Heavy metal pollution source identification and risk assessment system in floodplains based on machine learning

Abstract

The present invention discloses a system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning, and relates to the technical field of heavy metal pollution; flood prediction is performed through environmental data; flood circulation data of a target area is extracted from flood prediction data to determine the source of heavy metal pollution; diffusion analysis is performed on the source of heavy metal pollution to obtain the heavy metal concentration of floods in the target area; the heavy metal concentration of floods in the target area is used as an input parameter of a pre-trained machine learning model to obtain the heavy metal variation of soil in the target area; risk assessment is performed based on the original heavy metal concentration and heavy metal variation of soil in the target area to obtain the risk level of the target area. The method can identify the scope and intensity of heavy metal pollution under the influence of floods in advance, thereby providing decision-making basis for managers, allocating resources in advance, and reducing the impact of floods and the heavy metal pollution they bring on the environment and residents.

Description

Technical Field

[0001] The present invention relates to the technical field of heavy metal pollution, and in particular to a floodplain heavy metal pollution source identification and risk assessment system based on machine learning. Background Art

[0002] The sources of heavy metal pollution can be traced back to many aspects. Industrial activities are the main source of heavy metal pollution, especially mining, smelting, fertilizer and other industries. Heavy metal elements in industrial wastewater and solid waste, such as lead, cadmium, mercury, arsenic, etc., are discharged into the soil and water bodies, and slowly affect the non-direct discharge areas over time. However, when a flood occurs, the scouring and soaking of the flood will bring heavy metals into the floodplain, thereby increasing the speed and area of ​​pollution spread.

[0003] The traditional risk assessment method for heavy metal pollution in floodplains is mainly to collect soil samples after the flood and conduct qualitative and quantitative analysis. Then, control measures are taken according to the analysis results. The existing heavy metal pollution assessment and treatment are lagging, and effective and accurate preventive measures cannot be taken, resulting in a large waste of resources in the later stage of control. Summary of the invention

[0004] The purpose of the present invention is to solve the hysteresis problem mentioned in the above background technology, and to propose a floodplain heavy metal pollution source identification and risk assessment system based on machine learning.

[0005] The first aspect of the present invention provides a system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning, the system comprising:

[0006] The flood prediction module is used to obtain the environmental data of the target area as the input of the preset flood simulation model to obtain flood prediction data;

[0007] A heavy metal pollution source identification module is used to extract flood flow data of a target area from the flood prediction data and determine a set of heavy metal pollution sources;

[0008] The pollution diffusion module is used to perform diffusion analysis based on the heavy metal pollution source set to obtain the heavy metal concentration of flood in the target area;

[0009] The regional pollution change calculation module is used to use the heavy metal concentration of flood in the target area as the input parameter of the pre-trained machine learning model to obtain the change of heavy metals in the soil of the target area;

[0010] The regional risk assessment module is used to conduct risk assessment based on the original heavy metal concentration and heavy metal change in the soil of the target area to obtain the risk level of the target area.

[0011] Optionally, the flood simulation model is a coupled model combining hydrology and hydrodynamics;

[0012] The flood prediction module comprises:

[0013] A data acquisition module is used to acquire environmental data of a target area; the environmental data includes meteorological data, hydrological data and land data;

[0014] The data prediction module is used to use the environmental data as input parameters of a preset flood simulation model, simulate with a preset time step, and obtain flood prediction data; the flood prediction data includes the flood inundation range at different times, water level and flow rate at different locations.

[0015] Optionally, the heavy metal pollution source identification module includes:

[0016] The emission source data acquisition module is used to obtain the data of pollutant emission sources in the target area; the data of pollutant emission sources include the location of the emission source and the amount of pollutant emission;

[0017] A flood branch extraction module is used to extract the flow data of the flood spreading to the target area from the flood prediction data according to the geographical location of the target area; the flow data includes the flow path;

[0018] The pollution source determination module is used to add the pollutant emission sources that emit heavy metal pollution on the flow path to the heavy metal pollution source set.

