River network water pollution tracing method

Abstract

The invention discloses a river network water pollution tracing method which comprises the following steps: firstly, acquiring river network data, inputting the river network data into MIKE software, constructing a hydrodynamic model of a river network, and verifying the hydrodynamic model; dividing the whole river network into a plurality of river reaches by taking water quality monitoring points as nodes, and inputting river network data corresponding to each river reach into the verified hydrodynamic model to obtain simulation data of each river reach; preprocessing the simulation data of each river reach, inputting the preprocessed simulation data into a gating circulation unit, and constructing a regression model of each river reach; and tracing by using the regression model. According to the method, MIKE software is adopted to simulate the pollutant diffusion condition, so that data supplementation is carried out under the condition that monitoring data is sparse, pollutant tracing is carried out in combination with a regression model, hydrodynamic simulation data is used as input of the regression model, the precision and density of a data set can be improved, and the accuracy and the density of the data set are improved. And accurate positioning of the pollution source and the pollution source parameters is realized.

Description

Technical Field

[0001] The invention belongs to the field of river network water pollution source tracing, and in particular relates to a river network water pollution source tracing method. Background Art

[0002] At present, sewage is still a major challenge that needs to be solved among environmental pollution problems. If environmental pollution exceeds the environmental capacity, it will cause a devastating blow to the local ecosystem, and with the circulation of materials and energy, it will pose a huge threat to people's lives. Therefore, it is necessary to accurately locate the source of pollutants in the river network. If the source of pollutants in the river network is simply traced by manual investigation, it will not only be ineffective, but also the reliability and timeliness of the traceability results cannot be guaranteed. With the development of artificial intelligence technology, it is of great significance and value to trace the source of pollutants in the river network through big data analysis and artificial intelligence algorithms.

[0003] The invention patent with patent number CN202311274115.5 mentions a method for constructing a fluorescent fingerprint database of water pollution sources in industrial parks. The pollutant fluorescent fingerprint library can be constructed based on the different fluorescent characteristic peaks presented by different pollutants, which can help accurately identify pollutants and trace the corresponding factory emission sources.

[0004] However, although this invention has high traceability accuracy, it is costly to build a fingerprint library containing enough pollutants. Moreover, with the continuous development of industrialization and the continuous changes in the raw materials used by factories in industrial parks, the fingerprint library needs to be constantly updated, which requires a lot of manpower, material and financial resources. Therefore, this method can only play a good role when the pollutants of the pollutant-discharging enterprises in the traceability area are extremely different. If a certain industrial park concentrates multiple enterprises of the same type, this method is difficult to accurately trace the source.

[0005] To this end, the present invention proposes a river network water pollution tracing method based on MIKE software and gated circulation unit. Summary of the invention

[0006] The purpose of the present invention is to overcome the defects of the prior art and propose a method for tracing the source of water pollution in a river network. The MIKE software is combined with a gated loop unit and a regression model to reversely trace the source of the water pollution that has occurred. Specifically, the MIKE software is used to simulate the concentration data of the monitoring point under any circumstances, and then the gated loop unit is used to train the model to find the concentration corresponding regression relationship between the pollution source and the monitoring point in any river section.

