a tree model to determine a first dosing plan.

[0122] In some embodiments, step S105 of the present application includes:

[0123] According to the abnormal time interval and risk category corresponding to the current pollution risk area, the relevant target governance records are matched in the historical governance data;

[0124] Determine the remediation type of the target remediation record based on the pollution area characteristics and time distribution in the target remediation record;

[0125] Determine whether the pollution parameter level corresponding to the current pollution risk area exceeds the preset pollution threshold range;

[0126] If so, the initial dosing parameters are determined based on the current pollution parameter level, the treatment type of the target treatment record, and the preset agent matching standard library, and using the decision tree model;

[0127] The initial dosing parameters are dynamically adjusted according to environmental variables and drug efficacy feedback information to determine the first dosing plan.

[0128] Specifically, according to the abnormal time intervals and risk categories corresponding to the pollution risk areas in the current early warning data, the related target governance records are matched in the historical governance data. The governance type of the target governance record is determined based on the characteristics of the pollution area and the time distribution in the target governance record. For example, a target governance record shows that a certain area has experienced high nitrogen and phosphorus pollution many times between June and July in the past summer. The pollution level was recorded as a nitrogen content of 8.5 mg / L, far exceeding the pollution threshold of 5.0 mg / L. Through classification processing, the target governance record can be divided into two types of governance types: short-term emergency treatment and long-term prevention and control according to the time distribution and regional characteristics, providing a basis for subsequent agent selection.

[0129] The pollution parameter level is determined based on the water quality parameter values ​​corresponding to the current pollution risk area. For example, the nitrogen content in the western region of a lake reaches 7.2 mg / L, exceeding the pollution threshold of 5.0 mg / L. In this embodiment, the pollution threshold range can be a point value or a range of values, which is not limited in this embodiment.

[0130] When it is determined that the pollution parameter level corresponding to the current pollution risk area exceeds the preset pollution threshold range, a decision tree model is used to compare the current pollution parameter level and the treatment type recorded in the target treatment plan with the preset agent matching standards to determine the initial dosing parameters, including the dosing type, initial dosing dosage, and dosing time. The agent matching standard refers to a standard library that includes dispensing parameters corresponding to different pollution parameter levels and different treatment types. The dosing type and initial dosing dosage corresponding to the current pollution parameter level can be determined based on two types of treatment: short-term emergency treatment and long-term prevention and control. This allows the type of agent to be quickly identified, laying the foundation for subsequent dosage adjustments.

[0131] The initial dosing parameters are dynamically adjusted based on environmental variables and efficacy feedback information. Specifically, based on the preliminary agent selection results, the dosing parameters are dynamically adjusted in combination with environmental variables such as water temperature and flow rate, as well as feedback information such as the effect of previous dosing. For example, in a lake monitoring project, the water temperature in the current area is 25 degrees Celsius and the flow rate is slow. Based on these environmental variables, it is determined that the initial dose needs to be increased by 10%. At the same time, the dosing time is dynamically updated to ensure that the timing of dosing is accurate. For example, concentrated dosing is carried out from the 3rd to the 5th day during the peak pollution period. This dynamic adjustment method can adapt to real-time changes in lake water quality.

[0132] Furthermore, based on the dynamically adjusted dosing parameters and the distribution characteristics of the current pollution risk area (i.e., the area to be dosed), the system generates a first dosing plan. The first dosing plan includes at least the area to be dosed, dosing parameters, and dosing time. For example, in a lake monitoring project, pollution is mainly concentrated in the southeastern area of ​​the lake, and the water in this area is relatively shallow. The system will generate content including the specific dosing type (such as nitrogen removal agent) and dosing dosage (such as 50 kg per day), and recommend a plan to set up dosing points around the pollution risk area. This method of program integration can ensure that the dosing plan is highly matched with regional characteristics, improving the targeted treatment.

[0133] It's important to note that the seamless integration of record matching, reagent comparison, parameter adjustment, and solution integration across these steps significantly improves the overall efficiency from pollution identification to treatment plan generation. Especially in the context of dynamic changes in lake water quality, timely adjustments to dosing type and dosage can effectively control the risk of pollution spread and provide reliable support for water protection.

[0134] See Figure 2 , Figure 2 A flow chart of a water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to another embodiment of the present application is shown.

