A simulation device for nitrogen migration transformation

Abstract

The application provides a simulation device for nitrogen migration and transformation, which comprises an input unit, a simulation unit, a biological unit, a regulation unit and a data processing unit, the input unit is used to simulate the migration of nitrogen source input from a downstream river to a surrounding area; the simulation unit comprises a river channel simulation mechanism, a confluence triangular area simulation tank, a first side bank simulation tank of a main stream and a second side bank simulation tank of the main stream, the biological unit controls aerobic / anaerobic partition and nitrification / denitrification partition by regulating carbon source, oxygen and controlling soil structure / particle, the regulation unit is arranged in the confluence triangular area simulation tank, the first side bank simulation tank of the main stream and the second side bank simulation tank of the main stream, and is used to regulate dissolved oxygen, PH and conductivity, and the data processing unit is used to give a prediction result. The application solves the problem that a traditional water tank cannot simulate the interaction of two rivers, and improves the accuracy of simulation experiment.

Description

Technical Field

[0001] This invention belongs to the technical field of nitrogen migration and conversion simulation devices, and specifically relates to a nitrogen migration and conversion simulation device. Background Technology

[0002] Globally, 120 million tons of nitrogen fertilizer are applied to farmland annually, and when combined with organic fertilizers and biological nitrogen fixation, the total input reaches approximately 200 million tons. However, crops utilize less than 50% of this nitrogen fertilizer, and excess nitrogen enters water bodies through runoff, becoming a major source of pollution. Meanwhile, nitrogen oxides emitted by industry and traffic exhaust are significant sources of atmospheric nitrogen pollution, especially in rapidly urbanizing areas. The inter-basin zone refers to the water-saturated sediment layer beneath the riverbed and extending to the banks and sides of rivers and streams. As an area where river water and groundwater exchange and mix, the inter-basin zone possesses a certain natural purification capacity and is effective in removing pollutants from river water and groundwater.

[0003] The confluence zone is the junction of different river systems. Its specific hydrodynamic characteristics greatly influence the migration and distribution of pollutants, thereby further affecting microbial growth and biogeochemical processes. Due to differences in density and flow velocity between tributaries and main streams, low-velocity backflow or separation zones emerge. In these areas, eddies or countercurrents form, leading to energy loss, sediment deposition, and pollutant retention. Simultaneously, velocity differences between the two streams at their confluence create shear layers, resulting in enhanced turbulence and intense mixing. This phenomenon significantly impacts pollutant diffusion and sediment transport. Differences in dissolved oxygen and pH between the two rivers directly affect the composition and structure of microorganisms in the confluence zone, further influencing the migration and transformation of nutrients such as nitrogen and phosphorus. In river confluence zones, especially backflow zones, increased water retention time and decreased flow velocity promote pollutant deposition and retention. Increased pollutant or nutrient content in sediments makes these areas high-risk areas for eutrophication. Nutrient accumulation promotes phytoplankton growth, potentially leading to algal blooms and other ecological problems.

[0004] The confluence zone of rivers differs significantly from that of normal rivers, primarily in terms of hydrodynamic characteristics, chemical properties, ecological functions, and spatial extent. The hydrodynamic conditions at the confluence are complex, influenced by the flow velocities, discharges, and water level fluctuations of the two rivers. The exchange between surface water and groundwater is more frequent, potentially leading to multidirectional flows. Secondly, the chemical conditions are more diverse; the two rivers may carry different nutrients, dissolved ions, or pollutants, forming a mixing reaction zone at the confluence.

[0005] As a tool to assist scientific research, simulation devices not only require high flexibility in the simulation process and high accuracy in the simulation results, but also necessitate combining the large amounts of data simulated by physical models (simulation devices) with mathematical models to extract scientific elements from the data and solve the scientific problems being studied. Deeper research requires fitting past data to predict the future. Taking traditional nitrogen conversion prediction methods as an example, existing methods mainly include: traditional supervised learning methods (such as linear regression fitting, decision trees, and random forests), whose advantages are strong interpretability and quick and convenient handling of simple problems, but still have disadvantages such as easy overfitting and susceptibility to complex environments; and deep learning methods (Artificial Neural Networks (ANNs), Recurrent Neural Networks (RNNs) / Long Short-Term Memory Networks (LSTMs), which use large-sample machine learning to handle arbitrarily complex functions.