[0019] Optionally, the pollution diffusion module includes a river section division module, a calculation element determination module, a sequence calculation module, a termination judgment module and a concentration calculation module; wherein:

[0020] The river section division module is used to divide the flow path into multiple river sections according to the flood flow line speed and flow rate;

[0021] The calculation element determination module is used to take the starting position of each river section and the position of the heavy metal pollution source as calculation points to obtain a calculation queue; the position of the target area is the tail element of the calculation queue;

[0022] The sequential calculation module is used to pop up the head element of the calculation queue as the target calculation point;

[0023] The termination judgment module is used to enter the concentration calculation module if the calculation queue is not empty; otherwise, output the heavy metal concentration of the flood in the target area;

[0024] The concentration calculation module is used to perform diffusion analysis according to the heavy metal concentration of the flood passing through the target calculation point, obtain the heavy metal concentration of the flood reaching the next calculation point, and return to the sequential calculation module.

[0025] Optionally, the concentration calculation module includes:

[0026] A concentration updating module, for updating the heavy metal concentration after the flood passes through the target calculation point according to the pollutant emission data of the heavy metal pollution source if the target calculation point is a heavy metal pollution source;

[0027] Diffusion analysis module, used to calculate the concentration of heavy metals after floods pass through any river section: in, is the concentration of the jth heavy metal when the flood reaches the calculation point i+1; is the concentration of the jth heavy metal after the flood passes through the calculation point i; exp() is an exponential function with the natural constant e as the base; is the distance from calculation point i to calculation point i+1; is the linear velocity of the flood from calculation point i to calculation point i+1; and It is the constant coefficient for calculating the concentration of the j-th heavy metal.

[0028] Optionally, the regional pollution change calculation module includes:

[0029] A feature extraction module is used to extract flood characteristic parameters of the target area from the flood prediction data; the flood characteristic parameters include geological data of the target area, flood start time, flood duration, maximum water depth, time to reach the maximum water depth, and flood heavy metal concentration change curve;

[0030] The change calculation module is used to use the characteristic parameters as input of a pre-trained machine learning model to obtain the change of heavy metals in the soil of the target area.

[0031] Optionally, the regional risk assessment module includes:

[0032] The soil heavy metal concentration update module is used to obtain the heavy metal concentration of the target area after the flood based on the original heavy metal concentration and heavy metal change of the soil in the target area;

[0033] The risk index calculation module is used to calculate the Nemerow comprehensive pollution index and potential ecological risk index based on the heavy metal concentration in the target area after the flood;

[0034] The risk level determination module is used to determine the risk level of the target area according to the Nemerow comprehensive pollution index and the potential ecological risk index.

[0035] Beneficial effects of the present invention:

[0036] The present invention proposes a system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning, the system comprising: a flood prediction module, used to obtain environmental data of a target area as an input of a preset flood simulation model to obtain flood prediction data; a heavy metal pollution source identification module, used to extract flood circulation data of a target area from the flood prediction data to determine a set of heavy metal pollution sources; a pollution diffusion module, used to perform diffusion analysis based on the set of heavy metal pollution sources to obtain the heavy metal concentration of floods in the target area; a regional pollution change calculation module, used to use the heavy metal concentration of floods in the target area as an input parameter of a pre-trained machine learning model to obtain the heavy metal change amount of soil in the target area; and a regional risk assessment module, used to perform risk assessment based on the original heavy metal concentration and heavy metal change amount of soil in the target area to obtain the risk level of the target area.

[0037] By predicting flood distribution and analyzing the heavy metal pollution sources and their diffusion, we can get the pollution level when the flood spreads from the river to the target area. Then, we can calculate the change in soil heavy metal concentration in the target area under the influence of the flood through the machine learning model, and then conduct risk assessment. This method can identify the scope and intensity of heavy metal pollution under the influence of floods in advance, thus providing managers with a basis for decision-making and taking effective and accurate preventive measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present invention will be further described below in conjunction with the accompanying drawings.

[0039] Figure 1 An architecture diagram of a system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning is provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0040] 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.

[0041] The embodiment of the present invention provides a system for identifying and assessing the risk of heavy metal pollution sources in floodplains based on machine learning. Figure 1 , Figure 1 The present invention provides an architecture diagram of a system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning. The system includes the following modules:

[0042] The flood prediction module is used to obtain the environmental data of the target area as the input of the preset flood simulation model to obtain flood prediction data.

