Underground water level prediction method and device based on artificial intelligence

Abstract

The invention provides an underground water level prediction method and device based on artificial intelligence. According to the method, according to the length of each fracture area, the width of each fracture area, the height of each fracture area, the opening degree of fractures in each fracture area, the matrix permeability of each fracture area, the relative permeability of water, the viscosity coefficient of water and the volume coefficient of water, the thickness of each fracture area is calculated; determining the equivalent permeability of water in a target stratum area on the shortest path between the to-be-constructed area and the target river, a first detection water level of a target river, a first underground water burial depth of a to-be-constructed area, an equivalent permeability of water in a target stratum area on a shortest path between the to-be-constructed area and the target river, a maximum precipitation amount of the to-be-constructed area in a week in a flood season in historical years, and a historical highest water level of the target river are calculated to obtain a first detection water level of the target river, a first underground water burial depth of the to-be-constructed area and a second underground water burial depth of the to-be-constructed area. The shortest distance between the to-be-built area and the target river is input into the highest groundwater level prediction model, the first highest groundwater level of the to-be-built area is obtained, and the accuracy is high.

Description

Technical Field

[0001] This invention relates to the field of engineering building structure design technology, and in particular to an artificial intelligence-based method and device for predicting groundwater levels. Background Technology

[0002] Anti-buoyancy design refers to the design methods used in engineering structures to prevent the buoyancy force exerted on the structure by groundwater, water flow, etc., to ensure the stability and safety of the structure during use. It is an important part of basement or foundation design. Improper anti-buoyancy design can easily lead to engineering accidents, ranging from minor cracks and leaks to major issues like overall floating and tilting. The main purpose of anti-buoyancy design is to prevent structural damage due to buoyancy, thereby extending the service life of the structure and reducing maintenance costs. When designing building projects in areas with abundant groundwater, the determination of the anti-buoyancy design water level and foundation elevation directly affects structural safety. These two parameters need to be comprehensively considered in conjunction with geological conditions, hydrological characteristics, and building functions. The anti-buoyancy design water level refers to the water level of the dominant groundwater layer within the foundation masonry depth during the building's operation, and its value directly affects the calculation of anti-buoyancy design loads.

[0003] Currently, anti-buoyancy design is mainly based on the measured groundwater level in the area where the engineering structure needs to be constructed. However, for areas near rivers where engineering structures need to be constructed, various environmental factors can affect the reliability of anti-buoyancy design based on randomly measured groundwater levels, which still poses certain safety risks. Summary of the Invention

[0004] This invention provides an artificial intelligence-based groundwater level prediction method and apparatus to address the problem that the reliability of anti-buoyancy design based on randomly measured groundwater levels in areas where engineering structures need to be constructed is still low, and there are certain safety hazards.

[0005] In a first aspect, the present invention provides an artificial intelligence-based groundwater level prediction method, applied to a backend server. The method provided by the present invention includes: If there is no rainfall in the area to be constructed within one week, the system receives the first detected water level of the target river from the liquid level sensor and the first groundwater depth of the area to be constructed from the water level depth sensor. Obtain the maximum weekly rainfall during the flood season in the area to be constructed, the historical highest water level of the target river, the shortest distance between the area to be constructed and the target river, and the geological data of the target stratigraphic region along the shortest path between the area to be constructed and the target river. The target stratigraphic region includes the minimum cuboid region of multiple fracture regions, each fracture region being the minimum cuboid region containing one fracture. The geological data includes the length, width, height, fracture aperture, matrix permeability, relative water permeability, water viscosity coefficient, and water volume coefficient of each fracture region. Based on the length, width, height, crack opening, matrix permeability, relative water permeability, water viscosity coefficient, and water volume coefficient of each fracture region, determine the equivalent permeability of the target stratum region along the shortest path between the area to be constructed and the target river. The first detected water level of the target river, the first groundwater depth of the area to be constructed, the equivalent permeability of the target stratum along the shortest path between the area to be constructed and the target river, the maximum weekly rainfall during the flood season in the area to be constructed in historical years, the historical highest water level of the target river, and the shortest distance between the area to be constructed and the target river are input into a pre-trained highest groundwater level prediction model to obtain the first highest groundwater level of the area to be constructed. The shallower the groundwater depth, the higher the groundwater level. The highest groundwater level prediction model is trained by inputting multiple first training samples into a first neural network. Each first training sample includes the detected water level of the target river in the case of no rainfall in the historical period, the first groundwater depth of the historical area to be constructed, the equivalent permeability of the target stratum along the shortest path between the historical area to be constructed and the historical target river, the maximum weekly rainfall during the flood season in the historical period of the historical area to be constructed, the historical highest water level of the historical target river, the shortest distance between the historical area to be constructed and the historical target river, and the corresponding actual first highest groundwater level of the historical area to be constructed.

