Water environment treatment method and platform based on water environment mathematical model and artificial intelligence

Abstract

The invention discloses a water environment treatment method and platform based on a water environment mathematical model and artificial intelligence, and belongs to the technical field of water environment treatment, and the method comprises the following steps: S1, obtaining water quality data including water depth, water level, flow velocity, flow direction, temperature, salinity, dissolved oxygen, pH value, chemical oxygen demand, ammonia nitrogen, total phosphorus and turbidity; s2, constructing a water environment mathematical model and an artificial intelligence model; the method specifically comprises the following steps: constructing a water environment mathematical model by using water quality data; the water environment mathematical model is used for simulating a hydrodynamic process and spatial and temporal distribution of water quality parameters; the method comprises the following steps: firstly, selecting a basic artificial intelligence model, then carrying out training learning on the artificial intelligence model by using water quality data, and finally obtaining a trained artificial intelligence model; s3, analyzing a water environment by using the artificial intelligence model to obtain spatial and temporal distribution characteristics and dominant factors of water quality parameters; and S4, making a water environment treatment scheme according to the spatial and temporal distribution characteristics of the water quality parameters and the dominant factors.

Description

Technical Field

[0001] The present invention belongs to the technical field of water environment management, and in particular relates to a water environment management method and platform based on a water environment mathematical model and artificial intelligence. Background Art

[0002] As humans pay more and more attention to water environmental issues, water pollution and water shortage have become major challenges affecting human survival and development. The increasing discharge of industrial wastewater, agricultural runoff and urban sewage has led to a large number of harmful substances entering the water body, exacerbating the water pollution problem. These pollutants include heavy metals, organic pollutants and pathogenic microorganisms, which not only damage aquatic ecosystems, but also threaten human health through food chains and water cycles. Especially in developing countries and emerging economies, with the acceleration of industrialization and the imperfection of environmental protection laws and regulations, water pollution problems are more prominent. At the same time, extreme weather events caused by global climate change, such as long-term droughts and severe floods, have further exacerbated the unbalanced distribution of water resources. Many regions, especially densely populated and rapidly developing cities, are facing the dilemma of water shortage. This shortage not only affects the basic living needs of residents, but also has an impact on agricultural irrigation and industrial production, which in turn has a profound impact on social and economic development.

[0003] Traditional water environment governance methods mainly rely on manual experience and conventional monitoring methods, but there are problems such as delayed response, low efficiency and limited coverage. The shortcomings of these methods are mainly reflected in the following aspects: First, they rely too much on manual experience and lack scientific basis; second, the monitoring methods are lagging and the data timeliness is poor; third, the monitoring coverage is limited, and it is difficult to fully reflect the water environment status; fourth, the high cost and high manpower demand limit its application on a large scale, especially in economically underdeveloped areas; fifth, the governance measures often lack systematicity and comprehensiveness, and it is difficult to fundamentally solve the problem. Traditional governance methods are difficult to provide effective solutions when dealing with complex water environment problems. For example, conventional monitoring usually relies on fixed monitoring points and periodic sampling and analysis. Data acquisition has a certain lag, and it is difficult to capture the dynamic changes of the water environment in real time, resulting in the inability of governance measures to respond to emergencies in a timely manner. In addition, the layout and number of monitoring points are limited, making it difficult to cover the entire water body range, and there are blind spots and information missing. High costs and high manpower requirements further limit the application of these methods on a larger scale. Traditional governance methods are often single and local, lack systematicity and comprehensiveness, and are difficult to fundamentally solve the problem. These shortcomings indicate that traditional water environment governance methods are no longer able to cope with increasingly complex and severe water environment problems. Therefore, it is urgent to introduce new technical means and methods to improve governance efficiency and effectiveness and meet the current and future needs of water environment governance.

[0004] In order to meet these challenges, we need to adopt more advanced technologies and methods, such as Internet of Things technology, big data analysis and artificial intelligence. These new technologies can achieve real-time monitoring and intelligent early warning of the water environment, and improve the timeliness and accuracy of governance. By deploying a large number of sensors and monitoring equipment, we can achieve comprehensive coverage of water bodies, timely detect pollution sources and pollution incidents, and take effective countermeasures. In addition, big data analysis can help us better understand the complexity and dynamic changes of the water environment, and provide a basis for formulating scientific and reasonable governance strategies. Artificial intelligence technology can assist decision-making, optimize resource allocation, and improve governance efficiency. Through the application of these new technologies, we can achieve comprehensive management and sustainable governance of the water environment, ensure the rational use and protection of water resources, and provide a solid foundation for the sustainable development of human society. Summary of the invention

[0005] The purpose of the present invention is to meet actual needs and provide a water environment management method and platform based on water environment mathematical model and artificial intelligence. By constructing water environment mathematical model and artificial intelligence model, the water environment can be accurately simulated and predicted, providing accurate data support for water environment management.

[0006] The first object of the present invention is to provide a water environment management method based on a water environment mathematical model and artificial intelligence, comprising:

[0007] S1. Obtain water quality data, including water depth, water level, flow velocity, flow direction, temperature, salinity, dissolved oxygen, pH value, chemical oxygen demand, ammonia nitrogen, total phosphorus and turbidity;

[0008] S2. Construct water environment mathematical model and artificial intelligence model; specifically include:

[0009] The water quality data is used to construct a water environment mathematical model; the water environment mathematical model is used to simulate the hydrodynamic process and the temporal and spatial distribution of water quality parameters;

[0010] First, a basic AI model is selected, then the AI ​​model is trained and learned using water quality data, and finally a trained AI model is obtained;

[0011] S3. Analyze the water environment using the artificial intelligence model to obtain the spatiotemporal distribution characteristics and dominant factors of water quality parameters;

[0012] S4. Formulate water environment management plans based on the temporal and spatial distribution characteristics of water quality parameters and dominant factors.