[0006] Existing seepage flume simulation chambers mainly simulate nutrient migration in the inter-basin zone of a single main stream or tributary, reflecting limited geophysical processes. This invention proposes a nitrogen migration and transformation simulation device to address the scientific question of nitrogen migration and transformation in the confluence region of two rivers. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a nitrogen migration and conversion simulation device to address the shortcomings of the prior art mentioned above.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a nitrogen migration and transformation simulation device, comprising;

[0009] The input unit uses a water pump to input the prepared nitrogen simulation solution into the simulation unit to simulate the migration of nitrogen sources from downstream rivers to the surrounding areas.

[0010] The simulation unit includes a river channel simulation mechanism, a confluence triangle area simulation channel, a main stream first side bank simulation channel, and a main stream second side bank simulation channel;

[0011] The river simulation mechanism is used to simulate the main stream and tributaries of the river. The confluence triangle simulation tank is set at the confluence of the main stream and tributaries. The first side bank simulation tank and the second side bank simulation tank of the main stream are respectively set on both sides of the main stream.

[0012] The biological unit controls aerobic / anaerobic zones and nitrification / denitrification zones by regulating carbon sources, oxygen content, and soil structure / particle size. Dissolved oxygen in the water can be controlled by agitating the water in the tank or introducing inert gas, while samplers are buried in the aerobic / anaerobic zones to monitor nitrogen migration under different conditions. Secondly, the C / N ratio of the simulated solution can be adjusted from 10:1 to 20:1 to promote / limit denitrification behavior. When heterogeneous soil is used, the uneven distribution of porosity will simultaneously alter the transfer of oxygen and nutrients, thereby changing the distribution of microorganisms.

[0013] By regulating carbon sources and controlling soil structure / particles, the aerobic / anaerobic and nitrification / denitrification zones can be controlled. Carbon source regulation involves adjusting the carbon-nitrogen ratio in the simulated solution. Simulated solutions with different concentrations are added to the input unit, and the C / N ratio is adjusted from 10:1 to 20:1 to promote / limit denitrification behavior. The composition of nutrients in the simulated tank is controlled. Homogeneous soil is filled into the geological units within the simulated tank. Aerobic zones are formed at 0-10 cm from the surface of the geological units, while the remaining parts of the geological units constitute anaerobic zones. When heterogeneous soil is added, the uneven distribution of porosity simultaneously alters oxygen and nutrient transport, thereby changing the distribution of microorganisms. Pore water samplers are also embedded in both aerobic and anaerobic zones and connected to the output module. By analyzing changes in nitrogen compounds in the pore water, nitrogen migration under the two different environments can be monitored.

[0014] The control units are respectively installed in the confluence triangle region simulation tank, the main stream first side bank simulation tank, and the main stream second side bank simulation tank, and are used to control dissolved oxygen, pH and conductivity.

[0015] The data processing unit includes a monitoring terminal and a computer. The monitoring terminal is set up in the simulation unit and transmits the monitored data to the computer. The computer has a preset LSTM neural network model. The data is processed by the LSTM neural network model for deep learning and prediction results are given.

[0016] As a further explanation of the present invention, the river channel simulation mechanism includes a main flow channel, and the inlet end of the main flow channel is connected to a main flow tank.

[0017] A branch water trough, wherein the inlet end of the branch water trough is connected to a branch water tank;

[0018] The tributary water channel is obliquely connected to one side of the main water channel;

[0019] A collection trough is movably connected to the tail end of the main flow trough, and a wastewater tank is connected to the tail end of the collection trough.

[0020] As a further explanation of the present invention, flow valves are provided on the connection channels between the main stream water channel and the main stream water tank, as well as the connection channels between the tributary water channel and the tributary water tank. Multiple river sampling ports are also provided on the tributary water channel.

[0021] As a further explanation of the present invention, the intersection triangular region simulation tank includes a first movable water-proof baffle, a first water-permeable baffle, and a first glass base plate. Two first water-permeable baffles are provided, and the two first water-permeable baffles are rotatably connected to a torsion spring hinge to form a zigzag structure. The zigzag structure is installed on the first glass base plate, and the first movable water-proof baffle is movably connected to the open end of the zigzag structure to form a sandbox structure with an open top. Multiple adjustable bases are installed at the bottom end of the first glass base plate.

[0022] As a further explanation of the present invention, the angle of the zigzag structure can be adjusted within the range of 20 to 60°, and the length of the first movable water-proof baffle 30 is 30 to 60 cm.