[0007] To achieve the above object, the present invention provides the following technical solutions: A method for tracing the source of water pollution in a river network comprises the following steps: Step S1, obtaining river network data; the river network data includes river network geographic data, river network hydrological data, river channel patrol data, river network monitoring point monitoring data, etc.; Step S2, constructing a hydrodynamic model: inputting the river network data into the MIKE software to construct a hydrodynamic model of the entire river network; Step S3, simulating pollutant diffusion and verifying the model: taking any monitoring point as a pollution source, setting the pollution discharge duration, pollution discharge concentration and concentration change interval of the pollution source and inputting them into the hydrodynamic model, obtaining the simulated data of the water quality monitoring points around the pollution source, and verifying the hydrodynamic model with the real monitoring data corresponding to the monitoring point; Step S4, constructing a regression model: dividing the entire river network into multiple river sections with monitoring points as nodes, and then inputting the river network data corresponding to each river section into the hydrodynamic model to obtain simulation data of each river section; inputting the simulation data of each river section into the gated recurrent unit after preprocessing to construct a regression model for each river section; Step S5, tracing the source of water pollution in the river network: when any monitoring point detects that the pollutants exceed the standard, the crossover and mutation operations in the genetic algorithm are used to arbitrarily change and reorganize the pollution source parameters, and input them into the regression model corresponding to the river section where the monitoring point is located to obtain the predicted result of the pollutant concentration exceeding the standard at the current monitoring point; the difference between the predicted result and the actual exceeding value is used as the evaluation index, and the above steps are repeated. After multiple iterations, the optimal prediction result of the pollutant concentration exceeding the standard at the current monitoring point is returned, and the pollution source parameter corresponding to the optimal prediction result is the tracing result.

[0008] The pollution source parameters include the location of the pollution source, the start time of pollution source emission, the duration of emission, and the emission concentration.

[0009] Preferably: the river network data include river network geographic data, river network hydrological data, river patrol data, and river network water quality monitoring point monitoring data. Specifically, the river network geographic data include river channel longitude and latitude data, and artificially measured river channel cross-sectional structure data; the river network hydrological data include flow velocity, flow direction, water level and other data in different river sections; the river patrol data include background value data of four major pollutants in the river network collected by unmanned boats and drones; and the river network water quality monitoring point monitoring data include the detection values ​​of the major pollutants at the river network monitoring point at regular intervals.

[0010] Preferably: using MIKE ZERO software, inputting two-dimensional geographic data of the river network (latitude and longitude, river direction, width, etc.), hydrological data, cross-sectional structure, etc., constructing and verifying the hydrodynamic model of the entire river network; inputting specific characteristic parameters of pollution sources, using the hydrodynamic model to simulate the degree of pollution impact, the specific pollutant concentration data under any time and space conditions and any precision can be obtained, and the background value data of river network pollutants measured by unmanned boats and the actual detection data of monitoring points are used as reference indicators for the correctness of the simulation, avoiding the influence of mechanical failures of actual monitoring stations, long monitoring intervals and other factors on the accuracy of the trained model.

[0011] Preferably: the water quality monitoring points in the river network are used as division nodes to divide the entire river network into multiple river sections. Due to changes in factors such as the riverbed cross-section structure and water level, different parts of the river network present different flow rates and flow velocities. The same pollution source will also have very different pollution conditions in different parts of the river network. If the river network is not divided and a regression model is trained with the data of the entire river network, the model will inevitably try to meet the prediction accuracy of all river sections with different hydrological characteristics as much as possible, which may eventually result in only a few river sections with high accuracy or even the prediction accuracy of the entire river network is seriously insufficient.

[0012] Preferably: after dividing the river network into river sections, the gated recurrent unit is used to substitute the simulation data, and an independent regression model for each river channel is constructed, and a recurrent neural network is used to help find the corresponding relationship between the pollution source and the monitoring point value under special conditions in different river sections.

[0013] Preferably: the preprocessing in step S4 includes unifying time units, removing spacers, and identifying and removing outliers.

[0014] Preferably, the identification and removal of outliers adopts two methods: Z-score and IQR (interquartile range).

[0015] Preferably: introducing a genetic algorithm to randomly change the pollution source parameters (location, emission start time, emission duration, emission concentration, etc.) within an appropriate range can traverse the complete pollution source situation, avoid the gated cycle unit falling into the local optimum, and effectively improve the traceability and accuracy of the pollution source.

[0016] Preferably: in addition to being verified by real detection data, the simulation data of MIKE software also needs to be visualized and analyzed. By calculating the difference with the background value under the same time and space conditions and visualizing it, it can be determined whether the changes are consistent with the actual river geography and hydrological characteristics. In addition, it is necessary to analyze whether the gap between the data at different monitoring points is obvious enough and whether it can be used as an input data set for model training without affecting the accuracy of the final model.