[0135] In some embodiments, after step S105, the method of the present application further includes:

[0136] S106, performing simulation evaluation on the dosing effect of the first dosing scheme to obtain an evaluation result;

[0137] S107. If the evaluation result does not meet the expected effect, iteratively optimize the first dosing plan to generate a second dosing plan;

[0138] S108 : Based on the second dosing plan, generate a dosing instruction to control the dosing equipment to perform a dosing operation.

[0139] In this embodiment, the process of performing simulation evaluation on the dosing effect of the first dosing scheme in step S106 includes: performing environmental simulation on the first dosing scheme through a simulation system to determine simulated water quality index data.

[0140] In this embodiment, if the evaluation result in step S107 does not meet the expected effect, the first dosing plan is iteratively optimized to generate a second dosing plan, including:

[0141] Preliminarily correcting the dosing parameters of the first dosing scheme according to the water quality parameter values ​​in the simulated water quality index data and the preset simulation threshold range;

[0142] The dosing parameters after preliminary correction are simulated and optimized in combination with dynamic feedback data to determine the second dosing plan.

[0143] Specifically, a simulation is performed based on the first dosing plan and environmental variables of the current pollution risk area to generate simulated water quality index data. The simulated water quality index data includes the water quality parameter values ​​of the pollution risk area during the simulation time period. When the water quality parameter values ​​in the simulated water quality index data exceed the preset simulation threshold range, the system will make a preliminary correction to the dosing parameters of the first dosing plan. By performing a secondary simulation on the initially corrected dosing parameters and combining them with dynamic feedback data (such as real-time monitoring results of lake water bodies), the initially corrected dosing parameters are optimized until the simulated water quality index data meets the preset standards (that is, within the simulation threshold range), and a second dosing plan is generated. In this way, through simulation and verification principles, the dosing parameters are continuously optimized and adjusted to determine the adjustment direction of the dosing plan.

[0144] The core of the simulation evaluation process lies in simulating the changes in water quality parameters after drug addition. For example, in a lake management project, the first dosing plan calls for daily addition of 50 kg of nitrogen removal agent. Based on environmental variables such as the current water temperature of 28 degrees Celsius and a slow flow rate, the simulation system simulates a drop in nitrogen content from 7.2 mg / L to 5.8 mg / L three days after drug addition. This process helps predict drug addition effectiveness and provides data support for subsequent adjustments.

[0145] After obtaining preliminary simulated water quality index data, the water quality parameter values ​​in the simulated water quality index data are compared with the preset simulation threshold range. For example, taking nitrogen content monitoring as an example, the preset simulation threshold range is nitrogen content below 5.0 mg / L, but the simulation result shows that the nitrogen content is still 5.8 mg / L, exceeding the simulation threshold range. This simulation result is marked as substandard and triggers preliminary corrections to the dosing parameters. Preliminary adjustments include increasing the dosing dosage or adjusting the dosing time. This comparison and correction mechanism can quickly identify deficiencies in the solution and ensure that the dosing parameter configuration is more closely aligned with actual needs.

[0146] A secondary simulation is conducted on the initially revised dosing parameters, combined with dynamic feedback data (such as real-time lake water monitoring results) to verify whether the water quality improvement targets have been achieved. For example, in the aforementioned lake treatment project, assuming a daily nitrogen removal dosage of 60 kg after the initial revision, the secondary simulation results showed that the nitrogen content had dropped to 4.8 mg / L on the fifth day, meeting the threshold requirement. The verification process also needs to consider dynamic feedback data to confirm the stability of the water quality improvement, such as whether there will be a rebound in subsequent days. In this way, secondary simulation and verification can further refine the plan and ensure the sustainability of the treatment results.

[0147] Based on the final simulation evaluation results, adjustments are generated to document the degree to which the dosing effect matches the water quality improvement target. For example, in the aforementioned lake treatment project, if the final results show that nitrogen levels remain stable below 4.8 mg / L and pollution in the southeastern region of the lake is effectively controlled, the system will record the adjustment of increasing the dosing by 10 kg as key evidence and recommend that this configuration be prioritized under similar conditions. This recording and integration approach can provide a reference for future treatments, improving the scientific and targeted nature of plan development.