[0013] Preferably, the water environment mathematical model adopts EFDC as the basic framework, and defines key parameters through multiple input files; wherein: hydrodynamic parameters include model grid and water depth, boundary conditions, initial conditions and water body dynamic parameters; water quality parameters include the concentration and dynamic process of dissolved oxygen, nutrients and algae plankton; sediment parameters include the sedimentation rate of suspended matter, the resuspension process and the material exchange between the bottom mud and the water body; thermodynamic parameters include water temperature distribution and its heat exchange process with the atmosphere.

[0014] Preferably, the artificial intelligence model includes a convolutional neural network and a long short-term memory network; wherein:

[0015] The convolutional neural network input data includes the spatial distribution map of water quality parameters such as water temperature, dissolved oxygen, ammonia nitrogen, and total phosphorus; by performing deep learning on the spatial distribution map of water quality parameters, the spatial pattern and trend of water quality parameters are identified;

[0016] The data input to the long short-term memory network is the time series data of water temperature, pH value and dissolved oxygen obtained by continuous monitoring. By analyzing the time series data, the future change trend of water quality parameters is predicted.

[0017] Preferably, the spatial patterns and trends of the water quality parameters include the distribution range of pollution hot spots and the gradient changes in pollutant concentrations; the predicted future change trends of the water quality parameters include predictions of short-term fluctuations in dissolved oxygen levels or long-term increases or decreases in pollutant concentrations.

[0018] The second object of the present invention is to provide a water environment management platform based on a water environment mathematical model and artificial intelligence, comprising:

[0019] A water quality data acquisition module, including several sensors for acquiring water depth, water level, flow velocity, flow direction, temperature, salinity, dissolved oxygen, pH value, chemical oxygen demand, ammonia nitrogen, total phosphorus and turbidity parameters;

[0020] Model building modules, including:

[0021] The water quality data is used to construct a water environment mathematical model; the water environment mathematical model is used to simulate the hydrodynamic process and the temporal and spatial distribution of water quality parameters;

[0022] First, a basic AI model is selected, then the AI ​​model is trained and learned using water quality data, and finally a trained AI model is obtained;

[0023] An analysis module, which uses the artificial intelligence model to analyze the water environment to obtain the spatiotemporal distribution characteristics and dominant factors of water quality parameters;

[0024] The governance module formulates water environment governance plans based on the spatiotemporal distribution characteristics and dominant factors of water quality parameters.

[0025] Preferably, the water environment mathematical model adopts EFDC as the basic framework, and defines key parameters through multiple input files; wherein: hydrodynamic parameters include model grid and water depth, boundary conditions, initial conditions and water body dynamic parameters; water quality parameters include the concentration and dynamic process of dissolved oxygen, nutrients and algae plankton; sediment parameters include the sedimentation rate of suspended matter, the resuspension process and the material exchange between the bottom mud and the water body; thermodynamic parameters include water temperature distribution and its heat exchange process with the atmosphere.

[0026] Preferably, the artificial intelligence model includes a convolutional neural network and a long short-term memory network; wherein:

[0027] The convolutional neural network input data includes the spatial distribution map of water quality parameters such as water temperature, dissolved oxygen, ammonia nitrogen, and total phosphorus; by performing deep learning on the spatial distribution map of water quality parameters, the spatial pattern and trend of water quality parameters are identified;

[0028] The data input to the long short-term memory network is the time series data of water temperature, pH value and dissolved oxygen obtained by continuous monitoring. By analyzing the time series data, the future change trend of water quality parameters is predicted.

[0029] Preferably, the spatial patterns and trends of the water quality parameters include the distribution range of pollution hot spots and the gradient changes in pollutant concentrations; the predicted future change trends of the water quality parameters include predictions of short-term fluctuations in dissolved oxygen levels or long-term increases or decreases in pollutant concentrations.

[0030] The third object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned water environment management method based on water environment mathematical model and artificial intelligence.

[0031] The fourth object of the present invention is to provide a computer program product, including a computer program, which, when executed by a processor, implements the above-mentioned water environment management method based on water environment mathematical model and artificial intelligence.

[0032] The advantages and positive effects of this application are:

[0033] The present invention not only realizes the accurate simulation and prediction of water environment by deeply integrating water quality data, constructing precise water environment mathematical model and advanced artificial intelligence model, but also provides scientific basis and decision-making support for water environment management.

[0034] First of all, the mathematical model of water environment, as the cornerstone of the entire system, can not only depict the complex physical processes in the water body in detail, such as water flow movement, turbulence effects and interactions between water bodies, but also accurately simulate the diffusion and transformation of chemical substances, such as the migration, transformation and degradation of pollutants in water bodies. At the same time, it can also deeply explore biological processes, such as the evolution of ecosystems, the structure and function of aquatic communities, etc. When these simulation results are combined with artificial intelligence algorithms, especially deep learning technology, by mining hidden laws and trends from historical data, it can timely identify subtle signs of water quality changes, and win precious time and space for water environment governance.

[0035] Secondly, the real-time monitoring and data acquisition system of the present invention can realize all-round and all-weather monitoring of water quality parameters through a sensor network densely distributed throughout the water area. From water temperature, pH value to key indicators such as dissolved oxygen, ammonia nitrogen, total phosphorus, turbidity, etc., no one is missed. The rapid collection and transmission of these real-time data provides the management department with first-hand and accurate water environment information. They can grasp the latest developments of water quality at the first time, carry out targeted monitoring and intervention, thereby greatly improving the efficiency and effect of governance.

[0036] The intelligent analysis and optimization capabilities are another highlight of the present invention. With the help of the powerful force of artificial intelligence technology, it is possible to deeply mine and analyze massive amounts of water quality data and extract valuable information and knowledge from them. Through the application of machine learning algorithms, the dominant pollution factors and their spatiotemporal distribution characteristics can be accurately identified, providing strong support for the formulation of scientific and reasonable governance measures. At the same time, these algorithms also have the ability of self-learning and optimization, and can continuously improve their intelligence level with the continuous accumulation of data and the continuous iteration of models. This enables water environment governance measures to respond to various complex situations and challenges more accurately and effectively.

