A water pollution tracing method and system based on big data
By using the dynamic weight adjustment algorithm of big data and Bayesian network in water pollution traceability, the problem of inaccurate traceability results caused by single data source and static weight allocation in traditional methods is solved, and more accurate and flexible pollution source positioning is achieved. Improved traceability efficiency and adaptability.
Patent Information
- Application Number
- CN202411090374.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-09
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-08-09
AI Technical Summary
Traditional water pollution traceability methods rely on a single data source or static weight allocation, resulting in one-sided or inaccurate traceability results, unable to adapt to complex and changeable water pollution situations, and lack dynamic assessment of the timeliness and importance of data sources.
Adopt a dynamic weight adjustment algorithm based on big data, combine real-time and historical data changes, automatically adjust the weights of each data source and influencing factors, build a water pollution traceability model, use Bayesian network and gradient descent algorithm for iterative optimization, and improve traceability Accuracy and adaptability of results.
It significantly improves the accuracy and adaptability of traceability results, enhances traceability efficiency, can quickly and accurately track and locate pollution sources, and provides scientific basis for environmental management and emergency response.
Smart Images

Figure FHA0000007298860000031 
Figure FHA0000007298860000033
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data processing technology, and in particular to a water pollution source tracing method and system based on big data. Background Art
[0002] With the continuous advancement of environmental monitoring technology and the widespread application of big data technology, water pollution tracing, as an important research direction in the field of environmental protection, has received increasing attention.
[0003] Traditional water pollution source tracing methods often rely on a single data source or simple data fusion technology. On the one hand, a single data source may lead to one-sided or inaccurate tracing results due to the uneven distribution of monitoring sites and limited monitoring indicators. On the other hand, traditional water pollution source tracing methods often use a fixed weight distribution method when processing multi-source data, ignoring the dynamic change characteristics of different data sources at different times and situations. This static weight distribution method is difficult to adapt to complex and changeable water pollution situations, and easily leads to inaccurate or biased tracing results. In addition, traditional methods lack a dynamic evaluation mechanism for the timeliness and importance of data sources, and cannot automatically adjust weights according to real-time updates and historical change trends of data, thereby limiting the adaptability and flexibility of tracing methods.
[0004] For example, the invention patent application with patent application number 202110987964X discloses a full-process coupled simulation method for water-related links in a smart city; it mainly focuses on the acquisition of meteorological data and the simulation of hydrological and surface source data, as well as the generation of sewage treatment and water quality and water ecological data; it does not involve specific ways to trace the source of pollution.
[0005] For example, the invention patent application with patent application number 202211676491.2 discloses an air-space-ground integrated lake water quality tracing method, system, equipment and storage medium; it focuses on using remote sensing technology for water quality monitoring and tracing.
[0006] For example, the invention patent application with patent application number 202211137648.4 discloses a method and system for refined control of the water ecological environment in a small watershed; it focuses on the control of small watersheds, including investigations into the current status of the water ecological environment, division of watershed units, establishment of a pollution emission inventory, etc., rather than the traceability technology itself.
[0007] For example, the invention patent application with patent application number 202211513556.1 discloses a method for intelligent management of water environment in a river basin, which focuses on analyzing the relationship between pipe networks, the spatial vector relationship between sewage outlets and pollution sources, and the corresponding relationship between pollutant components. The weight calculation is based on the water quality early warning alarm characteristic factors and the characteristics of regional water environment pollution, coupled with the pollution source relationship weight calculation rules, to construct a water quality monitoring relationship matrix between sewage outlets and pollution sources. The weight calculation rules are used to evaluate the impact of pollution sources on specific outlets, and they rely more on manually set rules.