[0148] It should be noted that, from multiple perspectives, the collaborative work of simulation evaluation, data comparison, parameter optimization, and result integration can form a complete closed loop from effect prediction to program adjustment. The simulation system provides simulated water quality index data, identifies problems through comparison, optimizes dosing parameters, and ultimately accumulates experience for subsequent decision-making through integrated solutions. Suppose that in another time period, the nitrogen content of the lake suddenly rises to 8.0 mg / L. Based on the adjustment basis recorded in the previous period, the dosage can be quickly determined to be 65 kg and concentrated. This multi-link collaborative approach can significantly improve the adaptability and accuracy of the treatment plan, and provide a strong guarantee for the protection of lake water quality.

[0149] In step S108 , based on the second dosing plan, a dosing instruction is generated to control the dosing device to perform a dosing operation.

[0150] It should be noted that through the second dosing plan, the automated dosing equipment is driven to perform precise dosing operations, record the dosing data in real time, and form feedback data for subsequent monitoring, model updates, and plan adjustments, thereby continuously optimizing the water quality management process.

[0151] Specifically, the specific parameters of the dosing dosage and dosing time are extracted from the second dosing scheme to obtain a configuration instruction set for device driving. Through the configuration instruction set, the device control technology is used to send an execution signal to the automated dosing equipment. When the dosing operation is executed, the changes in the data of each dosing are recorded in real time to determine the content of the completed dosing log. According to the content of the dosing log, the dosing data is classified and sorted. According to the goal of water quality management, the various water quality parameters after dosing are compared with the corresponding preset threshold range to determine whether they meet the threshold range and obtain structured feedback data. If some water quality parameters in the feedback data do not reach the preset threshold range, they will be uploaded to the central database of the monitoring process. Combined with the goal of continuous optimization, the data update mechanism is triggered regularly to obtain the latest basis for program adjustment.

[0152] In the business field of lake water quality management, the second dosing plan can be refined and monitored in response to the precise delivery requirements of dosing operations.

[0153] First, break down the second dosing plan's specifications, such as dosage and interval, into instructions that the equipment can recognize. For example, in a lake management project, the second dosing plan specifies daily dosing of 40 kg of a phosphorus removal agent, every six hours. Converting this data into specific equipment-driven instructions, such as dosing 10 kg per dose for four times, ensures the equipment executes as planned.

[0154] The configuration instruction set is then converted into execution signals for automated equipment. For example, in a lake management system, the dosing equipment used is an automated pumping system. Based on the configuration instruction set, a signal is sent to the dosing equipment to ensure that the dosing starts at the specified time. The actual dosage and time of each dosing are recorded in real time, creating a detailed dosing log. This approach ensures operational accuracy and provides data support for subsequent analysis.

[0155] Data from the dosing log is categorized and collated, and compared to water quality management targets. For example, in a specific scenario, if the dosing log shows that the phosphorus content in lake water after a certain dosing was 0.6 mg / L, while the target value was 0.5 mg / L, this indicator would be marked as not meeting the standard, and a structured feedback data package containing the specific value and deviation would be generated. This organization method facilitates rapid identification of issues and provides a basis for subsequent adjustments.

[0156] If feedback data indicates that certain water quality parameters are not meeting standards, the relevant information can be uploaded to a central database. For example, if phosphorus levels consistently fail to reach target values ​​during a treatment campaign, the system will regularly trigger a data update mechanism to obtain the latest water quality monitoring data and environmental variables, such as changes in water temperature or rainfall, to serve as new basis for adjusting the dosing plan. This mechanism ensures that treatment plans are dynamically aligned with the actual environment.

[0157] It should be noted that from data analysis to integration and transmission and updating, each link is closely connected to form a closed-loop management system. This not only improves the accuracy of dosing operations, but also provides reliable data support for continuous optimization and ensures the scientific nature and traceability of the management process.