[0043] The heavy metal pollution source identification module is used to extract flood flow data in the target area from the flood prediction data and determine the set of heavy metal pollution sources.

[0044] The pollution diffusion module is used to perform diffusion analysis based on the set of heavy metal pollution sources to obtain the heavy metal concentration of floods in the target area.

[0045] The regional pollution change calculation module is used to use the heavy metal concentration of flood in the target area as the input parameter of the pre-trained machine learning model to obtain the change of heavy metals in the soil of the target area.

[0046] The regional risk assessment module is used to conduct risk assessment based on the original heavy metal concentration and heavy metal change in the soil of the target area to obtain the risk level of the target area.

[0047] Based on the machine learning-based floodplain heavy metal pollution source identification and risk assessment system provided by the embodiment of the present invention, by predicting the flood distribution and then analyzing the heavy metal pollution source and its pollution diffusion, the pollution degree when the flood spreads from the river channel to the target area can be obtained, and then the change in the soil heavy metal concentration in the target area under the influence of the flood can be calculated through the machine learning model, and then the risk assessment can be performed. This method can identify the scope and intensity of heavy metal pollution under the influence of floods in advance, thereby providing managers with a decision-making basis, allocating resources in advance, and reducing the impact of floods and the heavy metal pollution they bring on the environment and residents.

[0048] In one implementation, based on the risk assessment results, an efficient sedimentation device can be installed at a certain point in the flood path in a high-risk area to reduce the concentration of heavy metals in the flood, thereby reducing the pollution risk in the target area.

[0049] In one embodiment, the flood prediction module includes:

[0050] The data acquisition module is used to obtain the environmental data of the target area.

[0051] The data prediction module is used to use environmental data as input parameters of a preset flood simulation model, simulate at a preset time step, and obtain flood prediction data.

[0052] Among them, environmental data include meteorological data, hydrological data and land data; flood prediction data include flood inundation range at different times, water level and flow rate at different locations.

[0053] In one implementation, meteorological data includes rainfall, temperature, etc.; hydrological data covers river water level, flow, etc.; and land data involves topography, land use, etc.

[0054] In one implementation, the flood simulation model is a coupled model that combines hydrology and hydrodynamics, such as MIKE SHE and MIKE 11, HEC-HMS and HEC-RAS, etc. It can achieve efficient and accurate flood prediction and provide a data basis for subsequent heavy metal diffusion analysis.

[0055] In one embodiment, the heavy metal pollution source identification module includes:

[0056] The emission source data acquisition module is used to obtain data on pollutant emission sources in the target area.

[0057] The flood branch extraction module is used to extract the flow data of flood spreading to the target area from the flood prediction data according to the geographical location of the target area.

[0058] The pollution source determination module is used to add the pollutant emission sources that emit heavy metal pollution on the flow path to the heavy metal pollution source set.

[0059] Among them, the data on pollutant emission sources include the location of the emission sources and the amount of pollutant emissions; the circulation data include the circulation paths through which the flood reaches the target area.

[0060] In one implementation, the pollution sources are screened based on the flood flow path, ensuring a high correlation between the identified heavy metal pollution sources and the actual flood impact area, thereby improving the scientificity and accuracy of pollution source identification.

[0061] In one embodiment, the pollution diffusion module includes a river section division module, a calculation element determination module, a sequence calculation module, a termination judgment module and a concentration calculation module; wherein:

[0062] The river segment division module is used to divide the flow path into multiple river segments according to the flood flow line speed and flow rate.

[0063] The calculation element determination module is used to take the starting position of each river section and the position of the heavy metal pollution source as calculation points to obtain a calculation queue; the position of the target area is the tail element of the calculation queue.

[0064] The sequential calculation module is used to pop the head element of the calculation queue as the target calculation point.

[0065] The termination judgment module is used to enter the concentration calculation module if the calculation queue is not empty; otherwise, the heavy metal concentration of the flood in the target area is output.