[0008] Water pollution source tracing aims to identify and locate the source of pollution by collecting and analyzing various types of data in water bodies, so as to provide a scientific basis for pollution control and environmental management. However, how to effectively integrate these multi-source data and improve the accuracy of tracing by dynamically adjusting data weights is still a problem that needs to be solved urgently. Summary of the invention
[0009] The purpose of the present invention is to make up for the shortcomings of the prior art and to provide a water pollution source tracing method and system based on big data. It can analyze multi-source data in real time, combine the changing trends of historical data, and use a dynamic weight adjustment algorithm to automatically adjust the weights of different data sources and influencing factors, so as to more accurately track and locate the pollution source. It not only improves the accuracy and adaptability of the tracing results, but also significantly improves the tracing efficiency through automated and intelligent processing methods.
[0010] In order to solve the above technical problems, the present invention provides the following technical solutions:
[0011] On the one hand, a method for tracing the source of water pollution based on big data is provided, which comprises the following specific steps:
[0012] Step S100, data collection and preprocessing: water quality, meteorological and hydrological data are collected in real time through water quality monitoring stations and meteorological station equipment deployed in rivers and lakes, and the collected data are preprocessed;
[0013] Step S200, dynamic weight adjustment algorithm design: design a dynamic weight adjustment algorithm based on the changes of real-time data and historical data, and automatically calculate and adjust the weights of various data sources and influencing factors according to the update of real-time data and the change trend of historical data;
[0014] Step S300, constructing a water pollution source tracing model: Based on the adjusted weights, a water pollution source tracing model is constructed to track and locate the pollution source by simulating the diffusion path and speed parameters of pollutants and combining geographic information and meteorological conditions;
[0015] Step S400, verification and optimization of traceability results: comparing and verifying the traceability results with field investigations and expert opinions, and optimizing and adjusting the algorithms and models based on the verification results;
[0016] Step S500, display and application of tracing results: display the tracing results to users in the form of charts, including pollution diffusion maps and pollution source distribution maps, and apply them to the fields of environmental management, emergency response and pollution control.
[0017] Furthermore, step S100 uses a data standardization algorithm to pre-process the collected data, and the calculation formula is: ,in: are the values in the original dataset, is the mean of the original data set, is the standard deviation of the original data set, is the normalized data value.
[0018] Further, step S200 includes:
[0019] Step S201, preliminarily setting the weights of various data sources and influencing factors based on historical data and expert knowledge;
[0020] Step S202, acquiring the latest monitoring data in real time, analyzing the changing trend of data from each data source, and identifying anomalies or mutation points in the data;
[0021] Step S203, calculate the weight adjustment factors of each data source and influencing factor according to the data change amount, correlation and importance factors, calculate the new weight value according to the weight adjustment factor and the current weight, and continuously adjust the weight through multiple iterations.
[0022] Further, step S202 uses the ARIMA model algorithm to analyze the change trend of the data from each data source, and its calculation formula is: ,in, The data collected at the time The value of and are the indices of the autoregressive term and the moving average term, respectively. and are the coefficients of the autoregressive term and the moving average term, respectively. is the lag operator, is the order of the autoregressive term, is the order of the moving average term, is the error term, is a constant term, is the coefficient of the trend term, is the difference order, express of Order difference.
[0023] Further, the specific steps of adjusting the weight in step S203 are:
[0024] (1) Preliminarily set the weights of various data sources and influencing factors based on historical data and expert knowledge;
[0025] (2) Analyze the data changes of each data source or influencing factor, evaluate the correlation between each data source or influencing factor, determine the importance of each data source or influencing factor based on business needs or expert knowledge, and calculate a weight adjustment factor for each data source or influencing factor;
[0026] (3) Calculate the new weight value based on the calculated weight adjustment factor and the current weight;
[0027] (4) Setting the termination conditions of the iterative optimization process, including reaching the preset number of iterations, the weight change being less than the preset threshold, and the objective function value converging. In each iteration, the weight adjustment factor is recalculated and the weight value is updated;
[0028] (5) When the iterative optimization process ends, the final weight vector is output.
