Medical waste treatment monitoring method based on artificial intelligence

Abstract

The invention discloses a medical waste treatment monitoring method based on artificial intelligence, and relates to the technical field of medical waste treatment.The medical waste treatment monitoring method comprises the steps that the percolate diffusion rate, the gas emission amount and the soil humidity data are collected in real time, the diffusion rules of percolate and gas under different geological conditions are dynamically simulated through CFD model simulation, and the percolate and gas diffusion rate is calculated; generating a spatial distribution map of a high-risk area in the landfill; then combining a high-risk area spatial distribution map generated by CFD with historical operation data of a landfill to construct a space-time multi-dimensional data set, and overcoming the problems of single data and limited analysis in a traditional method; on the basis, a depth map network model ST-GCN based on space-time dynamics is utilized, a mode of combining time sequence modeling and space modeling is adopted, potential high-risk areas are identified, targeted optimization suggestions are generated, and compared with passive management of a traditional method, accurate identification and active optimization can be provided for the risk areas, and the risk areas can be effectively optimized. And the monitoring precision and the emergency efficiency are obviously improved.

Description

Technical Field

[0001] The present invention relates to the technical field of medical waste treatment, and in particular to a medical waste treatment monitoring method based on artificial intelligence. Background Art

[0002] The treatment and monitoring of medical waste are of great significance to public health and environmental protection. The treatment process includes classification, collection, transportation, storage and terminal disposal. However, since medical waste is infectious and toxic, leachate diffusion and gas migration may occur during the landfill process. If not handled properly, it will pollute the environment and pose a threat to residents' health.

[0003] At present, the monitoring method of medical waste treatment mainly relies on sampling at monitoring points and periodic inspections. This method is difficult to fully capture the dynamic changes in the treatment process and has certain limitations. To address this problem, some landfills use a double-layer anti-seepage system in their design to reduce the risk of leachate pollution to the surrounding environment. At the same time, monitoring wells and gas sampling ports are set up to collect groundwater and gas data for laboratory analysis, which reduces the direct hazards of medical waste to a certain extent.

[0004] However, this type of monitoring scheme relies more on passive monitoring and post-intervention, and it is difficult to provide timely warnings of sudden problems such as leachate leakage or excessive gas concentration. Therefore, the technology and means of medical waste treatment monitoring still need to be further optimized to achieve more efficient dynamic monitoring and risk prevention and control. Summary of the invention

[0005] In view of the above existing problems, the present invention is proposed.

[0006] The present invention provides an artificial intelligence-based medical waste treatment monitoring method to solve the problems of traditional medical waste treatment monitoring solutions that cannot adapt to dynamic needs and lack the ability to analyze long-term operating data.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] The embodiment of the present invention provides a medical waste treatment monitoring method based on artificial intelligence, which includes:

[0009] Step S1, deploying sensors in the landfill, including RFID tags, gas concentration sensors, and soil moisture sensors, to collect landfill data in real time, including leachate diffusion rate, gas emissions, and soil moisture;

[0010] Step S2, constructing a CFD model, inputting landfill data into the CFD model, simulating the migration path and diffusion rate of leachate under different geological conditions, and the diffusion law of landfill gas, distinguishing high-risk areas and medium- and low-risk areas within the landfill based on the migration path, diffusion rate and diffusion law, and generating a spatial distribution map of high-risk areas within the landfill;

[0011] Step S3, combining the spatial distribution map of high-risk areas within the landfill generated by CFD with the historical operation data of the landfill to construct a landfill data set;

[0012] Step S4, build a deep graph network model based on spatiotemporal dynamics based on the data set, train the model, use the trained deep graph network model to analyze the data set, identify potential risk areas, and generate optimization suggestions based on spatiotemporal features during the analysis process.

[0013] As a preferred solution of the medical waste treatment monitoring method based on artificial intelligence described in the present invention, the step of constructing the CFD model is:

[0014] The migration model of leachate is carried out, and the model formula is:

[0015]

[0016] Among them, v l is the velocity vector of the leachate, k is the permeability of the porous medium, μ l is the dynamic viscosity of the leachate, is the pore pressure gradient;

[0017] Gas diffusion modeling is performed, and the model formula is:

[0018]

[0019] Where C is the gas concentration, u g is the gas velocity vector, D g is the diffusion coefficient of the gas, R g is the rate of gas generation or absorption, is the time rate of change of gas concentration;

[0020] Add geological constraints:

[0021]

[0022] Among them, C l is the leachate concentration, φ is the porosity, R L is the leachate generation or absorption rate.