[0066] The concentration calculation module is used to perform diffusion analysis based on the heavy metal concentration of the flood passing through the target calculation point, obtain the heavy metal concentration of the flood reaching the next calculation point, and return to the sequential calculation module.

[0067] In one implementation, the calculation is performed by finite element method. The flood of a preset time volume is analyzed as an independent unit. For example, for a flood passing through a certain calculation point, the flood that passes for 15 minutes can be intercepted as an independent unit for separate calculation. Finally, the calculation results of the independent flood units within the duration are combined to obtain the final prediction data.

[0068] In one implementation, floods in different river sections have different flow speeds. Dividing the flow path into multiple river sections helps to more accurately simulate the diffusion of flood heavy metal pollutants in different river sections, thereby improving the accuracy of the simulation results.

[0069] In one embodiment, the concentration calculation module includes:

[0070] The concentration updating module is used to update the heavy metal concentration after the flood passes through the target calculation point according to the pollutant emission data of the heavy metal pollution source if the target calculation point is a heavy metal pollution source.

[0071] Diffusion analysis module, used to calculate the concentration of heavy metals after floods pass through any river section: in, is the concentration of the jth heavy metal when the flood reaches the calculation point i+1; is the concentration of the jth heavy metal after the flood passes through the calculation point i; exp() is an exponential function with the natural constant e as the base; is the distance from calculation point i to calculation point i+1; is the linear velocity of the flood from calculation point i to calculation point i+1; and It is the constant coefficient for calculating the concentration of the j-th heavy metal.

[0072] In one implementation, the concentration can be updated based on the flow velocity of the flood water passing through the emission source and the emission rate of the pollutants: Among them, FMC is the contribution of pollution sources to the heavy metal concentration of flood water, which can be added to the heavy metal concentration of flood water; is the pollutant emission rate of the pollution source, in grams per hour; It is the concentration of heavy metals in pollutants emitted by pollution sources, in grams per cubic meter; is the flood velocity in cubic meters per hour. This simplified formula assumes that heavy metals are evenly diffused in independent flood units within a preset time period, which greatly simplifies the calculation process and improves the calculation efficiency. It should be noted that the unit of the flow line velocity mentioned above is meters per second.

[0073] In one implementation, the constant coefficient in the formula is related to the physical and chemical properties of the heavy metal and the target river, and can be obtained by fitting based on the monitoring data of historical floods in the target river. For example, for heavy metal lead, 0.02. For heavy metal cadmium, 0.07.

[0074] In one embodiment, the regional pollution change calculation module includes:

[0075] The feature extraction module is used to extract the flood characteristic parameters of the target area from the flood prediction data.

[0076] The change calculation module is used to use the characteristic parameters as the input of the pre-trained machine learning model to obtain the change of heavy metals in the soil of the target area.

[0077] Among them, the flooding characteristic parameters include geological data of the target area, the start time of flooding, the duration of flooding, the maximum water depth, the time to reach the maximum water depth, and the flood heavy metal concentration change curve.

[0078] In one implementation, multiple flooding characteristic parameters of the target area are closely related to heavy metal deposition in the soil, making subsequent predictions more accurate.

[0079] In one implementation, the machine learning model can be a decision tree model, a random forest model, or a neural network model. Machine learning can learn the heavy metal variation patterns in a specific region from historical data, avoiding the linear assumptions in traditional methods and providing higher accuracy and flexibility.

[0080] In one embodiment, the regional risk assessment module includes:

[0081] The soil heavy metal concentration update module is used to obtain the heavy metal concentration of the target area after the flood based on the original heavy metal concentration and heavy metal change in the soil of the target area.

[0082] Risk index calculation module, used to calculate Nemerow comprehensive pollution index and potential ecological risk index based on the heavy metal concentration in the target area after the flood

[0083] The risk level determination module is used to determine the risk level of the target area based on the Nemerow comprehensive pollution index and the potential ecological risk index.

[0084] In one implementation, by determining the risk level of the target area, the system can help managers prioritize high-risk areas and take timely pollution prevention and control measures.