[0023] As a further explanation of the present invention, the first movable water-proof baffle is also provided with a plurality of first water outlets and a first water level control head. A pore water sampler is connected to the first water outlet. The collection end of the pore water sampler is set inside the sand box structure, and the output end of the pore water sampler is connected to the input end of the data processing unit.

[0024] As a further explanation of the present invention, the simulated channel on the first side bank of the main stream is specifically a cavity structure composed of two first water-proof baffles, one third water-proof baffle, a first plexiglass base plate, one third permeable baffle, and a water-proof baffle with a reserved opening. A sampling tube is connected to the second outlet of the third water-proof baffle, and the sampling tube is connected to the input end of the data processing unit.

[0025] The water-resistant baffle with a reserved opening is connected to multiple secondary water level control heads to simulate the spatial distribution of groundwater level on the riverbank.

[0026] As a further explanation of the present invention, the simulated channel on the second side bank of the main stream includes a second permeable baffle, a movable perforated baffle, two second water-proof baffles, a second movable water-proof baffle, and a second plexiglass bottom plate.

[0027] The second permeable baffle is located on the side near the confluence water tank. The second permeable baffle, the second water-proof baffle, the second movable water-proof baffle, and the second plexiglass base plate form the main cavity structure. The movable perforated baffle and the movable water-proof baffle are both equipped with pulleys on their movable bottom sides. The movable perforated baffle is located inside the main cavity structure. The second movable water-proof baffle is also provided with a second water outlet and a second water level control head.

[0028] As a further explanation of the present invention, the dissolved oxygen regulation in the regulation unit is accomplished by a stirrer, an inert gas cylinder and a dissolved oxygen sensor preset in the river simulation mechanism. The dissolved oxygen regulation is achieved by turning on the stirrer or injecting inert gas.

[0029] The pH control in the control unit is accomplished by adding acid or alkali reagents or buffers to the confluence triangle area simulation tank, the first side bank simulation tank of the main stream, and the second side bank simulation tank of the main stream, so as to control the pH of the water in the confluence triangle area simulation tank, the first side bank simulation tank of the main stream, and the second side bank simulation tank of the main stream at 5~10.

[0030] The conductivity regulation in the control unit is achieved by mixing NaCl solution with raw water and then adding it into the river simulation mechanism.

[0031] As a further explanation of the present invention, the monitoring terminal monitors the hydrological information within the simulation unit and classifies and organizes the monitored hydrological information according to the set indicators.

[0032] Compared with the prior art, the present invention has the following advantages:

[0033] This invention, through controllable hydraulic conditions, can realistically simulate the migration and transformation of nitrogen salts in actual confluence areas, solving the problem that traditional flumes cannot simulate the interaction between two rivers. Simultaneously, by changing the flow ratio of tributaries and the groundwater level, the simulation conditions are flexible and controllable, enhancing adaptability to the confluence of natural rivers and further improving the accuracy of the simulation experiments. Attached Figure Description

[0034] Figure 1 This is a top view of the overall structure of the invention;

[0035] Figure 2 This is a side view of the simulated trench in the triangular region of the present invention;

[0036] Figure 3 This is a front view of the simulated trench in the triangular region of the present invention;

[0037] Figure 4 This is a front view of the simulated channel on the first side bank of the main stream of the present invention;

[0038] Figure 5 This is a three-dimensional view of the simulated channel on the first side bank of the main stream of the present invention;

[0039] Figure 6 This is a three-dimensional view of the simulated channel on the second side bank of the main stream of the present invention;

[0040] Figure 7 This is a comparison chart of the actual and predicted values ​​of the data output by the training model for 26-28 days in the experimental example of this invention.

[0041] Explanation of reference numerals in the attached figures:

[0042] 1-Simulation tank on the first side bank of the main stream; 2-Main stream water tank; 3-Branch water tank; 4-Water pump; 5-Wastewater tank; 6-Flow valve; 7-Simulation tank of the confluence triangle area; 8-CMOS monitoring camera; 9-First outlet; 10-Adjustable base; 11-Gauze screen; 12-Pore water sampler; 13-First water level control head; 14-Water level control box; 15-Second outlet; 16-Second water level control head; 17-Second permeable baffle; 18-Waterproof baffle with reserved opening; 19-Modible perforated baffle; 20-Second waterproof baffle 21-Main channel water tank; 22-Tributary water tank; 23-Confluence water tank; 24-Main channel second side bank simulation tank; 25-River sampling port; 26-Sampling capillary tube; 27-Dissolved oxygen sensor; 28-Torsion spring hinge; 29-Pulley; 30-First movable water-proof baffle; 31-First permeable baffle; 31-First glass base plate; 33-Third water-proof baffle; 34-Third permeable baffle; 35-First water-proof baffle; 36-First plexiglass base plate; 37-Second movable water-proof baffle; 38-Second plexiglass base plate. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] like Figure 1 As shown, the present invention provides a technical solution: a simulation device for nitrogen migration and transformation, comprising;