[0023] As a preferred embodiment of the medical waste treatment monitoring method based on artificial intelligence described in the present invention, the steps of simulating the migration path and diffusion rate of leachate under different geological conditions and the diffusion law of landfill gas are as follows:

[0024] The landfill area is divided into three-dimensional grids, and the grid nodes and units are defined. The definition formula is:

[0025] ΔV i，j，k =Δx·Δy·Δz,

[0026] Where, ΔV i，j，k is the volume of a single grid unit, Δx, Δy, Δz are the dimensions of the grid unit in each direction;

[0027] Set the boundary conditions:

[0028] The pressure boundary is P(x, y, z), P(x, y, z) = P boundary ,

[0029] The concentration boundary is C(x, y, z), C(x, y, z) = C boundary ,

[0030] Among them, P boundary is the pressure value at the boundary, C boundary is the gas concentration value at the boundary;

[0031] The fluid state is updated by the time step Δt, and the update formula is:

[0032]

[0033] Among them, C n , C n+1 is the gas concentration at time steps n and n+1, Δt is the time step;

[0034] Calculate the risk index of leachate or gas based on the simulation results The calculation formula is:

[0035]

[0036] in, is the risk indicator, C threshold is the concentration risk threshold, P threshold is the pressure gradient risk threshold;

[0037] Based on risk indicators Based on the distribution of landfills, the landfill is divided into high-risk areas R>1 and medium- and low-risk areas R≤1, and a spatial distribution map is generated through three-dimensional visualization technology.

[0038] As a preferred solution of the medical waste treatment monitoring method based on artificial intelligence described in the present invention, the historical operation data includes historical leachate treatment volume, gas concentration change trend and anti-seepage system parameters used for waste treatment.

[0039] As a preferred solution of the medical waste treatment monitoring method based on artificial intelligence described in the present invention, the step of combining the spatial distribution map of high-risk areas inside the landfill generated by CFD with the historical operation data of the landfill to construct the landfill data set is as follows:

[0040] The spatial distribution map of high-risk areas generated by CFD is represented by a three-dimensional matrix, and each grid cell contains the risk index value. is, expressed as:

[0041]

[0042] in, is the risk index value at position (i, j, k), C i，j，k is the gas concentration at position (i, j, k), C threshold is the concentration risk threshold, is the pressure gradient at position (i, j, k), P threshold is the pressure gradient risk threshold;

[0043] By Satisfaction The grid cell set defines the high-risk area, and its three-dimensional spatial location is represented by the coordinates (x i ,y j , z k )express;

[0044] Extract features of historical operation data, including:

[0045] Leachate treatment volume history:

[0046] Q t =∫ V φC l dV,

[0047] Among them, Q t is the total leachate treatment volume at time t, φ is the porosity, C l is the leachate concentration, V is the total volume of the landfill grid,

[0048] Gas concentration change trend:

[0049] ΔC g (t) = C g (t)-C g (t-Δt), where ΔC g (t) is the change in gas concentration at time t, Cg (t) is the gas concentration at time t, Δt is the time interval,

[0050] Anti-seepage system parameters: including permeability k and porosity φ.

[0051] As a preferred solution of the medical waste treatment monitoring method based on artificial intelligence described in the present invention, in the step of constructing the landfill data set, the spatial characteristics of the high-risk area are fused with the historical operation data to form a spatiotemporal multidimensional data set, and the final data record of each grid unit is

[0052] in, is the feature vector of position (i, j, k), is the risk index value of position (i, j, k), Q t is the total leachate treatment volume at time t, ΔC g (t) is the change in gas concentration at time t, k is the permeability, φ is the porosity, x i ,y j , z k is the spatial coordinate of the grid cell;

[0053] Integrate the data of each grid cell into a multidimensional matrix The matrix dimension is N x , N y , N z , F:

[0054] N x , N y , N z is the number of divisions of the landfill grid in the x, y, and z directions, and F is the number of features in each unit;

[0055] Final dataset for:

[0056] Dataset Used for subsequent training of deep graph network models to cover the spatiotemporal dynamic characteristics of landfills.