[0006] In some implementations, the equivalent permeability of water in the target formation along the shortest path between the area to be constructed and the target river is determined based on the length, width, height, fracture aperture, matrix permeability, relative permeability, viscosity coefficient, and volume coefficient of each fracture region. This includes: According to the formula Determine the permeability of each fracture region, where, The length of each crack region, The width of each crack region, The height of each crack region; Permeability of each crack region; The average permeability of each fracture region is determined based on the permeability of each fracture region. According to the formula ( - Determine the equivalent permeability of the target stratigraphic region; among which, The equivalent permeability of the target formation area. The average aperture of multiple cracks, The average permeability of the target formation area. The matrix permeability of the target formation region, The fracture linear density of the target formation region.

[0007] In some embodiments, after obtaining the first highest groundwater level in the area to be constructed, the method provided by the present invention further includes: At multiple sampling times within a preset target duration, the system receives the second detected water level of the target river collected by the liquid level sensor, the second precipitation data sent by the meteorological server, and the second groundwater depth of the area to be constructed collected by the water level depth sensor. By merging the second detection water level, the second precipitation data, and the second groundwater depth at the same sampling time into a set of target data, multiple sets of target data are obtained. By fitting multiple sets of target data, an objective function model is obtained that represents the functional relationship between the second detected water level, the second precipitation data, and the second groundwater depth. The maximum precipitation and the historical highest water level are input into the objective function model to obtain the second highest groundwater level in the area to be developed. The first highest groundwater level is adjusted based on the second highest groundwater level.

[0008] In some implementations, adjusting the first highest groundwater level based on the second highest groundwater level includes: According to the formula The first highest groundwater level was revised, among which... The first weighting coefficient, It is the second weighting coefficient, and =1, This is the highest groundwater level before the correction. This is the second highest groundwater level. This is the revised highest groundwater level.

[0009] In some implementations, the following data are obtained: the maximum weekly rainfall in the area to be constructed during the flood season in historical years; the historical highest water level of the target river; the shortest distance between the area to be constructed and the target river; and the geological data of the target stratigraphic region along the shortest path between the area to be constructed and the target river, including: It receives the maximum weekly rainfall during the flood season in the area to be constructed from the meteorological server, the historical highest water level of the target river from the river monitoring server, and the shortest distance between the area to be constructed and the target river, as well as the geological data of the target stratum area from the terminal equipment.

[0010] Secondly, the present invention also provides an artificial intelligence-based groundwater level prediction device, configured on a backend server. The device provided by the present invention includes: The data receiving unit is used to receive the first detected water level of the target river collected by the liquid level sensor and the first groundwater depth of the area to be constructed collected by the water level depth sensor, provided that there has been no precipitation in the area to be constructed within one week. The data acquisition unit is used to acquire the maximum weekly rainfall during the flood season in the historical years for the area to be constructed, the historical highest water level of the target river, the shortest distance between the area to be constructed and the target river, and the geological data of the target stratigraphic region along the shortest path between the area to be constructed and the target river. The target stratigraphic region includes the minimum cuboid region of multiple fracture regions, each fracture region being the minimum cuboid region containing one fracture. The geological data includes the length, width, height, fracture aperture, matrix permeability, relative water permeability, water viscosity coefficient, and water volume coefficient of each fracture region. The equivalent permeability determination unit is used to determine the equivalent permeability of the target stratum region along the shortest path between the area to be constructed and the target river, based on the length of each fracture region, the width of each fracture region, the height of each fracture region, the opening of the fracture in each fracture region, the matrix permeability of each fracture region, the relative permeability of water, the viscosity coefficient of water, and the volume coefficient of water. The highest groundwater level determination unit is used to input the first detected water level of the target river, the first groundwater depth of the area to be constructed, the equivalent permeability of the target stratum area on the shortest path between the area to be constructed and the target river, the maximum weekly rainfall during the flood season in the area to be constructed in historical years, the historical highest water level of the target river, and the shortest distance between the area to be constructed and the target river into a pre-trained highest groundwater level prediction model to obtain the first highest groundwater level of the area to be constructed. The shallower the groundwater depth, the higher the groundwater level. The highest groundwater level prediction model is trained by inputting multiple first training samples into a first neural network. Each first training sample includes the detected water level of the historical target river in the case of no rainfall in a week, the first groundwater depth of the historical area to be constructed, the equivalent permeability of the target stratum area on the shortest path between the historical area to be constructed and the historical target river, the maximum weekly rainfall during the flood season in the historical year, the historical highest water level of the historical target river, the shortest distance between the historical area to be constructed and the historical target river, and the corresponding actual first highest groundwater level of the historical area to be constructed.