[0045] A simulation device for nitrogen migration and transformation includes: an input unit, a simulation unit, a control unit, and a data processing unit;

[0046] The input unit uses a water pump 4 to input the prepared nitrogen simulation solution into the simulation unit to simulate the migration of nitrogen source input downstream river to the surrounding area; the simulation unit includes a river channel simulation mechanism, a confluence triangle area simulation tank 7, a main stream first side bank simulation tank 1, and a main stream second side bank simulation tank 24;

[0047] The river simulation mechanism is used to simulate the main stream and tributaries of the river. The confluence triangle simulation tank 7 is set at the confluence of the main stream and tributaries. The first side bank simulation tank 1 and the second side bank simulation tank 24 of the main stream are respectively set on both sides of the main stream.

[0048] Biological units are used to control aerobic / anaerobic zones and nitrification / denitrification zones by regulating carbon sources, oxygen content, and soil structure / particles.

[0049] Dissolved oxygen in the water can be controlled by stirring the water or introducing inert gas at two points in the main flow tank. Simultaneously, samplers can be buried in the aerobic / anaerobic zones to monitor nitrogen migration under the two different environments. Secondly, the C / N ratio of the simulated solution can be adjusted from 10:1 to 20:1 to promote / limit denitrification. When heterogeneous soil is used, the uneven distribution of porosity will simultaneously alter the transfer of oxygen and nutrients, thereby changing the distribution of microorganisms.

[0050] By regulating carbon sources and controlling soil structure / particles, the aerobic / anaerobic and nitrification / denitrification zones can be controlled. Carbon source regulation involves adjusting the carbon-nitrogen ratio in the simulated solution. Simulated solutions with different concentrations are added to the input unit, and the C / N ratio is adjusted from 10:1 to 20:1 to promote / limit denitrification behavior and control the nutrient composition in the simulation tank. Homogeneous soil is filled into the geological units within the simulation tank. Aerobic zones are formed at 0-10 cm from the surface of the geological units, while the remaining parts of the geological units constitute anaerobic zones. When heterogeneous soil is filled, the uneven distribution of porosity simultaneously alters the transfer of oxygen and nutrients, thereby changing the distribution of microorganisms. Pore water samplers 12 are embedded in both the aerobic and anaerobic zones and connected to the output module. By analyzing changes in nitrogen compounds in the pore water, nitrogen migration under the two different environments can be monitored.

[0051] The control units are respectively installed in the confluence triangle region simulation tank 7, the main stream first side bank simulation tank 1, and the main stream second side bank simulation tank 24, and are used to control dissolved oxygen, pH and conductivity.

[0052] The data processing unit includes a monitoring terminal and a computer. The monitoring terminal is set up in the simulation unit and transmits the monitored data to the computer. The computer has a preset LSTM neural network model. The data is processed by the LSTM neural network model for deep learning and prediction results are given.

[0053] In this embodiment, the river simulation mechanism includes a main flow channel 21, the inlet end of the main flow channel 21 is connected to a main flow tank 2, and a flow valve 6 is provided on the communication passage between the main flow channel 21 and the main flow tank 2. The flow valve 6 can control the water flow from the main flow tank 2 to the main flow channel 21.

[0054] A branch water tank 22 is provided, with its inlet end connected to a branch water tank 3. A flow valve 6 is also provided on the connection between the branch water tank 22 and the branch water tank 3 to control the water flow from the branch water tank 3 into the branch water tank 22.

[0055] The tributary channel 22 is obliquely connected to one side of the main channel channel 21 to simulate the flow patterns of the main stream and tributaries in a river.

[0056] The main water channel 21 is movably connected to the end of the main water channel 21 via a confluence trough 23. The confluence trough 23 has a certain tilt angle, which can be freely adjusted via an adjustable base 10. The end of the confluence trough 23 is connected to a wastewater tank 5.