[0017] Technical effects and advantages of the present invention: 1. Compared with the prior art, the present invention adopts MIKE software to simulate the diffusion of pollutants, so as to supplement data under the condition of sparse monitoring data. The hydrodynamic model can be used to calculate the pollutant concentration under complex conditions, and the real data returned by the monitoring point sensor is combined as the verification of correctness, which overcomes the problem of insufficient frequency and accuracy of monitoring point collection and detection under real conditions. Using hydrodynamic simulation data as the input of the tracing model can improve the accuracy and density of the data set, and help the model better find the macro and micro correspondence between the pollution source and the monitoring point data.

[0018] 2. Compared with the prior art, the present invention introduces a gated circulation unit to trace the source of pollutants. By finding the corresponding relationship between the concentration and time between the pollution source and the monitoring point, the location of the pollution source is roughly determined by using the concentration and the time of arrival at the monitoring point. This is applicable even when many similar enterprises are gathered in the same industrial park. After the pollutant exceeding the standard alarm information is input, the entire tracing process is automatically completed by the system, which improves work efficiency and the reliability of tracing.

[0019] 3. The present invention introduces a genetic algorithm to randomly mutate and reorganize the pollution source information and then reinsert it into the model, and obtains the best tracing result after continuous iteration, which can effectively prevent the tracing model from falling into a local optimal solution and thus affecting the tracing result.

[0020] 4. The present invention divides the entire river network into river sections with monitoring points as nodes. Due to changes in factors such as the riverbed cross-section structure and water level, different parts of the river network present different flow rates and flow velocities. The same pollution source will also have very different pollution conditions in different parts of the river network. If a regression model is trained with data from the entire river network, the model will inevitably meet the prediction accuracy of all river sections with different hydrological characteristics as much as possible, which may eventually lead to only a few river sections with high accuracy, while the prediction accuracy of the entire river network is low. The present invention divides the river network as much as possible to divide and model the river sections with large differences in hydrological characteristics, effectively improving the accuracy of the regression model, and ultimately locating the pollution source more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The figure is an overall architecture diagram of a river network water pollution source tracing method based on MIKE software and gated circulation unit proposed in the present invention.

[0022] Figure 2 It is the comparison data between the simulation results and the actual results of the hydrodynamic model of the present invention.

[0023] Figure 3It is the visualization data of the simulation results of the hydrodynamic model of the present invention.

[0024] Figure 4 It is a schematic diagram of the training process of each river section source tracing model based on the gated recurrent unit applicable to the present invention.

[0025] Figure 5 It is a schematic diagram of the pollutant source tracing steps based on the gated cycle unit and the genetic algorithm applicable to the present invention.

[0026] Figure 6 This is the traceability result of the method of the present invention. DETAILED DESCRIPTION

[0027] 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.

[0028] Example Combination Figure 1 This embodiment proposes a river network water pollution source tracing method based on MIKE software and gated circulation unit, and the source tracing method includes the following steps: Step 1: Data collection: The data that needs to be collected include the river network data of the target area, including river network geographic data, river network hydrological data, river patrol data, and river network monitoring point monitoring data, specifically the longitude and latitude data of all rivers, riverbed cross-sectional structure data at certain intervals, flow direction, flow, and water level data of all river water bodies, complete river network pollutant background value data detected by unmanned boat patrols, and pollutant monitoring data of monitoring points in the river network (in this embodiment, the number of monitoring points is 20, and the types of pollutants are COD, TN, TP, and NH3-N).

[0029] Step 2: Establish a river network hydrodynamic model: Extract the longitude and latitude data of all rivers collected in step 1, the riverbed cross-sectional structure data at certain intervals, the flow direction, flow rate, and water level data of all river water bodies, input them into the MIKE software, and use the software to establish a hydrodynamic model.