[0158] In summary, multi-point sensing devices collect water quality parameters in real time, generating time-series water quality parameter data. Preprocessing this data, including data cleaning, yields a first water quality dataset, improving the accuracy and consistency of the underlying data. Time-series analysis of this first water quality dataset identifies water quality trends and generates a water quality trend feature set, providing critical data support for water quality prediction. A water quality state prediction model processes this water quality trend feature set to obtain predicted water quality parameter values ​​for future time periods, enabling trend prediction and facilitating the development of remediation plans. Anomaly analysis of predicted water quality parameter values ​​against preset parameter thresholds enables early warning monitoring. Based on early warning data and historical remediation data, a decision tree model is used to rapidly optimize dosing parameters, ultimately determining the initial dosing schedule. Simulations can then be used to evaluate the dosing effectiveness of the initial dosing schedule, with iterative optimization performed as necessary to ultimately drive automated equipment to execute precise dosing operations. This approach effectively improves the accuracy and timeliness of water pollution control, providing an intelligent solution for water environment management.

[0159] Based on the same inventive concept, an embodiment of the present application also provides a water quality parameter monitoring and intelligent dosing system based on the Internet of Things. The water quality parameter monitoring and intelligent dosing system based on the Internet of Things corresponds one-to-one to the water quality parameter monitoring and intelligent dosing system method based on the Internet of Things in the above embodiment.

[0160] See Figure 3 , Figure 3 The schematic diagram of the principle of a water quality parameter monitoring and intelligent dosing system based on the Internet of Things (IoT) is shown in an embodiment of the present application. The IoT-based water quality parameter monitoring and intelligent dosing system 200 of the present embodiment includes an acquisition and preprocessing module 201, a time analysis module 202, a state prediction module 203, an anomaly analysis module 204, and a solution generation module 205. The functional modules are described in detail as follows:

[0161] The acquisition and preprocessing module 201 is used to acquire time series water quality parameter data through multi-point sensing equipment, perform data preprocessing on the water quality parameter data, and obtain a first water quality data set;

[0162] A time analysis module 202 is used to perform time series analysis on the first water quality data set to obtain a water quality trend feature set;

[0163] The state prediction module 203 is used to input the water quality trend feature set into a preset water quality state prediction model to obtain the predicted water quality parameter value within a preset future time period;

[0164] Anomaly analysis module 204 is used to perform anomaly analysis on the predicted water quality parameter values ​​and the preset parameter threshold range to obtain early warning data, which includes pollution risk areas, abnormal time intervals and risk categories;

[0165] The plan generation module 205 is used to optimize and calculate the dosing parameters based on the early warning data and the historical treatment data through a tree model to obtain a first dosing plan.

[0166] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, a network interface and a database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store time series water quality parameter data, a first water quality data set, a water quality trend feature set, early warning data, and dosing plan data. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, the steps of implementing a water quality parameter monitoring and intelligent dosing method based on the Internet of Things in the above embodiment are implemented.

[0167] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of a water quality parameter monitoring and intelligent dosing method based on the Internet of Things in the above embodiment are implemented.

[0168] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0169] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A water quality parameter monitoring and intelligent dosing method based on the Internet of Things, characterized in that: include: Acquire time-series water quality parameter data through a multi-point sensing device, and perform data preprocessing on the water quality parameter data to obtain a first water quality data set; Performing time series analysis on the first water quality data set to obtain a water quality trend feature set; Inputting the water quality trend feature set into a preset water quality state prediction model to obtain predicted water quality parameter values ​​within a preset future time period; Performing an anomaly analysis on the predicted water quality parameter values ​​and the preset parameter threshold range to obtain early warning data, the early warning data including pollution risk areas, abnormal time intervals and risk categories; According to the early warning data and historical treatment data, the dosing parameters are optimized and calculated by a tree model to determine a first dosing plan; The time series analysis of the first water quality data set is performed to obtain a water quality trend feature set, including: Using a time series decomposition algorithm to perform stratification processing on the first water quality data set to obtain multiple data subsets with different change characteristics; determining trend slope data of the water quality parameters in the plurality of data subsets within each sliding time window; Performing correction processing according to the trend slope data and a preset slope threshold range; Performing multi-dimensional feature extraction on the trend slope data of the corrected water quality parameters to obtain a water quality trend feature set; Before determining the trend slope data of the water quality parameters in each time window in the plurality of data subsets, the method further includes: performing weighted processing on the water quality parameters in the plurality of data subsets according to preset environmental factor weights; The time series decomposition algorithm adopts the STL algorithm to decompose the time series data into trend, season and residual components.