[0057] As a preferred solution of the medical waste treatment monitoring method based on artificial intelligence described in the present invention, in which: in step S4, when a potential risk area is identified, the corresponding adjustment plan is initiated, including automatically adjusting the drainage speed of the leachate collection system or initiating backup protection measures.

[0058] As a preferred solution of the artificial intelligence-based medical waste treatment monitoring method described in the present invention, in which: in step S4, the deep graph network model generates optimization suggestions based on potential risk areas, including adjusting the layout of the anti-seepage system or optimizing the waste hierarchy.

[0059] As a preferred solution of the medical waste treatment monitoring method based on artificial intelligence described in the present invention, the step of constructing a deep graph network model based on spatiotemporal dynamics based on the data set is:

[0060] Building a deep graph network model based on spatiotemporal dynamics:

[0061] Landfill dataset Represented as a space-time graph in:

[0062] The node set is Each node represents a grid unit (i, j, k), and the node characteristics include risk indicators Historical DataQ t , ΔC g , spatial coordinate x i ,y j , z k and geological properties k, φ,

[0063] The edge set is ε={e p→q}, edge weight ω p→q Represents the distance correlation between nodes:

[0064]

[0065] The adjacency matrix is in

[0066] The spatiotemporal dynamic network ST-GCN based on graph convolution is used for modeling:

[0067] First, use the timing model GRU to capture time dynamics. The capture formula is:

[0068]

[0069] Then use the graph convolution layer GCN to learn spatial features, and the learning formula is:

[0070]

[0071] Finally, the time and space features are combined for spatiotemporal fusion, where: is the hidden state of the lth layer of the graph convolution at time step t, is the input feature matrix, W(L) is the trainable weight matrix of the lth layer, A is the adjacency matrix, and σ is the activation function.

[0072] As a preferred solution of the medical waste treatment monitoring method based on artificial intelligence described in the present invention, the steps of training the model, analyzing the data set using the trained deep graph network model, identifying potential risk areas, and generating optimization suggestions in combination with spatiotemporal features during the analysis process are as follows:

[0073] The loss function is designed to set the training objective of the model to minimize the deviation between the prediction and the actual risk area. The loss function formula is:

[0074]

[0075] in, is the risk value predicted by the model, is the true risk value, λ|W| 2 is the regularization term;

[0076] The trained ST-GCN model receives the input dataset and generates a predicted risk value for each node. like Then mark the node as a potential high-risk area;

[0077] According to the location of high-risk areas, the layout of the anti-seepage layer is replanned, and the thickness of the variable anti-seepage layer d is optimized. layer Minimize leachate leakage, the optimization formula is:

[0078]

[0079] The landfill structure is adjusted according to the distribution of risk areas to minimize the impact of the upper waste weight on the leachate flow in the lower layer. The optimization formula is:

[0080]

[0081] Among them, d layer is the thickness of the impermeable layer, h waste is the height of the waste layer.

[0082] The beneficial effects of the present invention are as follows: first, a variety of sensors are deployed in the landfill to collect data on the diffusion rate of leachate, gas emissions and soil moisture in real time, and through CFD model simulation, the diffusion laws of leachate and gas under different geological conditions are dynamically simulated to generate a spatial distribution map of high-risk areas inside the landfill; secondly, the spatial distribution map of high-risk areas generated by CFD is combined with the historical operation data of the landfill to construct a spatiotemporal multidimensional data set, which contains dynamic risk spatial characteristics, geological parameters and historical change trends, overcoming the problems of single data and limited analysis in traditional methods; on this basis, the deep graph network model ST-GCN based on spatiotemporal dynamics is used to identify potential high-risk areas through a combination of time series modeling and spatial modeling, and generate targeted optimization suggestions. Compared with the passive management of traditional methods, it can provide accurate identification and active optimization of risk areas, significantly improving monitoring accuracy and emergency efficiency;