[0030] Step 3: Simulation data generation: Since the detection interval of the monitoring point is too long (4 hours / time) and the detection data is not complete (less than 20% of the monitoring points are tested for all four pollutants and the detection equipment is fault-free), in this embodiment, the pollutant parameters of the specific situation are substituted into the hydrodynamic model to simulate the simulation data, which overcomes the problem of insufficient real data; in this embodiment, any monitoring point is taken as an imaginary pollution source, and the simulation parameters are: the discharge volume is 500 m 3 / d, the discharge time is 8:00~11:00 every day, the pollution source concentration is 0~20000 mg / L, and the simulation concentration data density is 200 mg / L (the simulation parameters are set according to the nature of the pollutant discharge enterprises in the study area, the relationship between water body fluidity and time, and the concentration density required for the source tracing model, and can be changed according to actual conditions). The above data are input into the hydrodynamic model to obtain the water quality pollution simulation data of four monitoring points in different tributaries in the surrounding area, which is used to establish the source tracing model. Figure 2 , using real monitoring data to verify the accuracy of the simulation data set. In actual situations, the simulation data error values ​​of river network water quality monitoring points are all less than 15%; combined with Figure 3 ,Four rivers with large differences in hydrological characteristics were selected, and the correspondence between the simulated concentration and time of the water quality monitoring points was visualized and analyzed. The results show that the simulation data of each river is consistent with the actual geographical and hydrological characteristics of the corresponding river, and there is an obvious gap between the data of different monitoring points, which can be used as the input data set for subsequent model training.

[0031] Step 4: Establishment of source tracing model: Before establishing the model, the river network needs to be divided. This embodiment uses the water quality monitoring points in the river network as the division nodes to divide the entire river network into multiple independent river sections. Due to changes in factors such as the riverbed section structure and water level, different parts of the river network present different flow rates and flow velocities. Using the data of each river section to train a unique regression model can prevent the model from sacrificing some river sections to meet the overall river network prediction accuracy. This embodiment selects a gated recurrent unit to construct a source tracing model, combined with Figure 4 The training process of the source tracing model for each river section is as follows: the water quality simulation data of the divided river channel is preprocessed by unifying the time unit, removing the interval symbol, using the Z score and IQR to identify and remove outliers, and the preprocessed data set is divided into a training set (80%) and a test set (20%). After the operation is completed, it is substituted into the entry control cycle unit to build an independent regression model for each river channel. If the model effect is good, the model will be saved. If the effect is not good, the model parameters will be modified to obtain good simulation accuracy.

[0032] The formula for the Z-score is: Among them, X represents the original data, represents the mean of the original data, and σ represents the standard deviation of the original data.

[0033] The formula for IQR is: Among them, Q1 represents the lower quartile value, that is, the 25th% number after all the values ​​in the sample are arranged from small to large, and Q3 represents the upper quartile value, that is, the 75th% number after all the values ​​in the sample are arranged from small to large.

[0034] The core formula of the gated recurrent unit is: Among them, z t 、r t are the outputs of the update gate and the reset gate, respectively, t is the input at time t, represents the candidate hidden state at time t, h t-1 、h t represents the hidden state at the corresponding moment, [] represents the connection between two vectors, W z , W r represents the weight matrix, represents the sigmoid function, represents the tanh activation function, represents the weight matrix, Represents the matrix product.

[0035] Step 5: Start the traceability model: Combine Figure 5 , the steps for pollutant tracing are as follows: (1) The monitoring point detects that the pollutant concentration in the area exceeds the standard, returns the information, and triggers the traceability model.

[0036] (2) The source tracing model randomly generates a pollution source parameter. In this embodiment, the pollution source parameter includes four items: pollution source location, pollution start time (0:00-8:00), pollution discharge duration (1-4 hours), and pollution discharge concentration (0-20000 mg / L). The pollution source location is randomly selected from the river section between the excessive monitoring point and the nearest monitoring point of its upstream tributaries. The remaining three parameters need to be determined based on the nature of the pollutant discharge enterprise, the relationship between water body fluidity and time, and the simulation density of the simulation data.

[0037] (3) Identify the pollution source parameters generated in step (2) and substitute them into the regression model of the river section for prediction. Return a predicted value of pollutants exceeding the standard corresponding to the pollution source parameters. Subtract the predicted value from the actual detected pollutants exceeding the standard to obtain a quantitative value of the correctness of the source tracing. If the difference is too large, it proves that the pollution source is very different from the actual pollution source.