[0057] Flow valves 6 are installed on the connection channels between the main flow channel 21 and the main flow tank 2, as well as the connection channels between the tributary channel 22 and the tributary tank 3. Multiple river sampling ports 25 are also installed on the tributary channel 22 for monitoring the water quality in the simulated river.

[0058] In this embodiment, the intersection triangular region simulation tank 7 includes a first movable water-proof baffle 30, a first water-permeable baffle 31, and a first glass base plate 32. Two first water-permeable baffles 31 are provided, and the two first water-permeable baffles 31 are rotatably connected to the torsion spring hinge 28 to form a zigzag structure. The zigzag structure is installed on the first glass base plate 32, and the first movable water-proof baffle 30 is movably connected to the open end of the zigzag structure to form a sandbox structure with an open top. Multiple adjustable bases 10 are installed at the bottom of the first glass base plate 32, and the height can be adjusted by adjusting the adjustable bases 10.

[0059] As one possible implementation in this embodiment, the angle of the zigzag structure can be adjusted within the range of 20~60°, and the length of the first movable water-proof baffle 30 is 30~60cm.

[0060] The first movable baffle 30 is also provided with multiple first water outlets 9 and first water level control heads 13. A pore water sampler 12 is connected to the first water outlet 9. A mesh 11 is provided on the inner end of the pore water sampler 12 to prevent mud and sand from being washed into the pore water sampler 12. The collection end of the pore water sampler 12 is set inside the sand box structure. The output end of the pore water sampler 12 is connected to the input end of the data processing unit.

[0061] The main stream first side bank simulation tank 1 is specifically a cavity structure with a water level control box 14, consisting of two first water-proof baffles 35, a third water-proof baffle 33, a first plexiglass base plate 36, a third permeable baffle 34, and a water-proof baffle 18 with a reserved opening. A sampling tube 26 is connected to the second outlet 15 of the third water-proof baffle 33, and the sampling tube 26 is connected to the input end of the data processing unit.

[0062] The water-resistant baffle 18 with a reserved opening is connected to multiple second water level control heads 16, which are used to simulate the spatial distribution of groundwater level on the riverbank.

[0063] The main stream second side bank simulation channel 24 includes a second permeable baffle 17, a movable perforated baffle 19, two second water-proof baffles 20, a second movable water-proof baffle 37, and a second plexiglass base plate 38.

[0064] The second permeable baffle 17 is located on the side near the confluence water tank 23. The second permeable baffle 17, the second water-proof baffle 20, the second movable water-proof baffle 37, and the second plexiglass base plate 38 form a main cavity structure with a water level control box 14. The movable perforated baffle 19 and the movable water-proof baffle 37 are both equipped with pulleys 29 at their movable bottom ends. The movable perforated baffle 19 is located inside the main cavity structure. The second movable water-proof baffle 37 is also provided with a second outlet 15 and a second water level control head 16.

[0065] The dissolved oxygen regulation in the regulation unit is accomplished by a stirrer, an inert gas cylinder, and a dissolved oxygen sensor 27 pre-installed in the river simulation mechanism. The dissolved oxygen regulation is achieved by turning on the stirrer or injecting inert gas.

[0066] The pH control in the control unit is accomplished by adding acid or alkali reagents or buffers to the confluence triangle area simulation tank 7, the main stream first side bank simulation tank 1, and the main stream second side bank simulation tank 24, so as to control the pH of the water in the confluence triangle area simulation tank 7, the main stream first side bank simulation tank 1, and the main stream second side bank simulation tank 24 at 5~10.

[0067] The conductivity regulation in the control unit is achieved by mixing NaCl solution with raw water and then adding it into the river simulation mechanism.

[0068] The monitoring terminal monitors the hydrological information within the simulation unit and categorizes and organizes the monitored hydrological information according to the set indicators. The hydrological information includes the content of ammonia nitrogen, nitrate nitrogen, COD, pH, dissolved oxygen, conductivity, and DOC in the water, as well as the water flow velocity and direction. Simultaneously, it records environmental information such as soil particle size and temperature.

[0069] In the experimental example, a high concentration of nitrogen was simulated being discharged into the main channel, with the nitrogen concentration in the tributaries being the measured value of the river. The nitrogen simulation solution for the main channel was prepared as follows: 980 mg / L carbon source CH3COONa (400 mg / L as C), 220 mg / L NaNO3 (48 mg / L as N), and 6 mg / L NH4Cl (5 mg / L as N).