[0083] The present invention has emergency response capabilities. After identifying potential risk areas, it can adjust the drainage speed of the leachate collection system in real time, start backup protection measures, and optimize the anti-seepage system layout and waste hierarchy to further reduce pollution risks; thereby improving the operating safety and resource utilization efficiency of the landfill. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0085] Figure 1 Schematic diagram of the process of medical waste treatment monitoring method in Example 1. DETAILED DESCRIPTION

[0086] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0087] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0088] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0089] Example 1, reference Figure 1 , this embodiment provides a medical waste treatment monitoring method based on artificial intelligence, comprising the following steps:

[0090] Step S1, deploying sensors in the landfill, including RFID tags, gas concentration sensors, and soil moisture sensors, to collect landfill data in real time, including leachate diffusion rate, gas emissions, and soil moisture;

[0091] Step S2, constructing a CFD model, inputting landfill data into the CFD model, simulating the migration path and diffusion rate of leachate under different geological conditions, and the diffusion law of landfill gas, distinguishing high-risk areas and medium- and low-risk areas within the landfill based on the migration path, diffusion rate and diffusion law, and generating a spatial distribution map of high-risk areas within the landfill;

[0092] The steps to build a CFD model are:

[0093] The migration model of leachate is carried out, and the model formula is:

[0094]

[0095] Among them, v L is the velocity vector of the leachate, k is the permeability of the porous medium, μ L is the dynamic viscosity of the leachate, is the pore pressure gradient;

[0096] Gas diffusion modeling is performed, and the model formula is:

[0097]

[0098] Where C is the gas concentration, u g is the gas velocity vector, D g is the diffusion coefficient of the gas, R g is the rate of gas generation or absorption, is the time rate of change of gas concentration;

[0099] Add geological constraints:

[0100]

[0101] Among them, C lis the leachate concentration, φ is the porosity, R L is the leachate generation or absorption rate,

[0102] The steps to simulate the migration path, diffusion rate of leachate under different geological conditions, and the diffusion law of landfill gas are as follows:

[0103] The landfill area is divided into three-dimensional grids, and the grid nodes and units are defined. The definition formula is:

[0104] ΔV i，j，k =Δx·Δy·Δz,

[0105] Where, ΔV i，j，k is the volume of a single grid unit, Δx, Δy, Δz are the dimensions of the grid unit in each direction;

[0106] Set the boundary conditions:

[0107] The pressure boundary is P(x, y, z), P(x, y, z) = P boundary ,

[0108] The concentration boundary is C(x, y, z), C(x, y, z) = C boundary ,

[0109] Among them, P boundary is the pressure value at the boundary, C boundary is the gas concentration value at the boundary;

[0110] The fluid state is updated by the time step Δt, and the update formula is:

[0111]

[0112] Among them, C n , C n+1 is the gas concentration at time steps n and n+1, Δt is the time step;

[0113] Calculate the risk index of leachate or gas based on the simulation results The calculation formula is:

[0114]

[0115] in, is the risk indicator, C threshold is the concentration risk threshold, P threshold is the pressure gradient risk threshold;

[0116] Based on risk indicators The distribution of landfills is divided into high-risk areas and medium- and low-risk areas And generate spatial distribution maps through three-dimensional visualization technology;

[0117] Specifically, the geological condition modeling is combined with risk indicators, the porous media characteristics and CFD model are integrated, the modeling of leachate and gas diffusion laws is strengthened, the multi-dimensional dynamic diffusion behavior is simulated through time step iteration, and a real-time early warning basis is provided. At the same time, the risk indicator function is introduced to achieve quantitative division and intuitive display of risk areas.

[0118] Step S3, combining the spatial distribution map of high-risk areas within the landfill generated by CFD with the historical operation data of the landfill to construct a landfill data set;

[0119] Historical operating data include historical leachate treatment volume, gas concentration trends, and parameters of the anti-seepage system used for waste treatment;

[0120] The steps to construct the landfill dataset are to combine the spatial distribution map of high-risk areas within the landfill generated by CFD with the historical operation data of the landfill.