[0011] In some implementations, the equivalent permeability determination unit is specifically used to determine the equivalent permeability based on a formula. Determine the permeability of each fracture region, where, The length of each crack region, The width of each crack region, The height of each crack region; The permeability of each fracture region is determined; based on the permeability of each fracture region, the average permeability of each fracture region is determined; according to the formula... ( - Determine the equivalent permeability of the target stratigraphic region; among which, The equivalent permeability of the target formation area. The average aperture of multiple cracks, The average permeability of the target formation area. The matrix permeability of the target formation region, The fracture linear density of the target formation region.

[0012] In some implementations, the data receiving unit further includes receiving, at multiple sampling times during a preset target duration, the second detected water level of the target river collected by the liquid level sensor, the second precipitation data sent by the meteorological server, and the second groundwater depth of the area to be constructed collected by the water level depth sensor. The device provided by the present invention further includes: a data fusion unit, used to fuse the second detected water level, the second precipitation data and the second groundwater depth that are sampled at the same time into a set of target data, thereby obtaining multiple sets of target data; The function model fitting unit is used to fit multiple sets of target data to obtain a target function model that represents the functional relationship between the second detected water level, the second precipitation data and the second groundwater depth. The highest groundwater level determination unit is also used to input the maximum precipitation and the historical highest water level into the objective function model to obtain the second highest groundwater level in the area to be constructed. The device provided by the present invention further includes: a highest groundwater level correction unit, used to correct the first highest groundwater level according to the second highest groundwater level.

[0013] In some implementations, the highest groundwater level correction unit is specifically used to calculate... The first highest groundwater level was revised, among which... The first weighting coefficient, It is the second weighting coefficient, and =1, This is the highest groundwater level before the correction. This is the second highest groundwater level. This is the revised highest groundwater level.

[0014] In some implementations, the data acquisition unit is specifically used to receive the maximum weekly rainfall of the area to be constructed during the flood season in historical years from the meteorological server, the historical highest water level of the target river from the river monitoring server, and the shortest distance between the area to be constructed and the target river and the geological data of the target stratum area from the terminal device.

[0015] The present invention provides an artificial intelligence-based groundwater level prediction method and device, which can obtain the maximum weekly rainfall of the area to be constructed during the flood season in historical years, the historical highest water level of the target river, the shortest distance between the area to be constructed and the target river, and the geological data of the target stratigraphic area on the shortest path between the area to be constructed and the target river. Furthermore, based on the length, width, height, crack opening, matrix permeability, relative water permeability, viscosity coefficient, and volume coefficient of each crack region, the equivalent permeability of the target stratum along the shortest path between the construction area and the target river is determined. The equivalent permeability serves as a crucial parameter for predicting the highest groundwater level, ensuring that the resulting first highest groundwater level in the construction area is more accurately aligned with the actual topographic features. The first detected water level of the target river, the first groundwater depth in the construction area, the equivalent permeability of the target stratum along the shortest path between the construction area and the target river, the maximum weekly rainfall during the flood season in historical years for the construction area, the historical highest water level of the target river, and the shortest distance between the construction area and the target river are input into a pre-trained highest groundwater level prediction model to obtain the first highest groundwater level in the construction area. This model incorporates parameters from multiple dimensions, further enhancing accuracy. This ensures that the reliability of the anti-buoyancy design based on the highest groundwater level in the area to be constructed is high, thus improving the safety of the anti-buoyancy design. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating an artificial intelligence-based groundwater level prediction method provided in an embodiment of the present invention; Figure 2 A schematic diagram illustrating the positional relationship between the target geological region, the target river, and the area to be constructed, provided for an embodiment of the present invention. Figure 3 This is a structural schematic diagram of the crack region provided in an embodiment of the present invention.

[0018] Attached map labels: 201-Target river; 202-Area to be constructed; 203-Target stratigraphic area; 204-Fractured area; 205-Fractured area; 206-Non-fractured area. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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 made by those skilled in the art under the guidance of these embodiments are within the scope of protection of the present invention.

[0020] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0021] Please see Figure 1 This invention provides an artificial intelligence-based groundwater level prediction method, applied to a backend server. For example... Figure 1 As shown, the method provided in this embodiment of the invention includes: S101: If there is no precipitation in the construction area 202 within one week, receive the first detected water level of the target river 201 collected by the liquid level sensor, and receive the first groundwater depth of the construction area 202 collected by the water level depth sensor.

[0022] For example, the terminal device can receive the maximum weekly rainfall during the flood season in historical years for the area 202 to be constructed from a meteorological server, the historical highest water level of the target river 201 from a river monitoring server, and the shortest distance between the area 202 to be constructed and the target river 201, as well as geological data of the target stratigraphic region 203 from a terminal device. The shortest distance between the area 202 to be constructed and the target river 201 can be pre-measured by staff using a distance measuring instrument and stored in the terminal device (e.g., an office computer), and the geological data of the target stratigraphic region 203 can be measured by staff using geological surveying equipment and stored in the terminal device (e.g., an office computer).