[0070] Before the experiment, the confluence angle of the main stream and tributaries and the inclination angle of the main stream were determined, and the nitrogen simulation liquid was output at a constant rate through the flow valve.

[0071] The main stream velocity was determined to be 1 m / s, the tributary velocity to be 0.5 m / s, the tributary-to-main stream velocity ratio to be 2, the flow rate ratio to be 10:1, and the confluence angle to be 30°. After determining the confluence angle, each simulation tank was installed as follows: Figure 1 The location is such that a CMOS camera 8 is installed on the upper part of the main stream channel 21, the tributary channel 22 and the confluence channel 23, and the dissolved oxygen, pH and conductivity in the confluence triangle area simulation channel 7, the first side bank simulation channel 1 of the main stream and the second side bank simulation channel 24 of the main stream are adjusted by a computer control and regulation unit.

[0072] The experiment used a non-uniform soil setup. The backfill soil was air-dried sediment and sieved. In the confluence triangular area simulation tank 7, the first side bank simulation tank 1 of the main stream, and the second side bank simulation tank 24 of the main stream, soil samples from the 0-300mm, 300-500mm, and 500-700mm layers of undisturbed riverbed sediments in the inter-zone were placed according to the layering. The layers were filled with fine sand, clay, and coarse sand, respectively. At the same time, pore water samplers 12 were buried in each layer. The samplers were placed in series and connected to the simulation tanks via rubber tubes. A mesh screen 11 was installed inside to prevent silt from being washed in. A sampling tube 26 was connected to the outlet and connected to an automatic sampling device via a peristaltic pump. The peristaltic pump was set to have an interval of 20 minutes and an injection volume of 100ml. The sample was then automatically fed into a water quality analyzer for analysis and measurement of various chemical indicators. To simulate the interactive process, a fixed water head in the water level control head 16 was opened in the simulation tanks on the left and right banks of the main stream, and the water head h0 was recorded. The water head was then connected to a water pump through a rubber tube to send the seepage water into the wastewater tank.

[0073] The experiment used an intelligent water quality analyzer with an output module to measure NH4+ in real time. 3+ NO 3- The content of COD, DOC, and conductivity.

[0074] Furthermore, the flow velocity in the confluence area was observed and recorded using a COMS camera 8 to observe whether backflow zones or eddies appeared, and the water levels h1, h2, and h3 of each river segment were determined, along with the head difference h1-h0. The simulation device operated continuously for three months. Optionally, a filter membrane was installed at the sampling port, and the migration and transformation patterns of nitrogen nutrients under different soil and rock structures were analyzed by detecting the structure of each nitrifying / denitrifying bacterial community in the filter membrane. Through data analysis, in the vertically heterogeneous structures of fine sand, clay, and coarse sand:

[0075] The surface layer (0-10cm) has good aeration and is the main nitrification zone, producing NH4+. 4+ NO 3- Rapid conversion;

[0076] The middle layer (10-20cm) is rich in water and gradually depletes oxygen, making it a source of NO. 2-Accumulation and denitrification start-up zone;

[0077] The deeper layers (>20cm) are in an anaerobic state and are the main zone for denitrification, producing NO. 3- Completely transformed, NH 4+ Due to its strong adsorption properties, it easily accumulates. Taking nitrate nitrogen in the triangular catchment area as the observation object, specific data are shown in Table 1 below:

[0078] Table 1 Nitrate content in the triangular confluence area

[0079]

[0080] Furthermore, real-time monitoring of changes in various indicator data is conducted. Optionally, after the data stabilizes, key variables can be changed to further simulate nitrogen migration after changes in biological units, such as changes in nitrogen interactions in the confluence zone under non-constant environmental conditions. Optionally, dissolved oxygen gradient: 5~10 mg / L, pH fluctuation: 5~10, organic matter load (C / N ratio) range: 10~20 can be set. The changes in the content of various forms of nitrogen in the three simulation chambers are the focus of observation. Taking an increase of 1 mg / L dissolved oxygen concentration and pH adjustment to 8 as an example, nitrate nitrogen is the key observation target; specific data are shown in Tables 2~4 below:

[0081] Table 2 Nitrogen and Nitrogen Variation in the Triangular Confluence Zone

[0082]

[0083] Table 3. Nitrogen and Nitrogen Variation in the Simulated Tank on the Left Bank

[0084]

[0085] Table 4. Nitrogen and Nitrogen Variation in the Right Bank Simulated Tank

[0086]

[0087] As shown above, nitrate nitrogen levels initially increase and then decrease in the deltaic confluence area. On the left bank of the river, nitrate nitrogen content shows no significant change on the left side of the confluence area, while on the right bank, it gradually increases after approximately 10 simulated days. The right bank is most affected, with the degree of impact decreasing from upstream to downstream. Increasing pH and dissolved oxygen levels inhibits denitrification to some extent, leading to increased nitrate nitrogen formation and thus a slight increase in nitrate nitrogen content.