[0121] The spatial distribution map of high-risk areas generated by CFD is represented by a three-dimensional matrix, and each grid cell contains the risk index value. is, expressed as:

[0122]

[0123] in, is the risk index value at position (i, j, k), C i，j，k is the gas concentration at position (i, j, k), C threshold is the concentration risk threshold, is the pressure gradient at position (i, j, k), P threshold is the pressure gradient risk threshold;

[0124] By Satisfaction The grid cell set defines the high-risk area, and its three-dimensional spatial location is represented by the coordinates (x i ,y j , z k )express;

[0125] Extract features of historical operation data, including:

[0126] Leachate treatment volume history:

[0127] Q t =∫ V φC l dV,

[0128] Among them, Q t is the total leachate treatment volume at time t, φ is the porosity, Cl is the leachate concentration, V is the total volume of the landfill grid,

[0129] Gas concentration change trend:

[0130] ΔC g (t) = C g (t)-C g (t-Δt), where ΔC g (t) is the change in gas concentration at time t, C g (t) is the gas concentration at time t, Δt is the time interval,

[0131] Anti-seepage system parameters: including permeability k and porosity φ;

[0132] In the step of constructing the landfill dataset, the spatial characteristics of the high-risk areas are fused with the historical operation data to form a spatiotemporal multidimensional dataset. The final data record of each grid cell is

[0133] in, is the feature vector of position (i, j, k), is the risk index value of position (i, j, k), Q t is the total leachate treatment volume at time t, ΔC g (t) is the change in gas concentration at time t, k is the permeability, φ is the porosity, x i ,y j , z k is the spatial coordinate of the grid cell;

[0134] Integrate the data of each grid cell into a multidimensional matrix The matrix dimension is N x , N y , N z , F:

[0135] N x , N y , N z is the number of divisions of the landfill grid in the x, y, and z directions, and F is the number of features in each unit;

[0136] Final dataset for:

[0137] Dataset Used for subsequent training of deep graph network models to cover the spatiotemporal dynamic characteristics of landfills;

[0138] Specifically, the dynamic risk space characteristics of CFD are combined with historical operating data to form a complete spatiotemporal multidimensional data set, and physical parameters and dynamic change characteristics are integrated to provide comprehensive training input for machine learning models. At the same time, data is stored in a matrix form to provide more efficient access speed.

[0139] Step S4, constructing a deep graph network model based on spatiotemporal dynamics based on the data set, training the model, analyzing the data set using the trained deep graph network model, identifying potential risk areas, and generating optimization suggestions in combination with spatiotemporal features during the analysis process;

[0140] In step S4, when a potential risk area is identified, a corresponding adjustment plan is initiated, including automatically adjusting the drainage speed of the leachate collection system or initiating backup protection measures;

[0141] In step S4, the deep graph network model generates optimization suggestions based on the potential risk areas, including adjusting the layout of the anti-seepage system or optimizing the waste hierarchy;

[0142] The steps to build a deep graph network model based on spatiotemporal dynamics based on the dataset are:

[0143] Building a deep graph network model based on spatiotemporal dynamics:

[0144] Landfill dataset Represented as a space-time graph in:

[0145] The node set is Each node represents a grid unit (i, j, k), and the node characteristics include risk indicators Historical DataQ t , ΔC g , spatial coordinate x i ,y j , z k and geological properties k, φ,

[0146] The edge set is ε={e p→q}, edge weight ω p→q Represents the distance correlation between nodes:

[0147]

[0148] The adjacency matrix is in

[0149] The spatiotemporal dynamic network ST-GCN based on graph convolution is used for modeling:

[0150] First, use the timing model GRU to capture time dynamics. The capture formula is:

[0151]

[0152] Then use the graph convolution layer GCN to learn spatial features, and the learning formula is:

[0153]

[0154] Finally, the time and space features are combined for spatiotemporal fusion, where: is the hidden state of the lth layer of the graph convolution at time step t, is the input feature matrix, W(L) is the trainable weight matrix of the lth layer, is the adjacency matrix, σ is the activation function;

[0155] The steps of training the model, using the trained deep graph network model to analyze the data set, identifying potential risk areas, and generating optimization suggestions based on spatiotemporal features during the analysis are as follows:

[0156] The loss function is designed to set the training objective of the model to minimize the deviation between the prediction and the actual risk area. The loss function formula is:

[0157]