[0023] Understandably, the water level in the area to be constructed, 202, is affected by factors such as rainfall, the water level of nearby rivers, and the equivalent permeability of the strata.

[0024] S102: Obtain the maximum weekly precipitation of the construction area 202 during the flood season in historical years, the historical highest water level of the target river 201, the shortest distance between the construction area 202 and the target river 201, and the geological data of the target stratigraphic area 203 on the shortest path between the construction area 202 and the target river 201.

[0025] For example, assuming the flood season is from July to September each year, we can obtain the maximum weekly rainfall during July to September for each of the past 30 years to determine the week with the highest rainfall over those 30 years. The historical highest water level of the target river 201 mentioned above can be the peak water level of the target river 201 over the past 30 years.

[0026] For example, the positional relationship between the target stratigraphic region 203, the target river 201, and the area to be constructed 202 can be as follows: Figure 2 As shown. Additionally... Figure 2 It also includes the non-cracked region 206.

[0027] The target stratum region 203 includes a minimum cuboid region of multiple fracture regions 204. Each fracture region 204 is a minimum cuboid region containing one fracture 205. The geological data include the length of each fracture region 204, the width of each fracture region 204, the height of each fracture region 204, the aperture of the fracture 205 in each fracture region 204, the matrix permeability of each fracture region 204, the relative permeability of water, the viscosity coefficient of water, and the volume coefficient of water.

[0028] S103: Based on the length of each fracture region 204, the width of each fracture region 204, the height of each fracture region 204, the opening of the fracture 205 in each fracture region 204, the matrix permeability of each fracture region 204, the relative permeability of water, the viscosity coefficient of water, and the volume coefficient of water, determine the equivalent permeability of water in the target stratum region 203 on the shortest path between the area to be constructed 202 and the target river 201.

[0029] For example, the structure of crack region 204 can be as follows: Figure 3 As shown. Figure 3 The length of the crack region 204 is l, the width is b, and the height is h.

[0030] Furthermore, S103 can be specifically implemented as follows: Step 1: According to the formula The permeability of each fracture region 204 was determined, where, The length of each crack region is 204. With a width of 204 for each crack region, The height of each crack region is 204; Permeability of 204 for each crack region.

[0031] Step 2: Determine the average permeability of each crack region 204 based on the permeability of each crack region 204.

[0032] Step 3: According to the formula ( - The equivalent permeability of the target stratum region 203 was determined.

[0033] in, The equivalent permeability of the target formation region 203. The average aperture of multiple cracks is 205. The average permeability of the target formation region 203. The matrix permeability of the target formation region 203 is given. Linear density of fractures 205 in target stratigraphic region 203.

[0034] Understandably, based on steps 1-3 above, the equivalent permeability of the target formation region 203 can be accurately obtained.

[0035] S104: Input the first detected water level of the target river 201, the first groundwater depth of the construction area 202, the equivalent permeability of the target stratum region 203 on the shortest path between the construction area 202 and the target river 201, the maximum weekly rainfall of the construction area 202 during the flood season in historical years, the historical highest water level of the target river 201, and the shortest distance between the construction area 202 and the target river 201 into the pre-trained highest groundwater level prediction model to obtain the first highest groundwater level of the construction area 202.

[0036] Among them, the shallower the groundwater depth, the higher the groundwater level. The highest groundwater level prediction model is obtained by inputting multiple first training samples into the first neural network. Each first training sample includes the detected water level of the historical target river 201 in the case of no precipitation within a week, the first groundwater depth of the historical construction area 202, the equivalent permeability of the target stratum area 203 on the shortest path between the historical construction area 202 and the historical target river 201, the maximum precipitation in the historical construction area 202 during the flood season in historical years, the historical highest water level of the historical target river 201, the shortest distance between the historical construction area 202 and the historical target river 201, and the corresponding actual first highest groundwater level of the historical construction area 202.

[0037] The above should be noted as follows: the measured water level of the target river in a week without precipitation is the historical measured water level of the target river in a week without precipitation; the first groundwater depth of the historical construction area is the historical groundwater depth of the construction area; the equivalent permeability of the target stratum along the shortest path between the historical construction area and the historical target river is the historical equivalent permeability of the historical target stratum corresponding to the shortest path between the construction area and the target river; the maximum precipitation in the historical construction area during the flood season in a historical year during a week is the historical maximum precipitation in the construction area during the flood season in a historical year; the historical highest water level of the target river is the historical highest water level of the target river; the shortest distance between the historical construction area and the historical target river is the historical shortest distance between the construction area and the target river; and the actual first highest groundwater level of the historical construction area is the historical actual first highest groundwater level of the construction area.