[0037] Last but not least, the early warning and forecasting function of the present invention. Based on the powerful support of the big data analysis platform, it can integrate real-time monitoring data and historical data resources, and generate accurate early warning information through complex algorithm models. This information can not only help management departments respond to emergencies in a timely manner and reduce environmental risks; it can also provide the public with timely and accurate water quality information to remind them to pay attention to water quality changes and protect water resources safety. This undoubtedly lays a solid foundation for mankind to build a safer, healthier and more harmonious water environment ecosystem.

[0038] In summary, the present invention provides an efficient, intelligent and comprehensive solution for mankind through the deep integration and innovative application of multiple technologies. It not only significantly improves the scientificity, timeliness and effectiveness of water environment management, but also provides strong support and guarantee for addressing water shortage and pollution problems. In the future development, we expect this method to be more widely used and promoted to contribute more to the sustainable development of mankind. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a flow chart of a preferred embodiment of the present invention;

[0040] Figure 2 is a flow chart of EFDC modeling in a preferred embodiment of the present invention;

[0041] Figure 3 It is a weight prediction roadmap in a preferred embodiment of the present invention;

[0042] Figure 4 is a workflow diagram of the platform in a preferred embodiment of the present invention;

[0043] Figure 5 is a structural diagram of three lakes in a preferred embodiment of the present invention;

[0044] Figure 6 The numerical simulation results of Lake B in the preferred embodiment of the present invention are compared with the measured data;

[0045] Figure 7 The numerical simulation results of Lake X in the preferred embodiment of the present invention are compared with the measured data;

[0046] Figure 8 It is a comparison between the numerical simulation results of Lake N in the preferred embodiment of the present invention and the measured data;

[0047] Fig. 9 is a graph showing the flow velocity change at the gate connecting Lake B and Lake N and the flow velocity change at the gate connecting Lake B and Lake X in a preferred embodiment of the present invention;

[0048] Fig.10 is a graph showing the flow velocity change at the gate connecting Lake B and Lake N and the flow rate change at the gate connecting Lake B and Lake X in a preferred embodiment of the present invention;

[0049] Fig.11 This is a flow field diagram during water replenishment in a preferred embodiment of the present invention;

[0050] Fig.12 This is a flow field diagram near the connecting gate under water replenishment working condition in a preferred embodiment of the present invention;

[0051] Fig.13 It is a flow field diagram of the whole lake under static working condition in the preferred embodiment of the present invention;

[0052] Fig.14 is a flow field diagram of Lake X in a preferred embodiment of the present invention;

[0053] Fig.15 Schematic diagram of wind speed and wind direction in a preferred embodiment of the present invention;

[0054] Fig.16 This is a flow field diagram of a preferred embodiment of the present invention after being left to stand for 75 days;

[0055] Fig.17 This is a distribution diagram of water environment elements on the 5th day of water replenishment in a preferred embodiment of the present invention;

[0056] Fig.18 is a chlorophyll a distribution diagram on the 5th day in a preferred embodiment of the present invention;

[0057] Fig.19 This is a distribution diagram of water environment elements after standing for 30 days in a preferred embodiment of the present invention;

[0058] Fig. 20 This is the chlorophyll a distribution diagram on the 30th day in the preferred embodiment of the present invention. DETAILED DESCRIPTION

[0059] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] In order to solve the global water pollution and water shortage problems, as well as the shortcomings of traditional water environment management methods, the present invention provides a water environment management method based on water environment mathematical model and artificial intelligence. This method achieves efficient, intelligent and comprehensive water environment management through accurate simulation and prediction, real-time monitoring and data collection, intelligent analysis and optimization, comprehensive coverage and real-time monitoring, and effective early warning and forecasting. In order to further illustrate the structure of the present invention, the detailed description is as follows in conjunction with the attached drawings:

[0061] See also Figures 1 to 20 ,

[0062] A water environment management method based on a water environment mathematical model and artificial intelligence comprises the following steps:

[0063] S1. Obtain water quality data, including but not limited to water depth, water level, flow velocity, flow direction, temperature, salinity, dissolved oxygen, pH value, chemical oxygen demand (COD), ammonia nitrogen, total phosphorus and turbidity;

[0064] S2. Construct water environment mathematical model and artificial intelligence model; specifically include:

[0065] The water quality data is used to construct a water environment mathematical model; the water environment mathematical model is used to simulate the hydrodynamic process and the temporal and spatial distribution of water quality parameters;

[0066] First, a basic AI model is selected, then the AI ​​model is trained and learned using water quality data, and finally a trained AI model is obtained;

[0067] S3. Analyze the water environment using the artificial intelligence model to obtain the spatiotemporal distribution characteristics and dominant factors of water quality parameters;

[0068] S4. Formulate water environment management plans based on the temporal and spatial distribution characteristics of water quality parameters and dominant factors.

[0069] The present invention first establishes a water environment mathematical model to simulate the hydrodynamic process of the water body; the water environment mathematical model can use EFDC (Environmental Fluid Dynamics Code) as the basic framework, which is an advanced three-dimensional fluid dynamics model designed specifically for simulating water flow, water quality changes and ecosystem interactions. The construction of this model depends on a series of key parameters, including but not limited to water depth, water level, flow rate, flow direction, temperature, salinity, dissolved oxygen, pH value, chemical oxygen demand (COD), ammonia nitrogen, total phosphorus and turbidity. These parameters jointly determine the physical state and chemical composition of the water body, so that the EFDC model can accurately simulate the spatiotemporal distribution of hydrodynamic processes and water quality parameters.