[0029] Further, in step S203, a gradient descent algorithm is used for iterative optimization to obtain the weight vector to be optimized. Assigning initial values , according to the requirements, define the objective function , which is used to evaluate the weight vector The quality of , in each iteration, calculate the objective function At the current weight vector The gradient at , according to the formula of gradient descent method , update the weight vector , and obtain the new weight vector ,in, is the learning rate, repeat the iteration until the termination condition is met, and output the final weight vector , as the optimization result of the dynamic weight adjustment algorithm.
[0030] Further, step S300 includes:
[0031] Step S301, integrating multi-source data of water quality, meteorology and hydrology to form a unified data set, and preprocessing the integrated data;
[0032] Step S302, select a mathematical model according to the characteristics of the traceability problem, set the parameters of the model according to the data characteristics and algorithm requirements, train the model using historical data, and adjust the model parameters to optimize the traceability performance;
[0033] Step S303, input the real-time data into the trained model, use the model to track and locate the pollution source, and output the tracing results, including the location of the pollution source and the diffusion path information.
[0034] Further, in step S303, a Bayesian network algorithm is used; based on water quality, meteorological, hydrological data and the dependencies between them, a Bayesian network including the location of the pollution source and the monitoring data nodes is constructed, a priori probability is set for each node in the Bayesian network, water quality monitoring data is collected in real time, and it is used as input data to calculate the posterior probability of the location of the pollution source. , and its calculation formula is: ,in Indicates the location of the pollution source, It represents water quality monitoring data. According to the calculated posterior probability, it determines the location of the pollution source, performs source tracing, and outputs the source tracing results, including the location of the pollution source and the diffusion path information.
[0035] Furthermore, in step S400, the specific steps of optimizing and adjusting the model are:
[0036] (1) Determine the parameters that affect model performance based on the model type and specific application scenario, and set the value range for each key parameter;
[0037] (2) Use the grid search method to traverse all parameter combinations. For each parameter combination, retrain the model using the training set and evaluate its performance using the validation set, and record the performance indicators corresponding to each parameter combination;
[0038] (3) According to the performance indicators, select the optimal parameter combination, use the optimal parameter combination to retrain the model, and obtain the final optimized model.
[0039] On the other hand, a water pollution tracing system based on big data is provided, which includes the following components:
[0040] Data collection and preprocessing module: Through the water quality monitoring stations and meteorological station equipment deployed in rivers and lakes, multi-source data such as water quality, meteorology, and hydrology are collected in real time, and the collected data are preprocessed;
[0041] Dynamic weight adjustment algorithm module: Design and implement a dynamic weight adjustment algorithm based on the changes of real-time data and historical data. According to the update of real-time data and the change trend of historical data, the weight of each data source and influencing factor is automatically calculated and adjusted;
[0042] Water pollution source tracing model construction module: Based on the adjusted weights, a water pollution source tracing model is constructed. The water quality, meteorological and hydrological data are integrated by using multi-source data fusion technology for comprehensive analysis. The pollution source is tracked and located by simulating the diffusion path and velocity parameters of pollutants and combining geographic information and meteorological conditions.
[0043] Traceability result verification and optimization module: compare and verify the traceability results with field investigations and expert opinions, and optimize and adjust the algorithms and models based on the verification results;
[0044] Source tracing result display and application module: The source tracing results are displayed to users in a visual form, including pollution diffusion maps, pollution source distribution maps, etc., providing an intuitive display of source tracing results, and applying the results to environmental management, emergency response and pollution control fields.
[0045] Compared with the prior art, the present invention has the following beneficial effects:
[0046] The present invention proposes a dynamic weight adjustment algorithm based on the changes of real-time data and historical data, which helps to more accurately reflect the current water quality status and environmental changes. Multi-source data fusion technology is used to integrate water quality, meteorological and hydrological data for comprehensive analysis to improve the accuracy of tracing. In the construction of the tracing model, the Bayesian network algorithm is used to handle complex dependencies. The present invention emphasizes the display of tracing results in the form of charts, which improves the readability and practicality of the results.
[0047] The present invention designs a dynamic weight adjustment algorithm that can automatically adjust the weights of various data sources and influencing factors in real time according to the changing trends and importance of the data, effectively utilizing the advantages of real-time data and historical data, so that the tracing model can more accurately reflect the current water quality conditions and environmental changes, thereby improving the accuracy and timeliness of the tracing results.