[0038] For example, the highest groundwater level prediction model can be, but is not limited to, a CNN-LSTM hybrid model, which includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, an LSTM layer, a fully connected layer, and an output layer; the first convolutional layer includes 64 convolutional kernels with a kernel size of 3×3 and the activation function is ReLU; the first pooling layer is a 1D convolutional layer with a pooling window size of 2; the second convolutional layer includes 128 convolutional kernels with a kernel size of 3×3 and the activation function is ReLU; the second pooling layer is a 1D convolutional layer with a pooling window size of 2; the LSTM layer contains 64 neurons and only outputs the result of the last time step; the fully connected layer is a Dense layer containing 50 neurons and the activation function is ReLU; the output layer uses a linear activation function.

[0039] The training parameters for the highest groundwater level prediction model are: learning rate of 0.001, Adam optimizer, batch size of 32, number of training epochs of 100, and loss function of mean squared error (MSE) or mean absolute error (MAE).

[0040] In summary, the groundwater level prediction method and apparatus based on artificial intelligence provided by this invention can obtain the maximum weekly rainfall of the construction area 202 during the flood season in historical years, the historical highest water level of the target river 201, the shortest distance between the construction area 202 and the target river 201, and the geological data of the target stratum region 203 on the shortest path between the construction area 202 and the target river 201. Furthermore, based on the length of each fracture region 204, the width of each fracture region 204, the height of each fracture region 204, the opening of the fracture 205 in each fracture region 204, the matrix permeability of each fracture region 204, the relative permeability of water, the viscosity coefficient of water, and the volume coefficient of water, the equivalent permeability of water in the target stratum region 203 on the shortest path between the construction area 202 and the target river 201 can be determined. Equivalent permeability can serve as a crucial parameter for predicting the highest groundwater level. This allows the subsequent prediction of the first highest groundwater level in the construction area 202 to be more accurately aligned with the actual topographic features. Furthermore, the first detected water level of the target river 201, the first groundwater depth in the construction area 202, the equivalent permeability of the target stratum region 203 along the shortest path between the construction area 202 and the target river 201, the maximum weekly rainfall during the flood season in the construction area 202 in historical years, the historical highest water level of the target river 201, and the shortest distance between the construction area 202 and the target river 201 are input into a pre-trained highest groundwater level prediction model. This yields the first highest groundwater level in the construction area 202. The resulting first highest groundwater level in the construction area 202 references parameters from multiple dimensions, further improving accuracy. Consequently, the reliability of the anti-buoyancy design based on the obtained first highest groundwater level in the construction area 202 is also high, enhancing the safety of the anti-buoyancy design.

[0041] In addition, after S104 described above, the method provided in this embodiment of the invention may further include: Step A: At multiple sampling times within a preset target duration, receive the second detected water level of the target river 201 collected by the liquid level sensor, receive the second precipitation data sent by the meteorological server, and receive the second groundwater depth of the construction area 202 collected by the water level depth sensor.

[0042] Step B: Combine the second detection water level, the second precipitation data, and the second groundwater depth data that were sampled at the same time into a set of target data to obtain multiple sets of target data.

[0043] Step C: Fit multiple sets of target data to obtain an objective function model that represents the functional relationship between the second detected water level, the second precipitation data, and the second groundwater depth.

[0044] Step D: Input the maximum precipitation and the historical highest water level into the objective function model to obtain the second highest groundwater level in the area to be constructed, 202.

[0045] Step E: Adjust the first highest groundwater level based on the second highest groundwater level.

[0046] For example, according to the formula The first highest groundwater level was revised, among which... The first weighting coefficient, It is the second weighting coefficient, and =1, This is the highest groundwater level before the correction. This is the second highest groundwater level. This is the revised highest groundwater level.

[0047] Understandably, based on the steps A-D described above, the accuracy of the first highest groundwater level can be further improved.