[0070] See also Figure 2EFDC model is a three-dimensional hydrodynamic and water quality model for water environment simulation. Its core defines the key parameters of the model through multiple modules (inp files). The input files include control files and hydrodynamic and topographic data. The control files include the master file (efdc.inp), display file (show.inp) and restart file (restart.inp), which are used to set the basic operating parameters and display options of the model. The hydrodynamic and topographic data include grid files (cell.inp) and bottom water depth files (dxdy.inp) to ensure that the model accurately reflects the topography and structure of the water body. Next, the hydrodynamic model uses initial and boundary dynamic files such as salinity file (salt.inp), meteorological file (aser.inp), water level file (Pser.inp), water temperature file (Tser.inp), etc. to define the initial state and dynamic boundary conditions of physical parameters in the water body. In addition, the tracer column file (dye.inp), flow time series file (Qser.inp) and salinity time series file (Sser.inp) further describe the spatiotemporal changes of hydrodynamics. The water quality and water ecology model uses the water quality input file (Wq3dwc.inp), the open boundary time series file (Cwqsr(N).inp) and the water quality point source time series file (Wqpsl.inp) to simulate the changes in chemical and ecological processes in the water body, covering key water quality indicators such as chemical oxygen demand, dissolved oxygen, total phosphorus, and total nitrogen. After running the model, the user can analyze the output results. If the model does not meet expectations, the parameters can be adjusted and rerun until satisfactory results are obtained. Finally, the application result file generated by the model includes simulation results of key indicators such as water temperature, total phosphorus, total nitrogen, dissolved oxygen, and chlorophyll-a, providing a scientific basis for water environment management.

[0071] The present invention collects water quality data through sensors and uploads the data to the platform in real time; the distribution of sensors follows scientific laws and takes into account the spatial structure and characteristics of water bodies, such as distribution density and water depth, to ensure comprehensive coverage and representativeness of data. The data types collected by the sensors correspond to the parameters in the EFDC model, such as water temperature, dissolved oxygen, ammonia nitrogen and other key water quality indicators. These real-time collected data do not simply replace the model predictions, but are fused with the model data through advanced data assimilation technology to improve the model's prediction accuracy and ability to respond to actual environmental changes.

[0072] The present invention uses artificial intelligence algorithms to analyze the spatiotemporal distribution characteristics and dominant factors of water quality parameters; according to the analysis results, the water environment is warned and forecasted, and abnormal situations are discovered and handled in time; using artificial intelligence algorithms, especially machine learning and deep learning technologies, the module can identify the spatiotemporal distribution characteristics of water quality parameters and their dominant factors. In previous studies, parameters with larger values ​​are usually identified as the main influencing factors through processing methods such as data normalization. However, there is no consensus on the research conclusions of the main influencing factors of water bloom. To this end, this paper uses a method combining artificial intelligence algorithms with EFDC mechanism models to accurately determine the weight distribution of each factor. This method can more scientifically reveal the key driving factors affecting the formation of water blooms, thereby providing a more reliable basis for its effective prediction and governance. According to the analysis results, the system can provide real-time warning and forecasting of the water environment, and timely discover and handle water quality anomalies and potential risks. The warning and forecasting system is based on a big data analysis platform, integrating real-time monitoring data and historical data, and the generated warning information can be pushed to the management department in time, providing scientific basis and decision support for water environment governance. The artificial intelligence algorithm includes a convolutional neural network and a long short-term memory network, which are used to process spatiotemporal data and perform predictive analysis. Convolutional neural networks (CNNs) capture local features of image data through convolutional layers to achieve in-depth analysis of the spatial distribution of water quality parameters. Specifically, the data input to CNNs are spatial distribution maps of water quality parameters including water temperature, dissolved oxygen, ammonia nitrogen, total phosphorus, etc. These distribution maps provide the geographical distribution of each parameter in the water body. By deep learning these image data, CNNs can identify spatial patterns and trends of water quality parameters, such as the distribution range of pollution hotspots and the gradient changes in pollutant concentrations. As an advanced recurrent neural network, long short-term memory networks (LSTMs) are particularly good at processing and analyzing time series data and can capture and memorize long-term dependencies. In the application of water environment governance methods, LSTMs focus on analyzing the evolution trend of water quality parameters over time. The data input to LSTMs are historical records of parameters such as water temperature, pH value, dissolved oxygen, etc. obtained through continuous monitoring. These time series data provide LSTMs with rich temporal information. Through in-depth analysis of these data, LSTMs can predict future trends in water quality parameters, such as short-term fluctuations in dissolved oxygen levels or long-term increases and decreases in pollutant concentrations.

[0073] The present invention comprehensively utilizes environmental fluid mechanics mathematical models, artificial intelligence technology, and machine learning algorithms to optimize water environment governance measures; the early warning and forecasting system is based on a big data analysis platform, integrates water quality monitoring data and historical data, generates early warning information, and pushes it to the management department. Specifically, it includes: early warning and forecasting of the water environment based on the analysis results, and timely detection and handling of abnormal situations: the early warning and forecasting system is based on a big data analysis platform, integrating real-time monitoring data and historical data, and conducting comprehensive analysis. When the system detects abnormal water quality or predicts possible environmental problems, it automatically generates early warning information and notifies the management department and relevant personnel through various channels (such as text messages, emails, APP push, etc.) to ensure timely response and processing.

[0074] A water environment management platform based on a water environment mathematical model and artificial intelligence is used to implement the water environment management method based on a water environment mathematical model and artificial intelligence in the above embodiment, and the platform includes:

[0075] A water quality data acquisition module, including several sensors for acquiring water depth, water level, flow velocity, flow direction, temperature, salinity, dissolved oxygen, pH value, chemical oxygen demand, ammonia nitrogen, total phosphorus and turbidity parameters;

[0076] Model building modules, including:

[0077] The water quality data is used to construct a water environment mathematical model; the water environment mathematical model is used to simulate the hydrodynamic process and the temporal and spatial distribution of water quality parameters;

[0078] First, a basic AI model is selected, then the AI ​​model is trained and learned using water quality data, and finally a trained AI model is obtained;

[0079] An analysis module, which uses the artificial intelligence model to analyze the water environment to obtain the spatiotemporal distribution characteristics and dominant factors of water quality parameters;

[0080] The governance module formulates water environment governance plans based on the spatiotemporal distribution characteristics and dominant factors of water quality parameters.