[0048] The present invention greatly enhances the adaptability and flexibility of the present invention by designing a dynamic weight adjustment algorithm. When faced with sudden water pollution incidents or long-standing pollution sources, the present invention can quickly and accurately track and locate pollution sources by dynamically adjusting weights, thereby providing strong support for environmental management, emergency response and pollution control.
[0049] Compared with the traditional method that relies on a single data source, the present invention overcomes the problem of one-sided data by fusing multi-source data. The present invention solves the problem that the fixed weight allocation in the traditional method cannot adapt to the dynamic changes of data by adjusting the dynamic weight. The adaptability and flexibility of the traceability method are improved by integrating real-time data and using historical data; the traceability efficiency is improved and the interference of human factors is reduced by automated data processing and intelligent model construction.
[0050] In summary, the present invention provides a novel method in the field of water pollution source tracing, which improves the accuracy and efficiency of tracing by comprehensively utilizing multi-source data, dynamic weight adjustment, advanced statistical methods and automation technology, and overcomes the limitations of the existing technology. DETAILED DESCRIPTION
[0051] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. 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.
[0052] Embodiment 1
[0053] A water pollution source tracing method and system based on big data is applied to urban river pollution source tracing, which can quickly and accurately locate the pollution source and formulate effective governance measures to cope with the current urgent environmental challenges.
[0054] Step S100, data collection and preprocessing: high-precision water quality monitoring stations, meteorological stations and hydrological monitoring equipment are deployed along the river and at key nodes, including river estuaries, downstream of industrial areas, and domestic sewage discharge outlets. These devices can collect water quality parameters including pH value, dissolved oxygen, heavy metal content, meteorological parameters such as wind speed, wind direction, and temperature, and hydrological parameters such as flow rate and water level in real time. Through wireless transmission technology, the data of each monitoring station is transmitted to the data center in real time to ensure the timeliness and integrity of the data. The collected raw data is preprocessed using a data standardization algorithm. The specific steps include calculating the mean and standard deviation of the raw data set, and then converting the raw data into standardized data values according to the formula to eliminate the influence of different dimensions and magnitudes on data analysis. The calculation formula is: ,in: are the values in the original dataset, is the mean of the original data set, is the standard deviation of the original data set, is the normalized data value.
[0055] Step S200, dynamic weight adjustment algorithm design (step S201-step S203): Based on historical data and expert knowledge, preliminarily set the weights of each data source and influencing factor. These weights reflect the importance and contribution of each factor in the traceability process. Obtain the latest monitoring data in real time, and use the dynamic weight adjustment algorithm to analyze the changing trend of the data. By identifying anomalies or mutation points in the data, determine the possible location and diffusion direction of the pollution source, and calculate the weight adjustment factor of each data source and influencing factor based on factors such as data change, correlation, and importance. Then, use the gradient descent algorithm to perform multiple iterative optimizations. In each iteration, calculate a new weight value based on the current weight and weight adjustment factor until the weight change is less than the preset threshold or reaches the preset number of iterations, and finally output the optimized weight vector.
[0056] Step S300, water pollution source tracing model construction (step S301-step S303): Integrate the pre-processed multi-source data to form a unified data set, ensure the accuracy and completeness of the data through data cleaning and integration, and select the Bayesian network algorithm as the source tracing model. Specifically, based on water quality, meteorological, hydrological data and the dependencies between them, construct a Bayesian network containing pollution source locations and monitoring data nodes, set a priori probability for each node in the Bayesian network, collect water quality monitoring data in real time, and use it as input data to calculate the posterior probability of the pollution source location. , and its calculation formula is: ,in Indicates the location of the pollution source, It represents water quality monitoring data. According to the calculated posterior probability, the location of the pollution source is determined, and source tracing is performed. The source tracing results are output, including the location of the pollution source and diffusion path information. The real-time data is input into the trained model, and the model is used to simulate the diffusion path and speed of pollutants. The pollution source is tracked and located in combination with geographic information and meteorological conditions. The model outputs the tracing results, including the location of the pollution source, diffusion path, and impact range information.