[0085] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning, characterized in that: The system comprises: The flood prediction module is used to obtain the environmental data of the target area as the input of the preset flood simulation model to obtain flood prediction data; A heavy metal pollution source identification module is used to extract flood circulation data of a target area from the flood prediction data to determine a set of heavy metal pollution sources; the flood circulation data includes a circulation path; The pollution diffusion module is used to perform diffusion analysis based on the heavy metal pollution source set to obtain the heavy metal concentration of flood in the target area; The regional pollution change calculation module is used to use the heavy metal concentration of flood in the target area as the input parameter of the pre-trained machine learning model to obtain the change of heavy metals in the soil of the target area; The regional risk assessment module is used to conduct risk assessment based on the original heavy metal concentration and heavy metal change in the soil of the target area to obtain the risk level of the target area; The pollution diffusion module includes a river section division module, a calculation element determination module, a sequence calculation module, a termination judgment module and a concentration calculation module; wherein: The river section division module is used to divide the flow path into multiple river sections according to the flood flow line speed and flow rate; The calculation element determination module is used to take the starting position of each river section and the position of the heavy metal pollution source as calculation points to obtain a calculation queue; the position of the target area is the tail element of the calculation queue; The sequential calculation module is used to pop up the head element of the calculation queue as the target calculation point; The termination judgment module is used to enter the concentration calculation module if the calculation queue is not empty; otherwise, output the heavy metal concentration of the flood in the target area; The concentration calculation module is used to perform diffusion analysis based on the heavy metal concentration of the flood passing through the target calculation point, obtain the heavy metal concentration of the flood reaching the next calculation point, and return to the sequential calculation module; The concentration calculation module comprises: A concentration updating module, for updating the heavy metal concentration after the flood passes through the target calculation point according to the pollutant emission data of the heavy metal pollution source if the target calculation point is a heavy metal pollution source; Diffusion analysis module, used to calculate the concentration of heavy metals after floods pass through any river section: in, is the concentration of the jth heavy metal when the flood reaches the calculation point i+1; is the concentration of the jth heavy metal after the flood passes through the calculation point i; exp() is an exponential function with the natural constant e as the base; is the distance from calculation point i to calculation point i+1; is the linear velocity of the flood from calculation point i to calculation point i+1; and is the constant coefficient for the calculation of the j-th heavy metal concentration; The regional pollution change calculation module includes: A feature extraction module is used to extract flood characteristic parameters of the target area from the flood prediction data; the flood characteristic parameters include geological data of the target area, flood start time, flood duration, maximum water depth, time to reach the maximum water depth, and flood heavy metal concentration change curve; The change calculation module is used to use the characteristic parameters as input of a pre-trained machine learning model to obtain the change of heavy metals in the soil of the target area.

2. The system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning according to claim 1 is characterized in that: The flood simulation model is a coupled model that combines hydrology and hydrodynamics; The flood prediction module comprises: A data acquisition module is used to acquire environmental data of a target area; the environmental data includes meteorological data, hydrological data and land data; The data prediction module is used to use the environmental data as input parameters of a preset flood simulation model, simulate with a preset time step, and obtain flood prediction data; the flood prediction data includes the flood inundation range at different times, water level and flow rate at different locations.

3. The system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning according to claim 1, characterized in that: The heavy metal pollution source identification module includes: The emission source data acquisition module is used to obtain the data of pollutant emission sources in the target area; the data of pollutant emission sources include the location of the emission source and the amount of pollutant emission; A flood branch extraction module is used to extract the flow data of the flood spreading to the target area from the flood prediction data according to the geographical location of the target area; the flow data includes the flow path; The pollution source determination module is used to add the pollutant emission sources that emit heavy metal pollution on the flow path to the heavy metal pollution source set.

4. The system for identifying and assessing heavy metal pollution sources in floodplains based on machine learning according to claim 1, characterized in that: The regional risk assessment module includes: The soil heavy metal concentration update module is used to obtain the heavy metal concentration of the target area after the flood based on the original heavy metal concentration and the change of heavy metal in the soil of the target area; The risk index calculation module is used to calculate the Nemerow comprehensive pollution index and potential ecological risk index based on the heavy metal concentration in the target area after the flood; The risk level determination module is used to determine the risk level of the target area based on the Nemerow comprehensive pollution index and the potential ecological risk index.

Images