[0081] The water environment mathematical model adopts EFDC as the basic framework and defines key parameters through multiple input files; among them: hydrodynamic parameters include model grid and water depth, boundary conditions, initial conditions and water body dynamic parameters; water quality parameters include the concentration and dynamic process of dissolved oxygen, nutrients and algae plankton; sediment parameters include the sedimentation rate of suspended matter, resuspension process and material exchange between sediment and water body; thermodynamic parameters include water temperature distribution and its heat exchange process with the atmosphere.

[0082] The artificial intelligence model includes a convolutional neural network and a long short-term memory network; wherein:

[0083] The convolutional neural network input data includes the spatial distribution map of water quality parameters such as water temperature, dissolved oxygen, ammonia nitrogen, and total phosphorus; by performing deep learning on the spatial distribution map of water quality parameters, the spatial pattern and trend of water quality parameters are identified;

[0084] The data input to the long short-term memory network is the time series data of water temperature, pH value and dissolved oxygen obtained by continuous monitoring. By analyzing the time series data, the future change trend of water quality parameters is predicted.

[0085] The spatial patterns and trends of the water quality parameters include the distribution range of pollution hot spots and the gradient changes in pollutant concentrations; the predicted future change trends of the water quality parameters include predictions of short-term fluctuations in dissolved oxygen levels or long-term increases or decreases in pollutant concentrations.

[0086] Specific implementation cases

[0087] 1. Model establishment: The lake consists of Lake B, Lake N and Lake X, which contains a total of 8837 grids, such as Figure 5 As shown in the figure, the left figure is the grid division diagram of the lake area, and the right figure is the three-dimensional terrain processing diagram. At the connecting gate location, due to the complex interaction of water flow, conventional grids may not be able to fully reflect the details of hydrodynamics and water quality changes. Therefore, the model uses a denser grid layout in these key areas to simulate the hydrodynamics and water environment characteristics in these areas in a more detailed manner.

[0088] 2. Model verification: The model uses data from July 1, 2018 to August 10, 2018 for simulation and verification. From July 1 to 10 (1-10 days), the pump station of Lake B is opened to replenish the lake with a flow rate of 1m 3 / s, July 11 to August 10 (11-40 days), leave it still without replenishing water, and the water level will drop due to the influence of atmospheric evaporation, etc.

[0089] One point was selected in each lake for model validation, for a total of three points. Water level and water temperature were selected for validation by comparing the model simulation values ​​with the measured values. Figure 5 The left and right figures in the figure show the relevant comparison results to evaluate the accuracy of the model. Except for the water level, all data are measured 0.5m below the surface of the water body.

[0090] The iodine titration method was used to determine the dissolved oxygen (DO) in the sample, and the dichromate method was used to determine the chemical oxygen demand (COD) in the lake. The alkaline potassium persulfate digestion method was used to determine the total nitrogen (TN) content in the lake. For the determination of total phosphorus (TP), the potassium persulfate digestion method was used.

[0091] The numerical simulation results are compared with the measured data. Figure 6 - Figure 8 The comparison of COD, DO, TN, and TP at each point in Lake B, Lake X, and Lake N is shown.

[0092] The simulated and actual values ​​can be analyzed in detail based on the statistics of water quality variables in Table 1. These variables include water level, water temperature, dissolved oxygen (DO), total nitrogen (TN), chemical oxygen demand (COD), and total phosphorus (TP) at different locations.

[0093] Table 1 is the statistical table of simulated and actual water quality variable data

[0094]

[0095]

[0096] The simulation results of Lake B showed low root mean square error (RMSE) and mean absolute error (MAE) and high Nash coefficient (NSE), indicating that the model has high accuracy in predicting water level and hydration indices (such as dissolved oxygen, chemical oxygen demand, total nitrogen and total phosphorus). The simulation results of Lake X also showed high accuracy, especially in the simulation of water temperature and water level, with NSE values ​​close to 0.9, and the NSE values ​​of other hydration indices also performed well. The simulation results of Lake N also performed well in water level and water temperature, with NSE values ​​of 0.82 and 0.93 respectively, but in the simulation of total phosphorus, the NSE value was lower, which was 0.45 in this embodiment.

[0097] The hydrological simulation results of the three lakes of Lake B, Lake X and Lake N show that there are certain differences in the simulation effects of various variables. In terms of water level, the simulation root mean square errors of Lake B, Lake X and Lake N are 0.17m, 0.22m and 0.19m respectively, and the Nash coefficients are 0.80, 0.80 and 0.82 respectively, indicating that the water level simulation of Lake N is slightly better. In terms of water temperature, Lake N has the best simulation effect, with a Nash coefficient of 0.93 and the lowest RMSE of 0.37.

[0098] In the simulation of water quality parameters such as dissolved oxygen, chemical oxygen demand, total nitrogen and total phosphorus, Lake N performed well in chemical oxygen demand and total nitrogen, with Nash coefficients of 0.83 and 0.90, respectively, while the simulation effect of total phosphorus was slightly worse. The model simulated the water quality variables of the three lakes well, especially in variables such as water temperature, flow rate and chemical oxygen demand (COD). Lake N had the best simulation effect on most variables, especially water temperature and total nitrogen, with the highest NSE values, indicating that the model simulated these variables very accurately. However, in the simulation of total phosphorus, the effect of Lake N was poor, and the model parameters and boundary conditions needed to be further optimized.

[0099] 3. Analysis of factors affecting lake flow field

[0100] The entrance is located at Lake B. The circle in the figure is the location of the connecting gate. The connecting gate from Lake B to Lake N is 1#, and the connecting gate from Lake B to Lake X is 2#.