[0048] In addition, this embodiment of the invention also provides an artificial intelligence-based groundwater level prediction device, configured on a backend server. It should be noted that the basic principle and technical effects of the artificial intelligence-based groundwater level prediction device provided in this embodiment are the same as those in the above embodiments. For the sake of brevity, any parts not mentioned in this embodiment can be referred to the corresponding content in the above embodiments. The device provided in this embodiment includes a data receiving unit, a data acquisition unit, an equivalent permeability determination unit, and a maximum groundwater level determination unit, wherein... The data receiving unit is used to receive the first detected water level of the target river 201 collected by the liquid level sensor and the first groundwater depth of the area 202 to be constructed collected by the water level depth sensor, when there is no precipitation in the area 202 to be constructed within one week. The data acquisition unit is used to acquire the maximum weekly rainfall of the construction area 202 during the flood season in historical years, the historical highest water level of the target river 201, the shortest distance between the construction area 202 and the target river 201, and the geological data of the target stratigraphic region 203 on the shortest path between the construction area 202 and the target river 201. The target stratigraphic region 203 includes a minimum cuboid region of multiple fracture regions 204, each fracture region 204 being a minimum cuboid region containing one fracture 205. The geological data includes the length of each fracture region 204, the width of each fracture region 204, the height of each fracture region 204, the aperture of the fracture 205 in each fracture region 204, the matrix permeability of each fracture region 204, the relative permeability of water, the viscosity coefficient of water, and the volume coefficient of water. The equivalent permeability determination unit is used to determine the equivalent permeability of water in the target stratum region 203 on the shortest path between the area to be constructed 202 and the target river 201 based on the length of each fracture region 204, the width of each fracture region 204, the height of each fracture region 204, the aperture of the fracture 205 in each fracture region 204, the matrix permeability of each fracture region 204, the relative permeability of water, the viscosity coefficient of water, and the volume coefficient of water. The highest groundwater level determination unit is used to input the first detected water level of the target river 201, the first groundwater depth of the construction area 202, the equivalent permeability of the target stratum region 203 along the shortest path between the construction area 202 and the target river 201, the maximum weekly rainfall of the construction area 202 during the flood season in historical years, the historical highest water level of the target river 201, and the shortest distance between the construction area 202 and the target river 201 into a pre-trained highest groundwater level prediction model to obtain the first highest groundwater level of the construction area 202. The shallower the groundwater depth, the higher the groundwater level. The highest groundwater level prediction model is... The training is obtained by inputting multiple first training samples into the first neural network. Each first training sample includes the detected water level of the historical target river 201 under the condition of no precipitation in a week, the first groundwater depth of the historical construction area 202, the equivalent permeability of the target stratum area 203 on the shortest path between the historical construction area 202 and the historical target river 201, the maximum precipitation in a week during the flood season of the historical construction area 202 in a historical year, the historical highest water level of the historical target river 201, the shortest distance between the historical construction area 202 and the historical target river 201, and the corresponding actual first highest groundwater level of the historical construction area 202.

[0049] For example, the highest groundwater level prediction model can be, but is not limited to, a CNN-LSTM hybrid model, which includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, an LSTM layer, a fully connected layer, and an output layer; the first convolutional layer includes 64 convolutional kernels with a kernel size of 3×3 and the activation function is ReLU; the first pooling layer is a 1D convolutional layer with a pooling window size of 2; the second convolutional layer includes 128 convolutional kernels with a kernel size of 3×3 and the activation function is ReLU; the second pooling layer is a 1D convolutional layer with a pooling window size of 2; the LSTM layer contains 64 neurons and only outputs the result of the last time step; the fully connected layer is a Dense layer containing 50 neurons and the activation function is ReLU; the output layer uses a linear activation function.

[0050] The training parameters for the highest groundwater level prediction model are: learning rate of 0.001, Adam optimizer, batch size of 32, number of training epochs of 100, and loss function of mean squared error (MSE) or mean absolute error (MAE).

[0051] In some implementations, the equivalent permeability determination unit is specifically used to determine the equivalent permeability based on a formula. The permeability of each fracture region 204 was determined, where, The length of each crack region is 204. With a width of 204 for each crack region, The height of each crack region is 204; The permeability of each fracture region 204 is determined; based on the permeability of each fracture region 204, the average permeability of each fracture region 204 is determined; according to the formula... ( - The equivalent permeability of the target stratigraphic region 203 was determined; among which, The equivalent permeability of the target formation region 203. The average aperture of multiple cracks is 205. The average permeability of the target formation region 203. The matrix permeability of the target formation region 203 is given. Linear density of fractures 205 in target stratigraphic region 203.

[0052] In some implementations, the data receiving unit further includes receiving, at multiple sampling times during a preset target duration, the second detected water level of the target river 201 collected by the liquid level sensor, the second precipitation data sent by the meteorological server, and the second groundwater depth of the area to be constructed 202 collected by the water level depth sensor. The apparatus provided in this embodiment of the invention further includes: a data fusion unit, used to fuse second detected water level, second precipitation data and second groundwater depth with the same sampling time into a set of target data, to obtain multiple sets of target data; The function model fitting unit is used to fit multiple sets of target data to obtain a target function model that represents the functional relationship between the second detected water level, the second precipitation data and the second groundwater depth. The highest groundwater level determination unit is also used to input the maximum precipitation and the historical highest water level into the objective function model to obtain the second highest groundwater level in the area to be constructed, 202. The apparatus provided in this embodiment of the invention further includes: a highest groundwater level correction unit, used to correct the first highest groundwater level according to the second highest groundwater level.