2. The water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to claim 1 is characterized in that: The water quality parameter data include dissolved oxygen, turbidity, pH, nitrogen content and phosphorus content.

3. The water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to claim 1 is characterized in that: The training process of the water quality state prediction model includes: A water quality status prediction model is constructed based on a long short-term memory neural network, where the input layer is the historical water quality trend characteristics and the output layer is the predicted values ​​of water quality parameters; The mean square error between the predicted value of water quality parameters and the true value of water quality parameters is taken as the loss value; Based on the loss value, executing the back propagation algorithm to update the water quality state prediction model parameters and optimize the water quality state prediction model; When it is detected that the loss value of the model meets the conditions or the number of training times reaches the set upper limit, the training is considered to be completed and the trained model is obtained.

4. The water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to claim 1 is characterized in that: Anomaly analysis is performed on the predicted water quality parameter values ​​and the preset parameter threshold range to obtain early warning data, including: Determining abnormal data points and corresponding abnormal time intervals based on the predicted water quality parameter values ​​and preset parameter threshold ranges; Determine whether the abnormal duration corresponding to the abnormal data point exceeds a preset time threshold range; If so, the abnormal time distribution and abnormal parameter type of the abnormal data point are classified to obtain the risk category corresponding to the abnormal data point; Acquiring historical water quality records, wherein the historical water quality records include abnormal areas and their corresponding abnormal categories and abnormal time periods; The pollution risk area is determined based on the abnormal time interval and risk category of the abnormal data point in combination with the historical water quality records.

5. The water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to any one of claims 1 to 4, characterized in that: Based on the early warning data and historical treatment data, and by optimizing and calculating the dosing parameters through a tree model, a first dosing plan is obtained, including: According to the abnormal time interval and risk category corresponding to the current pollution risk area, the relevant target governance records are matched in the historical governance data; Determine the remediation type of the target remediation record based on the pollution area characteristics and time distribution in the target remediation record; Determine whether the pollution parameter level corresponding to the current pollution risk area exceeds the preset pollution threshold range; If so, a decision tree model is used to compare the target treatment record with the preset drug matching standard, and based on the treatment type, the initial dosing parameters are determined; The initial dosing parameters are dynamically adjusted according to environmental variables and drug efficacy feedback information to determine a first dosing plan.

6. The water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to any one of claims 1 to 4, characterized in that: Also includes: Performing a simulation evaluation on the dosing effect of the first dosing scheme to obtain an evaluation result; If the evaluation result does not meet the expected effect, iteratively optimize the first dosing plan to generate a second dosing plan; Based on the second dosing plan, a dosing instruction is generated to control the dosing device to perform a dosing operation.

7. The water quality parameter monitoring and intelligent dosing method based on the Internet of Things according to claim 6 is characterized in that: The process of simulating and evaluating the dosing effect of the first dosing scheme includes: performing environmental simulation on the first dosing scheme through a simulation system to determine simulated water quality index data; If the evaluation result does not meet the expected effect, the first dosing plan is iteratively optimized to generate a second dosing plan, including: Preliminarily revising the dosing parameters of the first dosing scheme according to the water quality parameter values ​​in the simulated water quality index data and a preset simulation threshold range; The dosing parameters after preliminary correction are simulated and optimized in combination with dynamic feedback data to determine the second dosing plan.

8. A water quality parameter monitoring and intelligent dosing system based on the Internet of Things, characterized in that: The method for implementing water quality parameter monitoring and intelligent dosing based on the Internet of Things as claimed in any one of claims 1 to 7 comprises: an acquisition and preprocessing module, configured to acquire time-series water quality parameter data through a multi-point sensing device, and perform data preprocessing on the water quality parameter data to obtain a first water quality data set; a time analysis module, configured to perform time series analysis on the first water quality data set to obtain a water quality trend feature set; A state prediction module is used to input the water quality trend feature set into a preset water quality state prediction model to obtain predicted water quality parameter values ​​within a preset future time period; An anomaly analysis module, configured to perform an anomaly analysis on the predicted water quality parameter values ​​and a preset parameter threshold range to obtain early warning data, wherein the early warning data includes pollution risk areas, abnormal time intervals, and risk categories; A scheme generation module is used to optimize and calculate the dosing parameters based on the early warning data and historical management data through a tree model to determine a first dosing scheme.

Images