[0101] 3.1. Distribution characteristics of lake flow fields

[0102] The flow velocity changes at the gate connecting Lake B and Lake N and the flow velocity changes at the gate connecting Lake B and Lake X, such as Fig. 9 As shown. During the water replenishment stage (1 to 10 days), the average flow rate of the 1# connecting gate is about 0.027m / s, and the average flow rate of the 2# connecting gate is about 0.0087m / s. When water flows from a narrow area into a wide water area through a connecting gate, a circulation often forms at the outlet. This may lead to the mixing and sedimentation of suspended particles and nutrients in the water body, increasing the risk of eutrophication of the water body. In order to alleviate this phenomenon, the connecting gate can be expanded and renovated, gates can be added to disperse the water flow, and the geometric structure of the connecting gate can be optimized to improve the flow field and further improve the water quality.

[0103] During the static period, wind had a significant impact on the flow of lake water, especially in the three lakes with different shapes, Lake B, Lake N and Lake X. Lake B, due to its relatively regular shape, formed a wide circulation on its lake surface under the action of wind, which promoted the mixing of lake water and gas exchange, and was beneficial to the uniform distribution and self-purification of water quality. The long and narrow shape of Lake N created two main circulation areas in its northern and southern regions. The northern circulation was stronger, which helped maintain the water quality in the area, while the southern circulation was relatively weak, which may lead to local eutrophication problems. In contrast, the complex shape and tortuous shoreline of Lake X restricted the formation of circulation, especially in the southern and western parts of the lake, which made Lake X more prone to problems of reduced dissolved oxygen and accumulation of pollutants during the static period, and the risk of eutrophication increased. Fig.13 As shown, the flow field tends to be stable after 39 days.

[0104] 3.2 The influence of lake shape on lake hydrodynamics

[0105] During the water replenishment process, the shape of the lake has a significant impact on the hydrodynamics. The water in the south of Lake B flows into the north of Lake N through the No. 1 connecting gate. The 90-day flow rate data shows that the maximum flow rate is 0.2213m / s and the minimum flow rate is 0.00043m / s. A circulation is formed at the entrance of Lake N, indicating that the shape of the lake has a great influence on the flow velocity distribution. A lower flow rate may lead to insufficient mixing of the water body and affect the water quality.

[0106] The other part of the water enters Lake X at a slower speed through the No. 2 connecting gate, with a smaller flow rate, the maximum is 0.08m / s, and the minimum is 0.0008m / s. Lake X is long and narrow with an irregular shape, and the flow field is unidirectional. The possibility of circulation is low, and the water body is not updated in time, which may have an adverse impact on water quality.

[0107] 3.3 Analysis of the influence of wind on flow field during static process

[0108] During the static period, wind is the main factor affecting water circulation and nutrient distribution. Wind not only stirs the water surface, but also affects the circulation of the water. During the period of 11 to 40 days, the wind speed ranged from 1.39 to 3.61 m / s, with an average wind speed of 2.297 m / s, and the wind direction was mainly from east-southwest to southwest-west. During the period of 60 to 90 days, the wind speed was stable between 1.10 and 3.51 m / s, with an average wind speed of 2.293 m / s, and the wind direction changed from east-southwest to north-west. The action of wind leads to the formation of water circulation, which affects the transport of dissolved oxygen and nutrients in the water, regulates water quality and the stability of the ecosystem. Strong winds can also stir nutrients and algae from the bottom of the lake to the surface of the water, promote the reproduction and growth of phytoplankton, may cause algal blooms, and have a negative impact on water quality. During the static process, Lake N and Lake B were prone to circulation due to their open water surface, while Lake X did not produce obvious circulation due to its poor fluidity.

[0109] 4. Analysis of factors affecting lake water environment:

[0110] 4.1. Spatial and temporal distribution characteristics of lake water quality:

[0111] (1) Spatial distribution characteristics of lake water environment parameters during water replenishment

[0112] The distribution of water on the 5th day is as follows Fig.17 During the water replenishment period, due to the introduction of water, DO increased significantly at the initial stage of water replenishment, and the concentrations of COD, total nitrogen, and total phosphorus decreased significantly at the initial stage of water replenishment.

[0113] When the water flow rate is 1m 3 / s, the DO of Lake B rose rapidly near the connecting gate, and the initial DO increased significantly, especially in the water inlet and the area with fast flow rate, where the DO concentration could reach 6.45mg / L. The COD of Lake B decreased significantly in the early stage, with the concentration at the water inlet and the area with fast flow rate being 21.78mg / L, and overall it decreased due to regular morphology and good fluidity. Total nitrogen (TN) increased in the early stage, reaching 2.3mg / L in the water inlet and the area with fast flow rate, and the overall distribution was uniform at about 0.89mg / L. Total phosphorus (TP) decreased significantly in the early stage and was evenly distributed, with a minimum of 0.016mg / L. Chlorophyll a was evenly distributed, with a concentration of 12.71ug / L, indicating that the water body of Lake B was in good health and had strong fluidity.

[0114] Due to poor fluidity, DO of Lake X decreased rapidly in the early stage of water replenishment, with an average concentration of 2.75 mg / L, which is prone to hypoxia. The reduction of COD was small, about 83.06 mg / L overall, lower at the connecting gate and higher at the distant point. The reduction of total nitrogen (TN) was small, and the poor fluidity led to a slow nitrogen conversion rate. The reduction of total phosphorus (TP) was also small, ranging from 0.25 to 0.45 mg / L. The average concentration of chlorophyll a was 11.27 ug / L, which indicates that the water ecology of Lake X may be affected by poor fluidity.