[0158] in, is the risk value predicted by the model, is the true risk value, λ|W| 2 is the regularization term;

[0159] The trained ST-GCN model receives the input dataset and generates a predicted risk value for each node. like Then mark the node as a potential high-risk area;

[0160] According to the location of high-risk areas, the layout of the anti-seepage layer is replanned, and the thickness of the variable anti-seepage layer d is optimized. layer Minimize leachate leakage, the optimization formula is:

[0161]

[0162] The landfill structure is adjusted according to the distribution of risk areas to minimize the impact of the upper waste weight on the leachate flow in the lower layer. The optimization formula is:

[0163]

[0164] Among them, d layer is the thickness of the impermeable layer, h waste is the height of the waste layer;

[0165] Specifically, the ST-GCN model is used to integrate the spatiotemporal dynamic characteristics of the landfill and identify potential risk areas. At the same time, deep learning and physical optimization methods are combined to generate quantitative optimization plans to improve the safety and efficiency of the landfill. The model prediction and dynamic adjustment are closely combined to achieve efficient linkage between real-time monitoring and emergency response.

[0166] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A medical waste treatment monitoring method based on artificial intelligence, characterized in that: include, Step S1, deploying sensors in the landfill, including RFID tags, gas concentration sensors, and soil moisture sensors, to collect landfill data in real time, including leachate diffusion rate, gas emissions, and soil moisture; Step S2, constructing a CFD model, inputting landfill data into the CFD model, simulating the migration path and diffusion rate of leachate under different geological conditions, and the diffusion law of landfill gas, distinguishing high-risk areas and medium- and low-risk areas within the landfill based on the migration path, diffusion rate and diffusion law, and generating a spatial distribution map of high-risk areas within the landfill; Step S3, combining the spatial distribution map of high-risk areas within the landfill generated by CFD with the historical operation data of the landfill to construct a landfill data set; Step S4, build a deep graph network model based on spatiotemporal dynamics based on the data set, train the model, use the trained deep graph network model to analyze the data set, identify potential risk areas, and generate optimization suggestions based on spatiotemporal features during the analysis process.

2. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 1, characterized in that: The steps of constructing the CFD model are: The migration model of leachate is carried out, and the model formula is: Among them, v l is the velocity vector of the leachate, k is the permeability of the porous medium, μ l is the dynamic viscosity of the leachate, is the pore pressure gradient; Gas diffusion modeling is performed, and the model formula is: Where C is the gas concentration, u g is the gas velocity vector, D g is the diffusion coefficient of the gas, R g is the rate of gas generation or absorption, is the time rate of change of gas concentration; Add geological constraints: Among them, C l is the leachate concentration, φ is the porosity, R l is the leachate generation or absorption rate.

3. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 2, characterized in that: The steps of simulating the migration path and diffusion rate of leachate under different geological conditions and the diffusion law of landfill gas are as follows: The landfill area is divided into three-dimensional grids, and the grid nodes and units are defined. The definition formula is: ΔV i,j,k =Δx·Δy·Δz, Where, ΔV i,j,k is the volume of a single grid unit, Δx, Δy, Δz are the dimensions of the grid unit in each direction; Set the boundary conditions: The pressure boundary is P(x,y,z), P(x,y,z)=P boundary , The concentration boundary is C(x,y,z), C(x,y,z)=C boundary , Among them, P boundary is the pressure value at the boundary, C boundary is the gas concentration value at the boundary; The fluid state is updated by the time step Δt, and the update formula is: Among them, C n ,C n+1 is the gas concentration at time steps n and n+1, Δt is the time step; Calculate the risk index of leachate or gas based on the simulation results The calculation formula is: in, is the risk indicator, C threshold is the concentration risk threshold, P threshold is the pressure gradient risk threshold; Based on risk indicators The distribution of landfills is divided into high-risk areas and medium- and low-risk areas And generate spatial distribution maps through three-dimensional visualization technology.

4. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 3, characterized in that: The historical operation data include historical leachate treatment volume, gas concentration change trend and anti-seepage system parameters used for waste treatment.

5. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 4, characterized in that: The steps of combining the spatial distribution map of high-risk areas within the landfill generated by CFD with the historical operation data of the landfill to construct the landfill data set are: The spatial distribution map of high-risk areas generated by CFD is represented by a three-dimensional matrix, and each grid cell contains the risk index value. is, expressed as: in, is the risk index value at position (i,j,k), C i,j,k is the gas concentration at position (i, j, k), C threshold is the concentration risk threshold, is the pressure gradient at position (i, j, k), P threshold is the pressure gradient risk threshold; By Satisfaction The grid cell set defines the high-risk area, and its three-dimensional spatial location is represented by the coordinates (x i ,y j ,z k )express; Extract features of historical operation data, including: Leachate treatment volume history: Q t =∫ V φC l dV, Among them, Q t is the total leachate treatment volume at time t, φ is the porosity, C l is the leachate concentration, V is the total volume of the landfill grid, Gas concentration change trend: ΔC g (t) = C g (t)-C g (t-Δt), where ΔC g (t) is the change in gas concentration at time t, C g (t) is the gas concentration at time t, Δt is the time interval, Anti-seepage system parameters: including permeability k and porosity φ.

6. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 5, characterized in that: In the step of constructing the landfill dataset, the spatial characteristics of the high-risk area are fused with the historical operation data to form a spatiotemporal multidimensional dataset. The final data record of each grid cell is in, is the feature vector of position (i,j,k), is the risk index value of position (i,j,k), Q t is the total leachate treatment volume at time t, ΔC g (t) is the change in gas concentration at time t, k is the permeability, φ is the porosity, x i ,y j ,z k is the spatial coordinate of the grid cell; Integrate the data of each grid cell into a multidimensional matrix The matrix dimension is N x ,N y ,N z ,F: N x ,N y ,N z is the number of divisions of the landfill grid in the x, y, and z directions, and F is the number of features in each unit; Final dataset for:

7. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 6, characterized in that: In step S4, when a potential risk area is identified, a corresponding adjustment plan is initiated, including automatically adjusting the drainage speed of the leachate collection system or initiating backup protection measures.

8. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 7, characterized in that: In step S4, the deep graph network model generates optimization suggestions based on the potential risk areas, including adjusting the layout of the anti-seepage system or optimizing the waste hierarchy.

9. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 8, characterized in that: The step of constructing a spatiotemporal dynamic deep graph network model based on the data set is: Building a deep graph network model based on spatiotemporal dynamics: Landfill dataset Represented as a space-time graph in: The node set is Each node represents a grid unit (i, j, k), and the node characteristics include risk indicators Historical DataQ t ,ΔC g , spatial coordinate x i ,y j ,z k and geological properties k,φ, The edge set is ε={e p→q }, edge weight ω p→q Represents the distance correlation between nodes: The adjacency matrix is in The spatiotemporal dynamic network ST-GCN based on graph convolution is used for modeling: First, use the timing model GRU to capture time dynamics. The capture formula is: Then use the graph convolution layer GCN to learn spatial features, and the learning formula is: Finally, the time and space features are combined for spatiotemporal fusion, where: is the hidden state of the lth layer of the graph convolution at time step t, is the input feature matrix, W (l) is the trainable weight matrix of layer l, is the adjacency matrix and σ is the activation function.

10. The medical waste treatment monitoring method based on artificial intelligence as claimed in claim 9, characterized in that: The steps of training the model, analyzing the data set using the trained deep graph network model, identifying potential risk areas, and generating optimization suggestions in combination with spatiotemporal features during the analysis are as follows: The loss function is designed to set the training objective of the model to minimize the deviation between the prediction and the actual risk area. The loss function formula is: in, is the risk value predicted by the model, is the true risk value, λ|W| 2 is the regularization term; The trained ST-GCN model receives the input dataset and generates a predicted risk value for each node. like Then mark the node as a potential high-risk area; According to the location of high-risk areas, the layout of the anti-seepage layer is replanned, and the thickness of the variable anti-seepage layer d is optimized. layer Minimize leachate leakage, the optimization formula is: The landfill structure is adjusted according to the distribution of risk areas to minimize the impact of the upper waste weight on the leachate flow in the lower layer. The optimization formula is: Among them, d layer is the thickness of the impermeable layer, h waste is the height of the waste layer.

Images