[0057] Step S400, verification and optimization of traceability results: the traceability results are compared and verified with the field investigation results and expert opinions. The accuracy and reliability of the traceability results are confirmed through on-site investigation, sample collection and analysis, etc. According to the verification results, the model parameters are adjusted using the grid search method. By traversing different parameter combinations, the optimal parameter configuration is found to further improve the traceability performance. The optimized model can more accurately reflect the diffusion law and traceability process of pollutants.
[0058] Step S500, display and application of tracing results: display the tracing results to environmental protection departments, scientific research institutions and the general public in an intuitive and easy-to-understand form, and clearly display the diffusion path, impact range and specific location of the pollution source through pollution diffusion maps, pollution source distribution maps and other charts. Apply the tracing results to the fields of environmental management, emergency response and pollution control, provide a scientific basis for decision makers, guide the formulation and implementation of pollution control work, and at the same time, through media publicity and education activities, enhance the public's awareness and participation in environmental protection, and jointly maintain the city's ecological environment.
[0059] Embodiment 2
[0060] Based on Example 1, this example describes in detail the design of a dynamic weight adjustment algorithm for a water pollution source tracing method based on big data, and improves the overall performance of the system and the accuracy of the tracing results by optimizing the design of the dynamic weight adjustment algorithm.
[0061] Collect relevant data on historical water pollution incidents, including water quality parameters (dissolved oxygen, pH value, heavy metal concentration), meteorological conditions (rainfall, wind speed, wind direction), hydrological data (river flow, water level changes) and expert experience and knowledge. Based on the above data and combined with expert experience, conduct a preliminary assessment of each data source and influencing factor, and set their initial weights. The weight setting should reflect the importance of each factor to the tracing of water pollution sources, and ensure that the sum of all initial weights is 1 to facilitate subsequent calculations and adjustments.
[0062] Through the sensor network and online monitoring system, the latest water quality, meteorological, hydrological and other monitoring data are obtained in real time. The acquired data is pre-processed by cleaning, denoising and interpolation to ensure the accuracy and integrity of the data. The ARIMA model algorithm is used to analyze the change trend of the data from each data source. The calculation formula is: ,in, The data collected at the time The value of and are the indices of the autoregressive term and the moving average term, respectively. and are the coefficients of the autoregressive term and the moving average term, respectively. is the lag operator, is the order of the autoregressive term, is the order of the moving average term, is the error term, is a constant term, is the coefficient of the trend term, is the difference order, express of Order difference, identifying anomalies or mutation points in the data, which may indicate potential water contamination events or data source failures.
[0063] Statistical indicators such as correlation coefficient and covariance are used to evaluate the correlation between various data sources and influencing factors. High correlation means that certain factors may jointly affect the results of water pollution source tracing. The importance of each factor in water pollution source tracing is evaluated based on business needs, expert knowledge and historical data. The importance weight reflects the relative contribution of different factors in the tracing process. The weight adjustment factor is calculated for each data source and influencing factor based on the data change, correlation analysis results and importance assessment. The adjustment factor can be a positive or negative number, indicating the extent to which the weight should be increased or decreased.
[0064] Assign initial values to the weight vector to be optimized. These initial values can be preliminary weights based on historical data or expert knowledge. According to the requirements of the traceability task, define the objective function to evaluate the quality of the weight vector. The objective function can be the accuracy of the traceability result, the prediction error of the model, or other related indicators. In each iteration, calculate the gradient of the objective function at the current weight vector. The gradient indicates the direction of the weight vector adjustment, that is, how to change the weight to minimize the objective function. Update the weight vector according to the formula of the gradient descent method. Specifically, the weight vector to be optimized is Assigning initial values , according to the requirements, define the objective function , which is used to evaluate the weight vector The quality of , in each iteration, calculate the objective function At the current weight vector The gradient at , according to the formula of gradient descent method , update the weight vector , and obtain the new weight vector , usually, the weight vector will move a certain step length in the opposite direction of the gradient, and set the termination conditions of the iterative optimization process, such as reaching the preset number of iterations, the weight change is less than the preset threshold, or the objective function value converges. When any of the termination conditions is met, the iteration is stopped. Among them, is the learning rate, repeat the iteration until the termination condition is met, and output the final weight vector , as the optimization result of the dynamic weight adjustment algorithm.