[0115] The DO of Lake N increased significantly at the beginning of water replenishment, and the highest value at the water replenishment port was 11.37 mg / L. Due to good fluidity, the DO content was high. COD decreased significantly in the early stage, with the lowest value at the water replenishment port being 57.70 mg / L and about 74.00 mg / L in the lake. Good fluidity caused COD to decrease with water replenishment, and COD was low at high flow rates. Total nitrogen (TN) increased in the early stage, and the highest value at the water replenishment port was 1.56 mg / L. Total phosphorus (TP) decreased in the early stage, with the lowest value at the water replenishment port being 0.167 mg / L and 0.36 mg / L in the middle. The distribution was related to the flow rate, with low values ​​at high flow rates and high values ​​at low flow rates. Chlorophyll a increased significantly when TN was first replenished, reaching 18.65 ug / L, and TN was highest at the water replenishment port. Monitoring these parameters helps predict chlorophyll a distribution and guide water management and algae control.

[0116] (2) Spatial distribution characteristics of lake water environment parameters when static

[0117] During the static process, the water environment parameters of each lake showed different spatial distribution characteristics, such as Fig.17 shown.

[0118] Although the dissolved oxygen (DO) concentration of Lake B is slowly decreasing, it is still able to maintain an appropriate DO level due to its regular morphology and good fluidity. At the same time, the chemical oxygen demand (COD) concentration of Lake B is also showing a slow upward trend, but thanks to its regular morphology and good fluidity, the organic pollution in the water body is kept at a low level. In addition, the total nitrogen (TN) and total phosphorus (TP) concentrations of Lake B are rising relatively slowly, which is also due to its good fluidity, keeping the nitrogen and phosphorus pollution levels at a low level. The distribution of chlorophyll a is also relatively uniform and stable, with a concentration of 11.63ug / L, which indicates that the water health of Lake B is relatively good.

[0119] In contrast, the DO concentration of Lake X decreased significantly during the static period, which was mainly due to its complex shoreline and slow water flow, resulting in a weakened self-purification capacity. In terms of COD, due to the poor mobility of Lake X, the COD concentration accumulated rapidly, and organic pollutants easily gathered and decomposed here, resulting in a high COD concentration. The TN and TP concentrations of Lake X were also easy to accumulate due to poor mobility, with high concentrations, and the conversion rate of nitrogen and phosphorus was slow. The distribution of chlorophyll a may be locally high due to poor flow and complex shoreline, with a maximum concentration of 15.32ug / L, which may have a certain impact on the water ecology.

[0120] During the static period, Lake N showed strong self-purification ability, DO decreased slowly, and was able to maintain a high DO level. In terms of COD, the COD concentration of Lake N only rose slowly, the decomposition of organic matter was relatively thorough, and the pollution level was low. The TN concentration of Lake N rose slowly, the self-purification ability was strong, and the nitrogen conversion process was more thorough, so the pollution level was low. Although the TP concentration increased, due to the strong self-purification ability, the deposition and release of phosphorus pollutants tended to balance, and the concentration was relatively low. Chlorophyll a remained at a low level in Lake N, at 11.13ug / L, which shows that the water body of Lake N is in good health and the self-purification mechanism is effective.

[0121] 5. Analysis and ranking of factors affecting water quality:

[0122] Algal bloom is a sudden outbreak of blue algae, green algae, and diatom microorganisms. Its growth is related to flow rate, temperature, COD, DO, TN, TP and other contents. The main factor affecting the chlorophyll a value is the accumulation of chlorophyll a. The entropy weight method is used to sort the factors that affect the chlorophyll a value.

[0123] 5.1. Steps of entropy weight method:

[0124] (1) Data standardization

[0125] First, each indicator is de-dimensionalized. Assume that m indicators are given: X1, X2, ..., X m ;

[0126] Where: X i ={x1,x2,···,x n}

[0127] Assume that the standardized values ​​of each indicator data are: Y1, Y2, ..., Y m ;

[0128] but:

[0129]

[0130] (2) Find the ratio of each indicator under each scheme

[0131] Assume that a certain primary indicator has m secondary indicators and has obtained n years of data, recorded as matrix Y ij Under the same indicator, calculate the proportion of each year's value to the total value, the formula is as follows:

[0132]

[0133] (3) Calculate the information entropy of each indicator

[0134] According to the definition of information entropy in information theory, the information entropy of a set of data is:

[0135] (If p ij =0, then E j =0)

[0136] (4) Determine the weight of each indicator

[0137] According to the calculation formula of information entropy, the information entropy of each indicator is calculated as E1, E2, …, Em.

[0138] 1. Calculate the weight of each indicator through information entropy

[0139] Here k refers to the number of indicators, that is, k=m.

[0140] 2. Calculate the weight D by calculating the information redundancy j =1-E j

[0141] Then calculate the indicator weights:

[0142] 3. Finally, calculate the comprehensive score of each solution

[0143] The weight of each indicator every day is used to compare the impact of each indicator on chlorophyll a. For example, the weight of the first day is: COD> ammonia nitrogen> dissolved oxygen> total nitrogen> total phosphorus> water temperature> water level. Therefore, COD on the first day has the largest impact on chlorophyll, and water level has the smallest impact. The same applies to the subsequent days.

[0144] In the environmental impact assessment, accurately determining the weight of each factor is crucial to the entire analysis process. In addition to the entropy weight method mentioned earlier, there are a variety of methods that can be used to calculate weights, each with its own specific application scenarios and advantages. For example, principal component analysis (PCA) helps identify the main sources of variation in the data by converting the original variables into fewer principal components. Factor analysis (FA) further reveals the potential connections between variables and simplifies the data structure by extracting common factors. The analytic hierarchy process (AHP) focuses on integrating expert judgments through pairwise comparisons and hierarchical structures to determine the relative importance of each factor.

[0145] When data is incomplete or uncertain, grey relational analysis (GRA) provides an effective solution by evaluating the degree of association between each factor and the reference value to determine the weight. The comprehensive evaluation rule is a more comprehensive technique that combines the advantages of multiple methods to adapt to different evaluation needs. Artificial neural networks (ANN) use their powerful learning ability to automatically identify and adjust weights, which is particularly suitable for dealing with complex nonlinear relationships. Finally, the fuzzy comprehensive evaluation method effectively handles the uncertainty and ambiguity in the evaluation through the principles of fuzzy mathematics.