[0053] In some implementations, the highest groundwater level correction unit is specifically used to calculate... The first highest groundwater level was revised, among which... The first weighting coefficient, It is the second weighting coefficient, and =1, This is the highest groundwater level before the correction. This is the second highest groundwater level. This is the revised highest groundwater level.

[0054] In some implementations, the data acquisition unit is specifically used to receive the maximum weekly precipitation of the area to be constructed 202 during the flood season in historical years from the meteorological server, the historical highest water level of the target river 201 from the river monitoring server, and the shortest distance between the area to be constructed 202 and the target river 201 and the geological data of the target stratigraphic region 203 from the terminal device.

[0055] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A groundwater level prediction method based on artificial intelligence, characterized in that, Applied to a backend server, the method includes: If there is no rainfall in the area to be constructed within one week, the system receives the first detected water level of the target river collected by the liquid level sensor and the first groundwater depth of the area to be constructed collected by the water level depth sensor. The geological data of the target stratigraphic region along the shortest path between the target area and the target river are obtained, including the maximum weekly rainfall during the flood season in historical years, the historical highest water level of the target river, the shortest distance between the target area and the target river, and the target stratigraphic region. The target stratigraphic region is a minimum cuboid region comprising multiple fracture regions, each fracture region being a minimum cuboid region containing one fracture. The geological data includes the length, width, height, fracture aperture, matrix permeability, relative water permeability, water viscosity coefficient, and water volume coefficient of each fracture region. The equivalent permeability of water in the target stratum region along the shortest path between the area to be constructed and the target river is determined based on the length, width, height, crack opening, matrix permeability, relative permeability, viscosity coefficient, and volume coefficient of each fracture region. The first detected water level of the target river, the first groundwater depth of the area to be constructed, the equivalent permeability of the target stratum along the shortest path between the area to be constructed and the target river, the maximum weekly rainfall during the flood season in the area to be constructed in historical years, the historical highest water level of the target river, and the shortest distance between the area to be constructed and the target river are input into a pre-trained highest groundwater level prediction model to obtain the first highest groundwater level of the area to be constructed. The shallower the groundwater depth, the higher the groundwater level. The highest groundwater level prediction model incorporates multiple first... The training samples are input into the first neural network for training. Each first training sample includes the detected water level of the historical target river in the case of no precipitation in a week, the first groundwater depth of the historical construction area, the equivalent permeability of the target stratum area on the shortest path between the historical construction area and the historical target river, the maximum precipitation in the historical construction area during the flood season in a historical year, the historical highest water level of the historical target river, the shortest distance between the historical construction area and the historical target river, and the corresponding actual first highest groundwater level of the historical construction area.

2. The method according to claim 1, characterized in that, The determination of the equivalent permeability of water in the target stratum region along the shortest path between the area to be constructed and the target river, based on the length, width, height, crack aperture, matrix permeability, relative permeability, viscosity coefficient, and volume coefficient of each fracture region, includes: According to the formula Determine the permeability of each of the said crack regions, wherein, The length of each of the said crack regions, The width of each of the crack regions, The height of each of the crack regions; The permeability of each of the aforementioned crack regions; The average permeability of each of the fracture regions is determined based on the permeability of each fracture region. According to the formula ( - Determine the equivalent permeability of the target formation region; wherein, The equivalent permeability of the target formation region. The average aperture of the multiple cracks, The average permeability of the target formation region. The matrix permeability of the target formation region. The fracture linear density is the value of the target formation region.

3. The method according to claim 1, characterized in that, After obtaining the first highest groundwater level in the area to be constructed, the method further includes: At multiple sampling times within a preset target duration, the system receives the second detected water level of the target river collected by the liquid level sensor, the second precipitation data sent by the meteorological server, and the second groundwater depth of the area to be constructed collected by the water level depth sensor. By merging the second detection water level, the second precipitation data, and the second groundwater depth at the same sampling time into a set of target data, multiple sets of target data are obtained. By fitting multiple sets of target data, a target function model is obtained that represents the functional relationship between the second detected water level, the second precipitation data, and the second groundwater depth. The maximum precipitation and the historical highest water level are input into the objective function model to obtain the second highest groundwater level in the area to be developed. The first highest groundwater level is adjusted based on the second highest groundwater level.

4. The method according to claim 3, characterized in that, The step of adjusting the first highest groundwater level based on the second highest groundwater level includes: According to the formula The first highest groundwater level is corrected, wherein... The first weighting coefficient, It is the second weighting coefficient, and =1, This is the highest groundwater level before the correction. This is the second highest groundwater level. This is the revised highest groundwater level.