[0065] The optimized weight vector is applied to the water pollution source tracing model to evaluate its impact on the tracing performance. The evaluation indicators include the accuracy of the tracing results, calculation time, and model stability. The differences in the tracing results before and after optimization are analyzed. It is identified which data sources and influencing factors have a significant impact on the tracing results. The optimized weight vector is formally applied to the water pollution source tracing model to improve the adaptability and accuracy of the model. The performance of the model is continuously monitored, new monitoring data is collected, and the algorithm design is further optimized as needed, which includes regularly re-evaluating the weights and adjusting the optimization strategy.
[0066] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.
Claims
1. A water pollution tracing method based on big data, It is characterized in that The method comprises the following specific steps: Step S100, data collection and preprocessing: water quality, meteorological and hydrological data are collected in real time through water quality monitoring stations and meteorological station equipment deployed in rivers and lakes, and the collected data are preprocessed; Step S200, dynamic weight adjustment algorithm design: design a dynamic weight adjustment algorithm based on the changes of real-time data and historical data, and automatically calculate and adjust the weights of various data sources and influencing factors according to the update of real-time data and the change trend of historical data; Step S200 includes: Step S201, preliminarily setting the weights of various data sources and influencing factors based on historical data and expert knowledge; Step S202, acquiring the latest monitoring data in real time, analyzing the changing trend of data from each data source, and identifying anomalies or mutation points in the data; Step S202 uses the ARIMA model algorithm to analyze the change trend of the data from each data source, and its calculation formula is: Among them, Y t is the value of the collected data at time t, i and j are the indices of the autoregressive term and the moving average term, respectively, i and θ j are the coefficients of the autoregressive term and the moving average term, L is the lag operator, p is the order of the autoregressive term, q is the order of the moving average term, and ε t is the error term, c is the constant term, δ is the coefficient of the trend term, d is the difference order, Δ k Y t-k Represents Y t The k-order difference of Step S203, calculating the weight adjustment factor of each data source and influencing factor according to the data change amount, relevance and importance factors, calculating the new weight value according to the weight adjustment factor and the current weight, and continuously adjusting the weight through multiple iterations; The specific steps of adjusting the weight in step S203 are: (1) Preliminarily set the weights of various data sources and influencing factors based on historical data and expert knowledge; (2) Analyze the data change of each data source or influencing factor, evaluate the correlation between each data source or influencing factor, determine the importance of each data source or influencing factor based on business needs or expert knowledge, and calculate a weight adjustment factor for each data source or influencing factor; (3) Calculate a new weight value based on the calculated weight adjustment factor and the current weight; (4) Setting the termination conditions of the iterative optimization process, including reaching a preset number of iterations, the weight change being less than a preset threshold, and the objective function value converging. In each iteration, the weight adjustment factor is recalculated and the weight value is updated; (5) When the iterative optimization process is completed, the final weight vector is output; In step S203, the gradient descent algorithm is used for iterative optimization to assign an initial value w to the weight vector w to be optimized. 0 , according to the requirements, define the objective function f(w), which is used to evaluate the quality of the weight vector w. In each iteration, calculate the objective function f(w) at the current weight vector (w t ) According to the formula of gradient descent method Update the weight vector w to get the new weight vector w t+1 , where η is the learning rate, and the iteration is repeated until the termination condition is met, and the final weight vector w is output as the optimization result of the dynamic weight adjustment algorithm; Step S300, constructing a water pollution source tracing model: constructing a water pollution source tracing model based on the adjusted weights, simulating the diffusion path and speed parameters of pollutants, and combining geographic information and meteorological conditions to track and locate the pollution source; Step S400, verification and optimization of traceability results: comparing and verifying the traceability results with field investigations and expert opinions, and optimizing and adjusting the algorithms and models based on the verification results; Step S500, display and application of tracing results: display the tracing results to users in the form of charts, including pollution diffusion maps and pollution source distribution maps, and apply them to the fields of environmental management, emergency response and pollution control.