[0038] (4) Using the crossover and mutation operations in the genetic algorithm, the pollution source parameters generated in step (2) are randomly changed and reorganized, and the range of the change and reorganization is consistent with the parameter range in step (2).

[0039] (5) Return the new pollution source parameters that have been mutated in step (4) to step (3).

[0040] (6) Set an iteration number. In this embodiment, it is set to 50 times. It can be modified according to the actual situation and the traceability accuracy. In the continuous iteration, repeat steps (3) to (5) and finally return a minimum difference between the predicted value and the true value. The pollution source parameter corresponding to the predicted value is the result of this pollution source tracing. Finally, the traceability result is returned and visualized. The results are shown in Figure 6 The traceability accuracy varies in different situations. In this embodiment, the traceability error of each parameter is about 5%.

[0041] The entire process is completed automatically by the system, which improves work efficiency and traceability accuracy.

[0042] The above embodiments are preferred implementations of the present invention, but the implementations of the present invention are not limited to the above embodiments. Any other modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A method for tracing the source of water pollution in a river network, characterized in that: The method comprises the following steps: Step S1, obtaining river network data; Step S2, constructing a hydrodynamic model: inputting the river network data into the MIKE software to construct a hydrodynamic model of the river network; Step S3, verifying the hydrodynamic model: taking any water quality monitoring point in the river network as the pollution source, setting the pollution discharge duration, pollution discharge concentration and concentration change interval of the pollution source and inputting them into the hydrodynamic model, obtaining the simulated data of the water quality monitoring points around the pollution source, and verifying the hydrodynamic model with the real monitoring data corresponding to the water quality monitoring points around the pollution source; Step S4, constructing a regression model: dividing the entire river network into multiple river sections with water quality monitoring points as nodes, and then inputting the river network data corresponding to each river section into the verified hydrodynamic model to obtain simulation data for each river section; inputting the simulation data of each river section into the gated recurrent unit after preprocessing to construct a regression model for each river section; Step S5, tracing the source of water pollution in the river network: when any water quality monitoring point detects that the pollutants exceed the standard, the crossover and mutation operations in the genetic algorithm are used to change and reorganize the pollution source parameters, and then input them into the regression model corresponding to the river section where the water quality monitoring point is located to obtain the predicted result of the pollutant concentration exceeding the standard at the current water quality monitoring point; the difference between the predicted result and the actual exceeding value is used as the evaluation index, and the above steps are repeated. After multiple iterations, the optimal predicted result of the pollutant concentration exceeding the standard at the current monitoring point is returned, and the pollution source parameter corresponding to the optimal predicted result is the tracing result.

2. The method for tracing the source of river water pollution according to claim 1 is characterized in that: The river network data includes river network geographic data, river network hydrological data, river channel patrol data, and river network water quality monitoring point monitoring data.

3. The method for tracing the source of river water pollution according to claim 1, characterized in that: The preprocessing in step S4 includes unifying time units, removing spacers, and identifying and removing outliers.

4. The method for tracing the source of river water pollution according to claim 1, characterized in that: The pollution source parameters include the location of the pollution source, the start time of pollution source emission, the duration of emission, and the emission concentration.

5. A river network water pollution tracing system for implementing the method described in any one of claims 1 to 4, characterized in that: Includes the following modules: Data collection module, used to obtain river network data; A hydrodynamic model building module is used to build a hydrodynamic model of a river network based on river network data; The regression model building module is used to divide the river network into multiple river sections, input the river network data corresponding to the river sections into the hydrodynamic model, and build the regression model of each river section after the gated circulation unit; The water pollution source tracing module uses regression models to predict pollution source parameters.

6. An electronic device, characterized in that: The invention comprises a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor executes the machine executable instructions to implement the method according to any one of claims 1 to 4.

7. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the method described in any one of claims 1 to 4.

Images