5. The method according to claim 1, characterized in that, The acquisition of geological data for the target stratigraphic region along the shortest path between the target construction area and the target river during the flood season in historical years includes: The system receives the maximum weekly rainfall during the flood season in the area to be constructed from the meteorological server, the historical highest water level of the target river from the river monitoring server, the shortest distance between the area to be constructed and the target river from the terminal device, and geological data of the target stratum area.

6. A groundwater level prediction device based on artificial intelligence, characterized in that, Configured on a backend server, the device includes: The data receiving unit is used to receive the first detected water level of the target river collected by the liquid level sensor and the first groundwater depth of the area to be constructed collected by the water level depth sensor, provided that there has been no precipitation in the area to be constructed within one week. The data acquisition unit is used to acquire the maximum weekly rainfall during the flood season in the historical years for the area to be constructed, the historical highest water level of the target river, the shortest distance between the area to be constructed and the target river, and the geological data of the target stratigraphic region along the shortest path between the area to be constructed and the target river. The target stratigraphic region includes a minimum cuboid region of multiple fracture regions, each fracture region being a minimum cuboid region containing one fracture. The geological data includes the length, width, height, fracture aperture, matrix permeability, relative water permeability, water viscosity coefficient, and water volume coefficient of each fracture region. An equivalent permeability determination unit is used to determine the equivalent permeability of water in a target stratum region along the shortest path between the area to be constructed and the target river, based on the length of each fracture region, the width of each fracture region, the height of each fracture region, the aperture of the fracture in each fracture region, the matrix permeability of each fracture region, the relative permeability of water, the viscosity coefficient of water, and the volume coefficient of water. The highest groundwater level determination unit is used to input the first detected water level of the target river, the first groundwater depth of the area to be constructed, the equivalent permeability of the target stratum along the shortest path between the area to be constructed and the target river, the maximum weekly rainfall during the flood season in the area to be constructed in historical years, the historical highest water level of the target river, and the shortest distance between the area to be constructed and the target river into a pre-trained highest groundwater level prediction model to obtain the first highest groundwater level of the area to be constructed. The shallower the groundwater depth, the higher the groundwater level. It is obtained by inputting multiple first training samples into the first neural network for training. Each first training sample includes the detected water level of the historical target river in the case of no precipitation in a week, the first groundwater depth of the historical construction area, the equivalent permeability of the target stratum area on the shortest path between the historical construction area and the historical target river, the maximum precipitation in the historical construction area during the flood season in a historical year, the historical highest water level of the historical target river, the shortest distance between the historical construction area and the historical target river, and the corresponding actual first highest groundwater level of the historical construction area.

7. The apparatus according to claim 6, characterized in that, The equivalent permeability determination unit is specifically used to determine the equivalent permeability based on the formula. Determine the permeability of each of the said crack regions, wherein, The length of each of the said crack regions, The width of each of the crack regions, The height of each of the crack regions; The permeability of each of the fracture regions; based on the permeability of each of the fracture regions, the average permeability of each of the fracture regions is determined; according to the formula... ( - Determine the equivalent permeability of the target formation region; wherein, The equivalent permeability of the target formation region. The average aperture of the multiple cracks, The average permeability of the target formation region. The matrix permeability of the target formation region. The fracture linear density is the value of the target formation region.

8. The apparatus according to claim 6, characterized in that, The data receiving unit further includes receiving, at multiple sampling times within a preset target duration, the second detected water level of the target river collected by the liquid level sensor, the second precipitation data sent by the meteorological server, and the second groundwater depth of the area to be constructed collected by the water level depth sensor. The device further includes: a data fusion unit, used to fuse the second detected water level, the second precipitation data and the second groundwater depth that are sampled at the same time into a set of target data, and obtain multiple sets of target data; The function model fitting unit is used to fit multiple sets of target data to obtain a target function model that characterizes the functional relationship between the second detected water level, the second precipitation data and the second groundwater depth. The highest groundwater level determination unit is also used to input the maximum precipitation and the historical highest water level into the objective function model to obtain the second highest groundwater level of the area to be constructed. The device further includes: a highest groundwater level correction unit, used to correct the first highest groundwater level according to the second highest groundwater level.

9. The apparatus according to claim 8, characterized in that, The highest groundwater level correction unit is specifically used to calculate... The first highest groundwater level is corrected, wherein... The first weighting coefficient, It is the second weighting coefficient, and =1, This is the highest groundwater level before the correction. This is the second highest groundwater level. This is the revised highest groundwater level.

10. The apparatus according to claim 6, characterized in that, The data acquisition unit is specifically used to receive the maximum weekly rainfall of the area to be constructed during the flood season in historical years from the meteorological server, the historical highest water level of the target river from the river monitoring server, and the shortest distance between the area to be constructed and the target river, as well as the geological data of the target stratum area from the terminal device.

Images