[0088] Taking three months of simulated data as an example, after data processing, the various data points are packaged into a dataset and classified into training, validation, and test sets. The long-term nitrogen migration trend is predicted using an LSTM neural network. The number of neurons in the input and output layers of the LSTM network is determined based on the size of the data sample.

[0089] The input layer has 12 neurons, the output layer has 1 neuron, the time step is set to 6, and training begins with the goal of predicting three days.

[0090] The main calculation formulas and gate control in LSTM are as follows:

[0091] The input gate determines how much of the current cell's memory state is output as the hidden state at the current time step:

[0092]

[0093] in:

[0094] i t The output of the input gate controls how much new information is written.

[0095] σ Sigmoid function

[0096] W i Gate weight matrix

[0097] b i : Input gate bias term

[0098] C t Candidate memory states, new information

[0099] W C Weight matrix of candidate memory states

[0100] b C Bias terms for candidate memory states

[0101] The forgetting gate determines how much of the previous memory information is retained at the current moment:

[0102]

[0103] in:

[0104] f t The output of the forget gate controls how many old memories are retained.

[0105] σ The sigmoid function outputs values ​​in the range (0, 1).

[0106] W f Weight matrix of the forget gate

[0107] The hidden state of the previous time step

[0108] x t : Input vector at the current time step

[0109] Vector concatenation operation

[0110] b f Bias term of the forget gate

[0111] The update unit, the "long-term memory" that runs throughout the entire sequence, is updated jointly by the forget gate and the input gate:

[0112]

[0113] in:

[0114] C t : The state of the memory cell at the current time step

[0115] The memory state of the previous time step

[0116] f t Forgot Gate Output

[0117] i t Input gate output

[0118] C t-1 Candidate memory states

[0119] The output gate determines how much of the current cell's memory state is output as the hidden state at the current time step:

[0120]

[0121] in:

[0122] o t Output gate output (controls the amount of output currently stored).

[0123] σ Sigmoid function

[0124] W o : Weight matrix of the output gate

[0125] b o : Output gate bias term

[0126] : Output gate bias term

[0127] The activation function is:

[0128]

[0129]

[0130] in:

[0131] σ(x) The output of the sigmoid function normalizes the input x.

[0132] x The input value is a scalar, vector, or matrix, a weighted input from the neuron.

[0133] For model validation, three metrics were used for evaluation: mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R²).

[0134]

[0135] in:

[0136] n Total number of samples, number of observations

[0137] y i The true value of the i-th sample

[0138] The predicted value of the i-th sample

[0139] The absolute value of the prediction error for the i-th sample.

[0140] Sum of squared residuals;

[0141] After training, predictions are made on both the training and test sets, and the predicted values ​​are then denormalized.

[0142] Taking the aforementioned triangular confluence area data as an example, the data from the first 20 days was used as the training set, and the data from days 20 to 25 was used as the validation set. The trained model was then used to output data from days 26 to 28. A comparison of the actual and predicted values ​​is shown below. Figure 7 :

[0143] The LSTM model's prediction results showed an MAE of 0.396, R² of 0.956, and RMSE of 0.508. Compared with the actual values, the predicted values ​​matched the actual values ​​in terms of trend, and the overall error was small. The R² of 0.956 indicates that the neural network training achieved a good fit.