[0146] These methods can be used alone or in combination to adapt to different data characteristics and decision-making needs. In the water environment governance method of the present invention, the most suitable weight determination method can be selected according to the specific circumstances of the water quality monitoring data to improve the accuracy of the assessment and the effectiveness of the governance measures. Through the comprehensive application of this method, the complex dynamics in the water environment can be more comprehensively understood and managed, providing solid decision support for the sustainable development of the water environment.

[0147] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When the use is implemented in whole or in part in the form of a computer program product, the computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL) or wireless (e.g., infrared, wireless, microwave, etc.) mode) to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state hard disk SolidState Disk (SSD)), etc.

[0148] The above is only a preferred embodiment of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A water environment management method based on water environment mathematical model and artificial intelligence, characterized in that: include: S1. Obtain water quality data, including water depth, water level, flow velocity, flow direction, temperature, salinity, dissolved oxygen, pH value, chemical oxygen demand, ammonia nitrogen, total phosphorus and turbidity; S2. Construct water environment mathematical model and artificial intelligence model; specifically include: The water quality data is used to construct a water environment mathematical model; the water environment mathematical model is used to simulate the hydrodynamic process and the temporal and spatial distribution of water quality parameters; First, a basic AI model is selected, then the AI ​​model is trained and learned using water quality data, and finally a trained AI model is obtained; S3. Analyze the water environment using the artificial intelligence model to obtain the spatiotemporal distribution characteristics and dominant factors of water quality parameters; S4. Formulate water environment management plans based on the temporal and spatial distribution characteristics of water quality parameters and dominant factors.

2. The water environment management method based on water environment mathematical model and artificial intelligence according to claim 1 is characterized in that: The water environment mathematical model adopts EFDC as the basic framework and defines key parameters through multiple input files; among them: hydrodynamic parameters include model grid and water depth, boundary conditions, initial conditions and water body dynamic parameters; water quality parameters include the concentration and dynamic process of dissolved oxygen, nutrients and algae plankton; sediment parameters include the sedimentation rate of suspended matter, resuspension process and material exchange between sediment and water body; thermodynamic parameters include water temperature distribution and its heat exchange process with the atmosphere.

3. The water environment management method based on water environment mathematical model and artificial intelligence according to claim 1 is characterized in that: The artificial intelligence model includes a convolutional neural network and a long short-term memory network; wherein: The convolutional neural network input data includes the spatial distribution map of water quality parameters such as water temperature, dissolved oxygen, ammonia nitrogen, and total phosphorus; by performing deep learning on the spatial distribution map of water quality parameters, the spatial pattern and trend of water quality parameters are identified; The data input to the long short-term memory network is the time series data of water temperature, pH value and dissolved oxygen obtained by continuous monitoring. By analyzing the time series data, the future change trend of water quality parameters is predicted.

4. The water environment management method based on water environment mathematical model and artificial intelligence according to claim 3 is characterized in that: The spatial patterns and trends of the water quality parameters include the distribution range of pollution hot spots and the gradient changes in pollutant concentrations; the predicted future change trends of the water quality parameters include predictions of short-term fluctuations in dissolved oxygen levels or long-term increases or decreases in pollutant concentrations.

5. A water environment management platform based on water environment mathematical model and artificial intelligence, characterized by: include: A water quality data acquisition module, including several sensors for acquiring water depth, water level, flow velocity, flow direction, temperature, salinity, dissolved oxygen, pH value, chemical oxygen demand, ammonia nitrogen, total phosphorus and turbidity parameters; Model building modules, including: The water quality data is used to construct a water environment mathematical model; the water environment mathematical model is used to simulate the hydrodynamic process and the temporal and spatial distribution of water quality parameters; First, a basic AI model is selected, then the AI ​​model is trained and learned using water quality data, and finally a trained AI model is obtained; An analysis module, which uses the artificial intelligence model to analyze the water environment to obtain the spatiotemporal distribution characteristics and dominant factors of water quality parameters; The governance module formulates water environment governance plans based on the spatiotemporal distribution characteristics and dominant factors of water quality parameters.

6. The water environment management platform based on water environment mathematical model and artificial intelligence according to claim 5 is characterized in that: The water environment mathematical model adopts EFDC as the basic framework and defines key parameters through multiple input files; among them: hydrodynamic parameters include model grid and water depth, boundary conditions, initial conditions and water body dynamic parameters; water quality parameters include the concentration and dynamic process of dissolved oxygen, nutrients and algae plankton; sediment parameters include the sedimentation rate of suspended matter, resuspension process and material exchange between sediment and water body; thermodynamic parameters include water temperature distribution and its heat exchange process with the atmosphere.

7. The water environment management platform based on water environment mathematical model and artificial intelligence according to claim 5 is characterized in that: The artificial intelligence model includes a convolutional neural network and a long short-term memory network; wherein: The convolutional neural network input data includes the spatial distribution map of water quality parameters such as water temperature, dissolved oxygen, ammonia nitrogen, and total phosphorus; by performing deep learning on the spatial distribution map of water quality parameters, the spatial pattern and trend of water quality parameters are identified; The data input to the long short-term memory network is the time series data of water temperature, pH value and dissolved oxygen obtained by continuous monitoring. By analyzing the time series data, the future change trend of water quality parameters is predicted.

8. The water environment management platform based on water environment mathematical model and artificial intelligence according to claim 7 is characterized in that: The spatial patterns and trends of the water quality parameters include the distribution range of pollution hot spots and the gradient changes in pollutant concentrations; the predicted future change trends of the water quality parameters include predictions of short-term fluctuations in dissolved oxygen levels or long-term increases or decreases in pollutant concentrations.

9. A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the water environment management method based on a water environment mathematical model and artificial intelligence as described in any one of claims 1 to 4.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the water environment management method based on the water environment mathematical model and artificial intelligence described in any one of claims 1 to 4 is implemented.

Images