2. According to the method for tracing the source of water pollution based on big data in claim 1, It is characterized in that Step S100 uses a data standardization algorithm to pre-process the collected data, and its calculation formula is: Where: X is the value in the original data set, μ is the mean of the original data set, σ is the standard deviation of the original data set, X norm is the normalized data value.
3. According to the method for tracing the source of water pollution based on big data in claim 1, It is characterized in that Step S300 includes: Step S301, integrating multi-source data of water quality, meteorology and hydrology to form a unified data set, and preprocessing the integrated data; Step S302, select a mathematical model according to the characteristics of the traceability problem, set the parameters of the model according to the data characteristics and algorithm requirements, train the model using historical data, and adjust the model parameters to optimize the traceability performance; Step S303, input the real-time data into the trained model, use the model to track and locate the pollution source, and output the tracing results, including the location of the pollution source and the diffusion path information.
4. According to the method of tracing the source of water pollution based on big data in claim 3, It is characterized in that In step S303, a Bayesian network algorithm is used; based on water quality, meteorological, hydrological data and their dependencies, a Bayesian network including pollution source locations and monitoring data nodes is constructed, a priori probability is set for each node in the Bayesian network, water quality monitoring data is collected in real time, and used as input data to calculate the posterior probability P(S|D 1 ,D 2 ,…,D n ), the calculation formula is: Where S represents the location of the pollution source, D 1 ,D 2 ,…,D n It represents water quality monitoring data. According to the calculated posterior probability, it determines the location of the pollution source, performs source tracing, and outputs the source tracing results, including the location of the pollution source and the diffusion path information.
5. According to the method for tracing the source of water pollution based on big data in claim 1, It is characterized in that In step S400, the specific steps of optimizing and adjusting the model are: (1) According to the model type and specific application scenario, determine the parameters that affect the model performance and set the value range for each key parameter; (2) Use the grid search method to traverse all parameter combinations. For each parameter combination, retrain the model using the training set and evaluate its performance using the validation set, and record the performance indicators corresponding to each parameter combination; (3) According to the performance indicators, the optimal parameter combination is selected, and the model is retrained using the optimal parameter combination to obtain the final optimized model.
6. According to the water pollution source tracing method of claim 1, a water pollution source tracing system based on big data is adopted. It is characterized in that The system consists of the following components: Data collection and preprocessing module: Through the water quality monitoring stations and meteorological station equipment deployed in rivers and lakes, water quality, meteorological, and hydrological multi-source data are collected in real time, and the collected data are preprocessed; Dynamic weight adjustment algorithm module: Design and implement a dynamic weight adjustment algorithm based on the changes of real-time data and historical data. According to the update of real-time data and the change trend of historical data, the weight of each data source and influencing factor is automatically calculated and adjusted; Water pollution source tracing model construction module: Based on the adjusted weights, a water pollution source tracing model is constructed. The water quality, meteorological and hydrological data are integrated by using multi-source data fusion technology for comprehensive analysis. The pollution source is tracked and located by simulating the diffusion path and velocity parameters of pollutants and combining geographic information and meteorological conditions. Traceability result verification and optimization module: compare and verify the traceability results with field investigations and expert opinions, and optimize and adjust the algorithms and models based on the verification results; Source tracing result display and application module: The source tracing results are displayed to users in a visual form, including pollution diffusion maps and pollution source distribution maps, providing an intuitive display of source tracing results, and applying the results to environmental management, emergency response and pollution control fields.