[0144] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0145] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A device for simulating nitrogen migration and transformation, characterized in that, include; The input unit uses a water pump to input the prepared nitrogen simulation solution into the simulation unit to simulate the migration of nitrogen sources from downstream rivers to the surrounding areas. The simulation unit includes a river channel simulation mechanism, a confluence triangle area simulation channel, a main stream first side bank simulation channel, and a main stream second side bank simulation channel; The river simulation mechanism is used to simulate the main stream and tributaries of the river. The confluence triangle simulation tank is set at the confluence of the main stream and tributaries. The first side bank simulation tank and the second side bank simulation tank of the main stream are respectively set on both sides of the main stream. The confluence triangle region simulation tank includes a first movable water-proof baffle, a first permeable baffle, and a first glass base plate. Two first permeable baffles are provided, rotatably connected to a torsion spring hinge to form a zigzag structure. The zigzag structure is mounted on the first glass base plate, and the first movable water-proof baffles are movably connected to the open end of the zigzag structure to form a sandbox structure with a top opening. Multiple adjustable bases are installed at the bottom of the first glass base plate. Specifically, the main stream first sidebank simulation tank consists of two first water-proof baffles, one third water-proof baffle, a first plexiglass base plate, one third permeable baffle, and a pre-reserved opening. The structure comprises a cavity with a water level control box, consisting of water-proof baffles. A sampling tube is connected to the second outlet of the third water-proof baffle, which is connected to the input of a data processing unit. Multiple second water level control heads are connected to the water-proof baffle with reserved openings to simulate the spatial distribution of groundwater levels along the riverbank. The main stream's second sidebank simulation trench includes a second permeable baffle, a movable perforated baffle, two second water-proof baffles, a second movable water-proof baffle, and a second plexiglass base plate. The second permeable baffle, the second water-proof baffle, the second movable water-proof baffle, and the second plexiglass base plate form the main cavity structure with a water level control box. Biological units, namely, controlling aerobic / anaerobic zones and nitrification / denitrification zones by regulating carbon sources, oxygen content, and soil structure / particles; The control units are respectively installed in the confluence triangle area simulation tank, the first side bank simulation tank of the main stream, and the second side bank simulation tank of the main stream. They are used to control dissolved oxygen, pH, and conductivity. Dissolved oxygen control is achieved through a stirrer, inert gas cylinder, and dissolved oxygen sensor pre-installed within the river simulation mechanism. This is done by turning on the stirrer or injecting inert gas. pH control is achieved by adding acid / base reagents or buffers to the confluence triangle area simulation tank, the first side bank simulation tank of the main stream, and the second side bank simulation tank of the main stream, maintaining the pH of the water in these tanks between 5 and 10. Conductivity control is achieved by mixing NaCl solution with the raw water and then adding it to the river simulation mechanism. The data processing unit includes a monitoring terminal and a computer. The monitoring terminal is set up in the simulation unit and transmits the monitored data to the computer. The computer has a preset LSTM neural network model. The data is processed by the LSTM neural network model for deep learning and prediction results are given.

2. The nitrogen migration and transformation simulation device according to claim 1, characterized in that, The river simulation mechanism includes a main flow channel, and the inlet end of the main flow channel is connected to a main flow tank. A branch water trough, wherein the inlet end of the branch water trough is connected to a branch water tank; The tributary water channel is obliquely connected to one side of the main water channel; A collection trough is movably connected to the tail end of the main flow trough, and a wastewater tank is connected to the tail end of the collection trough.

3. The nitrogen migration and transformation simulation device according to claim 2, characterized in that, Flow valves are installed on the connecting passages between the main stream water channel and the main stream water tank, as well as the connecting passages between the tributary water channels and the tributary water tanks. Multiple river sampling ports are also installed on the tributary water channels.

4. The nitrogen migration and transformation simulation device according to claim 1, characterized in that, The angle of the zigzag structure can be adjusted within 20~60°, and the length of the first movable water-proof baffle is 30~60cm.

5. The nitrogen migration and transformation simulation device according to claim 1, characterized in that, The first movable baffle plate is also provided with multiple first water outlets and first water level control heads. A pore water sampler is connected to the first water outlet. The collection end of the pore water sampler is set inside the sand box structure. The output end of the pore water sampler is connected to the input end of the data processing unit.

6. The nitrogen migration and transformation simulation device according to claim 2, characterized in that, The movable perforated baffle and the second movable water-blocking baffle are both equipped with pulleys on their movable bottom sides. The movable perforated baffle is set inside the main cavity structure. The second movable water-blocking baffle is also provided with a second water outlet and a second water level control head. The second permeable baffle is set on the side close to the confluence water tank.

7. The nitrogen migration and transformation simulation device according to claim 1, characterized in that, The monitoring terminal monitors the hydrological information within the simulation unit and categorizes and organizes the monitored hydrological information according to the set indicators.

Images